linux_dsm_epyc7002/kernel/events/uprobes.c

1689 lines
40 KiB
C
Raw Normal View History

uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
/*
* User-space Probes (UProbes)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*
* Copyright (C) IBM Corporation, 2008-2012
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
* Authors:
* Srikar Dronamraju
* Jim Keniston
* Copyright (C) 2011-2012 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com>
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
*/
#include <linux/kernel.h>
#include <linux/highmem.h>
#include <linux/pagemap.h> /* read_mapping_page */
#include <linux/slab.h>
#include <linux/sched.h>
#include <linux/rmap.h> /* anon_vma_prepare */
#include <linux/mmu_notifier.h> /* set_pte_at_notify */
#include <linux/swap.h> /* try_to_free_swap */
uprobes/core: Handle breakpoint and singlestep exceptions Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or a singlestep exception. When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters. Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups. In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64. Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe. Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost. Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete. Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing. Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc. Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too. Credits to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120313180011.29771.89027.sendpatchset@srdronam.in.ibm.com [ Performed various cleanliness edits ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-03-14 01:00:11 +07:00
#include <linux/ptrace.h> /* user_enable_single_step */
#include <linux/kdebug.h> /* notifier mechanism */
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
#include <linux/uprobes.h>
uprobes/core: Allocate XOL slots for uprobes use Uprobes executes the original instruction at a probed location out of line. For this, we allocate a page (per mm) upon the first uprobe hit, in the process user address space, divide it into slots that are used to store the actual instructions to be singlestepped. These slots are known as xol (execution out of line) slots. Care is taken to ensure that the allocation is in an unmapped area as close to the top of the user address space as possible, with appropriate permission settings to keep selinux like frameworks happy. Upon a uprobe hit, a free slot is acquired, and is released after the singlestep completes. Lots of improvements courtesy suggestions/inputs from Peter and Oleg. [ Folded a fix for build issue on powerpc fixed and reported by Stephen Rothwell. ] Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182631.10018.48175.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:31 +07:00
#define UINSNS_PER_PAGE (PAGE_SIZE/UPROBE_XOL_SLOT_BYTES)
#define MAX_UPROBE_XOL_SLOTS UINSNS_PER_PAGE
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
static struct rb_root uprobes_tree = RB_ROOT;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
static DEFINE_SPINLOCK(uprobes_treelock); /* serialize rbtree access */
#define UPROBES_HASH_SZ 13
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
/* serialize (un)register */
static struct mutex uprobes_mutex[UPROBES_HASH_SZ];
#define uprobes_hash(v) (&uprobes_mutex[((unsigned long)(v)) % UPROBES_HASH_SZ])
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
/* serialize uprobe->pending_list */
static struct mutex uprobes_mmap_mutex[UPROBES_HASH_SZ];
#define uprobes_mmap_hash(v) (&uprobes_mmap_mutex[((unsigned long)(v)) % UPROBES_HASH_SZ])
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
/*
* uprobe_events allows us to skip the uprobe_mmap if there are no uprobe
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
* events active at this time. Probably a fine grained per inode count is
* better?
*/
static atomic_t uprobe_events = ATOMIC_INIT(0);
/*
* Maintain a temporary per vma info that can be used to search if a vma
* has already been handled. This structure is introduced since extending
* vm_area_struct wasnt recommended.
*/
struct vma_info {
struct list_head probe_list;
struct mm_struct *mm;
loff_t vaddr;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
};
struct uprobe {
struct rb_node rb_node; /* node in the rb tree */
atomic_t ref;
struct rw_semaphore consumer_rwsem;
struct list_head pending_list;
struct uprobe_consumer *consumers;
struct inode *inode; /* Also hold a ref to inode */
loff_t offset;
int flags;
struct arch_uprobe arch;
};
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
/*
* valid_vma: Verify if the specified vma is an executable vma
* Relax restrictions while unregistering: vm_flags might have
* changed after breakpoint was inserted.
* - is_register: indicates if we are in register context.
* - Return 1 if the specified virtual address is in an
* executable vma.
*/
static bool valid_vma(struct vm_area_struct *vma, bool is_register)
{
if (!vma->vm_file)
return false;
if (!is_register)
return true;
if ((vma->vm_flags & (VM_HUGETLB|VM_READ|VM_WRITE|VM_EXEC|VM_SHARED))
== (VM_READ|VM_EXEC))
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return true;
return false;
}
static loff_t vma_address(struct vm_area_struct *vma, loff_t offset)
{
loff_t vaddr;
vaddr = vma->vm_start + offset;
vaddr -= vma->vm_pgoff << PAGE_SHIFT;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return vaddr;
}
/**
* __replace_page - replace page in vma by new page.
* based on replace_page in mm/ksm.c
*
* @vma: vma that holds the pte pointing to page
* @page: the cowed page we are replacing by kpage
* @kpage: the modified page we replace page by
*
* Returns 0 on success, -EFAULT on failure.
*/
static int __replace_page(struct vm_area_struct *vma, struct page *page, struct page *kpage)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
{
struct mm_struct *mm = vma->vm_mm;
pgd_t *pgd;
pud_t *pud;
pmd_t *pmd;
pte_t *ptep;
spinlock_t *ptl;
unsigned long addr;
int err = -EFAULT;
addr = page_address_in_vma(page, vma);
if (addr == -EFAULT)
goto out;
pgd = pgd_offset(mm, addr);
if (!pgd_present(*pgd))
goto out;
pud = pud_offset(pgd, addr);
if (!pud_present(*pud))
goto out;
pmd = pmd_offset(pud, addr);
if (!pmd_present(*pmd))
goto out;
ptep = pte_offset_map_lock(mm, pmd, addr, &ptl);
if (!ptep)
goto out;
get_page(kpage);
page_add_new_anon_rmap(kpage, vma, addr);
if (!PageAnon(page)) {
dec_mm_counter(mm, MM_FILEPAGES);
inc_mm_counter(mm, MM_ANONPAGES);
}
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
flush_cache_page(vma, addr, pte_pfn(*ptep));
ptep_clear_flush(vma, addr, ptep);
set_pte_at_notify(mm, addr, ptep, mk_pte(kpage, vma->vm_page_prot));
page_remove_rmap(page);
if (!page_mapped(page))
try_to_free_swap(page);
put_page(page);
pte_unmap_unlock(ptep, ptl);
err = 0;
out:
return err;
}
/**
* is_swbp_insn - check if instruction is breakpoint instruction.
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
* @insn: instruction to be checked.
* Default implementation of is_swbp_insn
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
* Returns true if @insn is a breakpoint instruction.
*/
bool __weak is_swbp_insn(uprobe_opcode_t *insn)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
{
return *insn == UPROBE_SWBP_INSN;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
}
/*
* NOTE:
* Expect the breakpoint instruction to be the smallest size instruction for
* the architecture. If an arch has variable length instruction and the
* breakpoint instruction is not of the smallest length instruction
* supported by that architecture then we need to modify read_opcode /
* write_opcode accordingly. This would never be a problem for archs that
* have fixed length instructions.
*/
/*
* write_opcode - write the opcode at a given virtual address.
* @auprobe: arch breakpointing information.
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
* @mm: the probed process address space.
* @vaddr: the virtual address to store the opcode.
* @opcode: opcode to be written at @vaddr.
*
* Called with mm->mmap_sem held (for read and with a reference to
* mm).
*
* For mm @mm, write the opcode at @vaddr.
* Return 0 (success) or a negative errno.
*/
static int write_opcode(struct arch_uprobe *auprobe, struct mm_struct *mm,
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
unsigned long vaddr, uprobe_opcode_t opcode)
{
struct page *old_page, *new_page;
struct address_space *mapping;
void *vaddr_old, *vaddr_new;
struct vm_area_struct *vma;
struct uprobe *uprobe;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
loff_t addr;
int ret;
/* Read the page with vaddr into memory */
ret = get_user_pages(NULL, mm, vaddr, 1, 0, 0, &old_page, &vma);
if (ret <= 0)
return ret;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
ret = -EINVAL;
/*
* We are interested in text pages only. Our pages of interest
* should be mapped for read and execute only. We desist from
* adding probes in write mapped pages since the breakpoints
* might end up in the file copy.
*/
if (!valid_vma(vma, is_swbp_insn(&opcode)))
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
goto put_out;
uprobe = container_of(auprobe, struct uprobe, arch);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
mapping = uprobe->inode->i_mapping;
if (mapping != vma->vm_file->f_mapping)
goto put_out;
addr = vma_address(vma, uprobe->offset);
if (vaddr != (unsigned long)addr)
goto put_out;
ret = -ENOMEM;
new_page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma, vaddr);
if (!new_page)
goto put_out;
__SetPageUptodate(new_page);
/*
* lock page will serialize against do_wp_page()'s
* PageAnon() handling
*/
lock_page(old_page);
/* copy the page now that we've got it stable */
vaddr_old = kmap_atomic(old_page);
vaddr_new = kmap_atomic(new_page);
memcpy(vaddr_new, vaddr_old, PAGE_SIZE);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
/* poke the new insn in, ASSUMES we don't cross page boundary */
vaddr &= ~PAGE_MASK;
BUG_ON(vaddr + UPROBE_SWBP_INSN_SIZE > PAGE_SIZE);
memcpy(vaddr_new + vaddr, &opcode, UPROBE_SWBP_INSN_SIZE);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
kunmap_atomic(vaddr_new);
kunmap_atomic(vaddr_old);
ret = anon_vma_prepare(vma);
if (ret)
goto unlock_out;
lock_page(new_page);
ret = __replace_page(vma, old_page, new_page);
unlock_page(new_page);
unlock_out:
unlock_page(old_page);
page_cache_release(new_page);
put_out:
put_page(old_page);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return ret;
}
/**
* read_opcode - read the opcode at a given virtual address.
* @mm: the probed process address space.
* @vaddr: the virtual address to read the opcode.
* @opcode: location to store the read opcode.
*
* Called with mm->mmap_sem held (for read and with a reference to
* mm.
*
* For mm @mm, read the opcode at @vaddr and store it in @opcode.
* Return 0 (success) or a negative errno.
*/
static int read_opcode(struct mm_struct *mm, unsigned long vaddr, uprobe_opcode_t *opcode)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
{
struct page *page;
void *vaddr_new;
int ret;
ret = get_user_pages(NULL, mm, vaddr, 1, 0, 1, &page, NULL);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
if (ret <= 0)
return ret;
lock_page(page);
vaddr_new = kmap_atomic(page);
vaddr &= ~PAGE_MASK;
memcpy(opcode, vaddr_new + vaddr, UPROBE_SWBP_INSN_SIZE);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
kunmap_atomic(vaddr_new);
unlock_page(page);
put_page(page);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return 0;
}
static int is_swbp_at_addr(struct mm_struct *mm, unsigned long vaddr)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
{
uprobe_opcode_t opcode;
int result;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
if (current->mm == mm) {
pagefault_disable();
result = __copy_from_user_inatomic(&opcode, (void __user*)vaddr,
sizeof(opcode));
pagefault_enable();
if (likely(result == 0))
goto out;
}
result = read_opcode(mm, vaddr, &opcode);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
if (result)
return result;
out:
if (is_swbp_insn(&opcode))
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return 1;
return 0;
}
/**
* set_swbp - store breakpoint at a given address.
* @auprobe: arch specific probepoint information.
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
* @mm: the probed process address space.
* @vaddr: the virtual address to insert the opcode.
*
* For mm @mm, store the breakpoint instruction at @vaddr.
* Return 0 (success) or a negative errno.
*/
int __weak set_swbp(struct arch_uprobe *auprobe, struct mm_struct *mm, unsigned long vaddr)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
{
int result;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
result = is_swbp_at_addr(mm, vaddr);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
if (result == 1)
return -EEXIST;
if (result)
return result;
return write_opcode(auprobe, mm, vaddr, UPROBE_SWBP_INSN);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
}
/**
* set_orig_insn - Restore the original instruction.
* @mm: the probed process address space.
* @auprobe: arch specific probepoint information.
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
* @vaddr: the virtual address to insert the opcode.
* @verify: if true, verify existance of breakpoint instruction.
*
* For mm @mm, restore the original opcode (opcode) at @vaddr.
* Return 0 (success) or a negative errno.
*/
int __weak
set_orig_insn(struct arch_uprobe *auprobe, struct mm_struct *mm, unsigned long vaddr, bool verify)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
{
if (verify) {
int result;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
result = is_swbp_at_addr(mm, vaddr);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
if (!result)
return -EINVAL;
if (result != 1)
return result;
}
return write_opcode(auprobe, mm, vaddr, *(uprobe_opcode_t *)auprobe->insn);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
}
static int match_uprobe(struct uprobe *l, struct uprobe *r)
{
if (l->inode < r->inode)
return -1;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
if (l->inode > r->inode)
return 1;
if (l->offset < r->offset)
return -1;
if (l->offset > r->offset)
return 1;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return 0;
}
static struct uprobe *__find_uprobe(struct inode *inode, loff_t offset)
{
struct uprobe u = { .inode = inode, .offset = offset };
struct rb_node *n = uprobes_tree.rb_node;
struct uprobe *uprobe;
int match;
while (n) {
uprobe = rb_entry(n, struct uprobe, rb_node);
match = match_uprobe(&u, uprobe);
if (!match) {
atomic_inc(&uprobe->ref);
return uprobe;
}
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
if (match < 0)
n = n->rb_left;
else
n = n->rb_right;
}
return NULL;
}
/*
* Find a uprobe corresponding to a given inode:offset
* Acquires uprobes_treelock
*/
static struct uprobe *find_uprobe(struct inode *inode, loff_t offset)
{
struct uprobe *uprobe;
unsigned long flags;
spin_lock_irqsave(&uprobes_treelock, flags);
uprobe = __find_uprobe(inode, offset);
spin_unlock_irqrestore(&uprobes_treelock, flags);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return uprobe;
}
static struct uprobe *__insert_uprobe(struct uprobe *uprobe)
{
struct rb_node **p = &uprobes_tree.rb_node;
struct rb_node *parent = NULL;
struct uprobe *u;
int match;
while (*p) {
parent = *p;
u = rb_entry(parent, struct uprobe, rb_node);
match = match_uprobe(uprobe, u);
if (!match) {
atomic_inc(&u->ref);
return u;
}
if (match < 0)
p = &parent->rb_left;
else
p = &parent->rb_right;
}
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
u = NULL;
rb_link_node(&uprobe->rb_node, parent, p);
rb_insert_color(&uprobe->rb_node, &uprobes_tree);
/* get access + creation ref */
atomic_set(&uprobe->ref, 2);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return u;
}
/*
* Acquire uprobes_treelock.
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
* Matching uprobe already exists in rbtree;
* increment (access refcount) and return the matching uprobe.
*
* No matching uprobe; insert the uprobe in rb_tree;
* get a double refcount (access + creation) and return NULL.
*/
static struct uprobe *insert_uprobe(struct uprobe *uprobe)
{
unsigned long flags;
struct uprobe *u;
spin_lock_irqsave(&uprobes_treelock, flags);
u = __insert_uprobe(uprobe);
spin_unlock_irqrestore(&uprobes_treelock, flags);
uprobes/core: Handle breakpoint and singlestep exceptions Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or a singlestep exception. When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters. Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups. In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64. Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe. Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost. Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete. Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing. Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc. Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too. Credits to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120313180011.29771.89027.sendpatchset@srdronam.in.ibm.com [ Performed various cleanliness edits ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-03-14 01:00:11 +07:00
/* For now assume that the instruction need not be single-stepped */
uprobe->flags |= UPROBE_SKIP_SSTEP;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return u;
}
static void put_uprobe(struct uprobe *uprobe)
{
if (atomic_dec_and_test(&uprobe->ref))
kfree(uprobe);
}
static struct uprobe *alloc_uprobe(struct inode *inode, loff_t offset)
{
struct uprobe *uprobe, *cur_uprobe;
uprobe = kzalloc(sizeof(struct uprobe), GFP_KERNEL);
if (!uprobe)
return NULL;
uprobe->inode = igrab(inode);
uprobe->offset = offset;
init_rwsem(&uprobe->consumer_rwsem);
INIT_LIST_HEAD(&uprobe->pending_list);
/* add to uprobes_tree, sorted on inode:offset */
cur_uprobe = insert_uprobe(uprobe);
/* a uprobe exists for this inode:offset combination */
if (cur_uprobe) {
kfree(uprobe);
uprobe = cur_uprobe;
iput(inode);
} else {
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
atomic_inc(&uprobe_events);
}
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return uprobe;
}
uprobes/core: Handle breakpoint and singlestep exceptions Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or a singlestep exception. When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters. Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups. In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64. Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe. Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost. Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete. Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing. Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc. Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too. Credits to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120313180011.29771.89027.sendpatchset@srdronam.in.ibm.com [ Performed various cleanliness edits ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-03-14 01:00:11 +07:00
static void handler_chain(struct uprobe *uprobe, struct pt_regs *regs)
{
struct uprobe_consumer *uc;
if (!(uprobe->flags & UPROBE_RUN_HANDLER))
return;
down_read(&uprobe->consumer_rwsem);
for (uc = uprobe->consumers; uc; uc = uc->next) {
if (!uc->filter || uc->filter(uc, current))
uc->handler(uc, regs);
}
up_read(&uprobe->consumer_rwsem);
}
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
/* Returns the previous consumer */
static struct uprobe_consumer *
consumer_add(struct uprobe *uprobe, struct uprobe_consumer *uc)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
{
down_write(&uprobe->consumer_rwsem);
uc->next = uprobe->consumers;
uprobe->consumers = uc;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
up_write(&uprobe->consumer_rwsem);
return uc->next;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
}
/*
* For uprobe @uprobe, delete the consumer @uc.
* Return true if the @uc is deleted successfully
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
* or return false.
*/
static bool consumer_del(struct uprobe *uprobe, struct uprobe_consumer *uc)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
{
struct uprobe_consumer **con;
bool ret = false;
down_write(&uprobe->consumer_rwsem);
for (con = &uprobe->consumers; *con; con = &(*con)->next) {
if (*con == uc) {
*con = uc->next;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
ret = true;
break;
}
}
up_write(&uprobe->consumer_rwsem);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return ret;
}
static int
__copy_insn(struct address_space *mapping, struct vm_area_struct *vma, char *insn,
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
unsigned long nbytes, unsigned long offset)
{
struct file *filp = vma->vm_file;
struct page *page;
void *vaddr;
unsigned long off1;
unsigned long idx;
if (!filp)
return -EINVAL;
idx = (unsigned long)(offset >> PAGE_CACHE_SHIFT);
off1 = offset &= ~PAGE_MASK;
/*
* Ensure that the page that has the original instruction is
* populated and in page-cache.
*/
page = read_mapping_page(mapping, idx, filp);
if (IS_ERR(page))
return PTR_ERR(page);
vaddr = kmap_atomic(page);
memcpy(insn, vaddr + off1, nbytes);
kunmap_atomic(vaddr);
page_cache_release(page);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return 0;
}
static int
copy_insn(struct uprobe *uprobe, struct vm_area_struct *vma, unsigned long addr)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
{
struct address_space *mapping;
unsigned long nbytes;
int bytes;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
addr &= ~PAGE_MASK;
nbytes = PAGE_SIZE - addr;
mapping = uprobe->inode->i_mapping;
/* Instruction at end of binary; copy only available bytes */
if (uprobe->offset + MAX_UINSN_BYTES > uprobe->inode->i_size)
bytes = uprobe->inode->i_size - uprobe->offset;
else
bytes = MAX_UINSN_BYTES;
/* Instruction at the page-boundary; copy bytes in second page */
if (nbytes < bytes) {
if (__copy_insn(mapping, vma, uprobe->arch.insn + nbytes,
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
bytes - nbytes, uprobe->offset + nbytes))
return -ENOMEM;
bytes = nbytes;
}
return __copy_insn(mapping, vma, uprobe->arch.insn, bytes, uprobe->offset);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
}
uprobes/core: Optimize probe hits with the help of a counter Maintain a per-mm counter: number of uprobes that are inserted on this process address space. This counter can be used at probe hit time to determine if we need a lookup in the uprobes rbtree. Everytime a probe gets inserted successfully, the probe count is incremented and everytime a probe gets removed, the probe count is decremented. The new uprobe_munmap hook ensures the count is correct on a unmap or remap of a region. We expect that once a uprobe_munmap() is called, the vma goes away. So uprobe_unregister() finding a probe to unregister would either mean unmap event hasnt occurred yet or a mmap event on the same executable file occured after a unmap event. Additionally, uprobe_mmap hook now also gets called: a. on every executable vma that is COWed at fork. b. a vma of interest is newly mapped; breakpoint insertion also happens at the required address. On process creation, make sure the probes count in the child is set correctly. Special cases that are taken care include: a. mremap b. VM_DONTCOPY vmas on fork() c. insertion/removal races in the parent during fork(). Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182646.10018.85805.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:46 +07:00
/*
* How mm->uprobes_state.count gets updated
* uprobe_mmap() increments the count if
* - it successfully adds a breakpoint.
* - it cannot add a breakpoint, but sees that there is a underlying
* breakpoint (via a is_swbp_at_addr()).
*
* uprobe_munmap() decrements the count if
* - it sees a underlying breakpoint, (via is_swbp_at_addr)
* (Subsequent uprobe_unregister wouldnt find the breakpoint
* unless a uprobe_mmap kicks in, since the old vma would be
* dropped just after uprobe_munmap.)
*
* uprobe_register increments the count if:
* - it successfully adds a breakpoint.
*
* uprobe_unregister decrements the count if:
* - it sees a underlying breakpoint and removes successfully.
* (via is_swbp_at_addr)
* (Subsequent uprobe_munmap wouldnt find the breakpoint
* since there is no underlying breakpoint after the
* breakpoint removal.)
*/
static int
install_breakpoint(struct uprobe *uprobe, struct mm_struct *mm,
struct vm_area_struct *vma, loff_t vaddr)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
{
unsigned long addr;
int ret;
/*
* If probe is being deleted, unregister thread could be done with
* the vma-rmap-walk through. Adding a probe now can be fatal since
* nobody will be able to cleanup. Also we could be from fork or
* mremap path, where the probe might have already been inserted.
* Hence behave as if probe already existed.
*/
if (!uprobe->consumers)
return -EEXIST;
addr = (unsigned long)vaddr;
if (!(uprobe->flags & UPROBE_COPY_INSN)) {
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
ret = copy_insn(uprobe, vma, addr);
if (ret)
return ret;
if (is_swbp_insn((uprobe_opcode_t *)uprobe->arch.insn))
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return -EEXIST;
ret = arch_uprobe_analyze_insn(&uprobe->arch, mm, addr);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
if (ret)
return ret;
uprobe->flags |= UPROBE_COPY_INSN;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
}
uprobes/core: Optimize probe hits with the help of a counter Maintain a per-mm counter: number of uprobes that are inserted on this process address space. This counter can be used at probe hit time to determine if we need a lookup in the uprobes rbtree. Everytime a probe gets inserted successfully, the probe count is incremented and everytime a probe gets removed, the probe count is decremented. The new uprobe_munmap hook ensures the count is correct on a unmap or remap of a region. We expect that once a uprobe_munmap() is called, the vma goes away. So uprobe_unregister() finding a probe to unregister would either mean unmap event hasnt occurred yet or a mmap event on the same executable file occured after a unmap event. Additionally, uprobe_mmap hook now also gets called: a. on every executable vma that is COWed at fork. b. a vma of interest is newly mapped; breakpoint insertion also happens at the required address. On process creation, make sure the probes count in the child is set correctly. Special cases that are taken care include: a. mremap b. VM_DONTCOPY vmas on fork() c. insertion/removal races in the parent during fork(). Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182646.10018.85805.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:46 +07:00
/*
* Ideally, should be updating the probe count after the breakpoint
* has been successfully inserted. However a thread could hit the
* breakpoint we just inserted even before the probe count is
* incremented. If this is the first breakpoint placed, breakpoint
* notifier might ignore uprobes and pass the trap to the thread.
* Hence increment before and decrement on failure.
*/
atomic_inc(&mm->uprobes_state.count);
ret = set_swbp(&uprobe->arch, mm, addr);
uprobes/core: Optimize probe hits with the help of a counter Maintain a per-mm counter: number of uprobes that are inserted on this process address space. This counter can be used at probe hit time to determine if we need a lookup in the uprobes rbtree. Everytime a probe gets inserted successfully, the probe count is incremented and everytime a probe gets removed, the probe count is decremented. The new uprobe_munmap hook ensures the count is correct on a unmap or remap of a region. We expect that once a uprobe_munmap() is called, the vma goes away. So uprobe_unregister() finding a probe to unregister would either mean unmap event hasnt occurred yet or a mmap event on the same executable file occured after a unmap event. Additionally, uprobe_mmap hook now also gets called: a. on every executable vma that is COWed at fork. b. a vma of interest is newly mapped; breakpoint insertion also happens at the required address. On process creation, make sure the probes count in the child is set correctly. Special cases that are taken care include: a. mremap b. VM_DONTCOPY vmas on fork() c. insertion/removal races in the parent during fork(). Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182646.10018.85805.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:46 +07:00
if (ret)
atomic_dec(&mm->uprobes_state.count);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return ret;
}
static void
remove_breakpoint(struct uprobe *uprobe, struct mm_struct *mm, loff_t vaddr)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
{
uprobes/core: Optimize probe hits with the help of a counter Maintain a per-mm counter: number of uprobes that are inserted on this process address space. This counter can be used at probe hit time to determine if we need a lookup in the uprobes rbtree. Everytime a probe gets inserted successfully, the probe count is incremented and everytime a probe gets removed, the probe count is decremented. The new uprobe_munmap hook ensures the count is correct on a unmap or remap of a region. We expect that once a uprobe_munmap() is called, the vma goes away. So uprobe_unregister() finding a probe to unregister would either mean unmap event hasnt occurred yet or a mmap event on the same executable file occured after a unmap event. Additionally, uprobe_mmap hook now also gets called: a. on every executable vma that is COWed at fork. b. a vma of interest is newly mapped; breakpoint insertion also happens at the required address. On process creation, make sure the probes count in the child is set correctly. Special cases that are taken care include: a. mremap b. VM_DONTCOPY vmas on fork() c. insertion/removal races in the parent during fork(). Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182646.10018.85805.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:46 +07:00
if (!set_orig_insn(&uprobe->arch, mm, (unsigned long)vaddr, true))
atomic_dec(&mm->uprobes_state.count);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
}
uprobes/core: Handle breakpoint and singlestep exceptions Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or a singlestep exception. When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters. Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups. In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64. Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe. Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost. Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete. Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing. Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc. Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too. Credits to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120313180011.29771.89027.sendpatchset@srdronam.in.ibm.com [ Performed various cleanliness edits ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-03-14 01:00:11 +07:00
/*
* There could be threads that have already hit the breakpoint. They
* will recheck the current insn and restart if find_uprobe() fails.
* See find_active_uprobe().
uprobes/core: Handle breakpoint and singlestep exceptions Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or a singlestep exception. When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters. Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups. In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64. Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe. Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost. Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete. Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing. Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc. Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too. Credits to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120313180011.29771.89027.sendpatchset@srdronam.in.ibm.com [ Performed various cleanliness edits ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-03-14 01:00:11 +07:00
*/
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
static void delete_uprobe(struct uprobe *uprobe)
{
unsigned long flags;
spin_lock_irqsave(&uprobes_treelock, flags);
rb_erase(&uprobe->rb_node, &uprobes_tree);
spin_unlock_irqrestore(&uprobes_treelock, flags);
iput(uprobe->inode);
put_uprobe(uprobe);
atomic_dec(&uprobe_events);
}
static struct vma_info *
__find_next_vma_info(struct address_space *mapping, struct list_head *head,
struct vma_info *vi, loff_t offset, bool is_register)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
{
struct prio_tree_iter iter;
struct vm_area_struct *vma;
struct vma_info *tmpvi;
unsigned long pgoff;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
int existing_vma;
loff_t vaddr;
pgoff = offset >> PAGE_SHIFT;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
vma_prio_tree_foreach(vma, &iter, &mapping->i_mmap, pgoff, pgoff) {
if (!valid_vma(vma, is_register))
continue;
existing_vma = 0;
vaddr = vma_address(vma, offset);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
list_for_each_entry(tmpvi, head, probe_list) {
if (tmpvi->mm == vma->vm_mm && tmpvi->vaddr == vaddr) {
existing_vma = 1;
break;
}
}
/*
* Another vma needs a probe to be installed. However skip
* installing the probe if the vma is about to be unlinked.
*/
if (!existing_vma && atomic_inc_not_zero(&vma->vm_mm->mm_users)) {
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
vi->mm = vma->vm_mm;
vi->vaddr = vaddr;
list_add(&vi->probe_list, head);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return vi;
}
}
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return NULL;
}
/*
* Iterate in the rmap prio tree and find a vma where a probe has not
* yet been inserted.
*/
static struct vma_info *
find_next_vma_info(struct address_space *mapping, struct list_head *head,
loff_t offset, bool is_register)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
{
struct vma_info *vi, *retvi;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
vi = kzalloc(sizeof(struct vma_info), GFP_KERNEL);
if (!vi)
return ERR_PTR(-ENOMEM);
mutex_lock(&mapping->i_mmap_mutex);
retvi = __find_next_vma_info(mapping, head, vi, offset, is_register);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
mutex_unlock(&mapping->i_mmap_mutex);
if (!retvi)
kfree(vi);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return retvi;
}
static int register_for_each_vma(struct uprobe *uprobe, bool is_register)
{
struct list_head try_list;
struct vm_area_struct *vma;
struct address_space *mapping;
struct vma_info *vi, *tmpvi;
struct mm_struct *mm;
loff_t vaddr;
int ret;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
mapping = uprobe->inode->i_mapping;
INIT_LIST_HEAD(&try_list);
ret = 0;
for (;;) {
vi = find_next_vma_info(mapping, &try_list, uprobe->offset, is_register);
if (!vi)
break;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
if (IS_ERR(vi)) {
ret = PTR_ERR(vi);
break;
}
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
mm = vi->mm;
down_write(&mm->mmap_sem);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
vma = find_vma(mm, (unsigned long)vi->vaddr);
if (!vma || !valid_vma(vma, is_register)) {
list_del(&vi->probe_list);
kfree(vi);
up_write(&mm->mmap_sem);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
mmput(mm);
continue;
}
vaddr = vma_address(vma, uprobe->offset);
if (vma->vm_file->f_mapping->host != uprobe->inode ||
vaddr != vi->vaddr) {
list_del(&vi->probe_list);
kfree(vi);
up_write(&mm->mmap_sem);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
mmput(mm);
continue;
}
if (is_register)
ret = install_breakpoint(uprobe, mm, vma, vi->vaddr);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
else
remove_breakpoint(uprobe, mm, vi->vaddr);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
up_write(&mm->mmap_sem);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
mmput(mm);
if (is_register) {
if (ret && ret == -EEXIST)
ret = 0;
if (ret)
break;
}
}
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
list_for_each_entry_safe(vi, tmpvi, &try_list, probe_list) {
list_del(&vi->probe_list);
kfree(vi);
}
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return ret;
}
static int __uprobe_register(struct uprobe *uprobe)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
{
return register_for_each_vma(uprobe, true);
}
static void __uprobe_unregister(struct uprobe *uprobe)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
{
if (!register_for_each_vma(uprobe, false))
delete_uprobe(uprobe);
/* TODO : cant unregister? schedule a worker thread */
}
/*
* uprobe_register - register a probe
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
* @inode: the file in which the probe has to be placed.
* @offset: offset from the start of the file.
* @uc: information on howto handle the probe..
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
*
* Apart from the access refcount, uprobe_register() takes a creation
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
* refcount (thro alloc_uprobe) if and only if this @uprobe is getting
* inserted into the rbtree (i.e first consumer for a @inode:@offset
* tuple). Creation refcount stops uprobe_unregister from freeing the
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
* @uprobe even before the register operation is complete. Creation
* refcount is released when the last @uc for the @uprobe
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
* unregisters.
*
* Return errno if it cannot successully install probes
* else return 0 (success)
*/
int uprobe_register(struct inode *inode, loff_t offset, struct uprobe_consumer *uc)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
{
struct uprobe *uprobe;
int ret;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
if (!inode || !uc || uc->next)
return -EINVAL;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
if (offset > i_size_read(inode))
return -EINVAL;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
ret = 0;
mutex_lock(uprobes_hash(inode));
uprobe = alloc_uprobe(inode, offset);
if (uprobe && !consumer_add(uprobe, uc)) {
ret = __uprobe_register(uprobe);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
if (ret) {
uprobe->consumers = NULL;
__uprobe_unregister(uprobe);
} else {
uprobe->flags |= UPROBE_RUN_HANDLER;
}
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
}
mutex_unlock(uprobes_hash(inode));
put_uprobe(uprobe);
return ret;
}
/*
* uprobe_unregister - unregister a already registered probe.
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
* @inode: the file in which the probe has to be removed.
* @offset: offset from the start of the file.
* @uc: identify which probe if multiple probes are colocated.
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
*/
void uprobe_unregister(struct inode *inode, loff_t offset, struct uprobe_consumer *uc)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
{
struct uprobe *uprobe;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
if (!inode || !uc)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return;
uprobe = find_uprobe(inode, offset);
if (!uprobe)
return;
mutex_lock(uprobes_hash(inode));
if (consumer_del(uprobe, uc)) {
if (!uprobe->consumers) {
__uprobe_unregister(uprobe);
uprobe->flags &= ~UPROBE_RUN_HANDLER;
}
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
}
mutex_unlock(uprobes_hash(inode));
if (uprobe)
put_uprobe(uprobe);
}
/*
* Of all the nodes that correspond to the given inode, return the node
* with the least offset.
*/
static struct rb_node *find_least_offset_node(struct inode *inode)
{
struct uprobe u = { .inode = inode, .offset = 0};
struct rb_node *n = uprobes_tree.rb_node;
struct rb_node *close_node = NULL;
struct uprobe *uprobe;
int match;
while (n) {
uprobe = rb_entry(n, struct uprobe, rb_node);
match = match_uprobe(&u, uprobe);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
if (uprobe->inode == inode)
close_node = n;
if (!match)
return close_node;
if (match < 0)
n = n->rb_left;
else
n = n->rb_right;
}
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return close_node;
}
/*
* For a given inode, build a list of probes that need to be inserted.
*/
static void build_probe_list(struct inode *inode, struct list_head *head)
{
struct uprobe *uprobe;
unsigned long flags;
struct rb_node *n;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
spin_lock_irqsave(&uprobes_treelock, flags);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
n = find_least_offset_node(inode);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
for (; n; n = rb_next(n)) {
uprobe = rb_entry(n, struct uprobe, rb_node);
if (uprobe->inode != inode)
break;
list_add(&uprobe->pending_list, head);
atomic_inc(&uprobe->ref);
}
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
spin_unlock_irqrestore(&uprobes_treelock, flags);
}
/*
* Called from mmap_region.
* called with mm->mmap_sem acquired.
*
* Return -ve no if we fail to insert probes and we cannot
* bail-out.
* Return 0 otherwise. i.e:
*
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
* - successful insertion of probes
* - (or) no possible probes to be inserted.
* - (or) insertion of probes failed but we can bail-out.
*/
int uprobe_mmap(struct vm_area_struct *vma)
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
{
struct list_head tmp_list;
struct uprobe *uprobe, *u;
struct inode *inode;
uprobes/core: Optimize probe hits with the help of a counter Maintain a per-mm counter: number of uprobes that are inserted on this process address space. This counter can be used at probe hit time to determine if we need a lookup in the uprobes rbtree. Everytime a probe gets inserted successfully, the probe count is incremented and everytime a probe gets removed, the probe count is decremented. The new uprobe_munmap hook ensures the count is correct on a unmap or remap of a region. We expect that once a uprobe_munmap() is called, the vma goes away. So uprobe_unregister() finding a probe to unregister would either mean unmap event hasnt occurred yet or a mmap event on the same executable file occured after a unmap event. Additionally, uprobe_mmap hook now also gets called: a. on every executable vma that is COWed at fork. b. a vma of interest is newly mapped; breakpoint insertion also happens at the required address. On process creation, make sure the probes count in the child is set correctly. Special cases that are taken care include: a. mremap b. VM_DONTCOPY vmas on fork() c. insertion/removal races in the parent during fork(). Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182646.10018.85805.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:46 +07:00
int ret, count;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
if (!atomic_read(&uprobe_events) || !valid_vma(vma, true))
return 0;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
inode = vma->vm_file->f_mapping->host;
if (!inode)
return 0;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
INIT_LIST_HEAD(&tmp_list);
mutex_lock(uprobes_mmap_hash(inode));
build_probe_list(inode, &tmp_list);
ret = 0;
uprobes/core: Optimize probe hits with the help of a counter Maintain a per-mm counter: number of uprobes that are inserted on this process address space. This counter can be used at probe hit time to determine if we need a lookup in the uprobes rbtree. Everytime a probe gets inserted successfully, the probe count is incremented and everytime a probe gets removed, the probe count is decremented. The new uprobe_munmap hook ensures the count is correct on a unmap or remap of a region. We expect that once a uprobe_munmap() is called, the vma goes away. So uprobe_unregister() finding a probe to unregister would either mean unmap event hasnt occurred yet or a mmap event on the same executable file occured after a unmap event. Additionally, uprobe_mmap hook now also gets called: a. on every executable vma that is COWed at fork. b. a vma of interest is newly mapped; breakpoint insertion also happens at the required address. On process creation, make sure the probes count in the child is set correctly. Special cases that are taken care include: a. mremap b. VM_DONTCOPY vmas on fork() c. insertion/removal races in the parent during fork(). Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182646.10018.85805.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:46 +07:00
count = 0;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
list_for_each_entry_safe(uprobe, u, &tmp_list, pending_list) {
loff_t vaddr;
list_del(&uprobe->pending_list);
if (!ret) {
vaddr = vma_address(vma, uprobe->offset);
uprobes/core: Optimize probe hits with the help of a counter Maintain a per-mm counter: number of uprobes that are inserted on this process address space. This counter can be used at probe hit time to determine if we need a lookup in the uprobes rbtree. Everytime a probe gets inserted successfully, the probe count is incremented and everytime a probe gets removed, the probe count is decremented. The new uprobe_munmap hook ensures the count is correct on a unmap or remap of a region. We expect that once a uprobe_munmap() is called, the vma goes away. So uprobe_unregister() finding a probe to unregister would either mean unmap event hasnt occurred yet or a mmap event on the same executable file occured after a unmap event. Additionally, uprobe_mmap hook now also gets called: a. on every executable vma that is COWed at fork. b. a vma of interest is newly mapped; breakpoint insertion also happens at the required address. On process creation, make sure the probes count in the child is set correctly. Special cases that are taken care include: a. mremap b. VM_DONTCOPY vmas on fork() c. insertion/removal races in the parent during fork(). Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182646.10018.85805.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:46 +07:00
if (vaddr < vma->vm_start || vaddr >= vma->vm_end) {
put_uprobe(uprobe);
continue;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
}
uprobes/core: Optimize probe hits with the help of a counter Maintain a per-mm counter: number of uprobes that are inserted on this process address space. This counter can be used at probe hit time to determine if we need a lookup in the uprobes rbtree. Everytime a probe gets inserted successfully, the probe count is incremented and everytime a probe gets removed, the probe count is decremented. The new uprobe_munmap hook ensures the count is correct on a unmap or remap of a region. We expect that once a uprobe_munmap() is called, the vma goes away. So uprobe_unregister() finding a probe to unregister would either mean unmap event hasnt occurred yet or a mmap event on the same executable file occured after a unmap event. Additionally, uprobe_mmap hook now also gets called: a. on every executable vma that is COWed at fork. b. a vma of interest is newly mapped; breakpoint insertion also happens at the required address. On process creation, make sure the probes count in the child is set correctly. Special cases that are taken care include: a. mremap b. VM_DONTCOPY vmas on fork() c. insertion/removal races in the parent during fork(). Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182646.10018.85805.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:46 +07:00
ret = install_breakpoint(uprobe, vma->vm_mm, vma, vaddr);
/* Ignore double add: */
if (ret == -EEXIST) {
ret = 0;
if (!is_swbp_at_addr(vma->vm_mm, vaddr))
continue;
/*
* Unable to insert a breakpoint, but
* breakpoint lies underneath. Increment the
* probe count.
*/
atomic_inc(&vma->vm_mm->uprobes_state.count);
}
if (!ret)
count++;
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
}
put_uprobe(uprobe);
}
mutex_unlock(uprobes_mmap_hash(inode));
uprobes/core: Optimize probe hits with the help of a counter Maintain a per-mm counter: number of uprobes that are inserted on this process address space. This counter can be used at probe hit time to determine if we need a lookup in the uprobes rbtree. Everytime a probe gets inserted successfully, the probe count is incremented and everytime a probe gets removed, the probe count is decremented. The new uprobe_munmap hook ensures the count is correct on a unmap or remap of a region. We expect that once a uprobe_munmap() is called, the vma goes away. So uprobe_unregister() finding a probe to unregister would either mean unmap event hasnt occurred yet or a mmap event on the same executable file occured after a unmap event. Additionally, uprobe_mmap hook now also gets called: a. on every executable vma that is COWed at fork. b. a vma of interest is newly mapped; breakpoint insertion also happens at the required address. On process creation, make sure the probes count in the child is set correctly. Special cases that are taken care include: a. mremap b. VM_DONTCOPY vmas on fork() c. insertion/removal races in the parent during fork(). Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182646.10018.85805.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:46 +07:00
if (ret)
atomic_sub(count, &vma->vm_mm->uprobes_state.count);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
return ret;
}
uprobes/core: Optimize probe hits with the help of a counter Maintain a per-mm counter: number of uprobes that are inserted on this process address space. This counter can be used at probe hit time to determine if we need a lookup in the uprobes rbtree. Everytime a probe gets inserted successfully, the probe count is incremented and everytime a probe gets removed, the probe count is decremented. The new uprobe_munmap hook ensures the count is correct on a unmap or remap of a region. We expect that once a uprobe_munmap() is called, the vma goes away. So uprobe_unregister() finding a probe to unregister would either mean unmap event hasnt occurred yet or a mmap event on the same executable file occured after a unmap event. Additionally, uprobe_mmap hook now also gets called: a. on every executable vma that is COWed at fork. b. a vma of interest is newly mapped; breakpoint insertion also happens at the required address. On process creation, make sure the probes count in the child is set correctly. Special cases that are taken care include: a. mremap b. VM_DONTCOPY vmas on fork() c. insertion/removal races in the parent during fork(). Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182646.10018.85805.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:46 +07:00
/*
* Called in context of a munmap of a vma.
*/
uprobes/core: Decrement uprobe count before the pages are unmapped Uprobes has a callback (uprobe_munmap()) in the unmap path to maintain the uprobes count. In the exit path this callback gets called in unlink_file_vma(). However by the time unlink_file_vma() is called, the pages would have been unmapped (in unmap_vmas()) and the task->rss_stat counts accounted (in zap_pte_range()). If the exiting process has probepoints, uprobe_munmap() checks if the breakpoint instruction was around before decrementing the probe count. This results in a file backed page being reread by uprobe_munmap() and hence it does not find the breakpoint. This patch fixes this problem by moving the callback to unmap_single_vma(). Since unmap_single_vma() may not unmap the complete vma, add start and end parameters to uprobe_munmap(). This bug became apparent courtesy of commit c3f0327f8e9d ("mm: add rss counters consistency check"). Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120411103527.23245.9835.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-04-11 17:35:27 +07:00
void uprobe_munmap(struct vm_area_struct *vma, unsigned long start, unsigned long end)
uprobes/core: Optimize probe hits with the help of a counter Maintain a per-mm counter: number of uprobes that are inserted on this process address space. This counter can be used at probe hit time to determine if we need a lookup in the uprobes rbtree. Everytime a probe gets inserted successfully, the probe count is incremented and everytime a probe gets removed, the probe count is decremented. The new uprobe_munmap hook ensures the count is correct on a unmap or remap of a region. We expect that once a uprobe_munmap() is called, the vma goes away. So uprobe_unregister() finding a probe to unregister would either mean unmap event hasnt occurred yet or a mmap event on the same executable file occured after a unmap event. Additionally, uprobe_mmap hook now also gets called: a. on every executable vma that is COWed at fork. b. a vma of interest is newly mapped; breakpoint insertion also happens at the required address. On process creation, make sure the probes count in the child is set correctly. Special cases that are taken care include: a. mremap b. VM_DONTCOPY vmas on fork() c. insertion/removal races in the parent during fork(). Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182646.10018.85805.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:46 +07:00
{
struct list_head tmp_list;
struct uprobe *uprobe, *u;
struct inode *inode;
if (!atomic_read(&uprobe_events) || !valid_vma(vma, false))
return;
if (!atomic_read(&vma->vm_mm->uprobes_state.count))
return;
inode = vma->vm_file->f_mapping->host;
if (!inode)
return;
INIT_LIST_HEAD(&tmp_list);
mutex_lock(uprobes_mmap_hash(inode));
build_probe_list(inode, &tmp_list);
list_for_each_entry_safe(uprobe, u, &tmp_list, pending_list) {
loff_t vaddr;
list_del(&uprobe->pending_list);
vaddr = vma_address(vma, uprobe->offset);
uprobes/core: Decrement uprobe count before the pages are unmapped Uprobes has a callback (uprobe_munmap()) in the unmap path to maintain the uprobes count. In the exit path this callback gets called in unlink_file_vma(). However by the time unlink_file_vma() is called, the pages would have been unmapped (in unmap_vmas()) and the task->rss_stat counts accounted (in zap_pte_range()). If the exiting process has probepoints, uprobe_munmap() checks if the breakpoint instruction was around before decrementing the probe count. This results in a file backed page being reread by uprobe_munmap() and hence it does not find the breakpoint. This patch fixes this problem by moving the callback to unmap_single_vma(). Since unmap_single_vma() may not unmap the complete vma, add start and end parameters to uprobe_munmap(). This bug became apparent courtesy of commit c3f0327f8e9d ("mm: add rss counters consistency check"). Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120411103527.23245.9835.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-04-11 17:35:27 +07:00
if (vaddr >= start && vaddr < end) {
uprobes/core: Optimize probe hits with the help of a counter Maintain a per-mm counter: number of uprobes that are inserted on this process address space. This counter can be used at probe hit time to determine if we need a lookup in the uprobes rbtree. Everytime a probe gets inserted successfully, the probe count is incremented and everytime a probe gets removed, the probe count is decremented. The new uprobe_munmap hook ensures the count is correct on a unmap or remap of a region. We expect that once a uprobe_munmap() is called, the vma goes away. So uprobe_unregister() finding a probe to unregister would either mean unmap event hasnt occurred yet or a mmap event on the same executable file occured after a unmap event. Additionally, uprobe_mmap hook now also gets called: a. on every executable vma that is COWed at fork. b. a vma of interest is newly mapped; breakpoint insertion also happens at the required address. On process creation, make sure the probes count in the child is set correctly. Special cases that are taken care include: a. mremap b. VM_DONTCOPY vmas on fork() c. insertion/removal races in the parent during fork(). Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182646.10018.85805.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:46 +07:00
/*
* An unregister could have removed the probe before
* unmap. So check before we decrement the count.
*/
if (is_swbp_at_addr(vma->vm_mm, vaddr) == 1)
atomic_dec(&vma->vm_mm->uprobes_state.count);
}
put_uprobe(uprobe);
}
mutex_unlock(uprobes_mmap_hash(inode));
}
uprobes/core: Allocate XOL slots for uprobes use Uprobes executes the original instruction at a probed location out of line. For this, we allocate a page (per mm) upon the first uprobe hit, in the process user address space, divide it into slots that are used to store the actual instructions to be singlestepped. These slots are known as xol (execution out of line) slots. Care is taken to ensure that the allocation is in an unmapped area as close to the top of the user address space as possible, with appropriate permission settings to keep selinux like frameworks happy. Upon a uprobe hit, a free slot is acquired, and is released after the singlestep completes. Lots of improvements courtesy suggestions/inputs from Peter and Oleg. [ Folded a fix for build issue on powerpc fixed and reported by Stephen Rothwell. ] Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182631.10018.48175.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:31 +07:00
/* Slot allocation for XOL */
static int xol_add_vma(struct xol_area *area)
{
struct mm_struct *mm;
int ret;
area->page = alloc_page(GFP_HIGHUSER);
if (!area->page)
return -ENOMEM;
ret = -EALREADY;
mm = current->mm;
down_write(&mm->mmap_sem);
if (mm->uprobes_state.xol_area)
goto fail;
ret = -ENOMEM;
/* Try to map as high as possible, this is only a hint. */
area->vaddr = get_unmapped_area(NULL, TASK_SIZE - PAGE_SIZE, PAGE_SIZE, 0, 0);
if (area->vaddr & ~PAGE_MASK) {
ret = area->vaddr;
goto fail;
}
ret = install_special_mapping(mm, area->vaddr, PAGE_SIZE,
VM_EXEC|VM_MAYEXEC|VM_DONTCOPY|VM_IO, &area->page);
if (ret)
goto fail;
smp_wmb(); /* pairs with get_xol_area() */
mm->uprobes_state.xol_area = area;
ret = 0;
fail:
up_write(&mm->mmap_sem);
if (ret)
__free_page(area->page);
return ret;
}
static struct xol_area *get_xol_area(struct mm_struct *mm)
{
struct xol_area *area;
area = mm->uprobes_state.xol_area;
smp_read_barrier_depends(); /* pairs with wmb in xol_add_vma() */
return area;
}
/*
* xol_alloc_area - Allocate process's xol_area.
* This area will be used for storing instructions for execution out of
* line.
*
* Returns the allocated area or NULL.
*/
static struct xol_area *xol_alloc_area(void)
{
struct xol_area *area;
area = kzalloc(sizeof(*area), GFP_KERNEL);
if (unlikely(!area))
return NULL;
area->bitmap = kzalloc(BITS_TO_LONGS(UINSNS_PER_PAGE) * sizeof(long), GFP_KERNEL);
if (!area->bitmap)
goto fail;
init_waitqueue_head(&area->wq);
if (!xol_add_vma(area))
return area;
fail:
kfree(area->bitmap);
kfree(area);
return get_xol_area(current->mm);
}
/*
* uprobe_clear_state - Free the area allocated for slots.
*/
void uprobe_clear_state(struct mm_struct *mm)
{
struct xol_area *area = mm->uprobes_state.xol_area;
if (!area)
return;
put_page(area->page);
kfree(area->bitmap);
kfree(area);
}
/*
* uprobe_reset_state - Free the area allocated for slots.
*/
void uprobe_reset_state(struct mm_struct *mm)
{
mm->uprobes_state.xol_area = NULL;
uprobes/core: Optimize probe hits with the help of a counter Maintain a per-mm counter: number of uprobes that are inserted on this process address space. This counter can be used at probe hit time to determine if we need a lookup in the uprobes rbtree. Everytime a probe gets inserted successfully, the probe count is incremented and everytime a probe gets removed, the probe count is decremented. The new uprobe_munmap hook ensures the count is correct on a unmap or remap of a region. We expect that once a uprobe_munmap() is called, the vma goes away. So uprobe_unregister() finding a probe to unregister would either mean unmap event hasnt occurred yet or a mmap event on the same executable file occured after a unmap event. Additionally, uprobe_mmap hook now also gets called: a. on every executable vma that is COWed at fork. b. a vma of interest is newly mapped; breakpoint insertion also happens at the required address. On process creation, make sure the probes count in the child is set correctly. Special cases that are taken care include: a. mremap b. VM_DONTCOPY vmas on fork() c. insertion/removal races in the parent during fork(). Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182646.10018.85805.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:46 +07:00
atomic_set(&mm->uprobes_state.count, 0);
uprobes/core: Allocate XOL slots for uprobes use Uprobes executes the original instruction at a probed location out of line. For this, we allocate a page (per mm) upon the first uprobe hit, in the process user address space, divide it into slots that are used to store the actual instructions to be singlestepped. These slots are known as xol (execution out of line) slots. Care is taken to ensure that the allocation is in an unmapped area as close to the top of the user address space as possible, with appropriate permission settings to keep selinux like frameworks happy. Upon a uprobe hit, a free slot is acquired, and is released after the singlestep completes. Lots of improvements courtesy suggestions/inputs from Peter and Oleg. [ Folded a fix for build issue on powerpc fixed and reported by Stephen Rothwell. ] Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182631.10018.48175.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:31 +07:00
}
/*
* - search for a free slot.
*/
static unsigned long xol_take_insn_slot(struct xol_area *area)
{
unsigned long slot_addr;
int slot_nr;
do {
slot_nr = find_first_zero_bit(area->bitmap, UINSNS_PER_PAGE);
if (slot_nr < UINSNS_PER_PAGE) {
if (!test_and_set_bit(slot_nr, area->bitmap))
break;
slot_nr = UINSNS_PER_PAGE;
continue;
}
wait_event(area->wq, (atomic_read(&area->slot_count) < UINSNS_PER_PAGE));
} while (slot_nr >= UINSNS_PER_PAGE);
slot_addr = area->vaddr + (slot_nr * UPROBE_XOL_SLOT_BYTES);
atomic_inc(&area->slot_count);
return slot_addr;
}
/*
* xol_get_insn_slot - If was not allocated a slot, then
* allocate a slot.
* Returns the allocated slot address or 0.
*/
static unsigned long xol_get_insn_slot(struct uprobe *uprobe, unsigned long slot_addr)
{
struct xol_area *area;
unsigned long offset;
void *vaddr;
area = get_xol_area(current->mm);
if (!area) {
area = xol_alloc_area();
if (!area)
return 0;
}
current->utask->xol_vaddr = xol_take_insn_slot(area);
/*
* Initialize the slot if xol_vaddr points to valid
* instruction slot.
*/
if (unlikely(!current->utask->xol_vaddr))
return 0;
current->utask->vaddr = slot_addr;
offset = current->utask->xol_vaddr & ~PAGE_MASK;
vaddr = kmap_atomic(area->page);
memcpy(vaddr + offset, uprobe->arch.insn, MAX_UINSN_BYTES);
kunmap_atomic(vaddr);
return current->utask->xol_vaddr;
}
/*
* xol_free_insn_slot - If slot was earlier allocated by
* @xol_get_insn_slot(), make the slot available for
* subsequent requests.
*/
static void xol_free_insn_slot(struct task_struct *tsk)
{
struct xol_area *area;
unsigned long vma_end;
unsigned long slot_addr;
if (!tsk->mm || !tsk->mm->uprobes_state.xol_area || !tsk->utask)
return;
slot_addr = tsk->utask->xol_vaddr;
if (unlikely(!slot_addr || IS_ERR_VALUE(slot_addr)))
return;
area = tsk->mm->uprobes_state.xol_area;
vma_end = area->vaddr + PAGE_SIZE;
if (area->vaddr <= slot_addr && slot_addr < vma_end) {
unsigned long offset;
int slot_nr;
offset = slot_addr - area->vaddr;
slot_nr = offset / UPROBE_XOL_SLOT_BYTES;
if (slot_nr >= UINSNS_PER_PAGE)
return;
clear_bit(slot_nr, area->bitmap);
atomic_dec(&area->slot_count);
if (waitqueue_active(&area->wq))
wake_up(&area->wq);
tsk->utask->xol_vaddr = 0;
}
}
uprobes/core: Handle breakpoint and singlestep exceptions Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or a singlestep exception. When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters. Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups. In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64. Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe. Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost. Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete. Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing. Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc. Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too. Credits to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120313180011.29771.89027.sendpatchset@srdronam.in.ibm.com [ Performed various cleanliness edits ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-03-14 01:00:11 +07:00
/**
* uprobe_get_swbp_addr - compute address of swbp given post-swbp regs
* @regs: Reflects the saved state of the task after it has hit a breakpoint
* instruction.
* Return the address of the breakpoint instruction.
*/
unsigned long __weak uprobe_get_swbp_addr(struct pt_regs *regs)
{
return instruction_pointer(regs) - UPROBE_SWBP_INSN_SIZE;
}
/*
* Called with no locks held.
* Called in context of a exiting or a exec-ing thread.
*/
void uprobe_free_utask(struct task_struct *t)
{
struct uprobe_task *utask = t->utask;
if (!utask)
return;
if (utask->active_uprobe)
put_uprobe(utask->active_uprobe);
uprobes/core: Allocate XOL slots for uprobes use Uprobes executes the original instruction at a probed location out of line. For this, we allocate a page (per mm) upon the first uprobe hit, in the process user address space, divide it into slots that are used to store the actual instructions to be singlestepped. These slots are known as xol (execution out of line) slots. Care is taken to ensure that the allocation is in an unmapped area as close to the top of the user address space as possible, with appropriate permission settings to keep selinux like frameworks happy. Upon a uprobe hit, a free slot is acquired, and is released after the singlestep completes. Lots of improvements courtesy suggestions/inputs from Peter and Oleg. [ Folded a fix for build issue on powerpc fixed and reported by Stephen Rothwell. ] Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182631.10018.48175.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:31 +07:00
xol_free_insn_slot(t);
uprobes/core: Handle breakpoint and singlestep exceptions Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or a singlestep exception. When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters. Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups. In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64. Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe. Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost. Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete. Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing. Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc. Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too. Credits to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120313180011.29771.89027.sendpatchset@srdronam.in.ibm.com [ Performed various cleanliness edits ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-03-14 01:00:11 +07:00
kfree(utask);
t->utask = NULL;
}
/*
* Called in context of a new clone/fork from copy_process.
*/
void uprobe_copy_process(struct task_struct *t)
{
t->utask = NULL;
}
/*
* Allocate a uprobe_task object for the task.
* Called when the thread hits a breakpoint for the first time.
*
* Returns:
* - pointer to new uprobe_task on success
* - NULL otherwise
*/
static struct uprobe_task *add_utask(void)
{
struct uprobe_task *utask;
utask = kzalloc(sizeof *utask, GFP_KERNEL);
if (unlikely(!utask))
return NULL;
utask->active_uprobe = NULL;
current->utask = utask;
return utask;
}
/* Prepare to single-step probed instruction out of line. */
static int
pre_ssout(struct uprobe *uprobe, struct pt_regs *regs, unsigned long vaddr)
{
uprobes/core: Allocate XOL slots for uprobes use Uprobes executes the original instruction at a probed location out of line. For this, we allocate a page (per mm) upon the first uprobe hit, in the process user address space, divide it into slots that are used to store the actual instructions to be singlestepped. These slots are known as xol (execution out of line) slots. Care is taken to ensure that the allocation is in an unmapped area as close to the top of the user address space as possible, with appropriate permission settings to keep selinux like frameworks happy. Upon a uprobe hit, a free slot is acquired, and is released after the singlestep completes. Lots of improvements courtesy suggestions/inputs from Peter and Oleg. [ Folded a fix for build issue on powerpc fixed and reported by Stephen Rothwell. ] Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182631.10018.48175.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:31 +07:00
if (xol_get_insn_slot(uprobe, vaddr) && !arch_uprobe_pre_xol(&uprobe->arch, regs))
return 0;
uprobes/core: Handle breakpoint and singlestep exceptions Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or a singlestep exception. When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters. Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups. In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64. Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe. Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost. Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete. Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing. Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc. Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too. Credits to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120313180011.29771.89027.sendpatchset@srdronam.in.ibm.com [ Performed various cleanliness edits ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-03-14 01:00:11 +07:00
return -EFAULT;
}
/*
* If we are singlestepping, then ensure this thread is not connected to
* non-fatal signals until completion of singlestep. When xol insn itself
* triggers the signal, restart the original insn even if the task is
* already SIGKILL'ed (since coredump should report the correct ip). This
* is even more important if the task has a handler for SIGSEGV/etc, The
* _same_ instruction should be repeated again after return from the signal
* handler, and SSTEP can never finish in this case.
*/
bool uprobe_deny_signal(void)
{
struct task_struct *t = current;
struct uprobe_task *utask = t->utask;
if (likely(!utask || !utask->active_uprobe))
return false;
WARN_ON_ONCE(utask->state != UTASK_SSTEP);
if (signal_pending(t)) {
spin_lock_irq(&t->sighand->siglock);
clear_tsk_thread_flag(t, TIF_SIGPENDING);
spin_unlock_irq(&t->sighand->siglock);
if (__fatal_signal_pending(t) || arch_uprobe_xol_was_trapped(t)) {
utask->state = UTASK_SSTEP_TRAPPED;
set_tsk_thread_flag(t, TIF_UPROBE);
set_tsk_thread_flag(t, TIF_NOTIFY_RESUME);
}
}
return true;
}
/*
* Avoid singlestepping the original instruction if the original instruction
* is a NOP or can be emulated.
*/
static bool can_skip_sstep(struct uprobe *uprobe, struct pt_regs *regs)
{
if (arch_uprobe_skip_sstep(&uprobe->arch, regs))
return true;
uprobe->flags &= ~UPROBE_SKIP_SSTEP;
return false;
}
uprobes: Teach find_active_uprobe() to provide the "is_swbp" info A separate patch to simplify the review, and for the documentation. The patch adds another "int *is_swbp" argument to find_active_uprobe(), so far its only caller doesn't use this info. With this patch find_active_uprobe() additionally does: - if find_vma() + ->vm_start check fails, *is_swbp = -EFAULT - otherwise, if valid_vma() + find_uprobe() fails, it holds the result of is_swbp_at_addr(), can be negative too. The latter is only possible if we raced with another thread which did munmap/etc after we hit this bp. IOW. If find_active_uprobe(&is_swbp) returns NULL, the caller can look at is_swbp to figure out whether the current insn is bp or not, or detect the race with another thread if it is negative. Note: I think that performance-wise this change is fine. This adds is_swbp_at_addr(), but only if we raced with uprobe_unregister() or if we hit the "normal" int3 but this mm has uprobes as well. And even in this case the slow read_opcode() path is very unlikely, this insn recently triggered do_int3(), __copy_from_user_inatomic() shouldn't fail in the likely case. Signed-off-by: Oleg Nesterov <oleg@redhat.com> Acked-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Anton Arapov <anton@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120529192914.GD8057@redhat.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-05-30 02:29:14 +07:00
static struct uprobe *find_active_uprobe(unsigned long bp_vaddr, int *is_swbp)
uprobes/core: Handle breakpoint and singlestep exceptions Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or a singlestep exception. When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters. Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups. In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64. Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe. Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost. Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete. Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing. Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc. Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too. Credits to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120313180011.29771.89027.sendpatchset@srdronam.in.ibm.com [ Performed various cleanliness edits ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-03-14 01:00:11 +07:00
{
struct mm_struct *mm = current->mm;
struct uprobe *uprobe = NULL;
uprobes/core: Handle breakpoint and singlestep exceptions Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or a singlestep exception. When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters. Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups. In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64. Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe. Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost. Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete. Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing. Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc. Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too. Credits to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120313180011.29771.89027.sendpatchset@srdronam.in.ibm.com [ Performed various cleanliness edits ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-03-14 01:00:11 +07:00
struct vm_area_struct *vma;
down_read(&mm->mmap_sem);
vma = find_vma(mm, bp_vaddr);
if (vma && vma->vm_start <= bp_vaddr) {
if (valid_vma(vma, false)) {
struct inode *inode;
loff_t offset;
uprobes/core: Handle breakpoint and singlestep exceptions Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or a singlestep exception. When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters. Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups. In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64. Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe. Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost. Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete. Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing. Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc. Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too. Credits to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120313180011.29771.89027.sendpatchset@srdronam.in.ibm.com [ Performed various cleanliness edits ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-03-14 01:00:11 +07:00
inode = vma->vm_file->f_mapping->host;
offset = bp_vaddr - vma->vm_start;
offset += (vma->vm_pgoff << PAGE_SHIFT);
uprobe = find_uprobe(inode, offset);
}
uprobes: Teach find_active_uprobe() to provide the "is_swbp" info A separate patch to simplify the review, and for the documentation. The patch adds another "int *is_swbp" argument to find_active_uprobe(), so far its only caller doesn't use this info. With this patch find_active_uprobe() additionally does: - if find_vma() + ->vm_start check fails, *is_swbp = -EFAULT - otherwise, if valid_vma() + find_uprobe() fails, it holds the result of is_swbp_at_addr(), can be negative too. The latter is only possible if we raced with another thread which did munmap/etc after we hit this bp. IOW. If find_active_uprobe(&is_swbp) returns NULL, the caller can look at is_swbp to figure out whether the current insn is bp or not, or detect the race with another thread if it is negative. Note: I think that performance-wise this change is fine. This adds is_swbp_at_addr(), but only if we raced with uprobe_unregister() or if we hit the "normal" int3 but this mm has uprobes as well. And even in this case the slow read_opcode() path is very unlikely, this insn recently triggered do_int3(), __copy_from_user_inatomic() shouldn't fail in the likely case. Signed-off-by: Oleg Nesterov <oleg@redhat.com> Acked-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Anton Arapov <anton@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120529192914.GD8057@redhat.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-05-30 02:29:14 +07:00
if (!uprobe)
*is_swbp = is_swbp_at_addr(mm, bp_vaddr);
} else {
*is_swbp = -EFAULT;
uprobes/core: Handle breakpoint and singlestep exceptions Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or a singlestep exception. When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters. Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups. In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64. Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe. Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost. Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete. Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing. Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc. Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too. Credits to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120313180011.29771.89027.sendpatchset@srdronam.in.ibm.com [ Performed various cleanliness edits ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-03-14 01:00:11 +07:00
}
up_read(&mm->mmap_sem);
return uprobe;
}
/*
* Run handler and ask thread to singlestep.
* Ensure all non-fatal signals cannot interrupt thread while it singlesteps.
*/
static void handle_swbp(struct pt_regs *regs)
{
struct uprobe_task *utask;
struct uprobe *uprobe;
unsigned long bp_vaddr;
int uninitialized_var(is_swbp);
bp_vaddr = uprobe_get_swbp_addr(regs);
uprobes: Teach find_active_uprobe() to provide the "is_swbp" info A separate patch to simplify the review, and for the documentation. The patch adds another "int *is_swbp" argument to find_active_uprobe(), so far its only caller doesn't use this info. With this patch find_active_uprobe() additionally does: - if find_vma() + ->vm_start check fails, *is_swbp = -EFAULT - otherwise, if valid_vma() + find_uprobe() fails, it holds the result of is_swbp_at_addr(), can be negative too. The latter is only possible if we raced with another thread which did munmap/etc after we hit this bp. IOW. If find_active_uprobe(&is_swbp) returns NULL, the caller can look at is_swbp to figure out whether the current insn is bp or not, or detect the race with another thread if it is negative. Note: I think that performance-wise this change is fine. This adds is_swbp_at_addr(), but only if we raced with uprobe_unregister() or if we hit the "normal" int3 but this mm has uprobes as well. And even in this case the slow read_opcode() path is very unlikely, this insn recently triggered do_int3(), __copy_from_user_inatomic() shouldn't fail in the likely case. Signed-off-by: Oleg Nesterov <oleg@redhat.com> Acked-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Anton Arapov <anton@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120529192914.GD8057@redhat.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-05-30 02:29:14 +07:00
uprobe = find_active_uprobe(bp_vaddr, &is_swbp);
uprobes/core: Handle breakpoint and singlestep exceptions Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or a singlestep exception. When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters. Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups. In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64. Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe. Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost. Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete. Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing. Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc. Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too. Credits to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120313180011.29771.89027.sendpatchset@srdronam.in.ibm.com [ Performed various cleanliness edits ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-03-14 01:00:11 +07:00
if (!uprobe) {
if (is_swbp > 0) {
/* No matching uprobe; signal SIGTRAP. */
send_sig(SIGTRAP, current, 0);
} else {
/*
* Either we raced with uprobe_unregister() or we can't
* access this memory. The latter is only possible if
* another thread plays with our ->mm. In both cases
* we can simply restart. If this vma was unmapped we
* can pretend this insn was not executed yet and get
* the (correct) SIGSEGV after restart.
*/
instruction_pointer_set(regs, bp_vaddr);
}
uprobes/core: Handle breakpoint and singlestep exceptions Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or a singlestep exception. When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters. Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups. In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64. Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe. Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost. Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete. Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing. Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc. Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too. Credits to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120313180011.29771.89027.sendpatchset@srdronam.in.ibm.com [ Performed various cleanliness edits ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-03-14 01:00:11 +07:00
return;
}
utask = current->utask;
if (!utask) {
utask = add_utask();
/* Cannot allocate; re-execute the instruction. */
if (!utask)
goto cleanup_ret;
}
utask->active_uprobe = uprobe;
handler_chain(uprobe, regs);
if (uprobe->flags & UPROBE_SKIP_SSTEP && can_skip_sstep(uprobe, regs))
goto cleanup_ret;
utask->state = UTASK_SSTEP;
if (!pre_ssout(uprobe, regs, bp_vaddr)) {
user_enable_single_step(current);
return;
}
cleanup_ret:
if (utask) {
utask->active_uprobe = NULL;
utask->state = UTASK_RUNNING;
}
if (uprobe) {
if (!(uprobe->flags & UPROBE_SKIP_SSTEP))
/*
* cannot singlestep; cannot skip instruction;
* re-execute the instruction.
*/
instruction_pointer_set(regs, bp_vaddr);
put_uprobe(uprobe);
}
}
/*
* Perform required fix-ups and disable singlestep.
* Allow pending signals to take effect.
*/
static void handle_singlestep(struct uprobe_task *utask, struct pt_regs *regs)
{
struct uprobe *uprobe;
uprobe = utask->active_uprobe;
if (utask->state == UTASK_SSTEP_ACK)
arch_uprobe_post_xol(&uprobe->arch, regs);
else if (utask->state == UTASK_SSTEP_TRAPPED)
arch_uprobe_abort_xol(&uprobe->arch, regs);
else
WARN_ON_ONCE(1);
put_uprobe(uprobe);
utask->active_uprobe = NULL;
utask->state = UTASK_RUNNING;
user_disable_single_step(current);
uprobes/core: Allocate XOL slots for uprobes use Uprobes executes the original instruction at a probed location out of line. For this, we allocate a page (per mm) upon the first uprobe hit, in the process user address space, divide it into slots that are used to store the actual instructions to be singlestepped. These slots are known as xol (execution out of line) slots. Care is taken to ensure that the allocation is in an unmapped area as close to the top of the user address space as possible, with appropriate permission settings to keep selinux like frameworks happy. Upon a uprobe hit, a free slot is acquired, and is released after the singlestep completes. Lots of improvements courtesy suggestions/inputs from Peter and Oleg. [ Folded a fix for build issue on powerpc fixed and reported by Stephen Rothwell. ] Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182631.10018.48175.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:31 +07:00
xol_free_insn_slot(current);
uprobes/core: Handle breakpoint and singlestep exceptions Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or a singlestep exception. When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters. Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups. In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64. Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe. Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost. Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete. Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing. Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc. Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too. Credits to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120313180011.29771.89027.sendpatchset@srdronam.in.ibm.com [ Performed various cleanliness edits ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-03-14 01:00:11 +07:00
spin_lock_irq(&current->sighand->siglock);
recalc_sigpending(); /* see uprobe_deny_signal() */
spin_unlock_irq(&current->sighand->siglock);
}
/*
* On breakpoint hit, breakpoint notifier sets the TIF_UPROBE flag. (and on
* subsequent probe hits on the thread sets the state to UTASK_BP_HIT) and
* allows the thread to return from interrupt.
*
* On singlestep exception, singlestep notifier sets the TIF_UPROBE flag and
* also sets the state to UTASK_SSTEP_ACK and allows the thread to return from
* interrupt.
*
* While returning to userspace, thread notices the TIF_UPROBE flag and calls
* uprobe_notify_resume().
*/
void uprobe_notify_resume(struct pt_regs *regs)
{
struct uprobe_task *utask;
utask = current->utask;
if (!utask || utask->state == UTASK_BP_HIT)
handle_swbp(regs);
else
handle_singlestep(utask, regs);
}
/*
* uprobe_pre_sstep_notifier gets called from interrupt context as part of
* notifier mechanism. Set TIF_UPROBE flag and indicate breakpoint hit.
*/
int uprobe_pre_sstep_notifier(struct pt_regs *regs)
{
struct uprobe_task *utask;
uprobes/core: Optimize probe hits with the help of a counter Maintain a per-mm counter: number of uprobes that are inserted on this process address space. This counter can be used at probe hit time to determine if we need a lookup in the uprobes rbtree. Everytime a probe gets inserted successfully, the probe count is incremented and everytime a probe gets removed, the probe count is decremented. The new uprobe_munmap hook ensures the count is correct on a unmap or remap of a region. We expect that once a uprobe_munmap() is called, the vma goes away. So uprobe_unregister() finding a probe to unregister would either mean unmap event hasnt occurred yet or a mmap event on the same executable file occured after a unmap event. Additionally, uprobe_mmap hook now also gets called: a. on every executable vma that is COWed at fork. b. a vma of interest is newly mapped; breakpoint insertion also happens at the required address. On process creation, make sure the probes count in the child is set correctly. Special cases that are taken care include: a. mremap b. VM_DONTCOPY vmas on fork() c. insertion/removal races in the parent during fork(). Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Anton Arapov <anton@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120330182646.10018.85805.sendpatchset@srdronam.in.ibm.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-03-31 01:26:46 +07:00
if (!current->mm || !atomic_read(&current->mm->uprobes_state.count))
/* task is currently not uprobed */
uprobes/core: Handle breakpoint and singlestep exceptions Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or a singlestep exception. When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters. Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups. In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64. Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe. Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost. Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete. Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing. Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc. Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too. Credits to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120313180011.29771.89027.sendpatchset@srdronam.in.ibm.com [ Performed various cleanliness edits ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-03-14 01:00:11 +07:00
return 0;
utask = current->utask;
if (utask)
utask->state = UTASK_BP_HIT;
set_thread_flag(TIF_UPROBE);
return 1;
}
/*
* uprobe_post_sstep_notifier gets called in interrupt context as part of notifier
* mechanism. Set TIF_UPROBE flag and indicate completion of singlestep.
*/
int uprobe_post_sstep_notifier(struct pt_regs *regs)
{
struct uprobe_task *utask = current->utask;
if (!current->mm || !utask || !utask->active_uprobe)
/* task is currently not uprobed */
return 0;
utask->state = UTASK_SSTEP_ACK;
set_thread_flag(TIF_UPROBE);
return 1;
}
static struct notifier_block uprobe_exception_nb = {
.notifier_call = arch_uprobe_exception_notify,
.priority = INT_MAX-1, /* notified after kprobes, kgdb */
};
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
static int __init init_uprobes(void)
{
int i;
for (i = 0; i < UPROBES_HASH_SZ; i++) {
mutex_init(&uprobes_mutex[i]);
mutex_init(&uprobes_mmap_mutex[i]);
}
uprobes/core: Handle breakpoint and singlestep exceptions Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or a singlestep exception. When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters. Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups. In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64. Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe. Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost. Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete. Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing. Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc. Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too. Credits to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120313180011.29771.89027.sendpatchset@srdronam.in.ibm.com [ Performed various cleanliness edits ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-03-14 01:00:11 +07:00
return register_die_notifier(&uprobe_exception_nb);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
}
uprobes/core: Handle breakpoint and singlestep exceptions Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or a singlestep exception. When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters. Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups. In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64. Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe. Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost. Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete. Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing. Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc. Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too. Credits to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com> Cc: Linux-mm <linux-mm@kvack.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120313180011.29771.89027.sendpatchset@srdronam.in.ibm.com [ Performed various cleanliness edits ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-03-14 01:00:11 +07:00
module_init(init_uprobes);
uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints Add uprobes support to the core kernel, with x86 support. This commit adds the kernel facilities, the actual uprobes user-space ABI and perf probe support comes in later commits. General design: Uprobes are maintained in an rb-tree indexed by inode and offset (the offset here is from the start of the mapping). For a unique (inode, offset) tuple, there can be at most one uprobe in the rb-tree. Since the (inode, offset) tuple identifies a unique uprobe, more than one user may be interested in the same uprobe. This provides the ability to connect multiple 'consumers' to the same uprobe. Each consumer defines a handler and a filter (optional). The 'handler' is run every time the uprobe is hit, if it matches the 'filter' criteria. The first consumer of a uprobe causes the breakpoint to be inserted at the specified address and subsequent consumers are appended to this list. On subsequent probes, the consumer gets appended to the existing list of consumers. The breakpoint is removed when the last consumer unregisters. For all other unregisterations, the consumer is removed from the list of consumers. Given a inode, we get a list of the mms that have mapped the inode. Do the actual registration if mm maps the page where a probe needs to be inserted/removed. We use a temporary list to walk through the vmas that map the inode. - The number of maps that map the inode, is not known before we walk the rmap and keeps changing. - extending vm_area_struct wasn't recommended, it's a size-critical data structure. - There can be more than one maps of the inode in the same mm. We add callbacks to the mmap methods to keep an eye on text vmas that are of interest to uprobes. When a vma of interest is mapped, we insert the breakpoint at the right address. Uprobe works by replacing the instruction at the address defined by (inode, offset) with the arch specific breakpoint instruction. We save a copy of the original instruction at the uprobed address. This is needed for: a. executing the instruction out-of-line (xol). b. instruction analysis for any subsequent fixups. c. restoring the instruction back when the uprobe is unregistered. We insert or delete a breakpoint instruction, and this breakpoint instruction is assumed to be the smallest instruction available on the platform. For fixed size instruction platforms this is trivially true, for variable size instruction platforms the breakpoint instruction is typically the smallest (often a single byte). Writing the instruction is done by COWing the page and changing the instruction during the copy, this even though most platforms allow atomic writes of the breakpoint instruction. This also mirrors the behaviour of a ptrace() memory write to a PRIVATE file map. The core worker is derived from KSM's replace_page() logic. In essence, similar to KSM: a. allocate a new page and copy over contents of the page that has the uprobed vaddr b. modify the copy and insert the breakpoint at the required address c. switch the original page with the copy containing the breakpoint d. flush page tables. replace_page() is being replicated here because of some minor changes in the type of pages and also because Hugh Dickins had plans to improve replace_page() for KSM specific work. Instruction analysis on x86 is based on instruction decoder and determines if an instruction can be probed and determines the necessary fixups after singlestep. Instruction analysis is done at probe insertion time so that we avoid having to repeat the same analysis every time a probe is hit. A lot of code here is due to the improvement/suggestions/inputs from Peter Zijlstra. Changelog: (v10): - Add code to clear REX.B prefix as suggested by Denys Vlasenko and Masami Hiramatsu. (v9): - Use insn_offset_modrm as suggested by Masami Hiramatsu. (v7): Handle comments from Peter Zijlstra: - Dont take reference to inode. (expect inode to uprobe_register to be sane). - Use PTR_ERR to set the return value. - No need to take reference to inode. - use PTR_ERR to return error value. - register and uprobe_unregister share code. (v5): - Modified del_consumer as per comments from Peter. - Drop reference to inode before dropping reference to uprobe. - Use i_size_read(inode) instead of inode->i_size. - Ensure uprobe->consumers is NULL, before __uprobe_unregister() is called. - Includes errno.h as recommended by Stephen Rothwell to fix a build issue on sparc defconfig - Remove restrictions while unregistering. - Earlier code leaked inode references under some conditions while registering/unregistering. - Continue the vma-rmap walk even if the intermediate vma doesnt meet the requirements. - Validate the vma found by find_vma before inserting/removing the breakpoint - Call del_consumer under mutex_lock. - Use hash locks. - Handle mremap. - Introduce find_least_offset_node() instead of close match logic in find_uprobe - Uprobes no more depends on MM_OWNER; No reference to task_structs while inserting/removing a probe. - Uses read_mapping_page instead of grab_cache_page so that the pages have valid content. - pass NULL to get_user_pages for the task parameter. - call SetPageUptodate on the new page allocated in write_opcode. - fix leaking a reference to the new page under certain conditions. - Include Instruction Decoder if Uprobes gets defined. - Remove const attributes for instruction prefix arrays. - Uses mm_context to know if the application is 32 bit. Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Also-written-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Christoph Hellwig <hch@infradead.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Roland McGrath <roland@hack.frob.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Anton Arapov <anton@redhat.com> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Denys Vlasenko <vda.linux@googlemail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linux-mm <linux-mm@kvack.org> Link: http://lkml.kernel.org/r/20120209092642.GE16600@linux.vnet.ibm.com [ Made various small edits to the commit log ] Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-02-09 16:26:42 +07:00
static void __exit exit_uprobes(void)
{
}
module_exit(exit_uprobes);