AuxXxilium/linux_dsm_epyc7002
1
0
Fork 0
mirror of https://github.com/AuxXxilium/linux_dsm_epyc7002.git synced 2024-12-26 08:15:03 +07:00
Code Issues Actions Packages Projects Releases Wiki Activity
7ff3a7621d
linux_dsm_epyc7002/arch/s390/kernel/entry.S

1563 lines
41 KiB
ArmAsm
Raw Normal View History

License cleanup: add SPDX GPL-2.0 license identifier to files with no license Many source files in the tree are missing licensing information, which makes it harder for compliance tools to determine the correct license. By default all files without license information are under the default license of the kernel, which is GPL version 2. Update the files which contain no license information with the 'GPL-2.0' SPDX license identifier. The SPDX identifier is a legally binding shorthand, which can be used instead of the full boiler plate text. This patch is based on work done by Thomas Gleixner and Kate Stewart and Philippe Ombredanne. How this work was done: Patches were generated and checked against linux-4.14-rc6 for a subset of the use cases: - file had no licensing information it it. - file was a */uapi/* one with no licensing information in it, - file was a */uapi/* one with existing licensing information, Further patches will be generated in subsequent months to fix up cases where non-standard license headers were used, and references to license had to be inferred by heuristics based on keywords. The analysis to determine which SPDX License Identifier to be applied to a file was done in a spreadsheet of side by side results from of the output of two independent scanners (ScanCode & Windriver) producing SPDX tag:value files created by Philippe Ombredanne. Philippe prepared the base worksheet, and did an initial spot review of a few 1000 files. The 4.13 kernel was the starting point of the analysis with 60,537 files assessed. Kate Stewart did a file by file comparison of the scanner results in the spreadsheet to determine which SPDX license identifier(s) to be applied to the file. She confirmed any determination that was not immediately clear with lawyers working with the Linux Foundation. Criteria used to select files for SPDX license identifier tagging was: - Files considered eligible had to be source code files. - Make and config files were included as candidates if they contained >5 lines of source - File already had some variant of a license header in it (even if <5 lines). All documentation files were explicitly excluded. The following heuristics were used to determine which SPDX license identifiers to apply. - when both scanners couldn't find any license traces, file was considered to have no license information in it, and the top level COPYING file license applied. For non */uapi/* files that summary was: SPDX license identifier # files ---------------------------------------------------|------- GPL-2.0 11139 and resulted in the first patch in this series. If that file was a */uapi/* path one, it was "GPL-2.0 WITH Linux-syscall-note" otherwise it was "GPL-2.0". Results of that was: SPDX license identifier # files ---------------------------------------------------|------- GPL-2.0 WITH Linux-syscall-note 930 and resulted in the second patch in this series. - if a file had some form of licensing information in it, and was one of the */uapi/* ones, it was denoted with the Linux-syscall-note if any GPL family license was found in the file or had no licensing in it (per prior point). Results summary: SPDX license identifier # files ---------------------------------------------------|------ GPL-2.0 WITH Linux-syscall-note 270 GPL-2.0+ WITH Linux-syscall-note 169 ((GPL-2.0 WITH Linux-syscall-note) OR BSD-2-Clause) 21 ((GPL-2.0 WITH Linux-syscall-note) OR BSD-3-Clause) 17 LGPL-2.1+ WITH Linux-syscall-note 15 GPL-1.0+ WITH Linux-syscall-note 14 ((GPL-2.0+ WITH Linux-syscall-note) OR BSD-3-Clause) 5 LGPL-2.0+ WITH Linux-syscall-note 4 LGPL-2.1 WITH Linux-syscall-note 3 ((GPL-2.0 WITH Linux-syscall-note) OR MIT) 3 ((GPL-2.0 WITH Linux-syscall-note) AND MIT) 1 and that resulted in the third patch in this series. - when the two scanners agreed on the detected license(s), that became the concluded license(s). - when there was disagreement between the two scanners (one detected a license but the other didn't, or they both detected different licenses) a manual inspection of the file occurred. - In most cases a manual inspection of the information in the file resulted in a clear resolution of the license that should apply (and which scanner probably needed to revisit its heuristics). - When it was not immediately clear, the license identifier was confirmed with lawyers working with the Linux Foundation. - If there was any question as to the appropriate license identifier, the file was flagged for further research and to be revisited later in time. In total, over 70 hours of logged manual review was done on the spreadsheet to determine the SPDX license identifiers to apply to the source files by Kate, Philippe, Thomas and, in some cases, confirmation by lawyers working with the Linux Foundation. Kate also obtained a third independent scan of the 4.13 code base from FOSSology, and compared selected files where the other two scanners disagreed against that SPDX file, to see if there was new insights. The Windriver scanner is based on an older version of FOSSology in part, so they are related. Thomas did random spot checks in about 500 files from the spreadsheets for the uapi headers and agreed with SPDX license identifier in the files he inspected. For the non-uapi files Thomas did random spot checks in about 15000 files. In initial set of patches against 4.14-rc6, 3 files were found to have copy/paste license identifier errors, and have been fixed to reflect the correct identifier. Additionally Philippe spent 10 hours this week doing a detailed manual inspection and review of the 12,461 patched files from the initial patch version early this week with: - a full scancode scan run, collecting the matched texts, detected license ids and scores - reviewing anything where there was a license detected (about 500+ files) to ensure that the applied SPDX license was correct - reviewing anything where there was no detection but the patch license was not GPL-2.0 WITH Linux-syscall-note to ensure that the applied SPDX license was correct This produced a worksheet with 20 files needing minor correction. This worksheet was then exported into 3 different .csv files for the different types of files to be modified. These .csv files were then reviewed by Greg. Thomas wrote a script to parse the csv files and add the proper SPDX tag to the file, in the format that the file expected. This script was further refined by Greg based on the output to detect more types of files automatically and to distinguish between header and source .c files (which need different comment types.) Finally Greg ran the script using the .csv files to generate the patches. Reviewed-by: Kate Stewart <kstewart@linuxfoundation.org> Reviewed-by: Philippe Ombredanne <pombredanne@nexb.com> Reviewed-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2017-11-01 21:07:57 +07:00
/* SPDX-License-Identifier: GPL-2.0 */
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
/*
* S390 low-level entry points.
*
s390/comments: unify copyright messages and remove file names Remove the file name from the comment at top of many files. In most cases the file name was wrong anyway, so it's rather pointless. Also unify the IBM copyright statement. We did have a lot of sightly different statements and wanted to change them one after another whenever a file gets touched. However that never happened. Instead people start to take the old/"wrong" statements to use as a template for new files. So unify all of them in one go. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
2012-07-20 16:15:04 +07:00
* Copyright IBM Corp. 1999, 2012
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
* Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com),
[S390] Whitespace cleanup. Huge s390 assembly files whitespace cleanup. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2006-09-28 21:56:37 +07:00
* Hartmut Penner (hp@de.ibm.com),
* Denis Joseph Barrow (djbarrow@de.ibm.com,barrow_dj@yahoo.com),
[PATCH] s390: improved machine check handling Improved machine check handling. Kernel is now able to receive machine checks while in kernel mode (system call, interrupt and program check handling). Also register validation is now performed. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-26 04:55:30 +07:00
* Heiko Carstens <heiko.carstens@de.ibm.com>
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
*/
[S390] Fix couple of section mismatches. Fix couple of section mismatches. And since we touch the code anyway change the IPL code to use C99 initializers. Cc: Michael Holzheu <holzheu@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2008-02-05 22:50:40 +07:00
#include <linux/init.h>
[S390] fix s390 assembler code alignments The alignment is missing for various global symbols in s390 assembly code. With a recent gcc and an instruction like stgrl this can lead to a specification exception if the instruction uses such a mis-aligned address. Specify the alignment explicitely and while add it define __ALIGN for s390 and use the ENTRY define to save some lines of code. Signed-off-by: Jan Glauber <jang@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-07-24 15:48:19 +07:00
#include <linux/linkage.h>
s390/entry.S: use assembler alternatives Replace the open coded alternatives for the BPOFF, BPON, BPENTER, and BPEXIT macros with the new magic from asm/alternatives-asm.h to make the code easier to read. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-03-26 20:27:36 +07:00
#include <asm/alternative-asm.h>
s390/exceptions: switch to relative exception table entries This is the s390 port of 70627654 "x86, extable: Switch to relative exception table entries". Reduces the size of our exception tables by 50% on 64 bit builds. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-09-05 18:26:11 +07:00
#include <asm/processor.h>
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#include <asm/cache.h>
s390/nmi: do register validation as early as possible The validation of the CPU registers in the machine check handler is currently split into two parts. The first part is done at the start of the low level mcck_int_handler function, this includes the CPU timer register and the general purpose registers. The second part is done a bit later in s390_do_machine_check for all the other registers, including the control registers, floating pointer control, vector or floating pointer registers, the access registers, the guarded storage registers, the TOD programmable registers and the clock comparator. This is working fine to far but in theory a future extensions could cause the C code to use registers that are not validated yet. A better approach is to validate all CPU registers in "safe" assembler code before any C function is called. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-10-12 18:24:48 +07:00
#include <asm/ctl_reg.h>
s390/clean-up: use CFI_* macros in entry.S Commit f19fbd5ed642 ("s390: introduce execute-trampolines for branches") introduces .cfi_* assembler directives. Instead of using the directives directly, use the macros from asm/dwarf.h. This also ensures that the dwarf debug information are created in the .debug_frame section. Fixes: f19fbd5ed642 ("s390: introduce execute-trampolines for branches") Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-02-19 17:27:09 +07:00
#include <asm/dwarf.h>
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#include <asm/errno.h>
#include <asm/ptrace.h>
#include <asm/thread_info.h>
kbuild: m68k,parisc,ppc,ppc64,s390,xtensa use generic asm-offsets.h support Delete obsoleted parts form arch makefiles and rename to asm-offsets.h Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
2005-09-10 01:57:26 +07:00
#include <asm/asm-offsets.h>
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#include <asm/unistd.h>
#include <asm/page.h>
s390/sigp: use sigp order code defines in assembly code Use sigp order code defines in assembly code as well. With this change all places that use sigp constants should have been converted to use self describing defines instead of directly using constants. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-06-04 20:05:43 +07:00
#include <asm/sigp.h>
s390: convert interrupt handling to use generic hardirq With the introduction of PCI it became apparent that s390 should convert to generic hardirqs as too many drivers do not have the correct dependency for GENERIC_HARDIRQS. On the architecture level s390 does not have irq lines. It has external interrupts, I/O interrupts and adapter interrupts. This patch hard-codes all external interrupts as irq #1, all I/O interrupts as irq #2 and all adapter interrupts as irq #3. The additional information from the lowcore associated with the interrupt is stored in the pt_regs of the interrupt frame, where the interrupt handler can pick it up. For PCI/MSI interrupts the adapter interrupt handler scans the relevant bit fields and calls generic_handle_irq with the virtual irq number for the MSI interrupt. Reviewed-by: Sebastian Ott <sebott@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-06-27 14:01:09 +07:00
#include <asm/irq.h>
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
#include <asm/vx-insn.h>
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
#include <asm/setup.h>
#include <asm/nmi.h>
s390: move exports to definitions Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-01-13 01:30:03 +07:00
#include <asm/export.h>
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
__PT_R0 = __PT_GPRS
__PT_R1 = __PT_GPRS + 8
__PT_R2 = __PT_GPRS + 16
__PT_R3 = __PT_GPRS + 24
__PT_R4 = __PT_GPRS + 32
__PT_R5 = __PT_GPRS + 40
__PT_R6 = __PT_GPRS + 48
__PT_R7 = __PT_GPRS + 56
__PT_R8 = __PT_GPRS + 64
__PT_R9 = __PT_GPRS + 72
__PT_R10 = __PT_GPRS + 80
__PT_R11 = __PT_GPRS + 88
__PT_R12 = __PT_GPRS + 96
__PT_R13 = __PT_GPRS + 104
__PT_R14 = __PT_GPRS + 112
__PT_R15 = __PT_GPRS + 120
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
s390/thread_info: get rid of THREAD_ORDER define We have the s390 specific THREAD_ORDER define and the THREAD_SIZE_ORDER define which is also used in common code. Both have exactly the same semantics. Therefore get rid of THREAD_ORDER and always use THREAD_SIZE_ORDER instead. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-11-14 20:39:16 +07:00
STACK_SHIFT = PAGE_SHIFT + THREAD_SIZE_ORDER
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
STACK_SIZE = 1 << STACK_SHIFT
s390: lowcore stack pointer offsets Store the stack pointers in the lowcore for the kernel stack, the async stack and the panic stack with the offset required for the first user. This avoids an unnecessary add instruction on the system call path. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-04-24 15:20:43 +07:00
STACK_INIT = STACK_SIZE - STACK_FRAME_OVERHEAD - __PT_SIZE
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
s390/uprobes: architecture backend for uprobes Signed-off-by: Jan Willeke <willeke@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-09-22 21:39:06 +07:00
_TIF_WORK = (_TIF_SIGPENDING | _TIF_NOTIFY_RESUME | _TIF_NEED_RESCHED | \
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/livepatching Pull livepatch updates from Jiri Kosina: - a per-task consistency model is being added for architectures that support reliable stack dumping (extending this, currently rather trivial set, is currently in the works). This extends the nature of the types of patches that can be applied by live patching infrastructure. The code stems from the design proposal made [1] back in November 2014. It's a hybrid of SUSE's kGraft and RH's kpatch, combining advantages of both: it uses kGraft's per-task consistency and syscall barrier switching combined with kpatch's stack trace switching. There are also a number of fallback options which make it quite flexible. Most of the heavy lifting done by Josh Poimboeuf with help from Miroslav Benes and Petr Mladek [1] https://lkml.kernel.org/r/20141107140458.GA21774@suse.cz - module load time patch optimization from Zhou Chengming - a few assorted small fixes * 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/livepatching: livepatch: add missing printk newlines livepatch: Cancel transition a safe way for immediate patches livepatch: Reduce the time of finding module symbols livepatch: make klp_mutex proper part of API livepatch: allow removal of a disabled patch livepatch: add /proc/<pid>/patch_state livepatch: change to a per-task consistency model livepatch: store function sizes livepatch: use kstrtobool() in enabled_store() livepatch: move patching functions into patch.c livepatch: remove unnecessary object loaded check livepatch: separate enabled and patched states livepatch/s390: add TIF_PATCH_PENDING thread flag livepatch/s390: reorganize TIF thread flag bits livepatch/powerpc: add TIF_PATCH_PENDING thread flag livepatch/x86: add TIF_PATCH_PENDING thread flag livepatch: create temporary klp_update_patch_state() stub x86/entry: define _TIF_ALLWORK_MASK flags explicitly stacktrace/x86: add function for detecting reliable stack traces
2017-05-03 08:24:16 +07:00
_TIF_UPROBE | _TIF_GUARDED_STORAGE | _TIF_PATCH_PENDING)
s390: split TIF bits into CIF, PIF and TIF bits The oi and ni instructions used in entry[64].S to set and clear bits in the thread-flags are not guaranteed to be atomic in regard to other CPUs. Split the TIF bits into CPU, pt_regs and thread-info specific bits. Updates on the TIF bits are done with atomic instructions, updates on CPU and pt_regs bits are done with non-atomic instructions. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-04-15 17:55:07 +07:00
_TIF_TRACE = (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_AUDIT | _TIF_SECCOMP | \
_TIF_SYSCALL_TRACEPOINT)
s390: restore address space when returning to user space Unbalanced set_fs usages (e.g. early exit from a function and a forgotten set_fs(USER_DS) call) may lead to a situation where the secondary asce is the kernel space asce when returning to user space. This would allow user space to modify kernel space at will. This would only be possible with the above mentioned kernel bug, however we can detect this and fix the secondary asce before returning to user space. Therefore a new TIF_ASCE_SECONDARY which is used within set_fs. When returning to user space check if TIF_ASCE_SECONDARY is set, which would indicate a bug. If it is set print a message to the console, fixup the secondary asce, and then return to user space. This is similar to what is being discussed for x86 and arm: "[RFC] syscalls: Restore address limit after a syscall". Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-02-17 14:13:28 +07:00
_CIF_WORK = (_CIF_MCCK_PENDING | _CIF_ASCE_PRIMARY | \
_CIF_ASCE_SECONDARY | _CIF_FPU)
s390/kvm: avoid global config of vm.alloc_pgste=1 The system control vm.alloc_pgste is used to control the size of the page tables, either 2K or 4K. The idea is that a KVM host sets the vm.alloc_pgste control to 1 which causes *all* new processes to run with 4K page tables. For a non-kvm system the control should stay off to save on memory used for page tables. Trouble is that distributions choose to set the control globally to be able to run KVM guests. This wastes memory on non-KVM systems. Introduce the PT_S390_PGSTE ELF segment type to "mark" the qemu executable with it. All executables with this (empty) segment in its ELF phdr array will be started with 4K page tables. Any executable without PT_S390_PGSTE will run with the default 2K page tables. This removes the need to set vm.alloc_pgste=1 for a KVM host and minimizes the waste of memory for page tables. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-06-07 19:10:24 +07:00
_PIF_WORK = (_PIF_PER_TRAP | _PIF_SYSCALL_RESTART)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
s390/lpp: use assembler alternatives for the LPP instruction With the new macros for CPU alternatives the MACHINE_FLAG_LPP check around the LPP instruction can be optimized. After this is done the flag can be removed. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-03-26 20:23:33 +07:00
_LPP_OFFSET = __LC_LPP
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
#define BASED(name) name-cleanup_critical(%r13)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
[PATCH] lockdep: irqtrace subsystem, s390 support irqtrace support for s390. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-07-03 14:24:46 +07:00
.macro TRACE_IRQS_ON
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
#ifdef CONFIG_TRACE_IRQFLAGS
[S390] improve irq tracing code in entry[64].S The system call path in entry[64].S is run with interrupts enabled. Remove the irq tracing check from the system call exit code. If a program check interrupted a context enabled for interrupts do a call to trace_irq_off_caller in the program check handler before branching to the system call exit code. Restructure the system call and io interrupt return code to avoid avoid the lpsw[e] to disable machine checks. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-05-17 15:00:02 +07:00
basr %r2,%r0
brasl %r14,trace_hardirqs_on_caller
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
#endif
[PATCH] lockdep: irqtrace subsystem, s390 support irqtrace support for s390. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-07-03 14:24:46 +07:00
.endm
.macro TRACE_IRQS_OFF
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
#ifdef CONFIG_TRACE_IRQFLAGS
[S390] improve irq tracing code in entry[64].S The system call path in entry[64].S is run with interrupts enabled. Remove the irq tracing check from the system call exit code. If a program check interrupted a context enabled for interrupts do a call to trace_irq_off_caller in the program check handler before branching to the system call exit code. Restructure the system call and io interrupt return code to avoid avoid the lpsw[e] to disable machine checks. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-05-17 15:00:02 +07:00
basr %r2,%r0
brasl %r14,trace_hardirqs_off_caller
[S390] Fix irq tracing and lockdep_sys_exit calls. Current support for TRACE_IRQFLAGS and lockdep_sys_exit is broken. IRQ flag tracing is broken for program checks. Even worse is that the newly introduced calls to lockdep_sys_exit are in the critical section code which is not supposed to call any C functions. In addition the checks if locks are still held are also done when returning to kernel code which is broken as well. Fix all this by disabling interrupts and machine checks at the exit paths and then do the appropriate checks and calls. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2007-11-20 17:13:32 +07:00
#endif
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
.endm
[S390] Fix irq tracing and lockdep_sys_exit calls. Current support for TRACE_IRQFLAGS and lockdep_sys_exit is broken. IRQ flag tracing is broken for program checks. Even worse is that the newly introduced calls to lockdep_sys_exit are in the critical section code which is not supposed to call any C functions. In addition the checks if locks are still held are also done when returning to kernel code which is broken as well. Fix all this by disabling interrupts and machine checks at the exit paths and then do the appropriate checks and calls. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2007-11-20 17:13:32 +07:00
.macro LOCKDEP_SYS_EXIT
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
#ifdef CONFIG_LOCKDEP
tm __PT_PSW+1(%r11),0x01 # returning to user ?
jz .+10
[S390] Fix irq tracing and lockdep_sys_exit calls. Current support for TRACE_IRQFLAGS and lockdep_sys_exit is broken. IRQ flag tracing is broken for program checks. Even worse is that the newly introduced calls to lockdep_sys_exit are in the critical section code which is not supposed to call any C functions. In addition the checks if locks are still held are also done when returning to kernel code which is broken as well. Fix all this by disabling interrupts and machine checks at the exit paths and then do the appropriate checks and calls. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2007-11-20 17:13:32 +07:00
brasl %r14,lockdep_sys_exit
[PATCH] lockdep: irqtrace subsystem, s390 support irqtrace support for s390. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-07-03 14:24:46 +07:00
#endif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
.endm
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
.macro CHECK_STACK stacksize,savearea
[S390] virtual cpu accounting vs. machine checks. If a machine checks interrupts the external or the i/o interrupt handler before they have completed the cpu time calculations, the accounting goes wrong. After the cpu returned from the machine check handler to the interrupted interrupt handler, a negative cpu time delta can occur. If the accumulated cpu time in lowcore is small enough this value can get negative as well. The next jiffy interrupt will pick up that negative value, shift it by 12 and add the now huge positive value to the cpu time of the process. To solve this the machine check handler is modified not to change any of the timestamps in the lowcore if the machine check interrupted kernel context. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2006-06-29 19:58:05 +07:00
#ifdef CONFIG_CHECK_STACK
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
tml %r15,\stacksize - CONFIG_STACK_GUARD
lghi %r14,\savearea
jz stack_overflow
[S390] virtual cpu accounting vs. machine checks. If a machine checks interrupts the external or the i/o interrupt handler before they have completed the cpu time calculations, the accounting goes wrong. After the cpu returned from the machine check handler to the interrupted interrupt handler, a negative cpu time delta can occur. If the accumulated cpu time in lowcore is small enough this value can get negative as well. The next jiffy interrupt will pick up that negative value, shift it by 12 and add the now huge positive value to the cpu time of the process. To solve this the machine check handler is modified not to change any of the timestamps in the lowcore if the machine check interrupted kernel context. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2006-06-29 19:58:05 +07:00
#endif
.endm
s390/nmi: use the normal asynchronous stack for machine checks If a machine checks is received while the CPU is in the kernel, only the s390_do_machine_check function will be called. The call to s390_handle_mcck is postponed until the CPU returns to user space. Because of this it is safe to use the asynchronous stack for machine checks even if the CPU is already handling an interrupt. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:28:14 +07:00
.macro SWITCH_ASYNC savearea,timer
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
tmhh %r8,0x0001 # interrupting from user ?
jnz 1f
lgr %r14,%r9
slg %r14,BASED(.Lcritical_start)
clg %r14,BASED(.Lcritical_length)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
jhe 0f
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lghi %r11,\savearea # inside critical section, do cleanup
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
brasl %r14,cleanup_critical
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
tmhh %r8,0x0001 # retest problem state after cleanup
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
jnz 1f
s390/nmi: use the normal asynchronous stack for machine checks If a machine checks is received while the CPU is in the kernel, only the s390_do_machine_check function will be called. The call to s390_handle_mcck is postponed until the CPU returns to user space. Because of this it is safe to use the asynchronous stack for machine checks even if the CPU is already handling an interrupt. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:28:14 +07:00
0: lg %r14,__LC_ASYNC_STACK # are we already on the async stack?
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
slgr %r14,%r15
s390/nmi: use the normal asynchronous stack for machine checks If a machine checks is received while the CPU is in the kernel, only the s390_do_machine_check function will be called. The call to s390_handle_mcck is postponed until the CPU returns to user space. Because of this it is safe to use the asynchronous stack for machine checks even if the CPU is already handling an interrupt. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:28:14 +07:00
srag %r14,%r14,STACK_SHIFT
s390/kernel: squeeze a few more cycles out of the system call handler Reorder the instructions of UPDATE_VTIME to improve superscalar execution, remove duplicate checks for problem-state from the asynchronous interrupt handlers, and move the check for problem-state from the synchronous exit path to the program check path as it is only needed for program checks inside the kernel. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:27:48 +07:00
jnz 2f
s390/nmi: use the normal asynchronous stack for machine checks If a machine checks is received while the CPU is in the kernel, only the s390_do_machine_check function will be called. The call to s390_handle_mcck is postponed until the CPU returns to user space. Because of this it is safe to use the asynchronous stack for machine checks even if the CPU is already handling an interrupt. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:28:14 +07:00
CHECK_STACK 1<<STACK_SHIFT,\savearea
s390: lowcore stack pointer offsets Store the stack pointers in the lowcore for the kernel stack, the async stack and the panic stack with the offset required for the first user. This avoids an unnecessary add instruction on the system call path. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-04-24 15:20:43 +07:00
aghi %r15,-(STACK_FRAME_OVERHEAD + __PT_SIZE)
s390/kernel: squeeze a few more cycles out of the system call handler Reorder the instructions of UPDATE_VTIME to improve superscalar execution, remove duplicate checks for problem-state from the asynchronous interrupt handlers, and move the check for problem-state from the synchronous exit path to the program check path as it is only needed for program checks inside the kernel. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:27:48 +07:00
j 3f
s390: store breaking event address only for program checks The principles of operations specifies that the breaking event address is stored to the address 0x110 in the prefix page only for program checks. The last branch in user space is lost as soon as a branch in kernel space is executed after e.g. an svc. This makes it impossible to accurately maintain the breaking event address for a user space process. Simplify the code, just copy the current breaking event address from 0x110 to the task structure for program checks from user space. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-01-25 18:54:17 +07:00
1: UPDATE_VTIME %r14,%r15,\timer
s390: run user space and KVM guests with modified branch prediction Define TIF_ISOLATE_BP and TIF_ISOLATE_BP_GUEST and add the necessary plumbing in entry.S to be able to run user space and KVM guests with limited branch prediction. To switch a user space process to limited branch prediction the s390_isolate_bp() function has to be call, and to run a vCPU of a KVM guest associated with the current task with limited branch prediction call s390_isolate_bp_guest(). Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:36:46 +07:00
BPENTER __TI_flags(%r12),_TIF_ISOLATE_BP
s390/nmi: use the normal asynchronous stack for machine checks If a machine checks is received while the CPU is in the kernel, only the s390_do_machine_check function will be called. The call to s390_handle_mcck is postponed until the CPU returns to user space. Because of this it is safe to use the asynchronous stack for machine checks even if the CPU is already handling an interrupt. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:28:14 +07:00
2: lg %r15,__LC_ASYNC_STACK # load async stack
s390/kernel: squeeze a few more cycles out of the system call handler Reorder the instructions of UPDATE_VTIME to improve superscalar execution, remove duplicate checks for problem-state from the asynchronous interrupt handlers, and move the check for problem-state from the synchronous exit path to the program check path as it is only needed for program checks inside the kernel. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:27:48 +07:00
3: la %r11,STACK_FRAME_OVERHEAD(%r15)
[S390] Whitespace cleanup. Huge s390 assembly files whitespace cleanup. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2006-09-28 21:56:37 +07:00
.endm
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
s390/kernel: squeeze a few more cycles out of the system call handler Reorder the instructions of UPDATE_VTIME to improve superscalar execution, remove duplicate checks for problem-state from the asynchronous interrupt handlers, and move the check for problem-state from the synchronous exit path to the program check path as it is only needed for program checks inside the kernel. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:27:48 +07:00
.macro UPDATE_VTIME w1,w2,enter_timer
lg \w1,__LC_EXIT_TIMER
lg \w2,__LC_LAST_UPDATE_TIMER
slg \w1,\enter_timer
slg \w2,__LC_EXIT_TIMER
alg \w1,__LC_USER_TIMER
alg \w2,__LC_SYSTEM_TIMER
stg \w1,__LC_USER_TIMER
stg \w2,__LC_SYSTEM_TIMER
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
mvc __LC_LAST_UPDATE_TIMER(8),\enter_timer
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
.endm
[S390] cleanup lowcore access from program checks Read all required fields for program checks from the lowcore in the first level interrupt handler in entry[64].S. If the context that caused the fault was enabled for interrupts we can now re-enable the irqs in entry[64].S. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-10-25 21:10:37 +07:00
.macro REENABLE_IRQS
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stg %r8,__LC_RETURN_PSW
ni __LC_RETURN_PSW,0xbf
ssm __LC_RETURN_PSW
[S390] cleanup lowcore access from program checks Read all required fields for program checks from the lowcore in the first level interrupt handler in entry[64].S. If the context that caused the fault was enabled for interrupts we can now re-enable the irqs in entry[64].S. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-10-25 21:10:37 +07:00
.endm
s390/time: always use stckf instead of stck if available The store clock fast instruction saves a couple of instructions compared to the store clock instruction. Always use stckf instead of stck if it is available. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Reviewed-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-05-09 21:27:39 +07:00
.macro STCK savearea
s390/time: simply Kconfig dependency Use HAVE_MARCH_Z9_109_FEATURES to figure out if stckf is available at compile time. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-05-14 15:35:22 +07:00
#ifdef CONFIG_HAVE_MARCH_Z9_109_FEATURES
s390/time: always use stckf instead of stck if available The store clock fast instruction saves a couple of instructions compared to the store clock instruction. Always use stckf instead of stck if it is available. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Reviewed-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-05-09 21:27:39 +07:00
.insn s,0xb27c0000,\savearea # store clock fast
#else
.insn s,0xb2050000,\savearea # store clock
#endif
.endm
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
/*
* The TSTMSK macro generates a test-under-mask instruction by
* calculating the memory offset for the specified mask value.
* Mask value can be any constant. The macro shifts the mask
* value to calculate the memory offset for the test-under-mask
* instruction.
*/
.macro TSTMSK addr, mask, size=8, bytepos=0
.if (\bytepos < \size) && (\mask >> 8)
.if (\mask & 0xff)
.error "Mask exceeds byte boundary"
.endif
TSTMSK \addr, "(\mask >> 8)", \size, "(\bytepos + 1)"
.exitm
.endif
.ifeq \mask
.error "Mask must not be zero"
.endif
off = \size - \bytepos - 1
tm off+\addr, \mask
.endm
s390: add options to change branch prediction behaviour for the kernel Add the PPA instruction to the system entry and exit path to switch the kernel to a different branch prediction behaviour. The instructions are added via CPU alternatives and can be disabled with the "nospec" or the "nobp=0" kernel parameter. If the default behaviour selected with CONFIG_KERNEL_NOBP is set to "n" then the "nobp=1" parameter can be used to enable the changed kernel branch prediction. Acked-by: Cornelia Huck <cohuck@redhat.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:11:45 +07:00
.macro BPOFF
s390/entry.S: use assembler alternatives Replace the open coded alternatives for the BPOFF, BPON, BPENTER, and BPEXIT macros with the new magic from asm/alternatives-asm.h to make the code easier to read. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-03-26 20:27:36 +07:00
ALTERNATIVE "", ".long 0xb2e8c000", 82
s390: add options to change branch prediction behaviour for the kernel Add the PPA instruction to the system entry and exit path to switch the kernel to a different branch prediction behaviour. The instructions are added via CPU alternatives and can be disabled with the "nospec" or the "nobp=0" kernel parameter. If the default behaviour selected with CONFIG_KERNEL_NOBP is set to "n" then the "nobp=1" parameter can be used to enable the changed kernel branch prediction. Acked-by: Cornelia Huck <cohuck@redhat.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:11:45 +07:00
.endm
.macro BPON
s390/entry.S: use assembler alternatives Replace the open coded alternatives for the BPOFF, BPON, BPENTER, and BPEXIT macros with the new magic from asm/alternatives-asm.h to make the code easier to read. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-03-26 20:27:36 +07:00
ALTERNATIVE "", ".long 0xb2e8d000", 82
s390: add options to change branch prediction behaviour for the kernel Add the PPA instruction to the system entry and exit path to switch the kernel to a different branch prediction behaviour. The instructions are added via CPU alternatives and can be disabled with the "nospec" or the "nobp=0" kernel parameter. If the default behaviour selected with CONFIG_KERNEL_NOBP is set to "n" then the "nobp=1" parameter can be used to enable the changed kernel branch prediction. Acked-by: Cornelia Huck <cohuck@redhat.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:11:45 +07:00
.endm
s390: run user space and KVM guests with modified branch prediction Define TIF_ISOLATE_BP and TIF_ISOLATE_BP_GUEST and add the necessary plumbing in entry.S to be able to run user space and KVM guests with limited branch prediction. To switch a user space process to limited branch prediction the s390_isolate_bp() function has to be call, and to run a vCPU of a KVM guest associated with the current task with limited branch prediction call s390_isolate_bp_guest(). Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:36:46 +07:00
.macro BPENTER tif_ptr,tif_mask
s390/entry.S: use assembler alternatives Replace the open coded alternatives for the BPOFF, BPON, BPENTER, and BPEXIT macros with the new magic from asm/alternatives-asm.h to make the code easier to read. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-03-26 20:27:36 +07:00
ALTERNATIVE "TSTMSK \tif_ptr,\tif_mask; jz .+8; .long 0xb2e8d000", \
"", 82
s390: run user space and KVM guests with modified branch prediction Define TIF_ISOLATE_BP and TIF_ISOLATE_BP_GUEST and add the necessary plumbing in entry.S to be able to run user space and KVM guests with limited branch prediction. To switch a user space process to limited branch prediction the s390_isolate_bp() function has to be call, and to run a vCPU of a KVM guest associated with the current task with limited branch prediction call s390_isolate_bp_guest(). Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:36:46 +07:00
.endm
.macro BPEXIT tif_ptr,tif_mask
TSTMSK \tif_ptr,\tif_mask
s390/entry.S: use assembler alternatives Replace the open coded alternatives for the BPOFF, BPON, BPENTER, and BPEXIT macros with the new magic from asm/alternatives-asm.h to make the code easier to read. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-03-26 20:27:36 +07:00
ALTERNATIVE "jz .+8; .long 0xb2e8c000", \
"jnz .+8; .long 0xb2e8d000", 82
s390: run user space and KVM guests with modified branch prediction Define TIF_ISOLATE_BP and TIF_ISOLATE_BP_GUEST and add the necessary plumbing in entry.S to be able to run user space and KVM guests with limited branch prediction. To switch a user space process to limited branch prediction the s390_isolate_bp() function has to be call, and to run a vCPU of a KVM guest associated with the current task with limited branch prediction call s390_isolate_bp_guest(). Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:36:46 +07:00
.endm
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
#ifdef CONFIG_EXPOLINE
.macro GEN_BR_THUNK name,reg,tmp
.section .text.\name,"axG",@progbits,\name,comdat
.globl \name
.hidden \name
.type \name,@function
\name:
s390/clean-up: use CFI_* macros in entry.S Commit f19fbd5ed642 ("s390: introduce execute-trampolines for branches") introduces .cfi_* assembler directives. Instead of using the directives directly, use the macros from asm/dwarf.h. This also ensures that the dwarf debug information are created in the .debug_frame section. Fixes: f19fbd5ed642 ("s390: introduce execute-trampolines for branches") Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-02-19 17:27:09 +07:00
CFI_STARTPROC
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
#ifdef CONFIG_HAVE_MARCH_Z10_FEATURES
exrl 0,0f
#else
larl \tmp,0f
ex 0,0(\tmp)
#endif
j .
0: br \reg
s390/clean-up: use CFI_* macros in entry.S Commit f19fbd5ed642 ("s390: introduce execute-trampolines for branches") introduces .cfi_* assembler directives. Instead of using the directives directly, use the macros from asm/dwarf.h. This also ensures that the dwarf debug information are created in the .debug_frame section. Fixes: f19fbd5ed642 ("s390: introduce execute-trampolines for branches") Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-02-19 17:27:09 +07:00
CFI_ENDPROC
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
.endm
GEN_BR_THUNK __s390x_indirect_jump_r1use_r9,%r9,%r1
GEN_BR_THUNK __s390x_indirect_jump_r1use_r14,%r14,%r1
GEN_BR_THUNK __s390x_indirect_jump_r11use_r14,%r14,%r11
.macro BASR_R14_R9
0: brasl %r14,__s390x_indirect_jump_r1use_r9
.pushsection .s390_indirect_branches,"a",@progbits
.long 0b-.
.popsection
.endm
.macro BR_R1USE_R14
0: jg __s390x_indirect_jump_r1use_r14
.pushsection .s390_indirect_branches,"a",@progbits
.long 0b-.
.popsection
.endm
.macro BR_R11USE_R14
0: jg __s390x_indirect_jump_r11use_r14
.pushsection .s390_indirect_branches,"a",@progbits
.long 0b-.
.popsection
.endm
#else /* CONFIG_EXPOLINE */
.macro BASR_R14_R9
basr %r14,%r9
.endm
.macro BR_R1USE_R14
br %r14
.endm
.macro BR_R11USE_R14
br %r14
.endm
#endif /* CONFIG_EXPOLINE */
[S390] add kprobes annotations Add kprobes annotations to get the massive 'probe kernel.function("*") {}' stress test working. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-01-05 18:47:25 +07:00
.section .kprobes.text, "ax"
s390: have unique symbol for __switch_to address After linking there are several symbols for the same address that the __switch_to symbol points to. E.g.: 000000000089b9c0 T __kprobes_text_start 000000000089b9c0 T __lock_text_end 000000000089b9c0 T __lock_text_start 000000000089b9c0 T __sched_text_end 000000000089b9c0 T __switch_to When disassembling with "objdump -d" this results in a missing __switch_to function. It would be named __kprobes_text_start instead. To unconfuse objdump add a nop in front of the kprobes text section. That way __switch_to appears again. Obviously this solution is sort of a hack, since it also depends on link order if this works or not. However it is the best I can come up with for now. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-06-30 17:40:25 +07:00
.Ldummy:
/*
* This nop exists only in order to avoid that __switch_to starts at
* the beginning of the kprobes text section. In that case we would
* have several symbols at the same address. E.g. objdump would take
* an arbitrary symbol name when disassembling this code.
* With the added nop in between the __switch_to symbol is unique
* again.
*/
nop 0
[S390] add kprobes annotations Add kprobes annotations to get the massive 'probe kernel.function("*") {}' stress test working. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-01-05 18:47:25 +07:00
s390: add options to change branch prediction behaviour for the kernel Add the PPA instruction to the system entry and exit path to switch the kernel to a different branch prediction behaviour. The instructions are added via CPU alternatives and can be disabled with the "nospec" or the "nobp=0" kernel parameter. If the default behaviour selected with CONFIG_KERNEL_NOBP is set to "n" then the "nobp=1" parameter can be used to enable the changed kernel branch prediction. Acked-by: Cornelia Huck <cohuck@redhat.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:11:45 +07:00
ENTRY(__bpon)
.globl __bpon
BPON
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
BR_R1USE_R14
s390: add options to change branch prediction behaviour for the kernel Add the PPA instruction to the system entry and exit path to switch the kernel to a different branch prediction behaviour. The instructions are added via CPU alternatives and can be disabled with the "nospec" or the "nobp=0" kernel parameter. If the default behaviour selected with CONFIG_KERNEL_NOBP is set to "n" then the "nobp=1" parameter can be used to enable the changed kernel branch prediction. Acked-by: Cornelia Huck <cohuck@redhat.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:11:45 +07:00
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
/*
* Scheduler resume function, called by switch_to
* gpr2 = (task_struct *) prev
* gpr3 = (task_struct *) next
* Returns:
* gpr2 = prev
*/
[S390] fix s390 assembler code alignments The alignment is missing for various global symbols in s390 assembly code. With a recent gcc and an instruction like stgrl this can lead to a specification exception if the instruction uses such a mis-aligned address. Specify the alignment explicitely and while add it define __ALIGN for s390 and use the ENTRY define to save some lines of code. Signed-off-by: Jan Glauber <jang@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-07-24 15:48:19 +07:00
ENTRY(__switch_to)
s390: fix race on TIF_MCCK_PENDING There is a small race window in the __switch_to code in regard to the transfer of the TIF_MCCK_PENDING bit from the previous to the next task. The bit is transferred before the task struct pointer and the thread-info pointer for the next task has been stored to lowcore. If a machine check sets the TIF_MCCK_PENDING bit between the transfer code and the store of current/thread_info the bit is still set for the previous task. And if the previous task has terminated it can get lost. The effect is that a pending CRW is not retrieved until the next machine checks sets TIF_MCCK_PENDING. To fix this reorder __switch_to to first store the task struct and thread-info pointer and then do the transfer of the bit. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-05-15 14:20:06 +07:00
stmg %r6,%r15,__SF_GPRS(%r15) # store gprs of prev task
s390: rework __switch_to() to allow larger task_struct offsets If GCC_PLUGIN_RANDSTRUCT is enabled the members of task_struct will be shuffled around. The offsets of the "pid" and "stack" members within task_struct may not necessarily fit into 12 bits anymore, which causes compile errors within __switch_to, since instructions are used, which only have a 12 bit displacement field. Therefore rework __switch_to, to allow for larger offsets. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-11-16 20:54:04 +07:00
lghi %r4,__TASK_stack
lghi %r1,__TASK_thread
lg %r5,0(%r4,%r3) # start of kernel stack of next
stg %r15,__THREAD_ksp(%r1,%r2) # store kernel stack of prev
s390: fix race on TIF_MCCK_PENDING There is a small race window in the __switch_to code in regard to the transfer of the TIF_MCCK_PENDING bit from the previous to the next task. The bit is transferred before the task struct pointer and the thread-info pointer for the next task has been stored to lowcore. If a machine check sets the TIF_MCCK_PENDING bit between the transfer code and the store of current/thread_info the bit is still set for the previous task. And if the previous task has terminated it can get lost. The effect is that a pending CRW is not retrieved until the next machine checks sets TIF_MCCK_PENDING. To fix this reorder __switch_to to first store the task struct and thread-info pointer and then do the transfer of the bit. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-05-15 14:20:06 +07:00
lgr %r15,%r5
s390: lowcore stack pointer offsets Store the stack pointers in the lowcore for the kernel stack, the async stack and the panic stack with the offset required for the first user. This avoids an unnecessary add instruction on the system call path. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-04-24 15:20:43 +07:00
aghi %r15,STACK_INIT # end of kernel stack of next
s390: fix race on TIF_MCCK_PENDING There is a small race window in the __switch_to code in regard to the transfer of the TIF_MCCK_PENDING bit from the previous to the next task. The bit is transferred before the task struct pointer and the thread-info pointer for the next task has been stored to lowcore. If a machine check sets the TIF_MCCK_PENDING bit between the transfer code and the store of current/thread_info the bit is still set for the previous task. And if the previous task has terminated it can get lost. The effect is that a pending CRW is not retrieved until the next machine checks sets TIF_MCCK_PENDING. To fix this reorder __switch_to to first store the task struct and thread-info pointer and then do the transfer of the bit. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-05-15 14:20:06 +07:00
stg %r3,__LC_CURRENT # store task struct of next
stg %r15,__LC_KERNEL_STACK # store end of kernel stack
s390: rework __switch_to() to allow larger task_struct offsets If GCC_PLUGIN_RANDSTRUCT is enabled the members of task_struct will be shuffled around. The offsets of the "pid" and "stack" members within task_struct may not necessarily fit into 12 bits anymore, which causes compile errors within __switch_to, since instructions are used, which only have a 12 bit displacement field. Therefore rework __switch_to, to allow for larger offsets. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-11-16 20:54:04 +07:00
lg %r15,__THREAD_ksp(%r1,%r3) # load kernel stack of next
aghi %r3,__TASK_pid
mvc __LC_CURRENT_PID(4,%r0),0(%r3) # store pid of next
s390: split TIF bits into CIF, PIF and TIF bits The oi and ni instructions used in entry[64].S to set and clear bits in the thread-flags are not guaranteed to be atomic in regard to other CPUs. Split the TIF bits into CPU, pt_regs and thread-info specific bits. Updates on the TIF bits are done with atomic instructions, updates on CPU and pt_regs bits are done with non-atomic instructions. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-04-15 17:55:07 +07:00
lmg %r6,%r15,__SF_GPRS(%r15) # load gprs of next task
s390/lpp: use assembler alternatives for the LPP instruction With the new macros for CPU alternatives the MACHINE_FLAG_LPP check around the LPP instruction can be optimized. After this is done the flag can be removed. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-03-26 20:23:33 +07:00
ALTERNATIVE "", ".insn s,0xb2800000,_LPP_OFFSET", 40
BR_R1USE_R14
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.L__critical_start:
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
#if IS_ENABLED(CONFIG_KVM)
/*
* sie64a calling convention:
* %r2 pointer to sie control block
* %r3 guest register save area
*/
ENTRY(sie64a)
stmg %r6,%r14,__SF_GPRS(%r15) # save kernel registers
s390: run user space and KVM guests with modified branch prediction Define TIF_ISOLATE_BP and TIF_ISOLATE_BP_GUEST and add the necessary plumbing in entry.S to be able to run user space and KVM guests with limited branch prediction. To switch a user space process to limited branch prediction the s390_isolate_bp() function has to be call, and to run a vCPU of a KVM guest associated with the current task with limited branch prediction call s390_isolate_bp_guest(). Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:36:46 +07:00
lg %r12,__LC_CURRENT
s390/kvm: improve stack frame constants in entry.S The code in sie64a uses the stack frame passed to the function to store some temporary data in the empty1 array (see struct stack_frame in asm/processor.h. Replace the __SF_EMPTY+x constants with a properly defined offset: s/__SF_EMPTY/__SF_SIE_CONTROL/, s/__SF_EMPTY+8/__SF_SIE_SAVEAREA/, s/__SF_EMPTY+16/__SF_SIE_REASON/, s/__SF_EMPTY+24/__SF_SIE_FLAGS/. Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-03-20 19:33:43 +07:00
stg %r2,__SF_SIE_CONTROL(%r15) # save control block pointer
stg %r3,__SF_SIE_SAVEAREA(%r15) # save guest register save area
xc __SF_SIE_REASON(8,%r15),__SF_SIE_REASON(%r15) # reason code = 0
mvc __SF_SIE_FLAGS(8,%r15),__TI_flags(%r12) # copy thread flags
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __LC_CPU_FLAGS,_CIF_FPU # load guest fp/vx registers ?
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
jno .Lsie_load_guest_gprs
brasl %r14,load_fpu_regs # load guest fp/vx regs
.Lsie_load_guest_gprs:
lmg %r0,%r13,0(%r3) # load guest gprs 0-13
lg %r14,__LC_GMAP # get gmap pointer
ltgr %r14,%r14
jz .Lsie_gmap
lctlg %c1,%c1,__GMAP_ASCE(%r14) # load primary asce
.Lsie_gmap:
s390/kvm: improve stack frame constants in entry.S The code in sie64a uses the stack frame passed to the function to store some temporary data in the empty1 array (see struct stack_frame in asm/processor.h. Replace the __SF_EMPTY+x constants with a properly defined offset: s/__SF_EMPTY/__SF_SIE_CONTROL/, s/__SF_EMPTY+8/__SF_SIE_SAVEAREA/, s/__SF_EMPTY+16/__SF_SIE_REASON/, s/__SF_EMPTY+24/__SF_SIE_FLAGS/. Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-03-20 19:33:43 +07:00
lg %r14,__SF_SIE_CONTROL(%r15) # get control block pointer
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
oi __SIE_PROG0C+3(%r14),1 # we are going into SIE now
tm __SIE_PROG20+3(%r14),3 # last exit...
jnz .Lsie_skip
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __LC_CPU_FLAGS,_CIF_FPU
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
jo .Lsie_skip # exit if fp/vx regs changed
s390/kvm: improve stack frame constants in entry.S The code in sie64a uses the stack frame passed to the function to store some temporary data in the empty1 array (see struct stack_frame in asm/processor.h. Replace the __SF_EMPTY+x constants with a properly defined offset: s/__SF_EMPTY/__SF_SIE_CONTROL/, s/__SF_EMPTY+8/__SF_SIE_SAVEAREA/, s/__SF_EMPTY+16/__SF_SIE_REASON/, s/__SF_EMPTY+24/__SF_SIE_FLAGS/. Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-03-20 19:33:43 +07:00
BPEXIT __SF_SIE_FLAGS(%r15),(_TIF_ISOLATE_BP|_TIF_ISOLATE_BP_GUEST)
s390/nmi: s390: New low level handling for machine check happening in guest Add the logic to check if the machine check happens when the guest is running. If yes, set the exit reason -EINTR in the machine check's interrupt handler. Refactor s390_do_machine_check to avoid panicing the host for some kinds of machine checks which happen when guest is running. Reinject the instruction processing damage's machine checks including Delayed Access Exception instead of damaging the host if it happens in the guest because it could be caused by improper update on TLB entry or other software case and impacts the guest only. Signed-off-by: QingFeng Hao <haoqf@linux.vnet.ibm.com> Acked-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com> Acked-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
2017-06-07 16:30:42 +07:00
.Lsie_entry:
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
sie 0(%r14)
s390: add options to change branch prediction behaviour for the kernel Add the PPA instruction to the system entry and exit path to switch the kernel to a different branch prediction behaviour. The instructions are added via CPU alternatives and can be disabled with the "nospec" or the "nobp=0" kernel parameter. If the default behaviour selected with CONFIG_KERNEL_NOBP is set to "n" then the "nobp=1" parameter can be used to enable the changed kernel branch prediction. Acked-by: Cornelia Huck <cohuck@redhat.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:11:45 +07:00
.Lsie_exit:
BPOFF
s390/kvm: improve stack frame constants in entry.S The code in sie64a uses the stack frame passed to the function to store some temporary data in the empty1 array (see struct stack_frame in asm/processor.h. Replace the __SF_EMPTY+x constants with a properly defined offset: s/__SF_EMPTY/__SF_SIE_CONTROL/, s/__SF_EMPTY+8/__SF_SIE_SAVEAREA/, s/__SF_EMPTY+16/__SF_SIE_REASON/, s/__SF_EMPTY+24/__SF_SIE_FLAGS/. Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-03-20 19:33:43 +07:00
BPENTER __SF_SIE_FLAGS(%r15),(_TIF_ISOLATE_BP|_TIF_ISOLATE_BP_GUEST)
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
.Lsie_skip:
ni __SIE_PROG0C+3(%r14),0xfe # no longer in SIE
lctlg %c1,%c1,__LC_USER_ASCE # load primary asce
.Lsie_done:
# some program checks are suppressing. C code (e.g. do_protection_exception)
s390/kvm: do not rely on the ILC on kvm host protection fauls For most cases a protection exception in the host (e.g. copy on write or dirty tracking) on the sie instruction will indicate an instruction length of 4. Turns out that there are some corner cases (e.g. runtime instrumentation) where this is not necessarily true and the ILC is unpredictable. Let's replace our 4 byte rewind_pad with 3 byte nops to prepare for all possible ILCs. Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Cc: stable@vger.kernel.org Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-05-15 19:11:03 +07:00
# will rewind the PSW by the ILC, which is often 4 bytes in case of SIE. There
# are some corner cases (e.g. runtime instrumentation) where ILC is unpredictable.
# Other instructions between sie64a and .Lsie_done should not cause program
# interrupts. So lets use 3 nops as a landing pad for all possible rewinds.
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
# See also .Lcleanup_sie
s390/kvm: do not rely on the ILC on kvm host protection fauls For most cases a protection exception in the host (e.g. copy on write or dirty tracking) on the sie instruction will indicate an instruction length of 4. Turns out that there are some corner cases (e.g. runtime instrumentation) where this is not necessarily true and the ILC is unpredictable. Let's replace our 4 byte rewind_pad with 3 byte nops to prepare for all possible ILCs. Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Cc: stable@vger.kernel.org Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-05-15 19:11:03 +07:00
.Lrewind_pad6:
nopr 7
.Lrewind_pad4:
nopr 7
.Lrewind_pad2:
nopr 7
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
.globl sie_exit
sie_exit:
s390/kvm: improve stack frame constants in entry.S The code in sie64a uses the stack frame passed to the function to store some temporary data in the empty1 array (see struct stack_frame in asm/processor.h. Replace the __SF_EMPTY+x constants with a properly defined offset: s/__SF_EMPTY/__SF_SIE_CONTROL/, s/__SF_EMPTY+8/__SF_SIE_SAVEAREA/, s/__SF_EMPTY+16/__SF_SIE_REASON/, s/__SF_EMPTY+24/__SF_SIE_FLAGS/. Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-03-20 19:33:43 +07:00
lg %r14,__SF_SIE_SAVEAREA(%r15) # load guest register save area
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
stmg %r0,%r13,0(%r14) # save guest gprs 0-13
s390: scrub registers on kernel entry and KVM exit Clear all user space registers on entry to the kernel and all KVM guest registers on KVM guest exit if the register does not contain either a parameter or a result value. Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 19:27:30 +07:00
xgr %r0,%r0 # clear guest registers to
xgr %r1,%r1 # prevent speculative use
xgr %r2,%r2
xgr %r3,%r3
xgr %r4,%r4
xgr %r5,%r5
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
lmg %r6,%r14,__SF_GPRS(%r15) # restore kernel registers
s390/kvm: improve stack frame constants in entry.S The code in sie64a uses the stack frame passed to the function to store some temporary data in the empty1 array (see struct stack_frame in asm/processor.h. Replace the __SF_EMPTY+x constants with a properly defined offset: s/__SF_EMPTY/__SF_SIE_CONTROL/, s/__SF_EMPTY+8/__SF_SIE_SAVEAREA/, s/__SF_EMPTY+16/__SF_SIE_REASON/, s/__SF_EMPTY+24/__SF_SIE_FLAGS/. Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-03-20 19:33:43 +07:00
lg %r2,__SF_SIE_REASON(%r15) # return exit reason code
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
BR_R1USE_R14
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
.Lsie_fault:
lghi %r14,-EFAULT
s390/kvm: improve stack frame constants in entry.S The code in sie64a uses the stack frame passed to the function to store some temporary data in the empty1 array (see struct stack_frame in asm/processor.h. Replace the __SF_EMPTY+x constants with a properly defined offset: s/__SF_EMPTY/__SF_SIE_CONTROL/, s/__SF_EMPTY+8/__SF_SIE_SAVEAREA/, s/__SF_EMPTY+16/__SF_SIE_REASON/, s/__SF_EMPTY+24/__SF_SIE_FLAGS/. Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-03-20 19:33:43 +07:00
stg %r14,__SF_SIE_REASON(%r15) # set exit reason code
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
j sie_exit
s390/kvm: do not rely on the ILC on kvm host protection fauls For most cases a protection exception in the host (e.g. copy on write or dirty tracking) on the sie instruction will indicate an instruction length of 4. Turns out that there are some corner cases (e.g. runtime instrumentation) where this is not necessarily true and the ILC is unpredictable. Let's replace our 4 byte rewind_pad with 3 byte nops to prepare for all possible ILCs. Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Cc: stable@vger.kernel.org Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-05-15 19:11:03 +07:00
EX_TABLE(.Lrewind_pad6,.Lsie_fault)
EX_TABLE(.Lrewind_pad4,.Lsie_fault)
EX_TABLE(.Lrewind_pad2,.Lsie_fault)
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
EX_TABLE(sie_exit,.Lsie_fault)
s390: move exports to definitions Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-01-13 01:30:03 +07:00
EXPORT_SYMBOL(sie64a)
EXPORT_SYMBOL(sie_exit)
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
#endif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
/*
* SVC interrupt handler routine. System calls are synchronous events and
* are executed with interrupts enabled.
*/
[S390] fix s390 assembler code alignments The alignment is missing for various global symbols in s390 assembly code. With a recent gcc and an instruction like stgrl this can lead to a specification exception if the instruction uses such a mis-aligned address. Specify the alignment explicitely and while add it define __ALIGN for s390 and use the ENTRY define to save some lines of code. Signed-off-by: Jan Glauber <jang@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-07-24 15:48:19 +07:00
ENTRY(system_call)
[S390] Remove config options. On s390 we always want to run with precise cputime accounting. Remove the config options VIRT_TIMER and VIRT_CPU_ACCOUNTING. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2008-12-25 19:39:25 +07:00
stpt __LC_SYNC_ENTER_TIMER
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lsysc_stmg:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stmg %r8,%r15,__LC_SAVE_AREA_SYNC
s390: add options to change branch prediction behaviour for the kernel Add the PPA instruction to the system entry and exit path to switch the kernel to a different branch prediction behaviour. The instructions are added via CPU alternatives and can be disabled with the "nospec" or the "nobp=0" kernel parameter. If the default behaviour selected with CONFIG_KERNEL_NOBP is set to "n" then the "nobp=1" parameter can be used to enable the changed kernel branch prediction. Acked-by: Cornelia Huck <cohuck@redhat.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:11:45 +07:00
BPOFF
s390: move thread_info into task_struct This is the s390 variant of commit 15f4eae70d36 ("x86: Move thread_info into task_struct"). Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-11-08 17:08:26 +07:00
lg %r12,__LC_CURRENT
s390: store breaking event address only for program checks The principles of operations specifies that the breaking event address is stored to the address 0x110 in the prefix page only for program checks. The last branch in user space is lost as soon as a branch in kernel space is executed after e.g. an svc. This makes it impossible to accurately maintain the breaking event address for a user space process. Simplify the code, just copy the current breaking event address from 0x110 to the task structure for program checks from user space. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-01-25 18:54:17 +07:00
lghi %r13,__TASK_thread
s390: split TIF bits into CIF, PIF and TIF bits The oi and ni instructions used in entry[64].S to set and clear bits in the thread-flags are not guaranteed to be atomic in regard to other CPUs. Split the TIF bits into CPU, pt_regs and thread-info specific bits. Updates on the TIF bits are done with atomic instructions, updates on CPU and pt_regs bits are done with non-atomic instructions. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-04-15 17:55:07 +07:00
lghi %r14,_PIF_SYSCALL
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lsysc_per:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lg %r15,__LC_KERNEL_STACK
la %r11,STACK_FRAME_OVERHEAD(%r15) # pointer to pt_regs
s390/kernel: squeeze a few more cycles out of the system call handler Reorder the instructions of UPDATE_VTIME to improve superscalar execution, remove duplicate checks for problem-state from the asynchronous interrupt handlers, and move the check for problem-state from the synchronous exit path to the program check path as it is only needed for program checks inside the kernel. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:27:48 +07:00
.Lsysc_vtime:
s390: store breaking event address only for program checks The principles of operations specifies that the breaking event address is stored to the address 0x110 in the prefix page only for program checks. The last branch in user space is lost as soon as a branch in kernel space is executed after e.g. an svc. This makes it impossible to accurately maintain the breaking event address for a user space process. Simplify the code, just copy the current breaking event address from 0x110 to the task structure for program checks from user space. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-01-25 18:54:17 +07:00
UPDATE_VTIME %r8,%r9,__LC_SYNC_ENTER_TIMER
s390: run user space and KVM guests with modified branch prediction Define TIF_ISOLATE_BP and TIF_ISOLATE_BP_GUEST and add the necessary plumbing in entry.S to be able to run user space and KVM guests with limited branch prediction. To switch a user space process to limited branch prediction the s390_isolate_bp() function has to be call, and to run a vCPU of a KVM guest associated with the current task with limited branch prediction call s390_isolate_bp_guest(). Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:36:46 +07:00
BPENTER __TI_flags(%r12),_TIF_ISOLATE_BP
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stmg %r0,%r7,__PT_R0(%r11)
mvc __PT_R8(64,%r11),__LC_SAVE_AREA_SYNC
mvc __PT_PSW(16,%r11),__LC_SVC_OLD_PSW
[S390] cleanup trap handling Move the program interruption code and the translation exception identifier to the pt_regs structure as 'int_code' and 'int_parm_long' and make the first level interrupt handler in entry[64].S store the two values. That makes it possible to drop 'prot_addr' and 'trap_no' from the thread_struct and to reduce the number of arguments to a lot of functions. Finally un-inline do_trap. Overall this saves 5812 bytes in the .text section of the 64 bit kernel. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:18 +07:00
mvc __PT_INT_CODE(4,%r11),__LC_SVC_ILC
s390: split TIF bits into CIF, PIF and TIF bits The oi and ni instructions used in entry[64].S to set and clear bits in the thread-flags are not guaranteed to be atomic in regard to other CPUs. Split the TIF bits into CPU, pt_regs and thread-info specific bits. Updates on the TIF bits are done with atomic instructions, updates on CPU and pt_regs bits are done with non-atomic instructions. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-04-15 17:55:07 +07:00
stg %r14,__PT_FLAGS(%r11)
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lsysc_do_svc:
s390/entry.S: fix spurious zeroing of r0 when a system call is interrupted we might call the critical section cleanup handler that re-does some of the operations. When we are between .Lsysc_vtime and .Lsysc_do_svc we might also redo the saving of the problem state registers r0-r7: .Lcleanup_system_call: [...] 0: # update accounting time stamp mvc __LC_LAST_UPDATE_TIMER(8),__LC_SYNC_ENTER_TIMER # set up saved register r11 lg %r15,__LC_KERNEL_STACK la %r9,STACK_FRAME_OVERHEAD(%r15) stg %r9,24(%r11) # r11 pt_regs pointer # fill pt_regs mvc __PT_R8(64,%r9),__LC_SAVE_AREA_SYNC ---> stmg %r0,%r7,__PT_R0(%r9) The problem is now, that we might have already zeroed out r0. The fix is to move the zeroing of r0 after sysc_do_svc. Reported-by: Farhan Ali <alifm@linux.vnet.ibm.com> Fixes: 7041d28115e91 ("s390: scrub registers on kernel entry and KVM exit") Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-03-06 02:18:47 +07:00
# clear user controlled register to prevent speculative use
xgr %r0,%r0
s390: move sys_call_table and last_break from thread_info to thread_struct Move the last two architecture specific fields from the thread_info structure to the thread_struct. All that is left in thread_info is the flags field. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-11-08 18:33:38 +07:00
# load address of system call table
lg %r10,__THREAD_sysc_table(%r13,%r12)
[S390] cleanup trap handling Move the program interruption code and the translation exception identifier to the pt_regs structure as 'int_code' and 'int_parm_long' and make the first level interrupt handler in entry[64].S store the two values. That makes it possible to drop 'prot_addr' and 'trap_no' from the thread_struct and to reduce the number of arguments to a lot of functions. Finally un-inline do_trap. Overall this saves 5812 bytes in the .text section of the 64 bit kernel. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:18 +07:00
llgh %r8,__PT_INT_CODE+2(%r11)
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
slag %r8,%r8,2 # shift and test for svc 0
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jnz .Lsysc_nr_ok
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
# svc 0: system call number in %r1
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
llgfr %r1,%r1 # clear high word in r1
[S390] add breaking event address for user space Copy the last breaking event address from the lowcore to a new field in the thread_struct on each system entry. Add a new ptrace request PTRACE_GET_LAST_BREAK and a new utrace regset REGSET_LAST_BREAK to query the last breaking event. This is useful for debugging wild branches in user space code. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-05-17 15:00:05 +07:00
cghi %r1,NR_syscalls
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jnl .Lsysc_nr_ok
[S390] cleanup trap handling Move the program interruption code and the translation exception identifier to the pt_regs structure as 'int_code' and 'int_parm_long' and make the first level interrupt handler in entry[64].S store the two values. That makes it possible to drop 'prot_addr' and 'trap_no' from the thread_struct and to reduce the number of arguments to a lot of functions. Finally un-inline do_trap. Overall this saves 5812 bytes in the .text section of the 64 bit kernel. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:18 +07:00
sth %r1,__PT_INT_CODE+2(%r11)
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
slag %r8,%r1,2
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lsysc_nr_ok:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15)
stg %r2,__PT_ORIG_GPR2(%r11)
stg %r7,STACK_FRAME_OVERHEAD(%r15)
lgf %r9,0(%r8,%r10) # get system call add.
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __TI_flags(%r12),_TIF_TRACE
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jnz .Lsysc_tracesys
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
BASR_R14_R9 # call sys_xxxx
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stg %r2,__PT_R2(%r11) # store return value
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lsysc_return:
[S390] improve irq tracing code in entry[64].S The system call path in entry[64].S is run with interrupts enabled. Remove the irq tracing check from the system call exit code. If a program check interrupted a context enabled for interrupts do a call to trace_irq_off_caller in the program check handler before branching to the system call exit code. Restructure the system call and io interrupt return code to avoid avoid the lpsw[e] to disable machine checks. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-05-17 15:00:02 +07:00
LOCKDEP_SYS_EXIT
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lsysc_tif:
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __PT_FLAGS(%r11),_PIF_WORK
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jnz .Lsysc_work
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __TI_flags(%r12),_TIF_WORK
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jnz .Lsysc_work # check for work
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __LC_CPU_FLAGS,_CIF_WORK
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jnz .Lsysc_work
s390: run user space and KVM guests with modified branch prediction Define TIF_ISOLATE_BP and TIF_ISOLATE_BP_GUEST and add the necessary plumbing in entry.S to be able to run user space and KVM guests with limited branch prediction. To switch a user space process to limited branch prediction the s390_isolate_bp() function has to be call, and to run a vCPU of a KVM guest associated with the current task with limited branch prediction call s390_isolate_bp_guest(). Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:36:46 +07:00
BPEXIT __TI_flags(%r12),_TIF_ISOLATE_BP
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lsysc_restore:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lg %r14,__LC_VDSO_PER_CPU
lmg %r0,%r10,__PT_R0(%r11)
mvc __LC_RETURN_PSW(16),__PT_PSW(%r11)
s390/cputime: fix incorrect system time git commit c5328901aa1db134 "[S390] entry[64].S improvements" removed the update of the exit_timer lowcore field from the critical section cleanup of the .Lsysc_restore/.Lsysc_done and .Lio_restore/.Lio_done blocks. If the PSW is updated by the critical section cleanup to point to user space again, the interrupt entry code will do a vtime calculation after the cleanup completed with an exit_timer value which has *not* been updated. Due to this incorrect system time deltas are calculated. If an interrupt occured with an old PSW between .Lsysc_restore/.Lsysc_done or .Lio_restore/.Lio_done update __LC_EXIT_TIMER with the system entry time of the interrupt. Cc: stable@vger.kernel.org # 3.3+ Tested-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-05-02 18:36:00 +07:00
.Lsysc_exit_timer:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stpt __LC_EXIT_TIMER
mvc __VDSO_ECTG_BASE(16,%r14),__LC_EXIT_TIMER
lmg %r11,%r15,__PT_R11(%r11)
lpswe __LC_RETURN_PSW
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lsysc_done:
[S390] Fix irq tracing and lockdep_sys_exit calls. Current support for TRACE_IRQFLAGS and lockdep_sys_exit is broken. IRQ flag tracing is broken for program checks. Even worse is that the newly introduced calls to lockdep_sys_exit are in the critical section code which is not supposed to call any C functions. In addition the checks if locks are still held are also done when returning to kernel code which is broken as well. Fix all this by disabling interrupts and machine checks at the exit paths and then do the appropriate checks and calls. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2007-11-20 17:13:32 +07:00
[S390] cleanup sysc_work and io_work code Cleanup the #ifdef mess at io_work in entry[64].S and streamline the TIF work code of the system call and io exit path. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-05-17 15:00:01 +07:00
#
# One of the work bits is on. Find out which one.
#
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lsysc_work:
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __LC_CPU_FLAGS,_CIF_MCCK_PENDING
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jo .Lsysc_mcck_pending
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __TI_flags(%r12),_TIF_NEED_RESCHED
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jo .Lsysc_reschedule
s390/kvm: avoid global config of vm.alloc_pgste=1 The system control vm.alloc_pgste is used to control the size of the page tables, either 2K or 4K. The idea is that a KVM host sets the vm.alloc_pgste control to 1 which causes *all* new processes to run with 4K page tables. For a non-kvm system the control should stay off to save on memory used for page tables. Trouble is that distributions choose to set the control globally to be able to run KVM guests. This wastes memory on non-KVM systems. Introduce the PT_S390_PGSTE ELF segment type to "mark" the qemu executable with it. All executables with this (empty) segment in its ELF phdr array will be started with 4K page tables. Any executable without PT_S390_PGSTE will run with the default 2K page tables. This removes the need to set vm.alloc_pgste=1 for a KVM host and minimizes the waste of memory for page tables. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-06-07 19:10:24 +07:00
TSTMSK __PT_FLAGS(%r11),_PIF_SYSCALL_RESTART
jo .Lsysc_syscall_restart
s390/uprobes: architecture backend for uprobes Signed-off-by: Jan Willeke <willeke@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-09-22 21:39:06 +07:00
#ifdef CONFIG_UPROBES
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __TI_flags(%r12),_TIF_UPROBE
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jo .Lsysc_uprobe_notify
s390/uprobes: architecture backend for uprobes Signed-off-by: Jan Willeke <willeke@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-09-22 21:39:06 +07:00
#endif
s390: add a system call for guarded storage This adds a new system call to enable the use of guarded storage for user space processes. The system call takes two arguments, a command and pointer to a guarded storage control block: s390_guarded_storage(int command, struct gs_cb *gs_cb); The second argument is relevant only for the GS_SET_BC_CB command. The commands in detail: 0 - GS_ENABLE Enable the guarded storage facility for the current task. The initial content of the guarded storage control block will be all zeros. After the enablement the user space code can use load-guarded-storage-controls instruction (LGSC) to load an arbitrary control block. While a task is enabled the kernel will save and restore the current content of the guarded storage registers on context switch. 1 - GS_DISABLE Disables the use of the guarded storage facility for the current task. The kernel will cease to save and restore the content of the guarded storage registers, the task specific content of these registers is lost. 2 - GS_SET_BC_CB Set a broadcast guarded storage control block. This is called per thread and stores a specific guarded storage control block in the task struct of the current task. This control block will be used for the broadcast event GS_BROADCAST. 3 - GS_CLEAR_BC_CB Clears the broadcast guarded storage control block. The guarded- storage control block is removed from the task struct that was established by GS_SET_BC_CB. 4 - GS_BROADCAST Sends a broadcast to all thread siblings of the current task. Every sibling that has established a broadcast guarded storage control block will load this control block and will be enabled for guarded storage. The broadcast guarded storage control block is used up, a second broadcast without a refresh of the stored control block with GS_SET_BC_CB will not have any effect. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-01-26 20:10:34 +07:00
TSTMSK __TI_flags(%r12),_TIF_GUARDED_STORAGE
jo .Lsysc_guarded_storage
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __PT_FLAGS(%r11),_PIF_PER_TRAP
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jo .Lsysc_singlestep
livepatch/s390: add TIF_PATCH_PENDING thread flag Update a task's patch state when returning from a system call or user space interrupt, or after handling a signal. This greatly increases the chances of a patch operation succeeding. If a task is I/O bound, it can be patched when returning from a system call. If a task is CPU bound, it can be patched when returning from an interrupt. If a task is sleeping on a to-be-patched function, the user can send SIGSTOP and SIGCONT to force it to switch. Since there are two ways the syscall can be restarted on return from a signal handling process, it is important to clear the flag before do_signal() is called. Otherwise we could miss the migration if we used SIGSTOP/SIGCONT procedure or fake signal to migrate patching blocking tasks. If we place our hook to sysc_work label in entry before TIF_SIGPENDING is evaluated we kill two birds with one stone. The task is correctly migrated in all return paths from a syscall. Signed-off-by: Miroslav Benes <mbenes@suse.cz> Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com> Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2017-02-14 08:42:34 +07:00
#ifdef CONFIG_LIVEPATCH
TSTMSK __TI_flags(%r12),_TIF_PATCH_PENDING
jo .Lsysc_patch_pending # handle live patching just before
# signals and possible syscall restart
#endif
s390/kvm: avoid global config of vm.alloc_pgste=1 The system control vm.alloc_pgste is used to control the size of the page tables, either 2K or 4K. The idea is that a KVM host sets the vm.alloc_pgste control to 1 which causes *all* new processes to run with 4K page tables. For a non-kvm system the control should stay off to save on memory used for page tables. Trouble is that distributions choose to set the control globally to be able to run KVM guests. This wastes memory on non-KVM systems. Introduce the PT_S390_PGSTE ELF segment type to "mark" the qemu executable with it. All executables with this (empty) segment in its ELF phdr array will be started with 4K page tables. Any executable without PT_S390_PGSTE will run with the default 2K page tables. This removes the need to set vm.alloc_pgste=1 for a KVM host and minimizes the waste of memory for page tables. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-06-07 19:10:24 +07:00
TSTMSK __PT_FLAGS(%r11),_PIF_SYSCALL_RESTART
jo .Lsysc_syscall_restart
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __TI_flags(%r12),_TIF_SIGPENDING
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jo .Lsysc_sigpending
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __TI_flags(%r12),_TIF_NOTIFY_RESUME
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jo .Lsysc_notify_resume
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __LC_CPU_FLAGS,_CIF_FPU
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
jo .Lsysc_vxrs
s390: restore address space when returning to user space Unbalanced set_fs usages (e.g. early exit from a function and a forgotten set_fs(USER_DS) call) may lead to a situation where the secondary asce is the kernel space asce when returning to user space. This would allow user space to modify kernel space at will. This would only be possible with the above mentioned kernel bug, however we can detect this and fix the secondary asce before returning to user space. Therefore a new TIF_ASCE_SECONDARY which is used within set_fs. When returning to user space check if TIF_ASCE_SECONDARY is set, which would indicate a bug. If it is set print a message to the console, fixup the secondary asce, and then return to user space. This is similar to what is being discussed for x86 and arm: "[RFC] syscalls: Restore address limit after a syscall". Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-02-17 14:13:28 +07:00
TSTMSK __LC_CPU_FLAGS,(_CIF_ASCE_PRIMARY|_CIF_ASCE_SECONDARY)
jnz .Lsysc_asce
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
j .Lsysc_return # beware of critical section cleanup
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#
# _TIF_NEED_RESCHED is set, call schedule
[S390] Whitespace cleanup. Huge s390 assembly files whitespace cleanup. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2006-09-28 21:56:37 +07:00
#
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lsysc_reschedule:
larl %r14,.Lsysc_return
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
jg schedule
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
[PATCH] s390: improved machine check handling Improved machine check handling. Kernel is now able to receive machine checks while in kernel mode (system call, interrupt and program check handling). Also register validation is now performed. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-26 04:55:30 +07:00
#
s390: split TIF bits into CIF, PIF and TIF bits The oi and ni instructions used in entry[64].S to set and clear bits in the thread-flags are not guaranteed to be atomic in regard to other CPUs. Split the TIF bits into CPU, pt_regs and thread-info specific bits. Updates on the TIF bits are done with atomic instructions, updates on CPU and pt_regs bits are done with non-atomic instructions. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-04-15 17:55:07 +07:00
# _CIF_MCCK_PENDING is set, call handler
[PATCH] s390: improved machine check handling Improved machine check handling. Kernel is now able to receive machine checks while in kernel mode (system call, interrupt and program check handling). Also register validation is now performed. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-26 04:55:30 +07:00
#
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lsysc_mcck_pending:
larl %r14,.Lsysc_return
[S390] Whitespace cleanup. Huge s390 assembly files whitespace cleanup. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2006-09-28 21:56:37 +07:00
jg s390_handle_mcck # TIF bit will be cleared by handler
[PATCH] s390: improved machine check handling Improved machine check handling. Kernel is now able to receive machine checks while in kernel mode (system call, interrupt and program check handling). Also register validation is now performed. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-26 04:55:30 +07:00
s390/uaccess: rework uaccess code - fix locking issues The current uaccess code uses a page table walk in some circumstances, e.g. in case of the in atomic futex operations or if running on old hardware which doesn't support the mvcos instruction. However it turned out that the page table walk code does not correctly lock page tables when accessing page table entries. In other words: a different cpu may invalidate a page table entry while the current cpu inspects the pte. This may lead to random data corruption. Adding correct locking however isn't trivial for all uaccess operations. Especially copy_in_user() is problematic since that requires to hold at least two locks, but must be protected against ABBA deadlock when a different cpu also performs a copy_in_user() operation. So the solution is a different approach where we change address spaces: User space runs in primary address mode, or access register mode within vdso code, like it currently already does. The kernel usually also runs in home space mode, however when accessing user space the kernel switches to primary or secondary address mode if the mvcos instruction is not available or if a compare-and-swap (futex) instruction on a user space address is performed. KVM however is special, since that requires the kernel to run in home address space while implicitly accessing user space with the sie instruction. So we end up with: User space: - runs in primary or access register mode - cr1 contains the user asce - cr7 contains the user asce - cr13 contains the kernel asce Kernel space: - runs in home space mode - cr1 contains the user or kernel asce -> the kernel asce is loaded when a uaccess requires primary or secondary address mode - cr7 contains the user or kernel asce, (changed with set_fs()) - cr13 contains the kernel asce In case of uaccess the kernel changes to: - primary space mode in case of a uaccess (copy_to_user) and uses e.g. the mvcp instruction to access user space. However the kernel will stay in home space mode if the mvcos instruction is available - secondary space mode in case of futex atomic operations, so that the instructions come from primary address space and data from secondary space In case of kvm the kernel runs in home space mode, but cr1 gets switched to contain the gmap asce before the sie instruction gets executed. When the sie instruction is finished cr1 will be switched back to contain the user asce. A context switch between two processes will always load the kernel asce for the next process in cr1. So the first exit to user space is a bit more expensive (one extra load control register instruction) than before, however keeps the code rather simple. In sum this means there is no need to perform any error prone page table walks anymore when accessing user space. The patch seems to be rather large, however it mainly removes the the page table walk code and restores the previously deleted "standard" uaccess code, with a couple of changes. The uaccess without mvcos mode can be enforced with the "uaccess_primary" kernel parameter. Reported-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-03-21 16:42:25 +07:00
#
s390: remove all code using the access register mode The vdso code for the getcpu() and the clock_gettime() call use the access register mode to access the per-CPU vdso data page with the current code. An alternative to the complicated AR mode is to use the secondary space mode. This makes the vdso faster and quite a bit simpler. The downside is that the uaccess code has to be changed quite a bit. Which instructions are used depends on the machine and what kind of uaccess operation is requested. The instruction dictates which ASCE value needs to be loaded into %cr1 and %cr7. The different cases: * User copy with MVCOS for z10 and newer machines The MVCOS instruction can copy between the primary space (aka user) and the home space (aka kernel) directly. For set_fs(KERNEL_DS) the kernel ASCE is loaded into %cr1. For set_fs(USER_DS) the user space is already loaded in %cr1. * User copy with MVCP/MVCS for older machines To be able to execute the MVCP/MVCS instructions the kernel needs to switch to primary mode. The control register %cr1 has to be set to the kernel ASCE and %cr7 to either the kernel ASCE or the user ASCE dependent on set_fs(KERNEL_DS) vs set_fs(USER_DS). * Data access in the user address space for strnlen / futex To use "normal" instruction with data from the user address space the secondary space mode is used. The kernel needs to switch to primary mode, %cr1 has to contain the kernel ASCE and %cr7 either the user ASCE or the kernel ASCE, dependent on set_fs. To load a new value into %cr1 or %cr7 is an expensive operation, the kernel tries to be lazy about it. E.g. for multiple user copies in a row with MVCP/MVCS the replacement of the vdso ASCE in %cr7 with the user ASCE is done only once. On return to user space a CPU bit is checked that loads the vdso ASCE again. To enable and disable the data access via the secondary space two new functions are added, enable_sacf_uaccess and disable_sacf_uaccess. The fact that a context is in secondary space uaccess mode is stored in the mm_segment_t value for the task. The code of an interrupt may use set_fs as long as it returns to the previous state it got with get_fs with another call to set_fs. The code in finish_arch_post_lock_switch simply has to do a set_fs with the current mm_segment_t value for the task. For CPUs with MVCOS: CPU running in | %cr1 ASCE | %cr7 ASCE | --------------------------------------|-----------|-----------| user space | user | vdso | kernel, USER_DS, normal-mode | user | vdso | kernel, USER_DS, normal-mode, lazy | user | user | kernel, USER_DS, sacf-mode | kernel | user | kernel, KERNEL_DS, normal-mode | kernel | vdso | kernel, KERNEL_DS, normal-mode, lazy | kernel | kernel | kernel, KERNEL_DS, sacf-mode | kernel | kernel | For CPUs without MVCOS: CPU running in | %cr1 ASCE | %cr7 ASCE | --------------------------------------|-----------|-----------| user space | user | vdso | kernel, USER_DS, normal-mode | user | vdso | kernel, USER_DS, normal-mode lazy | kernel | user | kernel, USER_DS, sacf-mode | kernel | user | kernel, KERNEL_DS, normal-mode | kernel | vdso | kernel, KERNEL_DS, normal-mode, lazy | kernel | kernel | kernel, KERNEL_DS, sacf-mode | kernel | kernel | The lines with "lazy" refer to the state after a copy via the secondary space with a delayed reload of %cr1 and %cr7. There are three hardware address spaces that can cause a DAT exception, primary, secondary and home space. The exception can be related to four different fault types: user space fault, vdso fault, kernel fault, and the gmap faults. Dependent on the set_fs state and normal vs. sacf mode there are a number of fault combinations: 1) user address space fault via the primary ASCE 2) gmap address space fault via the primary ASCE 3) kernel address space fault via the primary ASCE for machines with MVCOS and set_fs(KERNEL_DS) 4) vdso address space faults via the secondary ASCE with an invalid address while running in secondary space in problem state 5) user address space fault via the secondary ASCE for user-copy based on the secondary space mode, e.g. futex_ops or strnlen_user 6) kernel address space fault via the secondary ASCE for user-copy with secondary space mode with set_fs(KERNEL_DS) 7) kernel address space fault via the primary ASCE for user-copy with secondary space mode with set_fs(USER_DS) on machines without MVCOS. 8) kernel address space fault via the home space ASCE Replace user_space_fault() with a new function get_fault_type() that can distinguish all four different fault types. With these changes the futex atomic ops from the kernel and the strnlen_user will get a little bit slower, as well as the old style uaccess with MVCP/MVCS. All user accesses based on MVCOS will be as fast as before. On the positive side, the user space vdso code is a lot faster and Linux ceases to use the complicated AR mode. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
2017-08-22 17:08:22 +07:00
# _CIF_ASCE_PRIMARY and/or _CIF_ASCE_SECONDARY set, load user space asce
s390/uaccess: rework uaccess code - fix locking issues The current uaccess code uses a page table walk in some circumstances, e.g. in case of the in atomic futex operations or if running on old hardware which doesn't support the mvcos instruction. However it turned out that the page table walk code does not correctly lock page tables when accessing page table entries. In other words: a different cpu may invalidate a page table entry while the current cpu inspects the pte. This may lead to random data corruption. Adding correct locking however isn't trivial for all uaccess operations. Especially copy_in_user() is problematic since that requires to hold at least two locks, but must be protected against ABBA deadlock when a different cpu also performs a copy_in_user() operation. So the solution is a different approach where we change address spaces: User space runs in primary address mode, or access register mode within vdso code, like it currently already does. The kernel usually also runs in home space mode, however when accessing user space the kernel switches to primary or secondary address mode if the mvcos instruction is not available or if a compare-and-swap (futex) instruction on a user space address is performed. KVM however is special, since that requires the kernel to run in home address space while implicitly accessing user space with the sie instruction. So we end up with: User space: - runs in primary or access register mode - cr1 contains the user asce - cr7 contains the user asce - cr13 contains the kernel asce Kernel space: - runs in home space mode - cr1 contains the user or kernel asce -> the kernel asce is loaded when a uaccess requires primary or secondary address mode - cr7 contains the user or kernel asce, (changed with set_fs()) - cr13 contains the kernel asce In case of uaccess the kernel changes to: - primary space mode in case of a uaccess (copy_to_user) and uses e.g. the mvcp instruction to access user space. However the kernel will stay in home space mode if the mvcos instruction is available - secondary space mode in case of futex atomic operations, so that the instructions come from primary address space and data from secondary space In case of kvm the kernel runs in home space mode, but cr1 gets switched to contain the gmap asce before the sie instruction gets executed. When the sie instruction is finished cr1 will be switched back to contain the user asce. A context switch between two processes will always load the kernel asce for the next process in cr1. So the first exit to user space is a bit more expensive (one extra load control register instruction) than before, however keeps the code rather simple. In sum this means there is no need to perform any error prone page table walks anymore when accessing user space. The patch seems to be rather large, however it mainly removes the the page table walk code and restores the previously deleted "standard" uaccess code, with a couple of changes. The uaccess without mvcos mode can be enforced with the "uaccess_primary" kernel parameter. Reported-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-03-21 16:42:25 +07:00
#
s390: restore address space when returning to user space Unbalanced set_fs usages (e.g. early exit from a function and a forgotten set_fs(USER_DS) call) may lead to a situation where the secondary asce is the kernel space asce when returning to user space. This would allow user space to modify kernel space at will. This would only be possible with the above mentioned kernel bug, however we can detect this and fix the secondary asce before returning to user space. Therefore a new TIF_ASCE_SECONDARY which is used within set_fs. When returning to user space check if TIF_ASCE_SECONDARY is set, which would indicate a bug. If it is set print a message to the console, fixup the secondary asce, and then return to user space. This is similar to what is being discussed for x86 and arm: "[RFC] syscalls: Restore address limit after a syscall". Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-02-17 14:13:28 +07:00
.Lsysc_asce:
s390: remove all code using the access register mode The vdso code for the getcpu() and the clock_gettime() call use the access register mode to access the per-CPU vdso data page with the current code. An alternative to the complicated AR mode is to use the secondary space mode. This makes the vdso faster and quite a bit simpler. The downside is that the uaccess code has to be changed quite a bit. Which instructions are used depends on the machine and what kind of uaccess operation is requested. The instruction dictates which ASCE value needs to be loaded into %cr1 and %cr7. The different cases: * User copy with MVCOS for z10 and newer machines The MVCOS instruction can copy between the primary space (aka user) and the home space (aka kernel) directly. For set_fs(KERNEL_DS) the kernel ASCE is loaded into %cr1. For set_fs(USER_DS) the user space is already loaded in %cr1. * User copy with MVCP/MVCS for older machines To be able to execute the MVCP/MVCS instructions the kernel needs to switch to primary mode. The control register %cr1 has to be set to the kernel ASCE and %cr7 to either the kernel ASCE or the user ASCE dependent on set_fs(KERNEL_DS) vs set_fs(USER_DS). * Data access in the user address space for strnlen / futex To use "normal" instruction with data from the user address space the secondary space mode is used. The kernel needs to switch to primary mode, %cr1 has to contain the kernel ASCE and %cr7 either the user ASCE or the kernel ASCE, dependent on set_fs. To load a new value into %cr1 or %cr7 is an expensive operation, the kernel tries to be lazy about it. E.g. for multiple user copies in a row with MVCP/MVCS the replacement of the vdso ASCE in %cr7 with the user ASCE is done only once. On return to user space a CPU bit is checked that loads the vdso ASCE again. To enable and disable the data access via the secondary space two new functions are added, enable_sacf_uaccess and disable_sacf_uaccess. The fact that a context is in secondary space uaccess mode is stored in the mm_segment_t value for the task. The code of an interrupt may use set_fs as long as it returns to the previous state it got with get_fs with another call to set_fs. The code in finish_arch_post_lock_switch simply has to do a set_fs with the current mm_segment_t value for the task. For CPUs with MVCOS: CPU running in | %cr1 ASCE | %cr7 ASCE | --------------------------------------|-----------|-----------| user space | user | vdso | kernel, USER_DS, normal-mode | user | vdso | kernel, USER_DS, normal-mode, lazy | user | user | kernel, USER_DS, sacf-mode | kernel | user | kernel, KERNEL_DS, normal-mode | kernel | vdso | kernel, KERNEL_DS, normal-mode, lazy | kernel | kernel | kernel, KERNEL_DS, sacf-mode | kernel | kernel | For CPUs without MVCOS: CPU running in | %cr1 ASCE | %cr7 ASCE | --------------------------------------|-----------|-----------| user space | user | vdso | kernel, USER_DS, normal-mode | user | vdso | kernel, USER_DS, normal-mode lazy | kernel | user | kernel, USER_DS, sacf-mode | kernel | user | kernel, KERNEL_DS, normal-mode | kernel | vdso | kernel, KERNEL_DS, normal-mode, lazy | kernel | kernel | kernel, KERNEL_DS, sacf-mode | kernel | kernel | The lines with "lazy" refer to the state after a copy via the secondary space with a delayed reload of %cr1 and %cr7. There are three hardware address spaces that can cause a DAT exception, primary, secondary and home space. The exception can be related to four different fault types: user space fault, vdso fault, kernel fault, and the gmap faults. Dependent on the set_fs state and normal vs. sacf mode there are a number of fault combinations: 1) user address space fault via the primary ASCE 2) gmap address space fault via the primary ASCE 3) kernel address space fault via the primary ASCE for machines with MVCOS and set_fs(KERNEL_DS) 4) vdso address space faults via the secondary ASCE with an invalid address while running in secondary space in problem state 5) user address space fault via the secondary ASCE for user-copy based on the secondary space mode, e.g. futex_ops or strnlen_user 6) kernel address space fault via the secondary ASCE for user-copy with secondary space mode with set_fs(KERNEL_DS) 7) kernel address space fault via the primary ASCE for user-copy with secondary space mode with set_fs(USER_DS) on machines without MVCOS. 8) kernel address space fault via the home space ASCE Replace user_space_fault() with a new function get_fault_type() that can distinguish all four different fault types. With these changes the futex atomic ops from the kernel and the strnlen_user will get a little bit slower, as well as the old style uaccess with MVCP/MVCS. All user accesses based on MVCOS will be as fast as before. On the positive side, the user space vdso code is a lot faster and Linux ceases to use the complicated AR mode. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
2017-08-22 17:08:22 +07:00
ni __LC_CPU_FLAGS+7,255-_CIF_ASCE_SECONDARY
lctlg %c7,%c7,__LC_VDSO_ASCE # load secondary asce
TSTMSK __LC_CPU_FLAGS,_CIF_ASCE_PRIMARY
jz .Lsysc_return
#ifndef CONFIG_HAVE_MARCH_Z10_FEATURES
tm __LC_STFLE_FAC_LIST+3,0x10 # has MVCOS ?
jnz .Lsysc_set_fs_fixup
s390: rename CIF_ASCE to CIF_ASCE_PRIMARY This is just a preparation patch in order to keep the "restore address space after syscall" patch small. Rename CIF_ASCE to CIF_ASCE_PRIMARY to be unique and specific when introducing a second CIF_ASCE_SECONDARY CIF flag. Suggested-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-02-17 14:12:30 +07:00
ni __LC_CPU_FLAGS+7,255-_CIF_ASCE_PRIMARY
s390/uaccess: rework uaccess code - fix locking issues The current uaccess code uses a page table walk in some circumstances, e.g. in case of the in atomic futex operations or if running on old hardware which doesn't support the mvcos instruction. However it turned out that the page table walk code does not correctly lock page tables when accessing page table entries. In other words: a different cpu may invalidate a page table entry while the current cpu inspects the pte. This may lead to random data corruption. Adding correct locking however isn't trivial for all uaccess operations. Especially copy_in_user() is problematic since that requires to hold at least two locks, but must be protected against ABBA deadlock when a different cpu also performs a copy_in_user() operation. So the solution is a different approach where we change address spaces: User space runs in primary address mode, or access register mode within vdso code, like it currently already does. The kernel usually also runs in home space mode, however when accessing user space the kernel switches to primary or secondary address mode if the mvcos instruction is not available or if a compare-and-swap (futex) instruction on a user space address is performed. KVM however is special, since that requires the kernel to run in home address space while implicitly accessing user space with the sie instruction. So we end up with: User space: - runs in primary or access register mode - cr1 contains the user asce - cr7 contains the user asce - cr13 contains the kernel asce Kernel space: - runs in home space mode - cr1 contains the user or kernel asce -> the kernel asce is loaded when a uaccess requires primary or secondary address mode - cr7 contains the user or kernel asce, (changed with set_fs()) - cr13 contains the kernel asce In case of uaccess the kernel changes to: - primary space mode in case of a uaccess (copy_to_user) and uses e.g. the mvcp instruction to access user space. However the kernel will stay in home space mode if the mvcos instruction is available - secondary space mode in case of futex atomic operations, so that the instructions come from primary address space and data from secondary space In case of kvm the kernel runs in home space mode, but cr1 gets switched to contain the gmap asce before the sie instruction gets executed. When the sie instruction is finished cr1 will be switched back to contain the user asce. A context switch between two processes will always load the kernel asce for the next process in cr1. So the first exit to user space is a bit more expensive (one extra load control register instruction) than before, however keeps the code rather simple. In sum this means there is no need to perform any error prone page table walks anymore when accessing user space. The patch seems to be rather large, however it mainly removes the the page table walk code and restores the previously deleted "standard" uaccess code, with a couple of changes. The uaccess without mvcos mode can be enforced with the "uaccess_primary" kernel parameter. Reported-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-03-21 16:42:25 +07:00
lctlg %c1,%c1,__LC_USER_ASCE # load primary asce
s390: remove all code using the access register mode The vdso code for the getcpu() and the clock_gettime() call use the access register mode to access the per-CPU vdso data page with the current code. An alternative to the complicated AR mode is to use the secondary space mode. This makes the vdso faster and quite a bit simpler. The downside is that the uaccess code has to be changed quite a bit. Which instructions are used depends on the machine and what kind of uaccess operation is requested. The instruction dictates which ASCE value needs to be loaded into %cr1 and %cr7. The different cases: * User copy with MVCOS for z10 and newer machines The MVCOS instruction can copy between the primary space (aka user) and the home space (aka kernel) directly. For set_fs(KERNEL_DS) the kernel ASCE is loaded into %cr1. For set_fs(USER_DS) the user space is already loaded in %cr1. * User copy with MVCP/MVCS for older machines To be able to execute the MVCP/MVCS instructions the kernel needs to switch to primary mode. The control register %cr1 has to be set to the kernel ASCE and %cr7 to either the kernel ASCE or the user ASCE dependent on set_fs(KERNEL_DS) vs set_fs(USER_DS). * Data access in the user address space for strnlen / futex To use "normal" instruction with data from the user address space the secondary space mode is used. The kernel needs to switch to primary mode, %cr1 has to contain the kernel ASCE and %cr7 either the user ASCE or the kernel ASCE, dependent on set_fs. To load a new value into %cr1 or %cr7 is an expensive operation, the kernel tries to be lazy about it. E.g. for multiple user copies in a row with MVCP/MVCS the replacement of the vdso ASCE in %cr7 with the user ASCE is done only once. On return to user space a CPU bit is checked that loads the vdso ASCE again. To enable and disable the data access via the secondary space two new functions are added, enable_sacf_uaccess and disable_sacf_uaccess. The fact that a context is in secondary space uaccess mode is stored in the mm_segment_t value for the task. The code of an interrupt may use set_fs as long as it returns to the previous state it got with get_fs with another call to set_fs. The code in finish_arch_post_lock_switch simply has to do a set_fs with the current mm_segment_t value for the task. For CPUs with MVCOS: CPU running in | %cr1 ASCE | %cr7 ASCE | --------------------------------------|-----------|-----------| user space | user | vdso | kernel, USER_DS, normal-mode | user | vdso | kernel, USER_DS, normal-mode, lazy | user | user | kernel, USER_DS, sacf-mode | kernel | user | kernel, KERNEL_DS, normal-mode | kernel | vdso | kernel, KERNEL_DS, normal-mode, lazy | kernel | kernel | kernel, KERNEL_DS, sacf-mode | kernel | kernel | For CPUs without MVCOS: CPU running in | %cr1 ASCE | %cr7 ASCE | --------------------------------------|-----------|-----------| user space | user | vdso | kernel, USER_DS, normal-mode | user | vdso | kernel, USER_DS, normal-mode lazy | kernel | user | kernel, USER_DS, sacf-mode | kernel | user | kernel, KERNEL_DS, normal-mode | kernel | vdso | kernel, KERNEL_DS, normal-mode, lazy | kernel | kernel | kernel, KERNEL_DS, sacf-mode | kernel | kernel | The lines with "lazy" refer to the state after a copy via the secondary space with a delayed reload of %cr1 and %cr7. There are three hardware address spaces that can cause a DAT exception, primary, secondary and home space. The exception can be related to four different fault types: user space fault, vdso fault, kernel fault, and the gmap faults. Dependent on the set_fs state and normal vs. sacf mode there are a number of fault combinations: 1) user address space fault via the primary ASCE 2) gmap address space fault via the primary ASCE 3) kernel address space fault via the primary ASCE for machines with MVCOS and set_fs(KERNEL_DS) 4) vdso address space faults via the secondary ASCE with an invalid address while running in secondary space in problem state 5) user address space fault via the secondary ASCE for user-copy based on the secondary space mode, e.g. futex_ops or strnlen_user 6) kernel address space fault via the secondary ASCE for user-copy with secondary space mode with set_fs(KERNEL_DS) 7) kernel address space fault via the primary ASCE for user-copy with secondary space mode with set_fs(USER_DS) on machines without MVCOS. 8) kernel address space fault via the home space ASCE Replace user_space_fault() with a new function get_fault_type() that can distinguish all four different fault types. With these changes the futex atomic ops from the kernel and the strnlen_user will get a little bit slower, as well as the old style uaccess with MVCP/MVCS. All user accesses based on MVCOS will be as fast as before. On the positive side, the user space vdso code is a lot faster and Linux ceases to use the complicated AR mode. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
2017-08-22 17:08:22 +07:00
j .Lsysc_return
.Lsysc_set_fs_fixup:
#endif
s390: restore address space when returning to user space Unbalanced set_fs usages (e.g. early exit from a function and a forgotten set_fs(USER_DS) call) may lead to a situation where the secondary asce is the kernel space asce when returning to user space. This would allow user space to modify kernel space at will. This would only be possible with the above mentioned kernel bug, however we can detect this and fix the secondary asce before returning to user space. Therefore a new TIF_ASCE_SECONDARY which is used within set_fs. When returning to user space check if TIF_ASCE_SECONDARY is set, which would indicate a bug. If it is set print a message to the console, fixup the secondary asce, and then return to user space. This is similar to what is being discussed for x86 and arm: "[RFC] syscalls: Restore address limit after a syscall". Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-02-17 14:13:28 +07:00
larl %r14,.Lsysc_return
jg set_fs_fixup
s390/uaccess: rework uaccess code - fix locking issues The current uaccess code uses a page table walk in some circumstances, e.g. in case of the in atomic futex operations or if running on old hardware which doesn't support the mvcos instruction. However it turned out that the page table walk code does not correctly lock page tables when accessing page table entries. In other words: a different cpu may invalidate a page table entry while the current cpu inspects the pte. This may lead to random data corruption. Adding correct locking however isn't trivial for all uaccess operations. Especially copy_in_user() is problematic since that requires to hold at least two locks, but must be protected against ABBA deadlock when a different cpu also performs a copy_in_user() operation. So the solution is a different approach where we change address spaces: User space runs in primary address mode, or access register mode within vdso code, like it currently already does. The kernel usually also runs in home space mode, however when accessing user space the kernel switches to primary or secondary address mode if the mvcos instruction is not available or if a compare-and-swap (futex) instruction on a user space address is performed. KVM however is special, since that requires the kernel to run in home address space while implicitly accessing user space with the sie instruction. So we end up with: User space: - runs in primary or access register mode - cr1 contains the user asce - cr7 contains the user asce - cr13 contains the kernel asce Kernel space: - runs in home space mode - cr1 contains the user or kernel asce -> the kernel asce is loaded when a uaccess requires primary or secondary address mode - cr7 contains the user or kernel asce, (changed with set_fs()) - cr13 contains the kernel asce In case of uaccess the kernel changes to: - primary space mode in case of a uaccess (copy_to_user) and uses e.g. the mvcp instruction to access user space. However the kernel will stay in home space mode if the mvcos instruction is available - secondary space mode in case of futex atomic operations, so that the instructions come from primary address space and data from secondary space In case of kvm the kernel runs in home space mode, but cr1 gets switched to contain the gmap asce before the sie instruction gets executed. When the sie instruction is finished cr1 will be switched back to contain the user asce. A context switch between two processes will always load the kernel asce for the next process in cr1. So the first exit to user space is a bit more expensive (one extra load control register instruction) than before, however keeps the code rather simple. In sum this means there is no need to perform any error prone page table walks anymore when accessing user space. The patch seems to be rather large, however it mainly removes the the page table walk code and restores the previously deleted "standard" uaccess code, with a couple of changes. The uaccess without mvcos mode can be enforced with the "uaccess_primary" kernel parameter. Reported-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-03-21 16:42:25 +07:00
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
#
# CIF_FPU is set, restore floating-point controls and floating-point registers.
#
.Lsysc_vxrs:
larl %r14,.Lsysc_return
jg load_fpu_regs
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#
signals: s390: renumber TIF_RESTORE_SIGMASK TIF_RESTORE_SIGMASK no longer needs to be in the _TIF_WORK_* masks. Those low bits are scarce, and are all used up now. Renumber TIF_RESTORE_SIGMASK to free one up. Signed-off-by: Roland McGrath <roland@redhat.com> Cc: Oleg Nesterov <oleg@tv-sign.ru> Cc: Ingo Molnar <mingo@elte.hu> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: "Luck, Tony" <tony.luck@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 14:53:08 +07:00
# _TIF_SIGPENDING is set, call do_signal
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lsysc_sigpending:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lgr %r2,%r11 # pass pointer to pt_regs
brasl %r14,do_signal
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __PT_FLAGS(%r11),_PIF_SYSCALL
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jno .Lsysc_return
s390: add no-execute support Bit 0x100 of a page table, segment table of region table entry can be used to disallow code execution for the virtual addresses associated with the entry. There is one tricky bit, the system call to return from a signal is part of the signal frame written to the user stack. With a non-executable stack this would stop working. To avoid breaking things the protection fault handler checks the opcode that caused the fault for 0x0a77 (sys_sigreturn) and 0x0aad (sys_rt_sigreturn) and injects a system call. This is preferable to the alternative solution with a stub function in the vdso because it works for vdso=off and statically linked binaries as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-03-22 16:54:24 +07:00
.Lsysc_do_syscall:
lghi %r13,__TASK_thread
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lmg %r2,%r7,__PT_R2(%r11) # load svc arguments
s390: add no-execute support Bit 0x100 of a page table, segment table of region table entry can be used to disallow code execution for the virtual addresses associated with the entry. There is one tricky bit, the system call to return from a signal is part of the signal frame written to the user stack. With a non-executable stack this would stop working. To avoid breaking things the protection fault handler checks the opcode that caused the fault for 0x0a77 (sys_sigreturn) and 0x0aad (sys_rt_sigreturn) and injects a system call. This is preferable to the alternative solution with a stub function in the vdso because it works for vdso=off and statically linked binaries as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-03-22 16:54:24 +07:00
lghi %r1,0 # svc 0 returns -ENOSYS
j .Lsysc_do_svc
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
[S390] ptrace changes * System call parameter and result access functions * Add tracehook calls * Split syscall_trace into two functions do_syscall_trace_enter and do_syscall_trace_exit Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2008-10-11 02:33:20 +07:00
#
# _TIF_NOTIFY_RESUME is set, call do_notify_resume
#
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lsysc_notify_resume:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lgr %r2,%r11 # pass pointer to pt_regs
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
larl %r14,.Lsysc_return
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
jg do_notify_resume
[S390] ptrace changes * System call parameter and result access functions * Add tracehook calls * Split syscall_trace into two functions do_syscall_trace_enter and do_syscall_trace_exit Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2008-10-11 02:33:20 +07:00
s390/uprobes: architecture backend for uprobes Signed-off-by: Jan Willeke <willeke@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-09-22 21:39:06 +07:00
#
# _TIF_UPROBE is set, call uprobe_notify_resume
#
#ifdef CONFIG_UPROBES
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lsysc_uprobe_notify:
s390/uprobes: architecture backend for uprobes Signed-off-by: Jan Willeke <willeke@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-09-22 21:39:06 +07:00
lgr %r2,%r11 # pass pointer to pt_regs
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
larl %r14,.Lsysc_return
s390/uprobes: architecture backend for uprobes Signed-off-by: Jan Willeke <willeke@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-09-22 21:39:06 +07:00
jg uprobe_notify_resume
#endif
s390: add a system call for guarded storage This adds a new system call to enable the use of guarded storage for user space processes. The system call takes two arguments, a command and pointer to a guarded storage control block: s390_guarded_storage(int command, struct gs_cb *gs_cb); The second argument is relevant only for the GS_SET_BC_CB command. The commands in detail: 0 - GS_ENABLE Enable the guarded storage facility for the current task. The initial content of the guarded storage control block will be all zeros. After the enablement the user space code can use load-guarded-storage-controls instruction (LGSC) to load an arbitrary control block. While a task is enabled the kernel will save and restore the current content of the guarded storage registers on context switch. 1 - GS_DISABLE Disables the use of the guarded storage facility for the current task. The kernel will cease to save and restore the content of the guarded storage registers, the task specific content of these registers is lost. 2 - GS_SET_BC_CB Set a broadcast guarded storage control block. This is called per thread and stores a specific guarded storage control block in the task struct of the current task. This control block will be used for the broadcast event GS_BROADCAST. 3 - GS_CLEAR_BC_CB Clears the broadcast guarded storage control block. The guarded- storage control block is removed from the task struct that was established by GS_SET_BC_CB. 4 - GS_BROADCAST Sends a broadcast to all thread siblings of the current task. Every sibling that has established a broadcast guarded storage control block will load this control block and will be enabled for guarded storage. The broadcast guarded storage control block is used up, a second broadcast without a refresh of the stored control block with GS_SET_BC_CB will not have any effect. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-01-26 20:10:34 +07:00
#
# _TIF_GUARDED_STORAGE is set, call guarded_storage_load
#
.Lsysc_guarded_storage:
lgr %r2,%r11 # pass pointer to pt_regs
larl %r14,.Lsysc_return
jg gs_load_bc_cb
livepatch/s390: add TIF_PATCH_PENDING thread flag Update a task's patch state when returning from a system call or user space interrupt, or after handling a signal. This greatly increases the chances of a patch operation succeeding. If a task is I/O bound, it can be patched when returning from a system call. If a task is CPU bound, it can be patched when returning from an interrupt. If a task is sleeping on a to-be-patched function, the user can send SIGSTOP and SIGCONT to force it to switch. Since there are two ways the syscall can be restarted on return from a signal handling process, it is important to clear the flag before do_signal() is called. Otherwise we could miss the migration if we used SIGSTOP/SIGCONT procedure or fake signal to migrate patching blocking tasks. If we place our hook to sysc_work label in entry before TIF_SIGPENDING is evaluated we kill two birds with one stone. The task is correctly migrated in all return paths from a syscall. Signed-off-by: Miroslav Benes <mbenes@suse.cz> Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com> Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2017-02-14 08:42:34 +07:00
#
# _TIF_PATCH_PENDING is set, call klp_update_patch_state
#
#ifdef CONFIG_LIVEPATCH
.Lsysc_patch_pending:
lg %r2,__LC_CURRENT # pass pointer to task struct
larl %r14,.Lsysc_return
jg klp_update_patch_state
#endif
s390: add a system call for guarded storage This adds a new system call to enable the use of guarded storage for user space processes. The system call takes two arguments, a command and pointer to a guarded storage control block: s390_guarded_storage(int command, struct gs_cb *gs_cb); The second argument is relevant only for the GS_SET_BC_CB command. The commands in detail: 0 - GS_ENABLE Enable the guarded storage facility for the current task. The initial content of the guarded storage control block will be all zeros. After the enablement the user space code can use load-guarded-storage-controls instruction (LGSC) to load an arbitrary control block. While a task is enabled the kernel will save and restore the current content of the guarded storage registers on context switch. 1 - GS_DISABLE Disables the use of the guarded storage facility for the current task. The kernel will cease to save and restore the content of the guarded storage registers, the task specific content of these registers is lost. 2 - GS_SET_BC_CB Set a broadcast guarded storage control block. This is called per thread and stores a specific guarded storage control block in the task struct of the current task. This control block will be used for the broadcast event GS_BROADCAST. 3 - GS_CLEAR_BC_CB Clears the broadcast guarded storage control block. The guarded- storage control block is removed from the task struct that was established by GS_SET_BC_CB. 4 - GS_BROADCAST Sends a broadcast to all thread siblings of the current task. Every sibling that has established a broadcast guarded storage control block will load this control block and will be enabled for guarded storage. The broadcast guarded storage control block is used up, a second broadcast without a refresh of the stored control block with GS_SET_BC_CB will not have any effect. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-01-26 20:10:34 +07:00
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#
s390: split TIF bits into CIF, PIF and TIF bits The oi and ni instructions used in entry[64].S to set and clear bits in the thread-flags are not guaranteed to be atomic in regard to other CPUs. Split the TIF bits into CPU, pt_regs and thread-info specific bits. Updates on the TIF bits are done with atomic instructions, updates on CPU and pt_regs bits are done with non-atomic instructions. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-04-15 17:55:07 +07:00
# _PIF_PER_TRAP is set, call do_per_trap
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lsysc_singlestep:
s390: split TIF bits into CIF, PIF and TIF bits The oi and ni instructions used in entry[64].S to set and clear bits in the thread-flags are not guaranteed to be atomic in regard to other CPUs. Split the TIF bits into CPU, pt_regs and thread-info specific bits. Updates on the TIF bits are done with atomic instructions, updates on CPU and pt_regs bits are done with non-atomic instructions. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-04-15 17:55:07 +07:00
ni __PT_FLAGS+7(%r11),255-_PIF_PER_TRAP
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lgr %r2,%r11 # pass pointer to pt_regs
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
larl %r14,.Lsysc_return
[S390] ptrace cleanup Overhaul program event recording and the code dealing with the ptrace user space interface. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-01-05 18:48:10 +07:00
jg do_per_trap
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
s390/kvm: avoid global config of vm.alloc_pgste=1 The system control vm.alloc_pgste is used to control the size of the page tables, either 2K or 4K. The idea is that a KVM host sets the vm.alloc_pgste control to 1 which causes *all* new processes to run with 4K page tables. For a non-kvm system the control should stay off to save on memory used for page tables. Trouble is that distributions choose to set the control globally to be able to run KVM guests. This wastes memory on non-KVM systems. Introduce the PT_S390_PGSTE ELF segment type to "mark" the qemu executable with it. All executables with this (empty) segment in its ELF phdr array will be started with 4K page tables. Any executable without PT_S390_PGSTE will run with the default 2K page tables. This removes the need to set vm.alloc_pgste=1 for a KVM host and minimizes the waste of memory for page tables. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-06-07 19:10:24 +07:00
#
# _PIF_SYSCALL_RESTART is set, repeat the current system call
#
.Lsysc_syscall_restart:
ni __PT_FLAGS+7(%r11),255-_PIF_SYSCALL_RESTART
lmg %r1,%r7,__PT_R1(%r11) # load svc arguments
lg %r2,__PT_ORIG_GPR2(%r11)
j .Lsysc_do_svc
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#
[S390] ptrace changes * System call parameter and result access functions * Add tracehook calls * Split syscall_trace into two functions do_syscall_trace_enter and do_syscall_trace_exit Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2008-10-11 02:33:20 +07:00
# call tracehook_report_syscall_entry/tracehook_report_syscall_exit before
# and after the system call
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lsysc_tracesys:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lgr %r2,%r11 # pass pointer to pt_regs
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
la %r3,0
[S390] cleanup trap handling Move the program interruption code and the translation exception identifier to the pt_regs structure as 'int_code' and 'int_parm_long' and make the first level interrupt handler in entry[64].S store the two values. That makes it possible to drop 'prot_addr' and 'trap_no' from the thread_struct and to reduce the number of arguments to a lot of functions. Finally un-inline do_trap. Overall this saves 5812 bytes in the .text section of the 64 bit kernel. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:18 +07:00
llgh %r0,__PT_INT_CODE+2(%r11)
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stg %r0,__PT_R2(%r11)
[S390] ptrace changes * System call parameter and result access functions * Add tracehook calls * Split syscall_trace into two functions do_syscall_trace_enter and do_syscall_trace_exit Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2008-10-11 02:33:20 +07:00
brasl %r14,do_syscall_trace_enter
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
lghi %r0,NR_syscalls
[S390] ptrace changes * System call parameter and result access functions * Add tracehook calls * Split syscall_trace into two functions do_syscall_trace_enter and do_syscall_trace_exit Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2008-10-11 02:33:20 +07:00
clgr %r0,%r2
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jnh .Lsysc_tracenogo
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
sllg %r8,%r2,2
lgf %r9,0(%r8,%r10)
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lsysc_tracego:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lmg %r3,%r7,__PT_R3(%r11)
stg %r7,STACK_FRAME_OVERHEAD(%r15)
lg %r2,__PT_ORIG_GPR2(%r11)
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
BASR_R14_R9 # call sys_xxx
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stg %r2,__PT_R2(%r11) # store return value
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lsysc_tracenogo:
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __TI_flags(%r12),_TIF_TRACE
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jz .Lsysc_return
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lgr %r2,%r11 # pass pointer to pt_regs
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
larl %r14,.Lsysc_return
[S390] ptrace changes * System call parameter and result access functions * Add tracehook calls * Split syscall_trace into two functions do_syscall_trace_enter and do_syscall_trace_exit Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2008-10-11 02:33:20 +07:00
jg do_syscall_trace_exit
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#
# a new process exits the kernel with ret_from_fork
#
[S390] fix s390 assembler code alignments The alignment is missing for various global symbols in s390 assembly code. With a recent gcc and an instruction like stgrl this can lead to a specification exception if the instruction uses such a mis-aligned address. Specify the alignment explicitely and while add it define __ALIGN for s390 and use the ENTRY define to save some lines of code. Signed-off-by: Jan Glauber <jang@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-07-24 15:48:19 +07:00
ENTRY(ret_from_fork)
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
la %r11,STACK_FRAME_OVERHEAD(%r15)
s390: move thread_info into task_struct This is the s390 variant of commit 15f4eae70d36 ("x86: Move thread_info into task_struct"). Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-11-08 17:08:26 +07:00
lg %r12,__LC_CURRENT
s390: fold kernel_thread_helper() into ret_from_fork() ... and don't bother with syscall return path in case of kernel threads. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2012-09-11 05:03:41 +07:00
brasl %r14,schedule_tail
TRACE_IRQS_ON
ssm __LC_SVC_NEW_PSW # reenable interrupts
s390: switch to saner kernel_execve() semantics Acked-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2012-10-12 02:30:14 +07:00
tm __PT_PSW+1(%r11),0x01 # forking a kernel thread ?
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jne .Lsysc_tracenogo
s390: switch to saner kernel_execve() semantics Acked-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2012-10-12 02:30:14 +07:00
# it's a kernel thread
lmg %r9,%r10,__PT_R9(%r11) # load gprs
s390: fold kernel_thread_helper() into ret_from_fork() ... and don't bother with syscall return path in case of kernel threads. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2012-09-11 05:03:41 +07:00
ENTRY(kernel_thread_starter)
la %r2,0(%r10)
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
BASR_R14_R9
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
j .Lsysc_tracenogo
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
/*
* Program check handler routine
*/
[S390] fix s390 assembler code alignments The alignment is missing for various global symbols in s390 assembly code. With a recent gcc and an instruction like stgrl this can lead to a specification exception if the instruction uses such a mis-aligned address. Specify the alignment explicitely and while add it define __ALIGN for s390 and use the ENTRY define to save some lines of code. Signed-off-by: Jan Glauber <jang@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-07-24 15:48:19 +07:00
ENTRY(pgm_check_handler)
[S390] Remove config options. On s390 we always want to run with precise cputime accounting. Remove the config options VIRT_TIMER and VIRT_CPU_ACCOUNTING. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2008-12-25 19:39:25 +07:00
stpt __LC_SYNC_ENTER_TIMER
s390: add options to change branch prediction behaviour for the kernel Add the PPA instruction to the system entry and exit path to switch the kernel to a different branch prediction behaviour. The instructions are added via CPU alternatives and can be disabled with the "nospec" or the "nobp=0" kernel parameter. If the default behaviour selected with CONFIG_KERNEL_NOBP is set to "n" then the "nobp=1" parameter can be used to enable the changed kernel branch prediction. Acked-by: Cornelia Huck <cohuck@redhat.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:11:45 +07:00
BPOFF
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stmg %r8,%r15,__LC_SAVE_AREA_SYNC
lg %r10,__LC_LAST_BREAK
s390: move thread_info into task_struct This is the s390 variant of commit 15f4eae70d36 ("x86: Move thread_info into task_struct"). Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-11-08 17:08:26 +07:00
lg %r12,__LC_CURRENT
s390/mm,kvm: improve detection of KVM guest faults The identification of guest fault currently relies on the PF_VCPU flag. This is set in guest_entry_irqoff and cleared in guest_exit_irqoff. Both functions are called by __vcpu_run, the PF_VCPU flag is set for quite a lot of kernel code outside of the guest execution. Replace the PF_VCPU scheme with the PIF_GUEST_FAULT in the pt_regs and make the program check handler code in entry.S set the bit only for exception that occurred between the .Lsie_gmap and .Lsie_done labels. Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
2017-10-05 13:44:26 +07:00
lghi %r11,0
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
larl %r13,cleanup_critical
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lmg %r8,%r9,__LC_PGM_OLD_PSW
tmhh %r8,0x0001 # test problem state bit
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
jnz 2f # -> fault in user space
#if IS_ENABLED(CONFIG_KVM)
s390/kvm: fix detection of guest machine checks The new detection code for guest machine checks added a check based on %r11 to .Lcleanup_sie to distinguish between normal asynchronous interrupts and machine checks. But the funtion is called from the program check handler as well with an undefined value in %r11. The effect is that all program exceptions pointing to the SIE instruction will set the CIF_MCCK_GUEST bit. The bit stays set for the CPU until the next machine check comes in which will incorrectly be interpreted as a guest machine check. The simplest fix is to stop using .Lcleanup_sie in the program check handler and duplicate a few instructions. Fixes: c929500d7a5a ("s390/nmi: s390: New low level handling for machine check happening in guest") Cc: <stable@vger.kernel.org> # v4.13+ Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-10-05 13:29:47 +07:00
# cleanup critical section for program checks in sie64a
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
lgr %r14,%r9
slg %r14,BASED(.Lsie_critical_start)
clg %r14,BASED(.Lsie_critical_length)
jhe 0f
s390/kvm: improve stack frame constants in entry.S The code in sie64a uses the stack frame passed to the function to store some temporary data in the empty1 array (see struct stack_frame in asm/processor.h. Replace the __SF_EMPTY+x constants with a properly defined offset: s/__SF_EMPTY/__SF_SIE_CONTROL/, s/__SF_EMPTY+8/__SF_SIE_SAVEAREA/, s/__SF_EMPTY+16/__SF_SIE_REASON/, s/__SF_EMPTY+24/__SF_SIE_FLAGS/. Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-03-20 19:33:43 +07:00
lg %r14,__SF_SIE_CONTROL(%r15) # get control block pointer
s390/kvm: fix detection of guest machine checks The new detection code for guest machine checks added a check based on %r11 to .Lcleanup_sie to distinguish between normal asynchronous interrupts and machine checks. But the funtion is called from the program check handler as well with an undefined value in %r11. The effect is that all program exceptions pointing to the SIE instruction will set the CIF_MCCK_GUEST bit. The bit stays set for the CPU until the next machine check comes in which will incorrectly be interpreted as a guest machine check. The simplest fix is to stop using .Lcleanup_sie in the program check handler and duplicate a few instructions. Fixes: c929500d7a5a ("s390/nmi: s390: New low level handling for machine check happening in guest") Cc: <stable@vger.kernel.org> # v4.13+ Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-10-05 13:29:47 +07:00
ni __SIE_PROG0C+3(%r14),0xfe # no longer in SIE
lctlg %c1,%c1,__LC_USER_ASCE # load primary asce
larl %r9,sie_exit # skip forward to sie_exit
s390/mm,kvm: improve detection of KVM guest faults The identification of guest fault currently relies on the PF_VCPU flag. This is set in guest_entry_irqoff and cleared in guest_exit_irqoff. Both functions are called by __vcpu_run, the PF_VCPU flag is set for quite a lot of kernel code outside of the guest execution. Replace the PF_VCPU scheme with the PIF_GUEST_FAULT in the pt_regs and make the program check handler code in entry.S set the bit only for exception that occurred between the .Lsie_gmap and .Lsie_done labels. Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
2017-10-05 13:44:26 +07:00
lghi %r11,_PIF_GUEST_FAULT
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
#endif
0: tmhh %r8,0x4000 # PER bit set in old PSW ?
jnz 1f # -> enabled, can't be a double fault
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
tm __LC_PGM_ILC+3,0x80 # check for per exception
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jnz .Lpgm_svcper # -> single stepped svc
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
1: CHECK_STACK STACK_SIZE,__LC_SAVE_AREA_SYNC
s390: lowcore stack pointer offsets Store the stack pointers in the lowcore for the kernel stack, the async stack and the panic stack with the offset required for the first user. This avoids an unnecessary add instruction on the system call path. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-04-24 15:20:43 +07:00
aghi %r15,-(STACK_FRAME_OVERHEAD + __PT_SIZE)
s390: fix in-kernel program checks A program check inside the kernel takes a slightly different path in entry.S compare to a normal user fault. A recent change moved the store of the breaking event address into the path taken for in-kernel program checks as well, but %r14 has not been setup to point to the correct location. A wild store is the consequence. Move the store of the breaking event address to the code path for user space faults. Fixes: 34525e1f7e8d ("s390: store breaking event address only for program checks") Reported-by: Michael Holzheu <holzheu@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-02-28 13:42:01 +07:00
j 4f
s390: store breaking event address only for program checks The principles of operations specifies that the breaking event address is stored to the address 0x110 in the prefix page only for program checks. The last branch in user space is lost as soon as a branch in kernel space is executed after e.g. an svc. This makes it impossible to accurately maintain the breaking event address for a user space process. Simplify the code, just copy the current breaking event address from 0x110 to the task structure for program checks from user space. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-01-25 18:54:17 +07:00
2: UPDATE_VTIME %r14,%r15,__LC_SYNC_ENTER_TIMER
s390: run user space and KVM guests with modified branch prediction Define TIF_ISOLATE_BP and TIF_ISOLATE_BP_GUEST and add the necessary plumbing in entry.S to be able to run user space and KVM guests with limited branch prediction. To switch a user space process to limited branch prediction the s390_isolate_bp() function has to be call, and to run a vCPU of a KVM guest associated with the current task with limited branch prediction call s390_isolate_bp_guest(). Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:36:46 +07:00
BPENTER __TI_flags(%r12),_TIF_ISOLATE_BP
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lg %r15,__LC_KERNEL_STACK
s390: move thread_info into task_struct This is the s390 variant of commit 15f4eae70d36 ("x86: Move thread_info into task_struct"). Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-11-08 17:08:26 +07:00
lgr %r14,%r12
s390: adapt entry.S to the move of thread_struct git commit 0c8c0f03e3a292e031596484275c14cf39c0ab7a "x86/fpu, sched: Dynamically allocate 'struct fpu'" moved the thread_struct to the end of the task_struct. This causes some of the offsets used in entry.S to overflow their instruction operand field. To fix this use aghi to create a dedicated pointer for the thread_struct. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-07-20 15:01:46 +07:00
aghi %r14,__TASK_thread # pointer to thread_struct
s390: add support for transactional memory Allow user-space processes to use transactional execution (TX). If the TX facility is available user space programs can use transactions for fine-grained serialization based on the data objects that are referenced during a transaction. This is useful for lockless data structures and speculative compiler optimizations. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-07-31 16:03:04 +07:00
lghi %r13,__LC_PGM_TDB
tm __LC_PGM_ILC+2,0x02 # check for transaction abort
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
jz 3f
s390: add support for transactional memory Allow user-space processes to use transactional execution (TX). If the TX facility is available user space programs can use transactions for fine-grained serialization based on the data objects that are referenced during a transaction. This is useful for lockless data structures and speculative compiler optimizations. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-07-31 16:03:04 +07:00
mvc __THREAD_trap_tdb(256,%r14),0(%r13)
s390: fix in-kernel program checks A program check inside the kernel takes a slightly different path in entry.S compare to a normal user fault. A recent change moved the store of the breaking event address into the path taken for in-kernel program checks as well, but %r14 has not been setup to point to the correct location. A wild store is the consequence. Move the store of the breaking event address to the code path for user space faults. Fixes: 34525e1f7e8d ("s390: store breaking event address only for program checks") Reported-by: Michael Holzheu <holzheu@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-02-28 13:42:01 +07:00
3: stg %r10,__THREAD_last_break(%r14)
s390/mm,kvm: improve detection of KVM guest faults The identification of guest fault currently relies on the PF_VCPU flag. This is set in guest_entry_irqoff and cleared in guest_exit_irqoff. Both functions are called by __vcpu_run, the PF_VCPU flag is set for quite a lot of kernel code outside of the guest execution. Replace the PF_VCPU scheme with the PIF_GUEST_FAULT in the pt_regs and make the program check handler code in entry.S set the bit only for exception that occurred between the .Lsie_gmap and .Lsie_done labels. Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
2017-10-05 13:44:26 +07:00
4: lgr %r13,%r11
la %r11,STACK_FRAME_OVERHEAD(%r15)
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stmg %r0,%r7,__PT_R0(%r11)
s390: scrub registers on kernel entry and KVM exit Clear all user space registers on entry to the kernel and all KVM guest registers on KVM guest exit if the register does not contain either a parameter or a result value. Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 19:27:30 +07:00
# clear user controlled registers to prevent speculative use
xgr %r0,%r0
xgr %r1,%r1
xgr %r2,%r2
xgr %r3,%r3
xgr %r4,%r4
xgr %r5,%r5
xgr %r6,%r6
xgr %r7,%r7
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
mvc __PT_R8(64,%r11),__LC_SAVE_AREA_SYNC
stmg %r8,%r9,__PT_PSW(%r11)
[S390] cleanup trap handling Move the program interruption code and the translation exception identifier to the pt_regs structure as 'int_code' and 'int_parm_long' and make the first level interrupt handler in entry[64].S store the two values. That makes it possible to drop 'prot_addr' and 'trap_no' from the thread_struct and to reduce the number of arguments to a lot of functions. Finally un-inline do_trap. Overall this saves 5812 bytes in the .text section of the 64 bit kernel. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:18 +07:00
mvc __PT_INT_CODE(4,%r11),__LC_PGM_ILC
mvc __PT_INT_PARM_LONG(8,%r11),__LC_TRANS_EXC_CODE
s390/mm,kvm: improve detection of KVM guest faults The identification of guest fault currently relies on the PF_VCPU flag. This is set in guest_entry_irqoff and cleared in guest_exit_irqoff. Both functions are called by __vcpu_run, the PF_VCPU flag is set for quite a lot of kernel code outside of the guest execution. Replace the PF_VCPU scheme with the PIF_GUEST_FAULT in the pt_regs and make the program check handler code in entry.S set the bit only for exception that occurred between the .Lsie_gmap and .Lsie_done labels. Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
2017-10-05 13:44:26 +07:00
stg %r13,__PT_FLAGS(%r11)
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stg %r10,__PT_ARGS(%r11)
tm __LC_PGM_ILC+3,0x80 # check for per exception
s390: fix in-kernel program checks A program check inside the kernel takes a slightly different path in entry.S compare to a normal user fault. A recent change moved the store of the breaking event address into the path taken for in-kernel program checks as well, but %r14 has not been setup to point to the correct location. A wild store is the consequence. Move the store of the breaking event address to the code path for user space faults. Fixes: 34525e1f7e8d ("s390: store breaking event address only for program checks") Reported-by: Michael Holzheu <holzheu@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-02-28 13:42:01 +07:00
jz 5f
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
tmhh %r8,0x0001 # kernel per event ?
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jz .Lpgm_kprobe
s390: split TIF bits into CIF, PIF and TIF bits The oi and ni instructions used in entry[64].S to set and clear bits in the thread-flags are not guaranteed to be atomic in regard to other CPUs. Split the TIF bits into CPU, pt_regs and thread-info specific bits. Updates on the TIF bits are done with atomic instructions, updates on CPU and pt_regs bits are done with non-atomic instructions. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-04-15 17:55:07 +07:00
oi __PT_FLAGS+7(%r11),_PIF_PER_TRAP
s390: add support for transactional memory Allow user-space processes to use transactional execution (TX). If the TX facility is available user space programs can use transactions for fine-grained serialization based on the data objects that are referenced during a transaction. This is useful for lockless data structures and speculative compiler optimizations. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-07-31 16:03:04 +07:00
mvc __THREAD_per_address(8,%r14),__LC_PER_ADDRESS
s390: rename and split lowcore field per_perc_atmid per_perc_atmid is currently a two-byte field that combines two fields, the PER code and the PER Addressing-and-Translation-Mode Identification (ATMID) Let's make them accessible indepently and also rename per_cause to per_code. Signed-off-by: Jens Freimann <jfrei@linux.vnet.ibm.com> Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
2014-02-26 22:32:46 +07:00
mvc __THREAD_per_cause(2,%r14),__LC_PER_CODE
mvc __THREAD_per_paid(1,%r14),__LC_PER_ACCESS_ID
s390: fix in-kernel program checks A program check inside the kernel takes a slightly different path in entry.S compare to a normal user fault. A recent change moved the store of the breaking event address into the path taken for in-kernel program checks as well, but %r14 has not been setup to point to the correct location. A wild store is the consequence. Move the store of the breaking event address to the code path for user space faults. Fixes: 34525e1f7e8d ("s390: store breaking event address only for program checks") Reported-by: Michael Holzheu <holzheu@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-02-28 13:42:01 +07:00
5: REENABLE_IRQS
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15)
[S390] Fix IRQ tracing in case of PER In case user space is single stepped (PER) the program check handler claims too early that IRQs are enabled on the return path. Subsequent checks will notice that the IRQ mask in the PSW and what lockdep thinks the IRQ mask should be do not correlate and therefore will print a warning to the console and disable lockdep. Fix this by doing all the work within the correct context. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-07-28 00:29:37 +07:00
larl %r1,pgm_check_table
[S390] cleanup trap handling Move the program interruption code and the translation exception identifier to the pt_regs structure as 'int_code' and 'int_parm_long' and make the first level interrupt handler in entry[64].S store the two values. That makes it possible to drop 'prot_addr' and 'trap_no' from the thread_struct and to reduce the number of arguments to a lot of functions. Finally un-inline do_trap. Overall this saves 5812 bytes in the .text section of the 64 bit kernel. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:18 +07:00
llgh %r10,__PT_INT_CODE+2(%r11)
nill %r10,0x007f
s390/traps: preinitialize program check table Preinitialize the program check table, so we can put it into the read-only data section. Also use only four byte entries for the table, since each program check handler resides within the first 2GB. Therefore this reduces the size of the table by 50% on 64 bit builds. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-10-18 23:10:06 +07:00
sll %r10,2
s390/kernel: squeeze a few more cycles out of the system call handler Reorder the instructions of UPDATE_VTIME to improve superscalar execution, remove duplicate checks for problem-state from the asynchronous interrupt handlers, and move the check for problem-state from the synchronous exit path to the program check path as it is only needed for program checks inside the kernel. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:27:48 +07:00
je .Lpgm_return
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
lgf %r9,0(%r10,%r1) # load address of handler routine
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lgr %r2,%r11 # pass pointer to pt_regs
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
BASR_R14_R9 # branch to interrupt-handler
s390/kernel: squeeze a few more cycles out of the system call handler Reorder the instructions of UPDATE_VTIME to improve superscalar execution, remove duplicate checks for problem-state from the asynchronous interrupt handlers, and move the check for problem-state from the synchronous exit path to the program check path as it is only needed for program checks inside the kernel. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:27:48 +07:00
.Lpgm_return:
LOCKDEP_SYS_EXIT
tm __PT_PSW+1(%r11),0x01 # returning to user ?
jno .Lsysc_restore
s390: add no-execute support Bit 0x100 of a page table, segment table of region table entry can be used to disallow code execution for the virtual addresses associated with the entry. There is one tricky bit, the system call to return from a signal is part of the signal frame written to the user stack. With a non-executable stack this would stop working. To avoid breaking things the protection fault handler checks the opcode that caused the fault for 0x0a77 (sys_sigreturn) and 0x0aad (sys_rt_sigreturn) and injects a system call. This is preferable to the alternative solution with a stub function in the vdso because it works for vdso=off and statically linked binaries as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-03-22 16:54:24 +07:00
TSTMSK __PT_FLAGS(%r11),_PIF_SYSCALL
jo .Lsysc_do_syscall
s390/kernel: squeeze a few more cycles out of the system call handler Reorder the instructions of UPDATE_VTIME to improve superscalar execution, remove duplicate checks for problem-state from the asynchronous interrupt handlers, and move the check for problem-state from the synchronous exit path to the program check path as it is only needed for program checks inside the kernel. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:27:48 +07:00
j .Lsysc_tif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
# PER event in supervisor state, must be kprobes
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lpgm_kprobe:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
REENABLE_IRQS
xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15)
lgr %r2,%r11 # pass pointer to pt_regs
brasl %r14,do_per_trap
s390/kernel: squeeze a few more cycles out of the system call handler Reorder the instructions of UPDATE_VTIME to improve superscalar execution, remove duplicate checks for problem-state from the asynchronous interrupt handlers, and move the check for problem-state from the synchronous exit path to the program check path as it is only needed for program checks inside the kernel. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:27:48 +07:00
j .Lpgm_return
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
[S390] add kprobes support. Signed-off-by: Michael Grundy <grundym@us.ibm.com> Signed-off-by: David Wilder <dwilder@us.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2006-09-20 20:58:39 +07:00
#
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
# single stepped system call
[S390] add kprobes support. Signed-off-by: Michael Grundy <grundym@us.ibm.com> Signed-off-by: David Wilder <dwilder@us.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2006-09-20 20:58:39 +07:00
#
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lpgm_svcper:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
mvc __LC_RETURN_PSW(8),__LC_SVC_NEW_PSW
s390/syscall: fix single stepped system calls Fix PER tracing of system calls after git commit 34525e1f7e8dc478 "s390: store breaking event address only for program checks" broke it. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-02-20 15:51:36 +07:00
lghi %r13,__TASK_thread
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
larl %r14,.Lsysc_per
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stg %r14,__LC_RETURN_PSW+8
s390: split TIF bits into CIF, PIF and TIF bits The oi and ni instructions used in entry[64].S to set and clear bits in the thread-flags are not guaranteed to be atomic in regard to other CPUs. Split the TIF bits into CPU, pt_regs and thread-info specific bits. Updates on the TIF bits are done with atomic instructions, updates on CPU and pt_regs bits are done with non-atomic instructions. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-04-15 17:55:07 +07:00
lghi %r14,_PIF_SYSCALL | _PIF_PER_TRAP
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
lpswe __LC_RETURN_PSW # branch to .Lsysc_per and enable irqs
[S390] add kprobes support. Signed-off-by: Michael Grundy <grundym@us.ibm.com> Signed-off-by: David Wilder <dwilder@us.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2006-09-20 20:58:39 +07:00
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
/*
* IO interrupt handler routine
*/
[S390] fix s390 assembler code alignments The alignment is missing for various global symbols in s390 assembly code. With a recent gcc and an instruction like stgrl this can lead to a specification exception if the instruction uses such a mis-aligned address. Specify the alignment explicitely and while add it define __ALIGN for s390 and use the ENTRY define to save some lines of code. Signed-off-by: Jan Glauber <jang@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-07-24 15:48:19 +07:00
ENTRY(io_int_handler)
s390/time: always use stckf instead of stck if available The store clock fast instruction saves a couple of instructions compared to the store clock instruction. Always use stckf instead of stck if it is available. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Reviewed-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-05-09 21:27:39 +07:00
STCK __LC_INT_CLOCK
[PATCH] improve idle cputime accounting Distinguish the cputime of the idle process where idle is actually using cpu cycles from the cputime where idle is sleeping on an enabled wait psw. The former is accounted as system time, the later as idle time. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2008-12-31 21:11:41 +07:00
stpt __LC_ASYNC_ENTER_TIMER
s390: add options to change branch prediction behaviour for the kernel Add the PPA instruction to the system entry and exit path to switch the kernel to a different branch prediction behaviour. The instructions are added via CPU alternatives and can be disabled with the "nospec" or the "nobp=0" kernel parameter. If the default behaviour selected with CONFIG_KERNEL_NOBP is set to "n" then the "nobp=1" parameter can be used to enable the changed kernel branch prediction. Acked-by: Cornelia Huck <cohuck@redhat.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:11:45 +07:00
BPOFF
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stmg %r8,%r15,__LC_SAVE_AREA_ASYNC
s390: move thread_info into task_struct This is the s390 variant of commit 15f4eae70d36 ("x86: Move thread_info into task_struct"). Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-11-08 17:08:26 +07:00
lg %r12,__LC_CURRENT
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
larl %r13,cleanup_critical
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lmg %r8,%r9,__LC_IO_OLD_PSW
s390/nmi: use the normal asynchronous stack for machine checks If a machine checks is received while the CPU is in the kernel, only the s390_do_machine_check function will be called. The call to s390_handle_mcck is postponed until the CPU returns to user space. Because of this it is safe to use the asynchronous stack for machine checks even if the CPU is already handling an interrupt. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:28:14 +07:00
SWITCH_ASYNC __LC_SAVE_AREA_ASYNC,__LC_ASYNC_ENTER_TIMER
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stmg %r0,%r7,__PT_R0(%r11)
s390: scrub registers on kernel entry and KVM exit Clear all user space registers on entry to the kernel and all KVM guest registers on KVM guest exit if the register does not contain either a parameter or a result value. Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 19:27:30 +07:00
# clear user controlled registers to prevent speculative use
xgr %r0,%r0
xgr %r1,%r1
xgr %r2,%r2
xgr %r3,%r3
xgr %r4,%r4
xgr %r5,%r5
xgr %r6,%r6
xgr %r7,%r7
xgr %r10,%r10
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
mvc __PT_R8(64,%r11),__LC_SAVE_AREA_ASYNC
stmg %r8,%r9,__PT_PSW(%r11)
s390/irq: store interrupt information in pt_regs Copy the interrupt parameters from the lowcore to the pt_regs structure in entry[64].S and reduce the arguments of the low level interrupt handler to the pt_regs pointer only. In addition move the test-pending-interrupt loop from do_IRQ to entry[64].S to make sure that interrupt information is always delivered via pt_regs. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-06-17 19:54:02 +07:00
mvc __PT_INT_CODE(12,%r11),__LC_SUBCHANNEL_ID
s390: split TIF bits into CIF, PIF and TIF bits The oi and ni instructions used in entry[64].S to set and clear bits in the thread-flags are not guaranteed to be atomic in regard to other CPUs. Split the TIF bits into CPU, pt_regs and thread-info specific bits. Updates on the TIF bits are done with atomic instructions, updates on CPU and pt_regs bits are done with non-atomic instructions. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-04-15 17:55:07 +07:00
xc __PT_FLAGS(8,%r11),__PT_FLAGS(%r11)
s390/udelay: make udelay have busy loop semantics When using systemtap it was observed that our udelay implementation is rather suboptimal if being called from a kprobe handler installed by systemtap. The problem observed when a kprobe was installed on lock_acquired(). When the probe was hit the kprobe handler did call udelay, which set up an (internal) timer and reenabled interrupts (only the clock comparator interrupt) and waited for the interrupt. This is an optimization to avoid that the cpu is busy looping while waiting that enough time passes. The problem is that the interrupt handler still does call irq_enter()/irq_exit() which then again can lead to a deadlock, since some accounting functions may take locks as well. If one of these locks is the same, which caused lock_acquired() to be called, we have a nice deadlock. This patch reworks the udelay code for the interrupts disabled case to immediately leave the low level interrupt handler when the clock comparator interrupt happens. That way no C code is being called and the deadlock cannot happen anymore. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Reviewed-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-08-15 16:42:21 +07:00
TSTMSK __LC_CPU_FLAGS,_CIF_IGNORE_IRQ
jo .Lio_restore
[PATCH] lockdep: irqtrace subsystem, s390 support irqtrace support for s390. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-07-03 14:24:46 +07:00
TRACE_IRQS_OFF
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15)
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lio_loop:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lgr %r2,%r11 # pass pointer to pt_regs
s390: convert interrupt handling to use generic hardirq With the introduction of PCI it became apparent that s390 should convert to generic hardirqs as too many drivers do not have the correct dependency for GENERIC_HARDIRQS. On the architecture level s390 does not have irq lines. It has external interrupts, I/O interrupts and adapter interrupts. This patch hard-codes all external interrupts as irq #1, all I/O interrupts as irq #2 and all adapter interrupts as irq #3. The additional information from the lowcore associated with the interrupt is stored in the pt_regs of the interrupt frame, where the interrupt handler can pick it up. For PCI/MSI interrupts the adapter interrupt handler scans the relevant bit fields and calls generic_handle_irq with the virtual irq number for the MSI interrupt. Reviewed-by: Sebastian Ott <sebott@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-06-27 14:01:09 +07:00
lghi %r3,IO_INTERRUPT
tm __PT_INT_CODE+8(%r11),0x80 # adapter interrupt ?
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jz .Lio_call
s390: convert interrupt handling to use generic hardirq With the introduction of PCI it became apparent that s390 should convert to generic hardirqs as too many drivers do not have the correct dependency for GENERIC_HARDIRQS. On the architecture level s390 does not have irq lines. It has external interrupts, I/O interrupts and adapter interrupts. This patch hard-codes all external interrupts as irq #1, all I/O interrupts as irq #2 and all adapter interrupts as irq #3. The additional information from the lowcore associated with the interrupt is stored in the pt_regs of the interrupt frame, where the interrupt handler can pick it up. For PCI/MSI interrupts the adapter interrupt handler scans the relevant bit fields and calls generic_handle_irq with the virtual irq number for the MSI interrupt. Reviewed-by: Sebastian Ott <sebott@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-06-27 14:01:09 +07:00
lghi %r3,THIN_INTERRUPT
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lio_call:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
brasl %r14,do_IRQ
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __LC_MACHINE_FLAGS,MACHINE_FLAG_LPAR
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jz .Lio_return
s390/irq: store interrupt information in pt_regs Copy the interrupt parameters from the lowcore to the pt_regs structure in entry[64].S and reduce the arguments of the low level interrupt handler to the pt_regs pointer only. In addition move the test-pending-interrupt loop from do_IRQ to entry[64].S to make sure that interrupt information is always delivered via pt_regs. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-06-17 19:54:02 +07:00
tpi 0
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jz .Lio_return
s390/irq: store interrupt information in pt_regs Copy the interrupt parameters from the lowcore to the pt_regs structure in entry[64].S and reduce the arguments of the low level interrupt handler to the pt_regs pointer only. In addition move the test-pending-interrupt loop from do_IRQ to entry[64].S to make sure that interrupt information is always delivered via pt_regs. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-06-17 19:54:02 +07:00
mvc __PT_INT_CODE(12,%r11),__LC_SUBCHANNEL_ID
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
j .Lio_loop
.Lio_return:
[S390] improve irq tracing code in entry[64].S The system call path in entry[64].S is run with interrupts enabled. Remove the irq tracing check from the system call exit code. If a program check interrupted a context enabled for interrupts do a call to trace_irq_off_caller in the program check handler before branching to the system call exit code. Restructure the system call and io interrupt return code to avoid avoid the lpsw[e] to disable machine checks. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-05-17 15:00:02 +07:00
LOCKDEP_SYS_EXIT
TRACE_IRQS_ON
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lio_tif:
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __TI_flags(%r12),_TIF_WORK
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jnz .Lio_work # there is work to do (signals etc.)
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __LC_CPU_FLAGS,_CIF_WORK
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jnz .Lio_work
.Lio_restore:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lg %r14,__LC_VDSO_PER_CPU
lmg %r0,%r10,__PT_R0(%r11)
mvc __LC_RETURN_PSW(16),__PT_PSW(%r11)
s390: add options to change branch prediction behaviour for the kernel Add the PPA instruction to the system entry and exit path to switch the kernel to a different branch prediction behaviour. The instructions are added via CPU alternatives and can be disabled with the "nospec" or the "nobp=0" kernel parameter. If the default behaviour selected with CONFIG_KERNEL_NOBP is set to "n" then the "nobp=1" parameter can be used to enable the changed kernel branch prediction. Acked-by: Cornelia Huck <cohuck@redhat.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:11:45 +07:00
tm __PT_PSW+1(%r11),0x01 # returning to user ?
jno .Lio_exit_kernel
s390: run user space and KVM guests with modified branch prediction Define TIF_ISOLATE_BP and TIF_ISOLATE_BP_GUEST and add the necessary plumbing in entry.S to be able to run user space and KVM guests with limited branch prediction. To switch a user space process to limited branch prediction the s390_isolate_bp() function has to be call, and to run a vCPU of a KVM guest associated with the current task with limited branch prediction call s390_isolate_bp_guest(). Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:36:46 +07:00
BPEXIT __TI_flags(%r12),_TIF_ISOLATE_BP
s390/cputime: fix incorrect system time git commit c5328901aa1db134 "[S390] entry[64].S improvements" removed the update of the exit_timer lowcore field from the critical section cleanup of the .Lsysc_restore/.Lsysc_done and .Lio_restore/.Lio_done blocks. If the PSW is updated by the critical section cleanup to point to user space again, the interrupt entry code will do a vtime calculation after the cleanup completed with an exit_timer value which has *not* been updated. Due to this incorrect system time deltas are calculated. If an interrupt occured with an old PSW between .Lsysc_restore/.Lsysc_done or .Lio_restore/.Lio_done update __LC_EXIT_TIMER with the system entry time of the interrupt. Cc: stable@vger.kernel.org # 3.3+ Tested-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-05-02 18:36:00 +07:00
.Lio_exit_timer:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stpt __LC_EXIT_TIMER
mvc __VDSO_ECTG_BASE(16,%r14),__LC_EXIT_TIMER
s390: add options to change branch prediction behaviour for the kernel Add the PPA instruction to the system entry and exit path to switch the kernel to a different branch prediction behaviour. The instructions are added via CPU alternatives and can be disabled with the "nospec" or the "nobp=0" kernel parameter. If the default behaviour selected with CONFIG_KERNEL_NOBP is set to "n" then the "nobp=1" parameter can be used to enable the changed kernel branch prediction. Acked-by: Cornelia Huck <cohuck@redhat.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:11:45 +07:00
.Lio_exit_kernel:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lmg %r11,%r15,__PT_R11(%r11)
lpswe __LC_RETURN_PSW
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lio_done:
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
[S390] s390: Optimize user and work TIF check On return from syscall or interrupt, we have to check if we return to userspace (likely) and if there is work todo (less likely) to decide if we handle the work. We can optimize this check: we first check for the less likely work case and then check for userspace. This patch is also a preparation for an additional patch, that fixes a bug in KVM dealing with cpu bound guests. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2008-05-07 14:22:52 +07:00
#
[S390] cleanup sysc_work and io_work code Cleanup the #ifdef mess at io_work in entry[64].S and streamline the TIF work code of the system call and io exit path. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-05-17 15:00:01 +07:00
# There is work todo, find out in which context we have been interrupted:
s390: split TIF bits into CIF, PIF and TIF bits The oi and ni instructions used in entry[64].S to set and clear bits in the thread-flags are not guaranteed to be atomic in regard to other CPUs. Split the TIF bits into CPU, pt_regs and thread-info specific bits. Updates on the TIF bits are done with atomic instructions, updates on CPU and pt_regs bits are done with non-atomic instructions. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-04-15 17:55:07 +07:00
# 1) if we return to user space we can do all _TIF_WORK work
[S390] cleanup sysc_work and io_work code Cleanup the #ifdef mess at io_work in entry[64].S and streamline the TIF work code of the system call and io exit path. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-05-17 15:00:01 +07:00
# 2) if we return to kernel code and kvm is enabled check if we need to
# modify the psw to leave SIE
# 3) if we return to kernel code and preemptive scheduling is enabled check
# the preemption counter and if it is zero call preempt_schedule_irq
# Before any work can be done, a switch to the kernel stack is required.
[S390] s390: Optimize user and work TIF check On return from syscall or interrupt, we have to check if we return to userspace (likely) and if there is work todo (less likely) to decide if we handle the work. We can optimize this check: we first check for the less likely work case and then check for userspace. This patch is also a preparation for an additional patch, that fixes a bug in KVM dealing with cpu bound guests. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2008-05-07 14:22:52 +07:00
#
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lio_work:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
tm __PT_PSW+1(%r11),0x01 # returning to user ?
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jo .Lio_work_user # yes -> do resched & signal
[S390] cleanup sysc_work and io_work code Cleanup the #ifdef mess at io_work in entry[64].S and streamline the TIF work code of the system call and io exit path. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-05-17 15:00:01 +07:00
#ifdef CONFIG_PREEMPT
[S390] s390: Optimize user and work TIF check On return from syscall or interrupt, we have to check if we return to userspace (likely) and if there is work todo (less likely) to decide if we handle the work. We can optimize this check: we first check for the less likely work case and then check for userspace. This patch is also a preparation for an additional patch, that fixes a bug in KVM dealing with cpu bound guests. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2008-05-07 14:22:52 +07:00
# check for preemptive scheduling
s390/preempt: move preempt_count to the lowcore Convert s390 to use a field in the struct lowcore for the CPU preemption count. It is a bit cheaper to access a lowcore field compared to a thread_info variable and it removes the depencency on a task related structure. bloat-o-meter on the vmlinux image for the default configuration (CONFIG_PREEMPT_NONE=y) reports a small reduction in text size: add/remove: 0/0 grow/shrink: 18/578 up/down: 228/-5448 (-5220) A larger improvement is achieved with the default configuration but with CONFIG_PREEMPT=y and CONFIG_DEBUG_PREEMPT=n: add/remove: 2/6 grow/shrink: 59/4477 up/down: 1618/-228762 (-227144) Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-10-25 17:21:44 +07:00
icm %r0,15,__LC_PREEMPT_COUNT
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jnz .Lio_restore # preemption is disabled
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __TI_flags(%r12),_TIF_NEED_RESCHED
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jno .Lio_restore
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
# switch to kernel stack
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lg %r1,__PT_R15(%r11)
aghi %r1,-(STACK_FRAME_OVERHEAD + __PT_SIZE)
mvc STACK_FRAME_OVERHEAD(__PT_SIZE,%r1),0(%r11)
xc __SF_BACKCHAIN(8,%r1),__SF_BACKCHAIN(%r1)
la %r11,STACK_FRAME_OVERHEAD(%r1)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
lgr %r15,%r1
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
# TRACE_IRQS_ON already done at .Lio_return, call
[S390] improve irq tracing code in entry[64].S The system call path in entry[64].S is run with interrupts enabled. Remove the irq tracing check from the system call exit code. If a program check interrupted a context enabled for interrupts do a call to trace_irq_off_caller in the program check handler before branching to the system call exit code. Restructure the system call and io interrupt return code to avoid avoid the lpsw[e] to disable machine checks. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-05-17 15:00:02 +07:00
# TRACE_IRQS_OFF to keep things symmetrical
TRACE_IRQS_OFF
brasl %r14,preempt_schedule_irq
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
j .Lio_return
[S390] improve irq tracing code in entry[64].S The system call path in entry[64].S is run with interrupts enabled. Remove the irq tracing check from the system call exit code. If a program check interrupted a context enabled for interrupts do a call to trace_irq_off_caller in the program check handler before branching to the system call exit code. Restructure the system call and io interrupt return code to avoid avoid the lpsw[e] to disable machine checks. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-05-17 15:00:02 +07:00
#else
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
j .Lio_restore
[S390] improve irq tracing code in entry[64].S The system call path in entry[64].S is run with interrupts enabled. Remove the irq tracing check from the system call exit code. If a program check interrupted a context enabled for interrupts do a call to trace_irq_off_caller in the program check handler before branching to the system call exit code. Restructure the system call and io interrupt return code to avoid avoid the lpsw[e] to disable machine checks. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-05-17 15:00:02 +07:00
#endif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
[S390] cleanup sysc_work and io_work code Cleanup the #ifdef mess at io_work in entry[64].S and streamline the TIF work code of the system call and io exit path. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-05-17 15:00:01 +07:00
#
# Need to do work before returning to userspace, switch to kernel stack
#
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lio_work_user:
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
lg %r1,__LC_KERNEL_STACK
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
mvc STACK_FRAME_OVERHEAD(__PT_SIZE,%r1),0(%r11)
xc __SF_BACKCHAIN(8,%r1),__SF_BACKCHAIN(%r1)
la %r11,STACK_FRAME_OVERHEAD(%r1)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
lgr %r15,%r1
[S390] cleanup sysc_work and io_work code Cleanup the #ifdef mess at io_work in entry[64].S and streamline the TIF work code of the system call and io exit path. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-05-17 15:00:01 +07:00
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#
# One of the work bits is on. Find out which one.
#
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lio_work_tif:
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __LC_CPU_FLAGS,_CIF_MCCK_PENDING
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jo .Lio_mcck_pending
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __TI_flags(%r12),_TIF_NEED_RESCHED
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jo .Lio_reschedule
livepatch/s390: add TIF_PATCH_PENDING thread flag Update a task's patch state when returning from a system call or user space interrupt, or after handling a signal. This greatly increases the chances of a patch operation succeeding. If a task is I/O bound, it can be patched when returning from a system call. If a task is CPU bound, it can be patched when returning from an interrupt. If a task is sleeping on a to-be-patched function, the user can send SIGSTOP and SIGCONT to force it to switch. Since there are two ways the syscall can be restarted on return from a signal handling process, it is important to clear the flag before do_signal() is called. Otherwise we could miss the migration if we used SIGSTOP/SIGCONT procedure or fake signal to migrate patching blocking tasks. If we place our hook to sysc_work label in entry before TIF_SIGPENDING is evaluated we kill two birds with one stone. The task is correctly migrated in all return paths from a syscall. Signed-off-by: Miroslav Benes <mbenes@suse.cz> Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com> Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2017-02-14 08:42:34 +07:00
#ifdef CONFIG_LIVEPATCH
TSTMSK __TI_flags(%r12),_TIF_PATCH_PENDING
jo .Lio_patch_pending
#endif
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __TI_flags(%r12),_TIF_SIGPENDING
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jo .Lio_sigpending
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __TI_flags(%r12),_TIF_NOTIFY_RESUME
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jo .Lio_notify_resume
s390: add a system call for guarded storage This adds a new system call to enable the use of guarded storage for user space processes. The system call takes two arguments, a command and pointer to a guarded storage control block: s390_guarded_storage(int command, struct gs_cb *gs_cb); The second argument is relevant only for the GS_SET_BC_CB command. The commands in detail: 0 - GS_ENABLE Enable the guarded storage facility for the current task. The initial content of the guarded storage control block will be all zeros. After the enablement the user space code can use load-guarded-storage-controls instruction (LGSC) to load an arbitrary control block. While a task is enabled the kernel will save and restore the current content of the guarded storage registers on context switch. 1 - GS_DISABLE Disables the use of the guarded storage facility for the current task. The kernel will cease to save and restore the content of the guarded storage registers, the task specific content of these registers is lost. 2 - GS_SET_BC_CB Set a broadcast guarded storage control block. This is called per thread and stores a specific guarded storage control block in the task struct of the current task. This control block will be used for the broadcast event GS_BROADCAST. 3 - GS_CLEAR_BC_CB Clears the broadcast guarded storage control block. The guarded- storage control block is removed from the task struct that was established by GS_SET_BC_CB. 4 - GS_BROADCAST Sends a broadcast to all thread siblings of the current task. Every sibling that has established a broadcast guarded storage control block will load this control block and will be enabled for guarded storage. The broadcast guarded storage control block is used up, a second broadcast without a refresh of the stored control block with GS_SET_BC_CB will not have any effect. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-01-26 20:10:34 +07:00
TSTMSK __TI_flags(%r12),_TIF_GUARDED_STORAGE
jo .Lio_guarded_storage
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __LC_CPU_FLAGS,_CIF_FPU
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
jo .Lio_vxrs
s390: restore address space when returning to user space Unbalanced set_fs usages (e.g. early exit from a function and a forgotten set_fs(USER_DS) call) may lead to a situation where the secondary asce is the kernel space asce when returning to user space. This would allow user space to modify kernel space at will. This would only be possible with the above mentioned kernel bug, however we can detect this and fix the secondary asce before returning to user space. Therefore a new TIF_ASCE_SECONDARY which is used within set_fs. When returning to user space check if TIF_ASCE_SECONDARY is set, which would indicate a bug. If it is set print a message to the console, fixup the secondary asce, and then return to user space. This is similar to what is being discussed for x86 and arm: "[RFC] syscalls: Restore address limit after a syscall". Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-02-17 14:13:28 +07:00
TSTMSK __LC_CPU_FLAGS,(_CIF_ASCE_PRIMARY|_CIF_ASCE_SECONDARY)
jnz .Lio_asce
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
j .Lio_return # beware of critical section cleanup
[S390] s390-kvm: leave sie context on work. Removes preemption requirement From: Martin Schwidefsky <schwidefsky@de.ibm.com> This patch fixes a bug with cpu bound guest on kvm-s390. Sometimes it was impossible to deliver a signal to a spinning guest. We used preemption as a circumvention. The preemption notifiers called vcpu_load, which checked for pending signals and triggered a host intercept. But even with preemption, a sigkill was not delivered immediately. This patch changes the low level host interrupt handler to check for the SIE instruction, if TIF_WORK is set. In that case we change the instruction pointer of the return PSW to rerun the vcpu_run loop. The kvm code sees an intercept reason 0 if that happens. This patch adds accounting for these types of intercept as well. The advantages: - works with and without preemption - signals are delivered immediately - much better host latencies without preemption Acked-by: Carsten Otte <cotte@de.ibm.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2008-05-07 14:22:53 +07:00
[PATCH] s390: improved machine check handling Improved machine check handling. Kernel is now able to receive machine checks while in kernel mode (system call, interrupt and program check handling). Also register validation is now performed. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-26 04:55:30 +07:00
#
s390: split TIF bits into CIF, PIF and TIF bits The oi and ni instructions used in entry[64].S to set and clear bits in the thread-flags are not guaranteed to be atomic in regard to other CPUs. Split the TIF bits into CPU, pt_regs and thread-info specific bits. Updates on the TIF bits are done with atomic instructions, updates on CPU and pt_regs bits are done with non-atomic instructions. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-04-15 17:55:07 +07:00
# _CIF_MCCK_PENDING is set, call handler
[PATCH] s390: improved machine check handling Improved machine check handling. Kernel is now able to receive machine checks while in kernel mode (system call, interrupt and program check handling). Also register validation is now performed. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-26 04:55:30 +07:00
#
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lio_mcck_pending:
# TRACE_IRQS_ON already done at .Lio_return
[S390] Fix IRQ tracing. If a machine check is pending and the external or I/O interrupt handler returns to userspace io_mcck_pending is going to call s390_handle_mcck. Before this happens a call to TRACE_IRQS_ON was already made since we know that we are going back to userspace and hence interrupts will be enabled. So there was an indication that interrupts are enabled while in reality they are still disabled. s390_handle_mcck will do a local_irq_save/restore pair and confuse lockdep which later complains about inconsistent irq tracing. To solve this just call trace_hardirqs_off before calling s390_handle_mcck and trace_hardirqs_on afterwards. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2007-07-27 17:29:18 +07:00
brasl %r14,s390_handle_mcck # TIF bit will be cleared by handler
[S390] improve irq tracing code in entry[64].S The system call path in entry[64].S is run with interrupts enabled. Remove the irq tracing check from the system call exit code. If a program check interrupted a context enabled for interrupts do a call to trace_irq_off_caller in the program check handler before branching to the system call exit code. Restructure the system call and io interrupt return code to avoid avoid the lpsw[e] to disable machine checks. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-05-17 15:00:02 +07:00
TRACE_IRQS_OFF
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
j .Lio_return
[PATCH] s390: improved machine check handling Improved machine check handling. Kernel is now able to receive machine checks while in kernel mode (system call, interrupt and program check handling). Also register validation is now performed. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-26 04:55:30 +07:00
s390/uaccess: rework uaccess code - fix locking issues The current uaccess code uses a page table walk in some circumstances, e.g. in case of the in atomic futex operations or if running on old hardware which doesn't support the mvcos instruction. However it turned out that the page table walk code does not correctly lock page tables when accessing page table entries. In other words: a different cpu may invalidate a page table entry while the current cpu inspects the pte. This may lead to random data corruption. Adding correct locking however isn't trivial for all uaccess operations. Especially copy_in_user() is problematic since that requires to hold at least two locks, but must be protected against ABBA deadlock when a different cpu also performs a copy_in_user() operation. So the solution is a different approach where we change address spaces: User space runs in primary address mode, or access register mode within vdso code, like it currently already does. The kernel usually also runs in home space mode, however when accessing user space the kernel switches to primary or secondary address mode if the mvcos instruction is not available or if a compare-and-swap (futex) instruction on a user space address is performed. KVM however is special, since that requires the kernel to run in home address space while implicitly accessing user space with the sie instruction. So we end up with: User space: - runs in primary or access register mode - cr1 contains the user asce - cr7 contains the user asce - cr13 contains the kernel asce Kernel space: - runs in home space mode - cr1 contains the user or kernel asce -> the kernel asce is loaded when a uaccess requires primary or secondary address mode - cr7 contains the user or kernel asce, (changed with set_fs()) - cr13 contains the kernel asce In case of uaccess the kernel changes to: - primary space mode in case of a uaccess (copy_to_user) and uses e.g. the mvcp instruction to access user space. However the kernel will stay in home space mode if the mvcos instruction is available - secondary space mode in case of futex atomic operations, so that the instructions come from primary address space and data from secondary space In case of kvm the kernel runs in home space mode, but cr1 gets switched to contain the gmap asce before the sie instruction gets executed. When the sie instruction is finished cr1 will be switched back to contain the user asce. A context switch between two processes will always load the kernel asce for the next process in cr1. So the first exit to user space is a bit more expensive (one extra load control register instruction) than before, however keeps the code rather simple. In sum this means there is no need to perform any error prone page table walks anymore when accessing user space. The patch seems to be rather large, however it mainly removes the the page table walk code and restores the previously deleted "standard" uaccess code, with a couple of changes. The uaccess without mvcos mode can be enforced with the "uaccess_primary" kernel parameter. Reported-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-03-21 16:42:25 +07:00
#
s390: restore address space when returning to user space Unbalanced set_fs usages (e.g. early exit from a function and a forgotten set_fs(USER_DS) call) may lead to a situation where the secondary asce is the kernel space asce when returning to user space. This would allow user space to modify kernel space at will. This would only be possible with the above mentioned kernel bug, however we can detect this and fix the secondary asce before returning to user space. Therefore a new TIF_ASCE_SECONDARY which is used within set_fs. When returning to user space check if TIF_ASCE_SECONDARY is set, which would indicate a bug. If it is set print a message to the console, fixup the secondary asce, and then return to user space. This is similar to what is being discussed for x86 and arm: "[RFC] syscalls: Restore address limit after a syscall". Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-02-17 14:13:28 +07:00
# _CIF_ASCE_PRIMARY and/or CIF_ASCE_SECONDARY set, load user space asce
s390/uaccess: rework uaccess code - fix locking issues The current uaccess code uses a page table walk in some circumstances, e.g. in case of the in atomic futex operations or if running on old hardware which doesn't support the mvcos instruction. However it turned out that the page table walk code does not correctly lock page tables when accessing page table entries. In other words: a different cpu may invalidate a page table entry while the current cpu inspects the pte. This may lead to random data corruption. Adding correct locking however isn't trivial for all uaccess operations. Especially copy_in_user() is problematic since that requires to hold at least two locks, but must be protected against ABBA deadlock when a different cpu also performs a copy_in_user() operation. So the solution is a different approach where we change address spaces: User space runs in primary address mode, or access register mode within vdso code, like it currently already does. The kernel usually also runs in home space mode, however when accessing user space the kernel switches to primary or secondary address mode if the mvcos instruction is not available or if a compare-and-swap (futex) instruction on a user space address is performed. KVM however is special, since that requires the kernel to run in home address space while implicitly accessing user space with the sie instruction. So we end up with: User space: - runs in primary or access register mode - cr1 contains the user asce - cr7 contains the user asce - cr13 contains the kernel asce Kernel space: - runs in home space mode - cr1 contains the user or kernel asce -> the kernel asce is loaded when a uaccess requires primary or secondary address mode - cr7 contains the user or kernel asce, (changed with set_fs()) - cr13 contains the kernel asce In case of uaccess the kernel changes to: - primary space mode in case of a uaccess (copy_to_user) and uses e.g. the mvcp instruction to access user space. However the kernel will stay in home space mode if the mvcos instruction is available - secondary space mode in case of futex atomic operations, so that the instructions come from primary address space and data from secondary space In case of kvm the kernel runs in home space mode, but cr1 gets switched to contain the gmap asce before the sie instruction gets executed. When the sie instruction is finished cr1 will be switched back to contain the user asce. A context switch between two processes will always load the kernel asce for the next process in cr1. So the first exit to user space is a bit more expensive (one extra load control register instruction) than before, however keeps the code rather simple. In sum this means there is no need to perform any error prone page table walks anymore when accessing user space. The patch seems to be rather large, however it mainly removes the the page table walk code and restores the previously deleted "standard" uaccess code, with a couple of changes. The uaccess without mvcos mode can be enforced with the "uaccess_primary" kernel parameter. Reported-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-03-21 16:42:25 +07:00
#
s390: restore address space when returning to user space Unbalanced set_fs usages (e.g. early exit from a function and a forgotten set_fs(USER_DS) call) may lead to a situation where the secondary asce is the kernel space asce when returning to user space. This would allow user space to modify kernel space at will. This would only be possible with the above mentioned kernel bug, however we can detect this and fix the secondary asce before returning to user space. Therefore a new TIF_ASCE_SECONDARY which is used within set_fs. When returning to user space check if TIF_ASCE_SECONDARY is set, which would indicate a bug. If it is set print a message to the console, fixup the secondary asce, and then return to user space. This is similar to what is being discussed for x86 and arm: "[RFC] syscalls: Restore address limit after a syscall". Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-02-17 14:13:28 +07:00
.Lio_asce:
s390: remove all code using the access register mode The vdso code for the getcpu() and the clock_gettime() call use the access register mode to access the per-CPU vdso data page with the current code. An alternative to the complicated AR mode is to use the secondary space mode. This makes the vdso faster and quite a bit simpler. The downside is that the uaccess code has to be changed quite a bit. Which instructions are used depends on the machine and what kind of uaccess operation is requested. The instruction dictates which ASCE value needs to be loaded into %cr1 and %cr7. The different cases: * User copy with MVCOS for z10 and newer machines The MVCOS instruction can copy between the primary space (aka user) and the home space (aka kernel) directly. For set_fs(KERNEL_DS) the kernel ASCE is loaded into %cr1. For set_fs(USER_DS) the user space is already loaded in %cr1. * User copy with MVCP/MVCS for older machines To be able to execute the MVCP/MVCS instructions the kernel needs to switch to primary mode. The control register %cr1 has to be set to the kernel ASCE and %cr7 to either the kernel ASCE or the user ASCE dependent on set_fs(KERNEL_DS) vs set_fs(USER_DS). * Data access in the user address space for strnlen / futex To use "normal" instruction with data from the user address space the secondary space mode is used. The kernel needs to switch to primary mode, %cr1 has to contain the kernel ASCE and %cr7 either the user ASCE or the kernel ASCE, dependent on set_fs. To load a new value into %cr1 or %cr7 is an expensive operation, the kernel tries to be lazy about it. E.g. for multiple user copies in a row with MVCP/MVCS the replacement of the vdso ASCE in %cr7 with the user ASCE is done only once. On return to user space a CPU bit is checked that loads the vdso ASCE again. To enable and disable the data access via the secondary space two new functions are added, enable_sacf_uaccess and disable_sacf_uaccess. The fact that a context is in secondary space uaccess mode is stored in the mm_segment_t value for the task. The code of an interrupt may use set_fs as long as it returns to the previous state it got with get_fs with another call to set_fs. The code in finish_arch_post_lock_switch simply has to do a set_fs with the current mm_segment_t value for the task. For CPUs with MVCOS: CPU running in | %cr1 ASCE | %cr7 ASCE | --------------------------------------|-----------|-----------| user space | user | vdso | kernel, USER_DS, normal-mode | user | vdso | kernel, USER_DS, normal-mode, lazy | user | user | kernel, USER_DS, sacf-mode | kernel | user | kernel, KERNEL_DS, normal-mode | kernel | vdso | kernel, KERNEL_DS, normal-mode, lazy | kernel | kernel | kernel, KERNEL_DS, sacf-mode | kernel | kernel | For CPUs without MVCOS: CPU running in | %cr1 ASCE | %cr7 ASCE | --------------------------------------|-----------|-----------| user space | user | vdso | kernel, USER_DS, normal-mode | user | vdso | kernel, USER_DS, normal-mode lazy | kernel | user | kernel, USER_DS, sacf-mode | kernel | user | kernel, KERNEL_DS, normal-mode | kernel | vdso | kernel, KERNEL_DS, normal-mode, lazy | kernel | kernel | kernel, KERNEL_DS, sacf-mode | kernel | kernel | The lines with "lazy" refer to the state after a copy via the secondary space with a delayed reload of %cr1 and %cr7. There are three hardware address spaces that can cause a DAT exception, primary, secondary and home space. The exception can be related to four different fault types: user space fault, vdso fault, kernel fault, and the gmap faults. Dependent on the set_fs state and normal vs. sacf mode there are a number of fault combinations: 1) user address space fault via the primary ASCE 2) gmap address space fault via the primary ASCE 3) kernel address space fault via the primary ASCE for machines with MVCOS and set_fs(KERNEL_DS) 4) vdso address space faults via the secondary ASCE with an invalid address while running in secondary space in problem state 5) user address space fault via the secondary ASCE for user-copy based on the secondary space mode, e.g. futex_ops or strnlen_user 6) kernel address space fault via the secondary ASCE for user-copy with secondary space mode with set_fs(KERNEL_DS) 7) kernel address space fault via the primary ASCE for user-copy with secondary space mode with set_fs(USER_DS) on machines without MVCOS. 8) kernel address space fault via the home space ASCE Replace user_space_fault() with a new function get_fault_type() that can distinguish all four different fault types. With these changes the futex atomic ops from the kernel and the strnlen_user will get a little bit slower, as well as the old style uaccess with MVCP/MVCS. All user accesses based on MVCOS will be as fast as before. On the positive side, the user space vdso code is a lot faster and Linux ceases to use the complicated AR mode. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
2017-08-22 17:08:22 +07:00
ni __LC_CPU_FLAGS+7,255-_CIF_ASCE_SECONDARY
lctlg %c7,%c7,__LC_VDSO_ASCE # load secondary asce
TSTMSK __LC_CPU_FLAGS,_CIF_ASCE_PRIMARY
jz .Lio_return
#ifndef CONFIG_HAVE_MARCH_Z10_FEATURES
tm __LC_STFLE_FAC_LIST+3,0x10 # has MVCOS ?
jnz .Lio_set_fs_fixup
s390: rename CIF_ASCE to CIF_ASCE_PRIMARY This is just a preparation patch in order to keep the "restore address space after syscall" patch small. Rename CIF_ASCE to CIF_ASCE_PRIMARY to be unique and specific when introducing a second CIF_ASCE_SECONDARY CIF flag. Suggested-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-02-17 14:12:30 +07:00
ni __LC_CPU_FLAGS+7,255-_CIF_ASCE_PRIMARY
s390/uaccess: rework uaccess code - fix locking issues The current uaccess code uses a page table walk in some circumstances, e.g. in case of the in atomic futex operations or if running on old hardware which doesn't support the mvcos instruction. However it turned out that the page table walk code does not correctly lock page tables when accessing page table entries. In other words: a different cpu may invalidate a page table entry while the current cpu inspects the pte. This may lead to random data corruption. Adding correct locking however isn't trivial for all uaccess operations. Especially copy_in_user() is problematic since that requires to hold at least two locks, but must be protected against ABBA deadlock when a different cpu also performs a copy_in_user() operation. So the solution is a different approach where we change address spaces: User space runs in primary address mode, or access register mode within vdso code, like it currently already does. The kernel usually also runs in home space mode, however when accessing user space the kernel switches to primary or secondary address mode if the mvcos instruction is not available or if a compare-and-swap (futex) instruction on a user space address is performed. KVM however is special, since that requires the kernel to run in home address space while implicitly accessing user space with the sie instruction. So we end up with: User space: - runs in primary or access register mode - cr1 contains the user asce - cr7 contains the user asce - cr13 contains the kernel asce Kernel space: - runs in home space mode - cr1 contains the user or kernel asce -> the kernel asce is loaded when a uaccess requires primary or secondary address mode - cr7 contains the user or kernel asce, (changed with set_fs()) - cr13 contains the kernel asce In case of uaccess the kernel changes to: - primary space mode in case of a uaccess (copy_to_user) and uses e.g. the mvcp instruction to access user space. However the kernel will stay in home space mode if the mvcos instruction is available - secondary space mode in case of futex atomic operations, so that the instructions come from primary address space and data from secondary space In case of kvm the kernel runs in home space mode, but cr1 gets switched to contain the gmap asce before the sie instruction gets executed. When the sie instruction is finished cr1 will be switched back to contain the user asce. A context switch between two processes will always load the kernel asce for the next process in cr1. So the first exit to user space is a bit more expensive (one extra load control register instruction) than before, however keeps the code rather simple. In sum this means there is no need to perform any error prone page table walks anymore when accessing user space. The patch seems to be rather large, however it mainly removes the the page table walk code and restores the previously deleted "standard" uaccess code, with a couple of changes. The uaccess without mvcos mode can be enforced with the "uaccess_primary" kernel parameter. Reported-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-03-21 16:42:25 +07:00
lctlg %c1,%c1,__LC_USER_ASCE # load primary asce
s390: remove all code using the access register mode The vdso code for the getcpu() and the clock_gettime() call use the access register mode to access the per-CPU vdso data page with the current code. An alternative to the complicated AR mode is to use the secondary space mode. This makes the vdso faster and quite a bit simpler. The downside is that the uaccess code has to be changed quite a bit. Which instructions are used depends on the machine and what kind of uaccess operation is requested. The instruction dictates which ASCE value needs to be loaded into %cr1 and %cr7. The different cases: * User copy with MVCOS for z10 and newer machines The MVCOS instruction can copy between the primary space (aka user) and the home space (aka kernel) directly. For set_fs(KERNEL_DS) the kernel ASCE is loaded into %cr1. For set_fs(USER_DS) the user space is already loaded in %cr1. * User copy with MVCP/MVCS for older machines To be able to execute the MVCP/MVCS instructions the kernel needs to switch to primary mode. The control register %cr1 has to be set to the kernel ASCE and %cr7 to either the kernel ASCE or the user ASCE dependent on set_fs(KERNEL_DS) vs set_fs(USER_DS). * Data access in the user address space for strnlen / futex To use "normal" instruction with data from the user address space the secondary space mode is used. The kernel needs to switch to primary mode, %cr1 has to contain the kernel ASCE and %cr7 either the user ASCE or the kernel ASCE, dependent on set_fs. To load a new value into %cr1 or %cr7 is an expensive operation, the kernel tries to be lazy about it. E.g. for multiple user copies in a row with MVCP/MVCS the replacement of the vdso ASCE in %cr7 with the user ASCE is done only once. On return to user space a CPU bit is checked that loads the vdso ASCE again. To enable and disable the data access via the secondary space two new functions are added, enable_sacf_uaccess and disable_sacf_uaccess. The fact that a context is in secondary space uaccess mode is stored in the mm_segment_t value for the task. The code of an interrupt may use set_fs as long as it returns to the previous state it got with get_fs with another call to set_fs. The code in finish_arch_post_lock_switch simply has to do a set_fs with the current mm_segment_t value for the task. For CPUs with MVCOS: CPU running in | %cr1 ASCE | %cr7 ASCE | --------------------------------------|-----------|-----------| user space | user | vdso | kernel, USER_DS, normal-mode | user | vdso | kernel, USER_DS, normal-mode, lazy | user | user | kernel, USER_DS, sacf-mode | kernel | user | kernel, KERNEL_DS, normal-mode | kernel | vdso | kernel, KERNEL_DS, normal-mode, lazy | kernel | kernel | kernel, KERNEL_DS, sacf-mode | kernel | kernel | For CPUs without MVCOS: CPU running in | %cr1 ASCE | %cr7 ASCE | --------------------------------------|-----------|-----------| user space | user | vdso | kernel, USER_DS, normal-mode | user | vdso | kernel, USER_DS, normal-mode lazy | kernel | user | kernel, USER_DS, sacf-mode | kernel | user | kernel, KERNEL_DS, normal-mode | kernel | vdso | kernel, KERNEL_DS, normal-mode, lazy | kernel | kernel | kernel, KERNEL_DS, sacf-mode | kernel | kernel | The lines with "lazy" refer to the state after a copy via the secondary space with a delayed reload of %cr1 and %cr7. There are three hardware address spaces that can cause a DAT exception, primary, secondary and home space. The exception can be related to four different fault types: user space fault, vdso fault, kernel fault, and the gmap faults. Dependent on the set_fs state and normal vs. sacf mode there are a number of fault combinations: 1) user address space fault via the primary ASCE 2) gmap address space fault via the primary ASCE 3) kernel address space fault via the primary ASCE for machines with MVCOS and set_fs(KERNEL_DS) 4) vdso address space faults via the secondary ASCE with an invalid address while running in secondary space in problem state 5) user address space fault via the secondary ASCE for user-copy based on the secondary space mode, e.g. futex_ops or strnlen_user 6) kernel address space fault via the secondary ASCE for user-copy with secondary space mode with set_fs(KERNEL_DS) 7) kernel address space fault via the primary ASCE for user-copy with secondary space mode with set_fs(USER_DS) on machines without MVCOS. 8) kernel address space fault via the home space ASCE Replace user_space_fault() with a new function get_fault_type() that can distinguish all four different fault types. With these changes the futex atomic ops from the kernel and the strnlen_user will get a little bit slower, as well as the old style uaccess with MVCP/MVCS. All user accesses based on MVCOS will be as fast as before. On the positive side, the user space vdso code is a lot faster and Linux ceases to use the complicated AR mode. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
2017-08-22 17:08:22 +07:00
j .Lio_return
.Lio_set_fs_fixup:
#endif
s390: restore address space when returning to user space Unbalanced set_fs usages (e.g. early exit from a function and a forgotten set_fs(USER_DS) call) may lead to a situation where the secondary asce is the kernel space asce when returning to user space. This would allow user space to modify kernel space at will. This would only be possible with the above mentioned kernel bug, however we can detect this and fix the secondary asce before returning to user space. Therefore a new TIF_ASCE_SECONDARY which is used within set_fs. When returning to user space check if TIF_ASCE_SECONDARY is set, which would indicate a bug. If it is set print a message to the console, fixup the secondary asce, and then return to user space. This is similar to what is being discussed for x86 and arm: "[RFC] syscalls: Restore address limit after a syscall". Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-02-17 14:13:28 +07:00
larl %r14,.Lio_return
jg set_fs_fixup
s390/uaccess: rework uaccess code - fix locking issues The current uaccess code uses a page table walk in some circumstances, e.g. in case of the in atomic futex operations or if running on old hardware which doesn't support the mvcos instruction. However it turned out that the page table walk code does not correctly lock page tables when accessing page table entries. In other words: a different cpu may invalidate a page table entry while the current cpu inspects the pte. This may lead to random data corruption. Adding correct locking however isn't trivial for all uaccess operations. Especially copy_in_user() is problematic since that requires to hold at least two locks, but must be protected against ABBA deadlock when a different cpu also performs a copy_in_user() operation. So the solution is a different approach where we change address spaces: User space runs in primary address mode, or access register mode within vdso code, like it currently already does. The kernel usually also runs in home space mode, however when accessing user space the kernel switches to primary or secondary address mode if the mvcos instruction is not available or if a compare-and-swap (futex) instruction on a user space address is performed. KVM however is special, since that requires the kernel to run in home address space while implicitly accessing user space with the sie instruction. So we end up with: User space: - runs in primary or access register mode - cr1 contains the user asce - cr7 contains the user asce - cr13 contains the kernel asce Kernel space: - runs in home space mode - cr1 contains the user or kernel asce -> the kernel asce is loaded when a uaccess requires primary or secondary address mode - cr7 contains the user or kernel asce, (changed with set_fs()) - cr13 contains the kernel asce In case of uaccess the kernel changes to: - primary space mode in case of a uaccess (copy_to_user) and uses e.g. the mvcp instruction to access user space. However the kernel will stay in home space mode if the mvcos instruction is available - secondary space mode in case of futex atomic operations, so that the instructions come from primary address space and data from secondary space In case of kvm the kernel runs in home space mode, but cr1 gets switched to contain the gmap asce before the sie instruction gets executed. When the sie instruction is finished cr1 will be switched back to contain the user asce. A context switch between two processes will always load the kernel asce for the next process in cr1. So the first exit to user space is a bit more expensive (one extra load control register instruction) than before, however keeps the code rather simple. In sum this means there is no need to perform any error prone page table walks anymore when accessing user space. The patch seems to be rather large, however it mainly removes the the page table walk code and restores the previously deleted "standard" uaccess code, with a couple of changes. The uaccess without mvcos mode can be enforced with the "uaccess_primary" kernel parameter. Reported-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-03-21 16:42:25 +07:00
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
#
# CIF_FPU is set, restore floating-point controls and floating-point registers.
#
.Lio_vxrs:
larl %r14,.Lio_return
jg load_fpu_regs
s390: add a system call for guarded storage This adds a new system call to enable the use of guarded storage for user space processes. The system call takes two arguments, a command and pointer to a guarded storage control block: s390_guarded_storage(int command, struct gs_cb *gs_cb); The second argument is relevant only for the GS_SET_BC_CB command. The commands in detail: 0 - GS_ENABLE Enable the guarded storage facility for the current task. The initial content of the guarded storage control block will be all zeros. After the enablement the user space code can use load-guarded-storage-controls instruction (LGSC) to load an arbitrary control block. While a task is enabled the kernel will save and restore the current content of the guarded storage registers on context switch. 1 - GS_DISABLE Disables the use of the guarded storage facility for the current task. The kernel will cease to save and restore the content of the guarded storage registers, the task specific content of these registers is lost. 2 - GS_SET_BC_CB Set a broadcast guarded storage control block. This is called per thread and stores a specific guarded storage control block in the task struct of the current task. This control block will be used for the broadcast event GS_BROADCAST. 3 - GS_CLEAR_BC_CB Clears the broadcast guarded storage control block. The guarded- storage control block is removed from the task struct that was established by GS_SET_BC_CB. 4 - GS_BROADCAST Sends a broadcast to all thread siblings of the current task. Every sibling that has established a broadcast guarded storage control block will load this control block and will be enabled for guarded storage. The broadcast guarded storage control block is used up, a second broadcast without a refresh of the stored control block with GS_SET_BC_CB will not have any effect. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-01-26 20:10:34 +07:00
#
# _TIF_GUARDED_STORAGE is set, call guarded_storage_load
#
.Lio_guarded_storage:
# TRACE_IRQS_ON already done at .Lio_return
ssm __LC_SVC_NEW_PSW # reenable interrupts
lgr %r2,%r11 # pass pointer to pt_regs
brasl %r14,gs_load_bc_cb
ssm __LC_PGM_NEW_PSW # disable I/O and ext. interrupts
TRACE_IRQS_OFF
j .Lio_return
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#
# _TIF_NEED_RESCHED is set, call schedule
[S390] Whitespace cleanup. Huge s390 assembly files whitespace cleanup. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2006-09-28 21:56:37 +07:00
#
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lio_reschedule:
# TRACE_IRQS_ON already done at .Lio_return
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
ssm __LC_SVC_NEW_PSW # reenable interrupts
[S390] Whitespace cleanup. Huge s390 assembly files whitespace cleanup. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2006-09-28 21:56:37 +07:00
brasl %r14,schedule # call scheduler
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
ssm __LC_PGM_NEW_PSW # disable I/O and ext. interrupts
[S390] Fix irq tracing and lockdep_sys_exit calls. Current support for TRACE_IRQFLAGS and lockdep_sys_exit is broken. IRQ flag tracing is broken for program checks. Even worse is that the newly introduced calls to lockdep_sys_exit are in the critical section code which is not supposed to call any C functions. In addition the checks if locks are still held are also done when returning to kernel code which is broken as well. Fix all this by disabling interrupts and machine checks at the exit paths and then do the appropriate checks and calls. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2007-11-20 17:13:32 +07:00
TRACE_IRQS_OFF
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
j .Lio_return
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
livepatch/s390: add TIF_PATCH_PENDING thread flag Update a task's patch state when returning from a system call or user space interrupt, or after handling a signal. This greatly increases the chances of a patch operation succeeding. If a task is I/O bound, it can be patched when returning from a system call. If a task is CPU bound, it can be patched when returning from an interrupt. If a task is sleeping on a to-be-patched function, the user can send SIGSTOP and SIGCONT to force it to switch. Since there are two ways the syscall can be restarted on return from a signal handling process, it is important to clear the flag before do_signal() is called. Otherwise we could miss the migration if we used SIGSTOP/SIGCONT procedure or fake signal to migrate patching blocking tasks. If we place our hook to sysc_work label in entry before TIF_SIGPENDING is evaluated we kill two birds with one stone. The task is correctly migrated in all return paths from a syscall. Signed-off-by: Miroslav Benes <mbenes@suse.cz> Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com> Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2017-02-14 08:42:34 +07:00
#
# _TIF_PATCH_PENDING is set, call klp_update_patch_state
#
#ifdef CONFIG_LIVEPATCH
.Lio_patch_pending:
lg %r2,__LC_CURRENT # pass pointer to task struct
larl %r14,.Lio_return
jg klp_update_patch_state
#endif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#
signals: s390: renumber TIF_RESTORE_SIGMASK TIF_RESTORE_SIGMASK no longer needs to be in the _TIF_WORK_* masks. Those low bits are scarce, and are all used up now. Renumber TIF_RESTORE_SIGMASK to free one up. Signed-off-by: Roland McGrath <roland@redhat.com> Cc: Oleg Nesterov <oleg@tv-sign.ru> Cc: Ingo Molnar <mingo@elte.hu> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: "Luck, Tony" <tony.luck@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 14:53:08 +07:00
# _TIF_SIGPENDING or is set, call do_signal
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lio_sigpending:
# TRACE_IRQS_ON already done at .Lio_return
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
ssm __LC_SVC_NEW_PSW # reenable interrupts
lgr %r2,%r11 # pass pointer to pt_regs
brasl %r14,do_signal
ssm __LC_PGM_NEW_PSW # disable I/O and ext. interrupts
[S390] Fix irq tracing and lockdep_sys_exit calls. Current support for TRACE_IRQFLAGS and lockdep_sys_exit is broken. IRQ flag tracing is broken for program checks. Even worse is that the newly introduced calls to lockdep_sys_exit are in the critical section code which is not supposed to call any C functions. In addition the checks if locks are still held are also done when returning to kernel code which is broken as well. Fix all this by disabling interrupts and machine checks at the exit paths and then do the appropriate checks and calls. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2007-11-20 17:13:32 +07:00
TRACE_IRQS_OFF
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
j .Lio_return
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
[S390] ptrace changes * System call parameter and result access functions * Add tracehook calls * Split syscall_trace into two functions do_syscall_trace_enter and do_syscall_trace_exit Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2008-10-11 02:33:20 +07:00
#
# _TIF_NOTIFY_RESUME or is set, call do_notify_resume
#
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lio_notify_resume:
# TRACE_IRQS_ON already done at .Lio_return
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
ssm __LC_SVC_NEW_PSW # reenable interrupts
lgr %r2,%r11 # pass pointer to pt_regs
brasl %r14,do_notify_resume
ssm __LC_PGM_NEW_PSW # disable I/O and ext. interrupts
[S390] ptrace changes * System call parameter and result access functions * Add tracehook calls * Split syscall_trace into two functions do_syscall_trace_enter and do_syscall_trace_exit Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2008-10-11 02:33:20 +07:00
TRACE_IRQS_OFF
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
j .Lio_return
[S390] ptrace changes * System call parameter and result access functions * Add tracehook calls * Split syscall_trace into two functions do_syscall_trace_enter and do_syscall_trace_exit Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2008-10-11 02:33:20 +07:00
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
/*
* External interrupt handler routine
*/
[S390] fix s390 assembler code alignments The alignment is missing for various global symbols in s390 assembly code. With a recent gcc and an instruction like stgrl this can lead to a specification exception if the instruction uses such a mis-aligned address. Specify the alignment explicitely and while add it define __ALIGN for s390 and use the ENTRY define to save some lines of code. Signed-off-by: Jan Glauber <jang@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-07-24 15:48:19 +07:00
ENTRY(ext_int_handler)
s390/time: always use stckf instead of stck if available The store clock fast instruction saves a couple of instructions compared to the store clock instruction. Always use stckf instead of stck if it is available. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Reviewed-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-05-09 21:27:39 +07:00
STCK __LC_INT_CLOCK
[PATCH] improve idle cputime accounting Distinguish the cputime of the idle process where idle is actually using cpu cycles from the cputime where idle is sleeping on an enabled wait psw. The former is accounted as system time, the later as idle time. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2008-12-31 21:11:41 +07:00
stpt __LC_ASYNC_ENTER_TIMER
s390: add options to change branch prediction behaviour for the kernel Add the PPA instruction to the system entry and exit path to switch the kernel to a different branch prediction behaviour. The instructions are added via CPU alternatives and can be disabled with the "nospec" or the "nobp=0" kernel parameter. If the default behaviour selected with CONFIG_KERNEL_NOBP is set to "n" then the "nobp=1" parameter can be used to enable the changed kernel branch prediction. Acked-by: Cornelia Huck <cohuck@redhat.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:11:45 +07:00
BPOFF
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stmg %r8,%r15,__LC_SAVE_AREA_ASYNC
s390: move thread_info into task_struct This is the s390 variant of commit 15f4eae70d36 ("x86: Move thread_info into task_struct"). Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-11-08 17:08:26 +07:00
lg %r12,__LC_CURRENT
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
larl %r13,cleanup_critical
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lmg %r8,%r9,__LC_EXT_OLD_PSW
s390/nmi: use the normal asynchronous stack for machine checks If a machine checks is received while the CPU is in the kernel, only the s390_do_machine_check function will be called. The call to s390_handle_mcck is postponed until the CPU returns to user space. Because of this it is safe to use the asynchronous stack for machine checks even if the CPU is already handling an interrupt. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:28:14 +07:00
SWITCH_ASYNC __LC_SAVE_AREA_ASYNC,__LC_ASYNC_ENTER_TIMER
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stmg %r0,%r7,__PT_R0(%r11)
s390: scrub registers on kernel entry and KVM exit Clear all user space registers on entry to the kernel and all KVM guest registers on KVM guest exit if the register does not contain either a parameter or a result value. Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 19:27:30 +07:00
# clear user controlled registers to prevent speculative use
xgr %r0,%r0
xgr %r1,%r1
xgr %r2,%r2
xgr %r3,%r3
xgr %r4,%r4
xgr %r5,%r5
xgr %r6,%r6
xgr %r7,%r7
xgr %r10,%r10
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
mvc __PT_R8(64,%r11),__LC_SAVE_AREA_ASYNC
stmg %r8,%r9,__PT_PSW(%r11)
s390/irq: store interrupt information in pt_regs Copy the interrupt parameters from the lowcore to the pt_regs structure in entry[64].S and reduce the arguments of the low level interrupt handler to the pt_regs pointer only. In addition move the test-pending-interrupt loop from do_IRQ to entry[64].S to make sure that interrupt information is always delivered via pt_regs. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-06-17 19:54:02 +07:00
lghi %r1,__LC_EXT_PARAMS2
mvc __PT_INT_CODE(4,%r11),__LC_EXT_CPU_ADDR
mvc __PT_INT_PARM(4,%r11),__LC_EXT_PARAMS
mvc __PT_INT_PARM_LONG(8,%r11),0(%r1)
s390: split TIF bits into CIF, PIF and TIF bits The oi and ni instructions used in entry[64].S to set and clear bits in the thread-flags are not guaranteed to be atomic in regard to other CPUs. Split the TIF bits into CPU, pt_regs and thread-info specific bits. Updates on the TIF bits are done with atomic instructions, updates on CPU and pt_regs bits are done with non-atomic instructions. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-04-15 17:55:07 +07:00
xc __PT_FLAGS(8,%r11),__PT_FLAGS(%r11)
s390/udelay: make udelay have busy loop semantics When using systemtap it was observed that our udelay implementation is rather suboptimal if being called from a kprobe handler installed by systemtap. The problem observed when a kprobe was installed on lock_acquired(). When the probe was hit the kprobe handler did call udelay, which set up an (internal) timer and reenabled interrupts (only the clock comparator interrupt) and waited for the interrupt. This is an optimization to avoid that the cpu is busy looping while waiting that enough time passes. The problem is that the interrupt handler still does call irq_enter()/irq_exit() which then again can lead to a deadlock, since some accounting functions may take locks as well. If one of these locks is the same, which caused lock_acquired() to be called, we have a nice deadlock. This patch reworks the udelay code for the interrupts disabled case to immediately leave the low level interrupt handler when the clock comparator interrupt happens. That way no C code is being called and the deadlock cannot happen anymore. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Reviewed-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-08-15 16:42:21 +07:00
TSTMSK __LC_CPU_FLAGS,_CIF_IGNORE_IRQ
jo .Lio_restore
[PATCH] lockdep: irqtrace subsystem, s390 support irqtrace support for s390. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-07-03 14:24:46 +07:00
TRACE_IRQS_OFF
s390: initialize backchain for ext_int_handler() To allow correct stack backtraces the backchain for the external interrupt handler is now initialized with zero like it is already done for example by io_int_handler(). Signed-off-by: Michael Holzheu <holzheu@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-05-09 21:27:35 +07:00
xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15)
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lgr %r2,%r11 # pass pointer to pt_regs
s390: convert interrupt handling to use generic hardirq With the introduction of PCI it became apparent that s390 should convert to generic hardirqs as too many drivers do not have the correct dependency for GENERIC_HARDIRQS. On the architecture level s390 does not have irq lines. It has external interrupts, I/O interrupts and adapter interrupts. This patch hard-codes all external interrupts as irq #1, all I/O interrupts as irq #2 and all adapter interrupts as irq #3. The additional information from the lowcore associated with the interrupt is stored in the pt_regs of the interrupt frame, where the interrupt handler can pick it up. For PCI/MSI interrupts the adapter interrupt handler scans the relevant bit fields and calls generic_handle_irq with the virtual irq number for the MSI interrupt. Reviewed-by: Sebastian Ott <sebott@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-06-27 14:01:09 +07:00
lghi %r3,EXT_INTERRUPT
brasl %r14,do_IRQ
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
j .Lio_return
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
[S390] rework idle code Whenever the cpu loads an enabled wait PSW it will appear as idle to the underlying host system. The code in default_idle calls vtime_stop_cpu which does the necessary voodoo to get the cpu time accounting right. The udelay code just loads an enabled wait PSW. To correct this rework the vtime_stop_cpu/vtime_start_cpu logic and move the difficult parts to entry[64].S, vtime_stop_cpu can now be called from anywhere and vtime_start_cpu is gone. The correction of the cpu time during wakeup from an enabled wait PSW is done with a critical section in entry[64].S. As vtime_start_cpu is gone, s390_idle_check can be removed as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:27 +07:00
/*
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
* Load idle PSW. The second "half" of this function is in .Lcleanup_idle.
[S390] rework idle code Whenever the cpu loads an enabled wait PSW it will appear as idle to the underlying host system. The code in default_idle calls vtime_stop_cpu which does the necessary voodoo to get the cpu time accounting right. The udelay code just loads an enabled wait PSW. To correct this rework the vtime_stop_cpu/vtime_start_cpu logic and move the difficult parts to entry[64].S, vtime_stop_cpu can now be called from anywhere and vtime_start_cpu is gone. The correction of the cpu time during wakeup from an enabled wait PSW is done with a critical section in entry[64].S. As vtime_start_cpu is gone, s390_idle_check can be removed as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:27 +07:00
*/
ENTRY(psw_idle)
s390/vtimer: rework virtual timer interface The current virtual timer interface is inherently per-cpu and hard to use. The sole user of the interface is appldata which uses it to execute a function after a specific amount of cputime has been used over all cpus. Rework the virtual timer interface to hook into the cputime accounting. This makes the interface independent from the CPU timer interrupts, and makes the virtual timers global as opposed to per-cpu. Overall the code is greatly simplified. The downside is that the accuracy is not as good as the original implementation, but it is still good enough for appldata. Reviewed-by: Jan Glauber <jang@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-07-20 16:15:08 +07:00
stg %r3,__SF_EMPTY(%r15)
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
larl %r1,.Lpsw_idle_lpsw+4
[S390] rework idle code Whenever the cpu loads an enabled wait PSW it will appear as idle to the underlying host system. The code in default_idle calls vtime_stop_cpu which does the necessary voodoo to get the cpu time accounting right. The udelay code just loads an enabled wait PSW. To correct this rework the vtime_stop_cpu/vtime_start_cpu logic and move the difficult parts to entry[64].S, vtime_stop_cpu can now be called from anywhere and vtime_start_cpu is gone. The correction of the cpu time during wakeup from an enabled wait PSW is done with a critical section in entry[64].S. As vtime_start_cpu is gone, s390_idle_check can be removed as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:27 +07:00
stg %r1,__SF_EMPTY+8(%r15)
s390/vtime: correct scaled cputime of partially idle CPUs The calculation for the SMT scaling factor for a hardware thread which has been partially idle needs to disregard the cycles spent by the other threads of the core while the thread is idle. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-09-18 21:41:36 +07:00
#ifdef CONFIG_SMP
larl %r1,smp_cpu_mtid
llgf %r1,0(%r1)
ltgr %r1,%r1
jz .Lpsw_idle_stcctm
.insn rsy,0xeb0000000017,%r1,5,__SF_EMPTY+16(%r15)
.Lpsw_idle_stcctm:
#endif
s390/spinlock: do not yield to a CPU in udelay/mdelay It does not make sense to try to relinquish the time slice with diag 0x9c to a CPU in a state that does not allow to schedule the CPU. The scenario where this can happen is a CPU waiting in udelay/mdelay while holding a spin-lock. Add a CIF bit to tag a CPU in enabled wait and use it to detect that the yield of a CPU will not be successful and skip the diagnose call. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-11-19 17:09:45 +07:00
oi __LC_CPU_FLAGS+7,_CIF_ENABLED_WAIT
s390: add options to change branch prediction behaviour for the kernel Add the PPA instruction to the system entry and exit path to switch the kernel to a different branch prediction behaviour. The instructions are added via CPU alternatives and can be disabled with the "nospec" or the "nobp=0" kernel parameter. If the default behaviour selected with CONFIG_KERNEL_NOBP is set to "n" then the "nobp=1" parameter can be used to enable the changed kernel branch prediction. Acked-by: Cornelia Huck <cohuck@redhat.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:11:45 +07:00
BPON
s390/vtimer: rework virtual timer interface The current virtual timer interface is inherently per-cpu and hard to use. The sole user of the interface is appldata which uses it to execute a function after a specific amount of cputime has been used over all cpus. Rework the virtual timer interface to hook into the cputime accounting. This makes the interface independent from the CPU timer interrupts, and makes the virtual timers global as opposed to per-cpu. Overall the code is greatly simplified. The downside is that the accuracy is not as good as the original implementation, but it is still good enough for appldata. Reviewed-by: Jan Glauber <jang@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-07-20 16:15:08 +07:00
STCK __CLOCK_IDLE_ENTER(%r2)
stpt __TIMER_IDLE_ENTER(%r2)
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lpsw_idle_lpsw:
[S390] rework idle code Whenever the cpu loads an enabled wait PSW it will appear as idle to the underlying host system. The code in default_idle calls vtime_stop_cpu which does the necessary voodoo to get the cpu time accounting right. The udelay code just loads an enabled wait PSW. To correct this rework the vtime_stop_cpu/vtime_start_cpu logic and move the difficult parts to entry[64].S, vtime_stop_cpu can now be called from anywhere and vtime_start_cpu is gone. The correction of the cpu time during wakeup from an enabled wait PSW is done with a critical section in entry[64].S. As vtime_start_cpu is gone, s390_idle_check can be removed as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:27 +07:00
lpswe __SF_EMPTY(%r15)
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
BR_R1USE_R14
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lpsw_idle_end:
[S390] rework idle code Whenever the cpu loads an enabled wait PSW it will appear as idle to the underlying host system. The code in default_idle calls vtime_stop_cpu which does the necessary voodoo to get the cpu time accounting right. The udelay code just loads an enabled wait PSW. To correct this rework the vtime_stop_cpu/vtime_start_cpu logic and move the difficult parts to entry[64].S, vtime_stop_cpu can now be called from anywhere and vtime_start_cpu is gone. The correction of the cpu time during wakeup from an enabled wait PSW is done with a critical section in entry[64].S. As vtime_start_cpu is gone, s390_idle_check can be removed as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:27 +07:00
s390/fpu: always enable the vector facility if it is available If the kernel detects that the s390 hardware supports the vector facility, it is enabled by default at an early stage. To force it off, use the novx kernel parameter. Note that there is a small time window, where the vector facility is enabled before it is forced to be off. With enabling the vector facility by default, the FPU save and restore functions can be improved. They do not longer require to manage expensive control register updates to enable or disable the vector enablement control for particular processes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-09-29 15:04:41 +07:00
/*
* Store floating-point controls and floating-point or vector register
* depending whether the vector facility is available. A critical section
* cleanup assures that the registers are stored even if interrupted for
* some other work. The CIF_FPU flag is set to trigger a lazy restore
* of the register contents at return from io or a system call.
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
*/
ENTRY(save_fpu_regs)
s390/kernel: remove save_fpu_regs() parameter and use __LC_CURRENT instead All calls to save_fpu_regs() specify the fpu structure of the current task pointer as parameter. The task pointer of the current task can also be retrieved from the CPU lowcore directly. Remove the parameter definition, load the __LC_CURRENT task pointer from the CPU lowcore, and rebase the FPU structure onto the task structure. Apply the same approach for the load_fpu_regs() function. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-29 21:43:06 +07:00
lg %r2,__LC_CURRENT
aghi %r2,__TASK_thread
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __LC_CPU_FLAGS,_CIF_FPU
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
jo .Lsave_fpu_regs_exit
s390/kernel: remove save_fpu_regs() parameter and use __LC_CURRENT instead All calls to save_fpu_regs() specify the fpu structure of the current task pointer as parameter. The task pointer of the current task can also be retrieved from the CPU lowcore directly. Remove the parameter definition, load the __LC_CURRENT task pointer from the CPU lowcore, and rebase the FPU structure onto the task structure. Apply the same approach for the load_fpu_regs() function. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-29 21:43:06 +07:00
stfpc __THREAD_FPU_fpc(%r2)
lg %r3,__THREAD_FPU_regs(%r2)
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __LC_MACHINE_FLAGS,MACHINE_FLAG_VX
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
jz .Lsave_fpu_regs_fp # no -> store FP regs
VSTM %v0,%v15,0,%r3 # vstm 0,15,0(3)
VSTM %v16,%v31,256,%r3 # vstm 16,31,256(3)
j .Lsave_fpu_regs_done # -> set CIF_FPU flag
.Lsave_fpu_regs_fp:
std 0,0(%r3)
std 1,8(%r3)
std 2,16(%r3)
std 3,24(%r3)
std 4,32(%r3)
std 5,40(%r3)
std 6,48(%r3)
std 7,56(%r3)
std 8,64(%r3)
std 9,72(%r3)
std 10,80(%r3)
std 11,88(%r3)
std 12,96(%r3)
std 13,104(%r3)
std 14,112(%r3)
std 15,120(%r3)
.Lsave_fpu_regs_done:
oi __LC_CPU_FLAGS+7,_CIF_FPU
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
.Lsave_fpu_regs_exit:
BR_R1USE_R14
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
.Lsave_fpu_regs_end:
s390: move exports to definitions Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-01-13 01:30:03 +07:00
EXPORT_SYMBOL(save_fpu_regs)
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
s390/fpu: always enable the vector facility if it is available If the kernel detects that the s390 hardware supports the vector facility, it is enabled by default at an early stage. To force it off, use the novx kernel parameter. Note that there is a small time window, where the vector facility is enabled before it is forced to be off. With enabling the vector facility by default, the FPU save and restore functions can be improved. They do not longer require to manage expensive control register updates to enable or disable the vector enablement control for particular processes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-09-29 15:04:41 +07:00
/*
* Load floating-point controls and floating-point or vector registers.
* A critical section cleanup assures that the register contents are
* loaded even if interrupted for some other work.
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
*
* There are special calling conventions to fit into sysc and io return work:
* %r15: <kernel stack>
* The function requires:
s390/fpu: always enable the vector facility if it is available If the kernel detects that the s390 hardware supports the vector facility, it is enabled by default at an early stage. To force it off, use the novx kernel parameter. Note that there is a small time window, where the vector facility is enabled before it is forced to be off. With enabling the vector facility by default, the FPU save and restore functions can be improved. They do not longer require to manage expensive control register updates to enable or disable the vector enablement control for particular processes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-09-29 15:04:41 +07:00
* %r4
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
*/
load_fpu_regs:
s390/kernel: remove save_fpu_regs() parameter and use __LC_CURRENT instead All calls to save_fpu_regs() specify the fpu structure of the current task pointer as parameter. The task pointer of the current task can also be retrieved from the CPU lowcore directly. Remove the parameter definition, load the __LC_CURRENT task pointer from the CPU lowcore, and rebase the FPU structure onto the task structure. Apply the same approach for the load_fpu_regs() function. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-29 21:43:06 +07:00
lg %r4,__LC_CURRENT
aghi %r4,__TASK_thread
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __LC_CPU_FLAGS,_CIF_FPU
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
jno .Lload_fpu_regs_exit
s390/kernel: remove save_fpu_regs() parameter and use __LC_CURRENT instead All calls to save_fpu_regs() specify the fpu structure of the current task pointer as parameter. The task pointer of the current task can also be retrieved from the CPU lowcore directly. Remove the parameter definition, load the __LC_CURRENT task pointer from the CPU lowcore, and rebase the FPU structure onto the task structure. Apply the same approach for the load_fpu_regs() function. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-29 21:43:06 +07:00
lfpc __THREAD_FPU_fpc(%r4)
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __LC_MACHINE_FLAGS,MACHINE_FLAG_VX
s390/kernel: remove save_fpu_regs() parameter and use __LC_CURRENT instead All calls to save_fpu_regs() specify the fpu structure of the current task pointer as parameter. The task pointer of the current task can also be retrieved from the CPU lowcore directly. Remove the parameter definition, load the __LC_CURRENT task pointer from the CPU lowcore, and rebase the FPU structure onto the task structure. Apply the same approach for the load_fpu_regs() function. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-29 21:43:06 +07:00
lg %r4,__THREAD_FPU_regs(%r4) # %r4 <- reg save area
s390/fpu: always enable the vector facility if it is available If the kernel detects that the s390 hardware supports the vector facility, it is enabled by default at an early stage. To force it off, use the novx kernel parameter. Note that there is a small time window, where the vector facility is enabled before it is forced to be off. With enabling the vector facility by default, the FPU save and restore functions can be improved. They do not longer require to manage expensive control register updates to enable or disable the vector enablement control for particular processes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-09-29 15:04:41 +07:00
jz .Lload_fpu_regs_fp # -> no VX, load FP regs
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
VLM %v0,%v15,0,%r4
VLM %v16,%v31,256,%r4
j .Lload_fpu_regs_done
.Lload_fpu_regs_fp:
ld 0,0(%r4)
ld 1,8(%r4)
ld 2,16(%r4)
ld 3,24(%r4)
ld 4,32(%r4)
ld 5,40(%r4)
ld 6,48(%r4)
ld 7,56(%r4)
ld 8,64(%r4)
ld 9,72(%r4)
ld 10,80(%r4)
ld 11,88(%r4)
ld 12,96(%r4)
ld 13,104(%r4)
ld 14,112(%r4)
ld 15,120(%r4)
.Lload_fpu_regs_done:
ni __LC_CPU_FLAGS+7,255-_CIF_FPU
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
.Lload_fpu_regs_exit:
BR_R1USE_R14
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
.Lload_fpu_regs_end:
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.L__critical_end:
[PATCH] s390: machine check handler bugs The new machine check handler still has a few bugs. 1) The system entry time has to be stored in the machine check handler, 2) the machine check return psw may not be stored at the usual place because it might overwrite the return psw of the interrupted context, 3) the return address for the call to s390_handle_mcck in the i/o interrupt handler is not correct, 4) the system call cleanup has to take the different save area of the machine check handler into account, 5) the machine check handler may not call UPDATE_VTIME before CREATE_STACK_FRAME, and 6) the io leave path needs a critical section cleanup to make sure that the TIF_MCCK_PENDING bit is really checked before switching back to user space. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-04 05:57:56 +07:00
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
/*
* Machine check handler routines
*/
[S390] fix s390 assembler code alignments The alignment is missing for various global symbols in s390 assembly code. With a recent gcc and an instruction like stgrl this can lead to a specification exception if the instruction uses such a mis-aligned address. Specify the alignment explicitely and while add it define __ALIGN for s390 and use the ENTRY define to save some lines of code. Signed-off-by: Jan Glauber <jang@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-07-24 15:48:19 +07:00
ENTRY(mcck_int_handler)
s390/time: always use stckf instead of stck if available The store clock fast instruction saves a couple of instructions compared to the store clock instruction. Always use stckf instead of stck if it is available. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Reviewed-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-05-09 21:27:39 +07:00
STCK __LC_MCCK_CLOCK
s390: add options to change branch prediction behaviour for the kernel Add the PPA instruction to the system entry and exit path to switch the kernel to a different branch prediction behaviour. The instructions are added via CPU alternatives and can be disabled with the "nospec" or the "nobp=0" kernel parameter. If the default behaviour selected with CONFIG_KERNEL_NOBP is set to "n" then the "nobp=1" parameter can be used to enable the changed kernel branch prediction. Acked-by: Cornelia Huck <cohuck@redhat.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:11:45 +07:00
BPOFF
s390/nmi: do register validation as early as possible The validation of the CPU registers in the machine check handler is currently split into two parts. The first part is done at the start of the low level mcck_int_handler function, this includes the CPU timer register and the general purpose registers. The second part is done a bit later in s390_do_machine_check for all the other registers, including the control registers, floating pointer control, vector or floating pointer registers, the access registers, the guarded storage registers, the TOD programmable registers and the clock comparator. This is working fine to far but in theory a future extensions could cause the C code to use registers that are not validated yet. A better approach is to validate all CPU registers in "safe" assembler code before any C function is called. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-10-12 18:24:48 +07:00
la %r1,4095 # validate r1
spt __LC_CPU_TIMER_SAVE_AREA-4095(%r1) # validate cpu timer
sckc __LC_CLOCK_COMPARATOR # validate comparator
lam %a0,%a15,__LC_AREGS_SAVE_AREA-4095(%r1) # validate acrs
lmg %r0,%r15,__LC_GPREGS_SAVE_AREA-4095(%r1)# validate gprs
s390: move thread_info into task_struct This is the s390 variant of commit 15f4eae70d36 ("x86: Move thread_info into task_struct"). Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-11-08 17:08:26 +07:00
lg %r12,__LC_CURRENT
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
larl %r13,cleanup_critical
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lmg %r8,%r9,__LC_MCK_OLD_PSW
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __LC_MCCK_CODE,MCCK_CODE_SYSTEM_DAMAGE
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jo .Lmcck_panic # yes -> rest of mcck code invalid
s390/nmi: do register validation as early as possible The validation of the CPU registers in the machine check handler is currently split into two parts. The first part is done at the start of the low level mcck_int_handler function, this includes the CPU timer register and the general purpose registers. The second part is done a bit later in s390_do_machine_check for all the other registers, including the control registers, floating pointer control, vector or floating pointer registers, the access registers, the guarded storage registers, the TOD programmable registers and the clock comparator. This is working fine to far but in theory a future extensions could cause the C code to use registers that are not validated yet. A better approach is to validate all CPU registers in "safe" assembler code before any C function is called. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-10-12 18:24:48 +07:00
TSTMSK __LC_MCCK_CODE,MCCK_CODE_CR_VALID
jno .Lmcck_panic # control registers invalid -> panic
la %r14,4095
lctlg %c0,%c15,__LC_CREGS_SAVE_AREA-4095(%r14) # validate ctl regs
ptlb
s390/nmi: avoid using long-displacement facility __LC_MCESAD is currently 4528 /* offsetof(struct lowcore, mcesad) */ that would require long-displacement facility for lg, which we don't have on z900. Fixes: 3037a52f9846 ("s390/nmi: do register validation as early as possible") Signed-off-by: Vasily Gorbik <gor@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-10-27 17:44:48 +07:00
lg %r11,__LC_MCESAD-4095(%r14) # extended machine check save area
s390/nmi: do register validation as early as possible The validation of the CPU registers in the machine check handler is currently split into two parts. The first part is done at the start of the low level mcck_int_handler function, this includes the CPU timer register and the general purpose registers. The second part is done a bit later in s390_do_machine_check for all the other registers, including the control registers, floating pointer control, vector or floating pointer registers, the access registers, the guarded storage registers, the TOD programmable registers and the clock comparator. This is working fine to far but in theory a future extensions could cause the C code to use registers that are not validated yet. A better approach is to validate all CPU registers in "safe" assembler code before any C function is called. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-10-12 18:24:48 +07:00
nill %r11,0xfc00 # MCESA_ORIGIN_MASK
TSTMSK __LC_CREGS_SAVE_AREA+16-4095(%r14),CR2_GUARDED_STORAGE
jno 0f
TSTMSK __LC_MCCK_CODE,MCCK_CODE_GS_VALID
jno 0f
.insn rxy,0xe3000000004d,0,__MCESA_GS_SAVE_AREA(%r11) # LGSC
0: l %r14,__LC_FP_CREG_SAVE_AREA-4095(%r14)
TSTMSK __LC_MCCK_CODE,MCCK_CODE_FC_VALID
jo 0f
sr %r14,%r14
0: sfpc %r14
TSTMSK __LC_MACHINE_FLAGS,MACHINE_FLAG_VX
jo 0f
lghi %r14,__LC_FPREGS_SAVE_AREA
ld %f0,0(%r14)
ld %f1,8(%r14)
ld %f2,16(%r14)
ld %f3,24(%r14)
ld %f4,32(%r14)
ld %f5,40(%r14)
ld %f6,48(%r14)
ld %f7,56(%r14)
ld %f8,64(%r14)
ld %f9,72(%r14)
ld %f10,80(%r14)
ld %f11,88(%r14)
ld %f12,96(%r14)
ld %f13,104(%r14)
ld %f14,112(%r14)
ld %f15,120(%r14)
j 1f
0: VLM %v0,%v15,0,%r11
VLM %v16,%v31,256,%r11
1: lghi %r14,__LC_CPU_TIMER_SAVE_AREA
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
mvc __LC_MCCK_ENTER_TIMER(8),0(%r14)
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __LC_MCCK_CODE,MCCK_CODE_CPU_TIMER_VALID
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
jo 3f
[S390] virtual cpu accounting vs. machine checks. If a machine checks interrupts the external or the i/o interrupt handler before they have completed the cpu time calculations, the accounting goes wrong. After the cpu returned from the machine check handler to the interrupted interrupt handler, a negative cpu time delta can occur. If the accumulated cpu time in lowcore is small enough this value can get negative as well. The next jiffy interrupt will pick up that negative value, shift it by 12 and add the now huge positive value to the cpu time of the process. To solve this the machine check handler is modified not to change any of the timestamps in the lowcore if the machine check interrupted kernel context. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2006-06-29 19:58:05 +07:00
la %r14,__LC_SYNC_ENTER_TIMER
clc 0(8,%r14),__LC_ASYNC_ENTER_TIMER
jl 0f
la %r14,__LC_ASYNC_ENTER_TIMER
0: clc 0(8,%r14),__LC_EXIT_TIMER
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
jl 1f
[S390] virtual cpu accounting vs. machine checks. If a machine checks interrupts the external or the i/o interrupt handler before they have completed the cpu time calculations, the accounting goes wrong. After the cpu returned from the machine check handler to the interrupted interrupt handler, a negative cpu time delta can occur. If the accumulated cpu time in lowcore is small enough this value can get negative as well. The next jiffy interrupt will pick up that negative value, shift it by 12 and add the now huge positive value to the cpu time of the process. To solve this the machine check handler is modified not to change any of the timestamps in the lowcore if the machine check interrupted kernel context. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2006-06-29 19:58:05 +07:00
la %r14,__LC_EXIT_TIMER
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
1: clc 0(8,%r14),__LC_LAST_UPDATE_TIMER
jl 2f
[S390] virtual cpu accounting vs. machine checks. If a machine checks interrupts the external or the i/o interrupt handler before they have completed the cpu time calculations, the accounting goes wrong. After the cpu returned from the machine check handler to the interrupted interrupt handler, a negative cpu time delta can occur. If the accumulated cpu time in lowcore is small enough this value can get negative as well. The next jiffy interrupt will pick up that negative value, shift it by 12 and add the now huge positive value to the cpu time of the process. To solve this the machine check handler is modified not to change any of the timestamps in the lowcore if the machine check interrupted kernel context. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2006-06-29 19:58:05 +07:00
la %r14,__LC_LAST_UPDATE_TIMER
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
2: spt 0(%r14)
[S390] idle time accounting vs. machine checks A machine check can interrupt the i/o and external interrupt handler anytime. If the machine check occurs while the interrupt handler is waking up from idle vtime_start_cpu can get executed a second time and the int_clock / async_enter_timer values in the lowcore get clobbered. This can confuse the cpu time accounting. To fix this problem two changes are needed. First the machine check handler has to use its own copies of int_clock and async_enter_timer, named mcck_clock and mcck_enter_timer. Second the nested execution of vtime_start_cpu has to be prevented. This is done in s390_idle_check by checking the wait bit in the program status word. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-05-17 15:00:03 +07:00
mvc __LC_MCCK_ENTER_TIMER(8),0(%r14)
s390/nmi: do register validation as early as possible The validation of the CPU registers in the machine check handler is currently split into two parts. The first part is done at the start of the low level mcck_int_handler function, this includes the CPU timer register and the general purpose registers. The second part is done a bit later in s390_do_machine_check for all the other registers, including the control registers, floating pointer control, vector or floating pointer registers, the access registers, the guarded storage registers, the TOD programmable registers and the clock comparator. This is working fine to far but in theory a future extensions could cause the C code to use registers that are not validated yet. A better approach is to validate all CPU registers in "safe" assembler code before any C function is called. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-10-12 18:24:48 +07:00
3: TSTMSK __LC_MCCK_CODE,MCCK_CODE_PSW_MWP_VALID
jno .Lmcck_panic
tmhh %r8,0x0001 # interrupting from user ?
jnz 4f
TSTMSK __LC_MCCK_CODE,MCCK_CODE_PSW_IA_VALID
jno .Lmcck_panic
4: SWITCH_ASYNC __LC_GPREGS_SAVE_AREA+64,__LC_MCCK_ENTER_TIMER
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lmcck_skip:
s390: critical section cleanup vs. machine checks The current machine check code uses the registers stored by the machine in the lowcore at __LC_GPREGS_SAVE_AREA as the registers of the interrupted context. The registers 0-7 of a user process can get clobbered if a machine checks interrupts the execution of a critical section in entry[64].S. The reason is that the critical section cleanup code may need to modify the PSW and the registers for the previous context to get to the end of a critical section. If registers 0-7 have to be replaced the relevant copy will be in the registers, which invalidates the copy in the lowcore. The machine check handler needs to explicitly store registers 0-7 to the stack. Cc: stable@vger.kernel.org Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-02-28 22:28:41 +07:00
lghi %r14,__LC_GPREGS_SAVE_AREA+64
stmg %r0,%r7,__PT_R0(%r11)
s390: scrub registers on kernel entry and KVM exit Clear all user space registers on entry to the kernel and all KVM guest registers on KVM guest exit if the register does not contain either a parameter or a result value. Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 19:27:30 +07:00
# clear user controlled registers to prevent speculative use
xgr %r0,%r0
xgr %r1,%r1
xgr %r2,%r2
xgr %r3,%r3
xgr %r4,%r4
xgr %r5,%r5
xgr %r6,%r6
xgr %r7,%r7
xgr %r10,%r10
s390: critical section cleanup vs. machine checks The current machine check code uses the registers stored by the machine in the lowcore at __LC_GPREGS_SAVE_AREA as the registers of the interrupted context. The registers 0-7 of a user process can get clobbered if a machine checks interrupts the execution of a critical section in entry[64].S. The reason is that the critical section cleanup code may need to modify the PSW and the registers for the previous context to get to the end of a critical section. If registers 0-7 have to be replaced the relevant copy will be in the registers, which invalidates the copy in the lowcore. The machine check handler needs to explicitly store registers 0-7 to the stack. Cc: stable@vger.kernel.org Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-02-28 22:28:41 +07:00
mvc __PT_R8(64,%r11),0(%r14)
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stmg %r8,%r9,__PT_PSW(%r11)
s390: split TIF bits into CIF, PIF and TIF bits The oi and ni instructions used in entry[64].S to set and clear bits in the thread-flags are not guaranteed to be atomic in regard to other CPUs. Split the TIF bits into CPU, pt_regs and thread-info specific bits. Updates on the TIF bits are done with atomic instructions, updates on CPU and pt_regs bits are done with non-atomic instructions. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-04-15 17:55:07 +07:00
xc __PT_FLAGS(8,%r11),__PT_FLAGS(%r11)
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15)
lgr %r2,%r11 # pass pointer to pt_regs
[PATCH] s390: improved machine check handling Improved machine check handling. Kernel is now able to receive machine checks while in kernel mode (system call, interrupt and program check handling). Also register validation is now performed. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-26 04:55:30 +07:00
brasl %r14,s390_do_machine_check
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
tm __PT_PSW+1(%r11),0x01 # returning to user ?
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jno .Lmcck_return
[PATCH] s390: improved machine check handling Improved machine check handling. Kernel is now able to receive machine checks while in kernel mode (system call, interrupt and program check handling). Also register validation is now performed. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-26 04:55:30 +07:00
lg %r1,__LC_KERNEL_STACK # switch to kernel stack
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
mvc STACK_FRAME_OVERHEAD(__PT_SIZE,%r1),0(%r11)
xc __SF_BACKCHAIN(8,%r1),__SF_BACKCHAIN(%r1)
la %r11,STACK_FRAME_OVERHEAD(%r1)
[PATCH] s390: improved machine check handling Improved machine check handling. Kernel is now able to receive machine checks while in kernel mode (system call, interrupt and program check handling). Also register validation is now performed. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-26 04:55:30 +07:00
lgr %r15,%r1
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
ssm __LC_PGM_NEW_PSW # turn dat on, keep irqs off
s390/entry: add assembler macro to conveniently tests under mask Various functions in entry.S perform test-under-mask instructions to test for particular bits in memory. Because test-under-mask uses a mask value of one byte, the mask value and the offset into the memory must be calculated manually. This easily introduces errors and is hard to review and read. Introduce the TSTMSK assembler macro to specify a mask constant and let the macro calculate the offset and the byte mask to generate a test-under-mask instruction. The benefit is that existing symbolic constants can now be used for tests. Also the macro checks for zero mask values and mask values that consist of multiple bytes. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-01 22:02:48 +07:00
TSTMSK __LC_CPU_FLAGS,_CIF_MCCK_PENDING
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
jno .Lmcck_return
[PATCH] lockdep: irqtrace subsystem, s390 support irqtrace support for s390. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-07-03 14:24:46 +07:00
TRACE_IRQS_OFF
[PATCH] s390: improved machine check handling Improved machine check handling. Kernel is now able to receive machine checks while in kernel mode (system call, interrupt and program check handling). Also register validation is now performed. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-26 04:55:30 +07:00
brasl %r14,s390_handle_mcck
[PATCH] lockdep: irqtrace subsystem, s390 support irqtrace support for s390. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-07-03 14:24:46 +07:00
TRACE_IRQS_ON
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lmcck_return:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lg %r14,__LC_VDSO_PER_CPU
lmg %r0,%r10,__PT_R0(%r11)
mvc __LC_RETURN_MCCK_PSW(16),__PT_PSW(%r11) # move return PSW
[S390] virtual cpu accounting vs. machine checks. If a machine checks interrupts the external or the i/o interrupt handler before they have completed the cpu time calculations, the accounting goes wrong. After the cpu returned from the machine check handler to the interrupted interrupt handler, a negative cpu time delta can occur. If the accumulated cpu time in lowcore is small enough this value can get negative as well. The next jiffy interrupt will pick up that negative value, shift it by 12 and add the now huge positive value to the cpu time of the process. To solve this the machine check handler is modified not to change any of the timestamps in the lowcore if the machine check interrupted kernel context. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2006-06-29 19:58:05 +07:00
tm __LC_RETURN_MCCK_PSW+1,0x01 # returning to user ?
jno 0f
s390: run user space and KVM guests with modified branch prediction Define TIF_ISOLATE_BP and TIF_ISOLATE_BP_GUEST and add the necessary plumbing in entry.S to be able to run user space and KVM guests with limited branch prediction. To switch a user space process to limited branch prediction the s390_isolate_bp() function has to be call, and to run a vCPU of a KVM guest associated with the current task with limited branch prediction call s390_isolate_bp_guest(). Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-16 13:36:46 +07:00
BPEXIT __TI_flags(%r12),_TIF_ISOLATE_BP
[S390] virtual cpu accounting vs. machine checks. If a machine checks interrupts the external or the i/o interrupt handler before they have completed the cpu time calculations, the accounting goes wrong. After the cpu returned from the machine check handler to the interrupted interrupt handler, a negative cpu time delta can occur. If the accumulated cpu time in lowcore is small enough this value can get negative as well. The next jiffy interrupt will pick up that negative value, shift it by 12 and add the now huge positive value to the cpu time of the process. To solve this the machine check handler is modified not to change any of the timestamps in the lowcore if the machine check interrupted kernel context. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2006-06-29 19:58:05 +07:00
stpt __LC_EXIT_TIMER
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
mvc __VDSO_ECTG_BASE(16,%r14),__LC_EXIT_TIMER
0: lmg %r11,%r15,__PT_R11(%r11)
lpswe __LC_RETURN_MCCK_PSW
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lmcck_panic:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lg %r15,__LC_PANIC_STACK
s390: fix machine check panic stack switch For system damage machine checks or machine checks due to invalid PSW fields the system will be stopped. In order to get an oops message out before killing the system the machine check handler branches to .Lmcck_panic, switches to the panic stack and then does the usual machine check handling. The switch to the panic stack is incomplete, the stack pointer in %r15 is replaced, but the pt_regs pointer in %r11 is not. The result is a program check which will kill the system in a slightly different way. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-12-02 19:29:22 +07:00
la %r11,STACK_FRAME_OVERHEAD(%r15)
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
j .Lmcck_skip
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
[S390] Add PSW restart shutdown trigger With this patch a new S390 shutdown trigger "restart" is added. If under z/VM "systerm restart" is entered or under the HMC the "PSW restart" button is pressed, the PSW located at 0 (31 bit) or 0x1a0 (64 bit) bit is loaded. Now we execute do_restart() that processes the restart action that is defined under /sys/firmware/shutdown_actions/on_restart. Currently the following actions are possible: reipl (default), stop, vmcmd, dump, and dump_reipl. Signed-off-by: Michael Holzheu <holzheu@linux.vnet.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
2011-08-03 21:44:19 +07:00
#
# PSW restart interrupt handler
#
[S390] rework smp code Define struct pcpu and merge some of the NR_CPUS arrays into it, including __cpu_logical_map, current_set and smp_cpu_state. Split smp related functions to those operating on physical cpus and the functions operating on a logical cpu number. Make the functions for physical cpus use a pointer to a struct pcpu. This hides the knowledge about cpu addresses in smp.c, entry[64].S and swsusp_asm64.S, thus remove the sigp.h header. The PSW restart mechanism is used to start secondary cpus, calling a function on an online cpu, calling a function on the ipl cpu, and for the nmi signal. Replace the different assembler functions with a single function restart_int_handler. The new entry point calls a function whose pointer is stored in the lowcore of the target cpu and it can wait for the source cpu to stop. This covers all existing use cases. Overall the code is now simpler and there are ~380 lines less code. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:26 +07:00
ENTRY(restart_int_handler)
s390/lpp: use assembler alternatives for the LPP instruction With the new macros for CPU alternatives the MACHINE_FLAG_LPP check around the LPP instruction can be optimized. After this is done the flag can be removed. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-03-26 20:23:33 +07:00
ALTERNATIVE "", ".insn s,0xb2800000,_LPP_OFFSET", 40
stg %r15,__LC_SAVE_AREA_RESTART
[S390] rework smp code Define struct pcpu and merge some of the NR_CPUS arrays into it, including __cpu_logical_map, current_set and smp_cpu_state. Split smp related functions to those operating on physical cpus and the functions operating on a logical cpu number. Make the functions for physical cpus use a pointer to a struct pcpu. This hides the knowledge about cpu addresses in smp.c, entry[64].S and swsusp_asm64.S, thus remove the sigp.h header. The PSW restart mechanism is used to start secondary cpus, calling a function on an online cpu, calling a function on the ipl cpu, and for the nmi signal. Replace the different assembler functions with a single function restart_int_handler. The new entry point calls a function whose pointer is stored in the lowcore of the target cpu and it can wait for the source cpu to stop. This covers all existing use cases. Overall the code is now simpler and there are ~380 lines less code. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:26 +07:00
lg %r15,__LC_RESTART_STACK
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
aghi %r15,-__PT_SIZE # create pt_regs on stack
[S390] rework smp code Define struct pcpu and merge some of the NR_CPUS arrays into it, including __cpu_logical_map, current_set and smp_cpu_state. Split smp related functions to those operating on physical cpus and the functions operating on a logical cpu number. Make the functions for physical cpus use a pointer to a struct pcpu. This hides the knowledge about cpu addresses in smp.c, entry[64].S and swsusp_asm64.S, thus remove the sigp.h header. The PSW restart mechanism is used to start secondary cpus, calling a function on an online cpu, calling a function on the ipl cpu, and for the nmi signal. Replace the different assembler functions with a single function restart_int_handler. The new entry point calls a function whose pointer is stored in the lowcore of the target cpu and it can wait for the source cpu to stop. This covers all existing use cases. Overall the code is now simpler and there are ~380 lines less code. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:26 +07:00
xc 0(__PT_SIZE,%r15),0(%r15)
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stmg %r0,%r14,__PT_R0(%r15)
mvc __PT_R15(8,%r15),__LC_SAVE_AREA_RESTART
mvc __PT_PSW(16,%r15),__LC_RST_OLD_PSW # store restart old psw
[S390] rework smp code Define struct pcpu and merge some of the NR_CPUS arrays into it, including __cpu_logical_map, current_set and smp_cpu_state. Split smp related functions to those operating on physical cpus and the functions operating on a logical cpu number. Make the functions for physical cpus use a pointer to a struct pcpu. This hides the knowledge about cpu addresses in smp.c, entry[64].S and swsusp_asm64.S, thus remove the sigp.h header. The PSW restart mechanism is used to start secondary cpus, calling a function on an online cpu, calling a function on the ipl cpu, and for the nmi signal. Replace the different assembler functions with a single function restart_int_handler. The new entry point calls a function whose pointer is stored in the lowcore of the target cpu and it can wait for the source cpu to stop. This covers all existing use cases. Overall the code is now simpler and there are ~380 lines less code. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:26 +07:00
aghi %r15,-STACK_FRAME_OVERHEAD # create stack frame on stack
xc 0(STACK_FRAME_OVERHEAD,%r15),0(%r15)
s390/smp: make absolute lowcore / cpu restart parameter accesses more robust Setting the cpu restart parameters is done in three different fashions: - directly setting the four parameters individually - copying the four parameters with memcpy (using 4 * sizeof(long)) - copying the four parameters using a private structure In addition code in entry*.S relies on a certain order of the restart members of struct _lowcore. Make all of this more robust to future changes by adding a mem_absolute_assign(dest, val) define, which assigns val to dest using absolute addressing mode. Also the load multiple instructions in entry*.S have been split into separate load instruction so the order of the struct _lowcore members doesn't matter anymore. In addition move the prototypes of memcpy_real/absolute from uaccess.h to processor.h. These memcpy* variants are not related to uaccess at all. string.h doesn't seem to match as well, so lets use processor.h. Also replace the eight byte array in struct _lowcore which represents a misaliged u64 with a u64. The compiler will always create code that handles the misaligned u64 correctly. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-06-05 14:59:52 +07:00
lg %r1,__LC_RESTART_FN # load fn, parm & source cpu
lg %r2,__LC_RESTART_DATA
lg %r3,__LC_RESTART_SOURCE
[S390] rework smp code Define struct pcpu and merge some of the NR_CPUS arrays into it, including __cpu_logical_map, current_set and smp_cpu_state. Split smp related functions to those operating on physical cpus and the functions operating on a logical cpu number. Make the functions for physical cpus use a pointer to a struct pcpu. This hides the knowledge about cpu addresses in smp.c, entry[64].S and swsusp_asm64.S, thus remove the sigp.h header. The PSW restart mechanism is used to start secondary cpus, calling a function on an online cpu, calling a function on the ipl cpu, and for the nmi signal. Replace the different assembler functions with a single function restart_int_handler. The new entry point calls a function whose pointer is stored in the lowcore of the target cpu and it can wait for the source cpu to stop. This covers all existing use cases. Overall the code is now simpler and there are ~380 lines less code. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:26 +07:00
ltgr %r3,%r3 # test source cpu address
jm 1f # negative -> skip source stop
s390/sigp: use sigp order code defines in assembly code Use sigp order code defines in assembly code as well. With this change all places that use sigp constants should have been converted to use self describing defines instead of directly using constants. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-06-04 20:05:43 +07:00
0: sigp %r4,%r3,SIGP_SENSE # sigp sense to source cpu
[S390] rework smp code Define struct pcpu and merge some of the NR_CPUS arrays into it, including __cpu_logical_map, current_set and smp_cpu_state. Split smp related functions to those operating on physical cpus and the functions operating on a logical cpu number. Make the functions for physical cpus use a pointer to a struct pcpu. This hides the knowledge about cpu addresses in smp.c, entry[64].S and swsusp_asm64.S, thus remove the sigp.h header. The PSW restart mechanism is used to start secondary cpus, calling a function on an online cpu, calling a function on the ipl cpu, and for the nmi signal. Replace the different assembler functions with a single function restart_int_handler. The new entry point calls a function whose pointer is stored in the lowcore of the target cpu and it can wait for the source cpu to stop. This covers all existing use cases. Overall the code is now simpler and there are ~380 lines less code. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:26 +07:00
brc 10,0b # wait for status stored
1: basr %r14,%r1 # call function
stap __SF_EMPTY(%r15) # store cpu address
llgh %r3,__SF_EMPTY(%r15)
s390/sigp: use sigp order code defines in assembly code Use sigp order code defines in assembly code as well. With this change all places that use sigp constants should have been converted to use self describing defines instead of directly using constants. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-06-04 20:05:43 +07:00
2: sigp %r4,%r3,SIGP_STOP # sigp stop to current cpu
[S390] rework smp code Define struct pcpu and merge some of the NR_CPUS arrays into it, including __cpu_logical_map, current_set and smp_cpu_state. Split smp related functions to those operating on physical cpus and the functions operating on a logical cpu number. Make the functions for physical cpus use a pointer to a struct pcpu. This hides the knowledge about cpu addresses in smp.c, entry[64].S and swsusp_asm64.S, thus remove the sigp.h header. The PSW restart mechanism is used to start secondary cpus, calling a function on an online cpu, calling a function on the ipl cpu, and for the nmi signal. Replace the different assembler functions with a single function restart_int_handler. The new entry point calls a function whose pointer is stored in the lowcore of the target cpu and it can wait for the source cpu to stop. This covers all existing use cases. Overall the code is now simpler and there are ~380 lines less code. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:26 +07:00
brc 2,2b
3: j 3b
[S390] Add PSW restart shutdown trigger With this patch a new S390 shutdown trigger "restart" is added. If under z/VM "systerm restart" is entered or under the HMC the "PSW restart" button is pressed, the PSW located at 0 (31 bit) or 0x1a0 (64 bit) bit is loaded. Now we execute do_restart() that processes the restart action that is defined under /sys/firmware/shutdown_actions/on_restart. Currently the following actions are possible: reipl (default), stop, vmcmd, dump, and dump_reipl. Signed-off-by: Michael Holzheu <holzheu@linux.vnet.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
2011-08-03 21:44:19 +07:00
[S390] add kprobes annotations Add kprobes annotations to get the massive 'probe kernel.function("*") {}' stress test working. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-01-05 18:47:25 +07:00
.section .kprobes.text, "ax"
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#ifdef CONFIG_CHECK_STACK
/*
* The synchronous or the asynchronous stack overflowed. We are dead.
* No need to properly save the registers, we are going to panic anyway.
* Setup a pt_regs so that show_trace can provide a good call trace.
*/
stack_overflow:
s390: lowcore stack pointer offsets Store the stack pointers in the lowcore for the kernel stack, the async stack and the panic stack with the offset required for the first user. This avoids an unnecessary add instruction on the system call path. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-04-24 15:20:43 +07:00
lg %r15,__LC_PANIC_STACK # change to panic stack
la %r11,STACK_FRAME_OVERHEAD(%r15)
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
stmg %r0,%r7,__PT_R0(%r11)
stmg %r8,%r9,__PT_PSW(%r11)
mvc __PT_R8(64,%r11),0(%r14)
stg %r10,__PT_ORIG_GPR2(%r11) # store last break to orig_gpr2
xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15)
lgr %r2,%r11 # pass pointer to pt_regs
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
jg kernel_stack_overflow
#endif
cleanup_critical:
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
#if IS_ENABLED(CONFIG_KVM)
clg %r9,BASED(.Lcleanup_table_sie) # .Lsie_gmap
jl 0f
clg %r9,BASED(.Lcleanup_table_sie+8)# .Lsie_done
jl .Lcleanup_sie
#endif
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
clg %r9,BASED(.Lcleanup_table) # system_call
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
jl 0f
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
clg %r9,BASED(.Lcleanup_table+8) # .Lsysc_do_svc
jl .Lcleanup_system_call
clg %r9,BASED(.Lcleanup_table+16) # .Lsysc_tif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
jl 0f
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
clg %r9,BASED(.Lcleanup_table+24) # .Lsysc_restore
jl .Lcleanup_sysc_tif
clg %r9,BASED(.Lcleanup_table+32) # .Lsysc_done
jl .Lcleanup_sysc_restore
clg %r9,BASED(.Lcleanup_table+40) # .Lio_tif
[S390] virtual cpu accounting vs. machine checks. If a machine checks interrupts the external or the i/o interrupt handler before they have completed the cpu time calculations, the accounting goes wrong. After the cpu returned from the machine check handler to the interrupted interrupt handler, a negative cpu time delta can occur. If the accumulated cpu time in lowcore is small enough this value can get negative as well. The next jiffy interrupt will pick up that negative value, shift it by 12 and add the now huge positive value to the cpu time of the process. To solve this the machine check handler is modified not to change any of the timestamps in the lowcore if the machine check interrupted kernel context. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2006-06-29 19:58:05 +07:00
jl 0f
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
clg %r9,BASED(.Lcleanup_table+48) # .Lio_restore
jl .Lcleanup_io_tif
clg %r9,BASED(.Lcleanup_table+56) # .Lio_done
jl .Lcleanup_io_restore
clg %r9,BASED(.Lcleanup_table+64) # psw_idle
[S390] rework idle code Whenever the cpu loads an enabled wait PSW it will appear as idle to the underlying host system. The code in default_idle calls vtime_stop_cpu which does the necessary voodoo to get the cpu time accounting right. The udelay code just loads an enabled wait PSW. To correct this rework the vtime_stop_cpu/vtime_start_cpu logic and move the difficult parts to entry[64].S, vtime_stop_cpu can now be called from anywhere and vtime_start_cpu is gone. The correction of the cpu time during wakeup from an enabled wait PSW is done with a critical section in entry[64].S. As vtime_start_cpu is gone, s390_idle_check can be removed as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:27 +07:00
jl 0f
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
clg %r9,BASED(.Lcleanup_table+72) # .Lpsw_idle_end
jl .Lcleanup_idle
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
clg %r9,BASED(.Lcleanup_table+80) # save_fpu_regs
jl 0f
clg %r9,BASED(.Lcleanup_table+88) # .Lsave_fpu_regs_end
jl .Lcleanup_save_fpu_regs
clg %r9,BASED(.Lcleanup_table+96) # load_fpu_regs
jl 0f
clg %r9,BASED(.Lcleanup_table+104) # .Lload_fpu_regs_end
jl .Lcleanup_load_fpu_regs
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
0: BR_R11USE_R14
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
.align 8
.Lcleanup_table:
.quad system_call
.quad .Lsysc_do_svc
.quad .Lsysc_tif
.quad .Lsysc_restore
.quad .Lsysc_done
.quad .Lio_tif
.quad .Lio_restore
.quad .Lio_done
.quad psw_idle
.quad .Lpsw_idle_end
.quad save_fpu_regs
.quad .Lsave_fpu_regs_end
.quad load_fpu_regs
.quad .Lload_fpu_regs_end
#if IS_ENABLED(CONFIG_KVM)
.Lcleanup_table_sie:
.quad .Lsie_gmap
.quad .Lsie_done
.Lcleanup_sie:
s390/nmi: s390: New low level handling for machine check happening in guest Add the logic to check if the machine check happens when the guest is running. If yes, set the exit reason -EINTR in the machine check's interrupt handler. Refactor s390_do_machine_check to avoid panicing the host for some kinds of machine checks which happen when guest is running. Reinject the instruction processing damage's machine checks including Delayed Access Exception instead of damaging the host if it happens in the guest because it could be caused by improper update on TLB entry or other software case and impacts the guest only. Signed-off-by: QingFeng Hao <haoqf@linux.vnet.ibm.com> Acked-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com> Acked-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
2017-06-07 16:30:42 +07:00
cghi %r11,__LC_SAVE_AREA_ASYNC #Is this in normal interrupt?
je 1f
slg %r9,BASED(.Lsie_crit_mcck_start)
clg %r9,BASED(.Lsie_crit_mcck_length)
jh 1f
oi __LC_CPU_FLAGS+7, _CIF_MCCK_GUEST
s390/kvm: improve stack frame constants in entry.S The code in sie64a uses the stack frame passed to the function to store some temporary data in the empty1 array (see struct stack_frame in asm/processor.h. Replace the __SF_EMPTY+x constants with a properly defined offset: s/__SF_EMPTY/__SF_SIE_CONTROL/, s/__SF_EMPTY+8/__SF_SIE_SAVEAREA/, s/__SF_EMPTY+16/__SF_SIE_REASON/, s/__SF_EMPTY+24/__SF_SIE_FLAGS/. Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-03-20 19:33:43 +07:00
1: BPENTER __SF_SIE_FLAGS(%r15),(_TIF_ISOLATE_BP|_TIF_ISOLATE_BP_GUEST)
lg %r9,__SF_SIE_CONTROL(%r15) # get control block pointer
s390/cpumf: rework program parameter setting to detect guest samples The program parameter can be used to mark hardware samples with some token. Previously, it was used to mark guest samples only. Improve the program parameter doubleword by combining two parts, the leftmost LPP part and the rightmost PID part. Set the PID part for processes by using the task PID. To distinguish host and guest samples for the kernel (PID part is zero), the guest must always set the program paramater to a non-zero value. Use the leftmost bit in the LPP part of the program parameter to be able to detect guest kernel samples. [brueckner@linux.vnet.ibm.com]: Split __LC_CURRENT and introduced __LC_LPP. Corrected __LC_CURRENT users and adjusted assembler parts. And updated the commit message accordingly. Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-10-06 23:06:15 +07:00
ni __SIE_PROG0C+3(%r9),0xfe # no longer in SIE
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
lctlg %c1,%c1,__LC_USER_ASCE # load primary asce
larl %r9,sie_exit # skip forward to sie_exit
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
BR_R11USE_R14
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
#endif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lcleanup_system_call:
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
# check if stpt has been executed
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
clg %r9,BASED(.Lcleanup_system_call_insn)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
jh 0f
mvc __LC_SYNC_ENTER_TIMER(8),__LC_ASYNC_ENTER_TIMER
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
cghi %r11,__LC_SAVE_AREA_ASYNC
[S390] idle time accounting vs. machine checks A machine check can interrupt the i/o and external interrupt handler anytime. If the machine check occurs while the interrupt handler is waking up from idle vtime_start_cpu can get executed a second time and the int_clock / async_enter_timer values in the lowcore get clobbered. This can confuse the cpu time accounting. To fix this problem two changes are needed. First the machine check handler has to use its own copies of int_clock and async_enter_timer, named mcck_clock and mcck_enter_timer. Second the nested execution of vtime_start_cpu has to be prevented. This is done in s390_idle_check by checking the wait bit in the program status word. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-05-17 15:00:03 +07:00
je 0f
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
mvc __LC_SYNC_ENTER_TIMER(8),__LC_MCCK_ENTER_TIMER
0: # check if stmg has been executed
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
clg %r9,BASED(.Lcleanup_system_call_insn+8)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
jh 0f
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
mvc __LC_SAVE_AREA_SYNC(64),0(%r11)
0: # check if base register setup + TIF bit load has been done
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
clg %r9,BASED(.Lcleanup_system_call_insn+16)
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
jhe 0f
s390: store breaking event address only for program checks The principles of operations specifies that the breaking event address is stored to the address 0x110 in the prefix page only for program checks. The last branch in user space is lost as soon as a branch in kernel space is executed after e.g. an svc. This makes it impossible to accurately maintain the breaking event address for a user space process. Simplify the code, just copy the current breaking event address from 0x110 to the task structure for program checks from user space. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-01-25 18:54:17 +07:00
# set up saved register r12 task struct pointer
stg %r12,32(%r11)
# set up saved register r13 __TASK_thread offset
mvc 40(8,%r11),BASED(.Lcleanup_system_call_const)
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
0: # check if the user time update has been done
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
clg %r9,BASED(.Lcleanup_system_call_insn+24)
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
jh 0f
lg %r15,__LC_EXIT_TIMER
slg %r15,__LC_SYNC_ENTER_TIMER
alg %r15,__LC_USER_TIMER
stg %r15,__LC_USER_TIMER
0: # check if the system time update has been done
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
clg %r9,BASED(.Lcleanup_system_call_insn+32)
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
jh 0f
lg %r15,__LC_LAST_UPDATE_TIMER
slg %r15,__LC_EXIT_TIMER
alg %r15,__LC_SYSTEM_TIMER
stg %r15,__LC_SYSTEM_TIMER
0: # update accounting time stamp
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
mvc __LC_LAST_UPDATE_TIMER(8),__LC_SYNC_ENTER_TIMER
s390: do not bypass BPENTER for interrupt system calls The system call path can be interrupted before the switch back to the standard branch prediction with BPENTER has been done. The critical section cleanup code skips forward to .Lsysc_do_svc and bypasses the BPENTER. In this case the kernel and all subsequent code will run with the limited branch prediction. Fixes: eacf67eb9b32 ("s390: run user space and KVM guests with modified branch prediction") Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-02-22 19:42:29 +07:00
BPENTER __TI_flags(%r12),_TIF_ISOLATE_BP
s390: store breaking event address only for program checks The principles of operations specifies that the breaking event address is stored to the address 0x110 in the prefix page only for program checks. The last branch in user space is lost as soon as a branch in kernel space is executed after e.g. an svc. This makes it impossible to accurately maintain the breaking event address for a user space process. Simplify the code, just copy the current breaking event address from 0x110 to the task structure for program checks from user space. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-01-25 18:54:17 +07:00
# set up saved register r11
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lg %r15,__LC_KERNEL_STACK
s390: lowcore stack pointer offsets Store the stack pointers in the lowcore for the kernel stack, the async stack and the panic stack with the offset required for the first user. This avoids an unnecessary add instruction on the system call path. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-04-24 15:20:43 +07:00
la %r9,STACK_FRAME_OVERHEAD(%r15)
stg %r9,24(%r11) # r11 pt_regs pointer
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
# fill pt_regs
s390: lowcore stack pointer offsets Store the stack pointers in the lowcore for the kernel stack, the async stack and the panic stack with the offset required for the first user. This avoids an unnecessary add instruction on the system call path. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-04-24 15:20:43 +07:00
mvc __PT_R8(64,%r9),__LC_SAVE_AREA_SYNC
stmg %r0,%r7,__PT_R0(%r9)
mvc __PT_PSW(16,%r9),__LC_SVC_OLD_PSW
mvc __PT_INT_CODE(4,%r9),__LC_SVC_ILC
s390: split TIF bits into CIF, PIF and TIF bits The oi and ni instructions used in entry[64].S to set and clear bits in the thread-flags are not guaranteed to be atomic in regard to other CPUs. Split the TIF bits into CPU, pt_regs and thread-info specific bits. Updates on the TIF bits are done with atomic instructions, updates on CPU and pt_regs bits are done with non-atomic instructions. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-04-15 17:55:07 +07:00
xc __PT_FLAGS(8,%r9),__PT_FLAGS(%r9)
mvi __PT_FLAGS+7(%r9),_PIF_SYSCALL
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
# setup saved register r15
stg %r15,56(%r11) # r15 stack pointer
# set new psw address and exit
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
larl %r9,.Lsysc_do_svc
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
BR_R11USE_R14
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lcleanup_system_call_insn:
[S390] Whitespace cleanup. Huge s390 assembly files whitespace cleanup. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2006-09-28 21:56:37 +07:00
.quad system_call
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.quad .Lsysc_stmg
.quad .Lsysc_per
s390/kernel: squeeze a few more cycles out of the system call handler Reorder the instructions of UPDATE_VTIME to improve superscalar execution, remove duplicate checks for problem-state from the asynchronous interrupt handlers, and move the check for problem-state from the synchronous exit path to the program check path as it is only needed for program checks inside the kernel. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:27:48 +07:00
.quad .Lsysc_vtime+36
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.quad .Lsysc_vtime+42
s390: store breaking event address only for program checks The principles of operations specifies that the breaking event address is stored to the address 0x110 in the prefix page only for program checks. The last branch in user space is lost as soon as a branch in kernel space is executed after e.g. an svc. This makes it impossible to accurately maintain the breaking event address for a user space process. Simplify the code, just copy the current breaking event address from 0x110 to the task structure for program checks from user space. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-01-25 18:54:17 +07:00
.Lcleanup_system_call_const:
.quad __TASK_thread
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lcleanup_sysc_tif:
larl %r9,.Lsysc_tif
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
BR_R11USE_R14
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lcleanup_sysc_restore:
s390/cputime: fix incorrect system time git commit c5328901aa1db134 "[S390] entry[64].S improvements" removed the update of the exit_timer lowcore field from the critical section cleanup of the .Lsysc_restore/.Lsysc_done and .Lio_restore/.Lio_done blocks. If the PSW is updated by the critical section cleanup to point to user space again, the interrupt entry code will do a vtime calculation after the cleanup completed with an exit_timer value which has *not* been updated. Due to this incorrect system time deltas are calculated. If an interrupt occured with an old PSW between .Lsysc_restore/.Lsysc_done or .Lio_restore/.Lio_done update __LC_EXIT_TIMER with the system entry time of the interrupt. Cc: stable@vger.kernel.org # 3.3+ Tested-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-05-02 18:36:00 +07:00
# check if stpt has been executed
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
clg %r9,BASED(.Lcleanup_sysc_restore_insn)
s390/cputime: fix incorrect system time git commit c5328901aa1db134 "[S390] entry[64].S improvements" removed the update of the exit_timer lowcore field from the critical section cleanup of the .Lsysc_restore/.Lsysc_done and .Lio_restore/.Lio_done blocks. If the PSW is updated by the critical section cleanup to point to user space again, the interrupt entry code will do a vtime calculation after the cleanup completed with an exit_timer value which has *not* been updated. Due to this incorrect system time deltas are calculated. If an interrupt occured with an old PSW between .Lsysc_restore/.Lsysc_done or .Lio_restore/.Lio_done update __LC_EXIT_TIMER with the system entry time of the interrupt. Cc: stable@vger.kernel.org # 3.3+ Tested-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-05-02 18:36:00 +07:00
jh 0f
mvc __LC_EXIT_TIMER(8),__LC_ASYNC_ENTER_TIMER
cghi %r11,__LC_SAVE_AREA_ASYNC
[S390] idle time accounting vs. machine checks A machine check can interrupt the i/o and external interrupt handler anytime. If the machine check occurs while the interrupt handler is waking up from idle vtime_start_cpu can get executed a second time and the int_clock / async_enter_timer values in the lowcore get clobbered. This can confuse the cpu time accounting. To fix this problem two changes are needed. First the machine check handler has to use its own copies of int_clock and async_enter_timer, named mcck_clock and mcck_enter_timer. Second the nested execution of vtime_start_cpu has to be prevented. This is done in s390_idle_check by checking the wait bit in the program status word. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-05-17 15:00:03 +07:00
je 0f
s390/cputime: fix incorrect system time git commit c5328901aa1db134 "[S390] entry[64].S improvements" removed the update of the exit_timer lowcore field from the critical section cleanup of the .Lsysc_restore/.Lsysc_done and .Lio_restore/.Lio_done blocks. If the PSW is updated by the critical section cleanup to point to user space again, the interrupt entry code will do a vtime calculation after the cleanup completed with an exit_timer value which has *not* been updated. Due to this incorrect system time deltas are calculated. If an interrupt occured with an old PSW between .Lsysc_restore/.Lsysc_done or .Lio_restore/.Lio_done update __LC_EXIT_TIMER with the system entry time of the interrupt. Cc: stable@vger.kernel.org # 3.3+ Tested-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-05-02 18:36:00 +07:00
mvc __LC_EXIT_TIMER(8),__LC_MCCK_ENTER_TIMER
0: clg %r9,BASED(.Lcleanup_sysc_restore_insn+8)
je 1f
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lg %r9,24(%r11) # get saved pointer to pt_regs
mvc __LC_RETURN_PSW(16),__PT_PSW(%r9)
mvc 0(64,%r11),__PT_R8(%r9)
lmg %r0,%r7,__PT_R0(%r9)
s390/cputime: fix incorrect system time git commit c5328901aa1db134 "[S390] entry[64].S improvements" removed the update of the exit_timer lowcore field from the critical section cleanup of the .Lsysc_restore/.Lsysc_done and .Lio_restore/.Lio_done blocks. If the PSW is updated by the critical section cleanup to point to user space again, the interrupt entry code will do a vtime calculation after the cleanup completed with an exit_timer value which has *not* been updated. Due to this incorrect system time deltas are calculated. If an interrupt occured with an old PSW between .Lsysc_restore/.Lsysc_done or .Lio_restore/.Lio_done update __LC_EXIT_TIMER with the system entry time of the interrupt. Cc: stable@vger.kernel.org # 3.3+ Tested-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-05-02 18:36:00 +07:00
1: lmg %r8,%r9,__LC_RETURN_PSW
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
BR_R11USE_R14
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lcleanup_sysc_restore_insn:
s390/cputime: fix incorrect system time git commit c5328901aa1db134 "[S390] entry[64].S improvements" removed the update of the exit_timer lowcore field from the critical section cleanup of the .Lsysc_restore/.Lsysc_done and .Lio_restore/.Lio_done blocks. If the PSW is updated by the critical section cleanup to point to user space again, the interrupt entry code will do a vtime calculation after the cleanup completed with an exit_timer value which has *not* been updated. Due to this incorrect system time deltas are calculated. If an interrupt occured with an old PSW between .Lsysc_restore/.Lsysc_done or .Lio_restore/.Lio_done update __LC_EXIT_TIMER with the system entry time of the interrupt. Cc: stable@vger.kernel.org # 3.3+ Tested-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-05-02 18:36:00 +07:00
.quad .Lsysc_exit_timer
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.quad .Lsysc_done - 4
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lcleanup_io_tif:
larl %r9,.Lio_tif
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
BR_R11USE_R14
[S390] fix io_return critical section cleanup If a machine check interrupts the io interrupt handler on one of the instructions between io_return and io_leave the critical section cleanup code will move the return psw to io_work_loop. By doing that the switch from the asynchronous interrupt stack to the process stack is skipped. If e.g. TIF_NEED_RESCHED is set things break because the scheduler is called with the asynchronous interrupts stack. Moving the psw back to io_return instead fixes the problem. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2010-04-09 18:43:00 +07:00
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lcleanup_io_restore:
s390/cputime: fix incorrect system time git commit c5328901aa1db134 "[S390] entry[64].S improvements" removed the update of the exit_timer lowcore field from the critical section cleanup of the .Lsysc_restore/.Lsysc_done and .Lio_restore/.Lio_done blocks. If the PSW is updated by the critical section cleanup to point to user space again, the interrupt entry code will do a vtime calculation after the cleanup completed with an exit_timer value which has *not* been updated. Due to this incorrect system time deltas are calculated. If an interrupt occured with an old PSW between .Lsysc_restore/.Lsysc_done or .Lio_restore/.Lio_done update __LC_EXIT_TIMER with the system entry time of the interrupt. Cc: stable@vger.kernel.org # 3.3+ Tested-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-05-02 18:36:00 +07:00
# check if stpt has been executed
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
clg %r9,BASED(.Lcleanup_io_restore_insn)
s390/cputime: fix incorrect system time git commit c5328901aa1db134 "[S390] entry[64].S improvements" removed the update of the exit_timer lowcore field from the critical section cleanup of the .Lsysc_restore/.Lsysc_done and .Lio_restore/.Lio_done blocks. If the PSW is updated by the critical section cleanup to point to user space again, the interrupt entry code will do a vtime calculation after the cleanup completed with an exit_timer value which has *not* been updated. Due to this incorrect system time deltas are calculated. If an interrupt occured with an old PSW between .Lsysc_restore/.Lsysc_done or .Lio_restore/.Lio_done update __LC_EXIT_TIMER with the system entry time of the interrupt. Cc: stable@vger.kernel.org # 3.3+ Tested-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-05-02 18:36:00 +07:00
jh 0f
mvc __LC_EXIT_TIMER(8),__LC_MCCK_ENTER_TIMER
0: clg %r9,BASED(.Lcleanup_io_restore_insn+8)
je 1f
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
lg %r9,24(%r11) # get saved r11 pointer to pt_regs
mvc __LC_RETURN_PSW(16),__PT_PSW(%r9)
mvc 0(64,%r11),__PT_R8(%r9)
lmg %r0,%r7,__PT_R0(%r9)
s390/cputime: fix incorrect system time git commit c5328901aa1db134 "[S390] entry[64].S improvements" removed the update of the exit_timer lowcore field from the critical section cleanup of the .Lsysc_restore/.Lsysc_done and .Lio_restore/.Lio_done blocks. If the PSW is updated by the critical section cleanup to point to user space again, the interrupt entry code will do a vtime calculation after the cleanup completed with an exit_timer value which has *not* been updated. Due to this incorrect system time deltas are calculated. If an interrupt occured with an old PSW between .Lsysc_restore/.Lsysc_done or .Lio_restore/.Lio_done update __LC_EXIT_TIMER with the system entry time of the interrupt. Cc: stable@vger.kernel.org # 3.3+ Tested-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-05-02 18:36:00 +07:00
1: lmg %r8,%r9,__LC_RETURN_PSW
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
BR_R11USE_R14
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lcleanup_io_restore_insn:
s390/cputime: fix incorrect system time git commit c5328901aa1db134 "[S390] entry[64].S improvements" removed the update of the exit_timer lowcore field from the critical section cleanup of the .Lsysc_restore/.Lsysc_done and .Lio_restore/.Lio_done blocks. If the PSW is updated by the critical section cleanup to point to user space again, the interrupt entry code will do a vtime calculation after the cleanup completed with an exit_timer value which has *not* been updated. Due to this incorrect system time deltas are calculated. If an interrupt occured with an old PSW between .Lsysc_restore/.Lsysc_done or .Lio_restore/.Lio_done update __LC_EXIT_TIMER with the system entry time of the interrupt. Cc: stable@vger.kernel.org # 3.3+ Tested-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-05-02 18:36:00 +07:00
.quad .Lio_exit_timer
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.quad .Lio_done - 4
[PATCH] s390: machine check handler bugs The new machine check handler still has a few bugs. 1) The system entry time has to be stored in the machine check handler, 2) the machine check return psw may not be stored at the usual place because it might overwrite the return psw of the interrupted context, 3) the return address for the call to s390_handle_mcck in the i/o interrupt handler is not correct, 4) the system call cleanup has to take the different save area of the machine check handler into account, 5) the machine check handler may not call UPDATE_VTIME before CREATE_STACK_FRAME, and 6) the io leave path needs a critical section cleanup to make sure that the TIF_MCCK_PENDING bit is really checked before switching back to user space. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-04 05:57:56 +07:00
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lcleanup_idle:
s390/spinlock: do not yield to a CPU in udelay/mdelay It does not make sense to try to relinquish the time slice with diag 0x9c to a CPU in a state that does not allow to schedule the CPU. The scenario where this can happen is a CPU waiting in udelay/mdelay while holding a spin-lock. Add a CIF bit to tag a CPU in enabled wait and use it to detect that the yield of a CPU will not be successful and skip the diagnose call. Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-11-19 17:09:45 +07:00
ni __LC_CPU_FLAGS+7,255-_CIF_ENABLED_WAIT
[S390] rework idle code Whenever the cpu loads an enabled wait PSW it will appear as idle to the underlying host system. The code in default_idle calls vtime_stop_cpu which does the necessary voodoo to get the cpu time accounting right. The udelay code just loads an enabled wait PSW. To correct this rework the vtime_stop_cpu/vtime_start_cpu logic and move the difficult parts to entry[64].S, vtime_stop_cpu can now be called from anywhere and vtime_start_cpu is gone. The correction of the cpu time during wakeup from an enabled wait PSW is done with a critical section in entry[64].S. As vtime_start_cpu is gone, s390_idle_check can be removed as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:27 +07:00
# copy interrupt clock & cpu timer
s390/vtimer: rework virtual timer interface The current virtual timer interface is inherently per-cpu and hard to use. The sole user of the interface is appldata which uses it to execute a function after a specific amount of cputime has been used over all cpus. Rework the virtual timer interface to hook into the cputime accounting. This makes the interface independent from the CPU timer interrupts, and makes the virtual timers global as opposed to per-cpu. Overall the code is greatly simplified. The downside is that the accuracy is not as good as the original implementation, but it is still good enough for appldata. Reviewed-by: Jan Glauber <jang@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-07-20 16:15:08 +07:00
mvc __CLOCK_IDLE_EXIT(8,%r2),__LC_INT_CLOCK
mvc __TIMER_IDLE_EXIT(8,%r2),__LC_ASYNC_ENTER_TIMER
[S390] rework idle code Whenever the cpu loads an enabled wait PSW it will appear as idle to the underlying host system. The code in default_idle calls vtime_stop_cpu which does the necessary voodoo to get the cpu time accounting right. The udelay code just loads an enabled wait PSW. To correct this rework the vtime_stop_cpu/vtime_start_cpu logic and move the difficult parts to entry[64].S, vtime_stop_cpu can now be called from anywhere and vtime_start_cpu is gone. The correction of the cpu time during wakeup from an enabled wait PSW is done with a critical section in entry[64].S. As vtime_start_cpu is gone, s390_idle_check can be removed as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:27 +07:00
cghi %r11,__LC_SAVE_AREA_ASYNC
je 0f
s390/vtimer: rework virtual timer interface The current virtual timer interface is inherently per-cpu and hard to use. The sole user of the interface is appldata which uses it to execute a function after a specific amount of cputime has been used over all cpus. Rework the virtual timer interface to hook into the cputime accounting. This makes the interface independent from the CPU timer interrupts, and makes the virtual timers global as opposed to per-cpu. Overall the code is greatly simplified. The downside is that the accuracy is not as good as the original implementation, but it is still good enough for appldata. Reviewed-by: Jan Glauber <jang@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-07-20 16:15:08 +07:00
mvc __CLOCK_IDLE_EXIT(8,%r2),__LC_MCCK_CLOCK
mvc __TIMER_IDLE_EXIT(8,%r2),__LC_MCCK_ENTER_TIMER
[S390] rework idle code Whenever the cpu loads an enabled wait PSW it will appear as idle to the underlying host system. The code in default_idle calls vtime_stop_cpu which does the necessary voodoo to get the cpu time accounting right. The udelay code just loads an enabled wait PSW. To correct this rework the vtime_stop_cpu/vtime_start_cpu logic and move the difficult parts to entry[64].S, vtime_stop_cpu can now be called from anywhere and vtime_start_cpu is gone. The correction of the cpu time during wakeup from an enabled wait PSW is done with a critical section in entry[64].S. As vtime_start_cpu is gone, s390_idle_check can be removed as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:27 +07:00
0: # check if stck & stpt have been executed
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
clg %r9,BASED(.Lcleanup_idle_insn)
[S390] rework idle code Whenever the cpu loads an enabled wait PSW it will appear as idle to the underlying host system. The code in default_idle calls vtime_stop_cpu which does the necessary voodoo to get the cpu time accounting right. The udelay code just loads an enabled wait PSW. To correct this rework the vtime_stop_cpu/vtime_start_cpu logic and move the difficult parts to entry[64].S, vtime_stop_cpu can now be called from anywhere and vtime_start_cpu is gone. The correction of the cpu time during wakeup from an enabled wait PSW is done with a critical section in entry[64].S. As vtime_start_cpu is gone, s390_idle_check can be removed as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:27 +07:00
jhe 1f
s390/vtimer: rework virtual timer interface The current virtual timer interface is inherently per-cpu and hard to use. The sole user of the interface is appldata which uses it to execute a function after a specific amount of cputime has been used over all cpus. Rework the virtual timer interface to hook into the cputime accounting. This makes the interface independent from the CPU timer interrupts, and makes the virtual timers global as opposed to per-cpu. Overall the code is greatly simplified. The downside is that the accuracy is not as good as the original implementation, but it is still good enough for appldata. Reviewed-by: Jan Glauber <jang@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-07-20 16:15:08 +07:00
mvc __CLOCK_IDLE_ENTER(8,%r2),__CLOCK_IDLE_EXIT(%r2)
mvc __TIMER_IDLE_ENTER(8,%r2),__TIMER_IDLE_EXIT(%r2)
s390/vtime: correct scaled cputime of partially idle CPUs The calculation for the SMT scaling factor for a hardware thread which has been partially idle needs to disregard the cycles spent by the other threads of the core while the thread is idle. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-09-18 21:41:36 +07:00
1: # calculate idle cycles
#ifdef CONFIG_SMP
clg %r9,BASED(.Lcleanup_idle_insn)
jl 3f
larl %r1,smp_cpu_mtid
llgf %r1,0(%r1)
ltgr %r1,%r1
jz 3f
.insn rsy,0xeb0000000017,%r1,5,__SF_EMPTY+80(%r15)
larl %r3,mt_cycles
ag %r3,__LC_PERCPU_OFFSET
la %r4,__SF_EMPTY+16(%r15)
2: lg %r0,0(%r3)
slg %r0,0(%r4)
alg %r0,64(%r4)
stg %r0,0(%r3)
la %r3,8(%r3)
la %r4,8(%r4)
brct %r1,2b
#endif
3: # account system time going idle
[S390] rework idle code Whenever the cpu loads an enabled wait PSW it will appear as idle to the underlying host system. The code in default_idle calls vtime_stop_cpu which does the necessary voodoo to get the cpu time accounting right. The udelay code just loads an enabled wait PSW. To correct this rework the vtime_stop_cpu/vtime_start_cpu logic and move the difficult parts to entry[64].S, vtime_stop_cpu can now be called from anywhere and vtime_start_cpu is gone. The correction of the cpu time during wakeup from an enabled wait PSW is done with a critical section in entry[64].S. As vtime_start_cpu is gone, s390_idle_check can be removed as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:27 +07:00
lg %r9,__LC_STEAL_TIMER
s390/vtimer: rework virtual timer interface The current virtual timer interface is inherently per-cpu and hard to use. The sole user of the interface is appldata which uses it to execute a function after a specific amount of cputime has been used over all cpus. Rework the virtual timer interface to hook into the cputime accounting. This makes the interface independent from the CPU timer interrupts, and makes the virtual timers global as opposed to per-cpu. Overall the code is greatly simplified. The downside is that the accuracy is not as good as the original implementation, but it is still good enough for appldata. Reviewed-by: Jan Glauber <jang@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-07-20 16:15:08 +07:00
alg %r9,__CLOCK_IDLE_ENTER(%r2)
[S390] rework idle code Whenever the cpu loads an enabled wait PSW it will appear as idle to the underlying host system. The code in default_idle calls vtime_stop_cpu which does the necessary voodoo to get the cpu time accounting right. The udelay code just loads an enabled wait PSW. To correct this rework the vtime_stop_cpu/vtime_start_cpu logic and move the difficult parts to entry[64].S, vtime_stop_cpu can now be called from anywhere and vtime_start_cpu is gone. The correction of the cpu time during wakeup from an enabled wait PSW is done with a critical section in entry[64].S. As vtime_start_cpu is gone, s390_idle_check can be removed as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:27 +07:00
slg %r9,__LC_LAST_UPDATE_CLOCK
stg %r9,__LC_STEAL_TIMER
s390/vtimer: rework virtual timer interface The current virtual timer interface is inherently per-cpu and hard to use. The sole user of the interface is appldata which uses it to execute a function after a specific amount of cputime has been used over all cpus. Rework the virtual timer interface to hook into the cputime accounting. This makes the interface independent from the CPU timer interrupts, and makes the virtual timers global as opposed to per-cpu. Overall the code is greatly simplified. The downside is that the accuracy is not as good as the original implementation, but it is still good enough for appldata. Reviewed-by: Jan Glauber <jang@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-07-20 16:15:08 +07:00
mvc __LC_LAST_UPDATE_CLOCK(8),__CLOCK_IDLE_EXIT(%r2)
[S390] rework idle code Whenever the cpu loads an enabled wait PSW it will appear as idle to the underlying host system. The code in default_idle calls vtime_stop_cpu which does the necessary voodoo to get the cpu time accounting right. The udelay code just loads an enabled wait PSW. To correct this rework the vtime_stop_cpu/vtime_start_cpu logic and move the difficult parts to entry[64].S, vtime_stop_cpu can now be called from anywhere and vtime_start_cpu is gone. The correction of the cpu time during wakeup from an enabled wait PSW is done with a critical section in entry[64].S. As vtime_start_cpu is gone, s390_idle_check can be removed as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:27 +07:00
lg %r9,__LC_SYSTEM_TIMER
alg %r9,__LC_LAST_UPDATE_TIMER
s390/vtimer: rework virtual timer interface The current virtual timer interface is inherently per-cpu and hard to use. The sole user of the interface is appldata which uses it to execute a function after a specific amount of cputime has been used over all cpus. Rework the virtual timer interface to hook into the cputime accounting. This makes the interface independent from the CPU timer interrupts, and makes the virtual timers global as opposed to per-cpu. Overall the code is greatly simplified. The downside is that the accuracy is not as good as the original implementation, but it is still good enough for appldata. Reviewed-by: Jan Glauber <jang@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-07-20 16:15:08 +07:00
slg %r9,__TIMER_IDLE_ENTER(%r2)
[S390] rework idle code Whenever the cpu loads an enabled wait PSW it will appear as idle to the underlying host system. The code in default_idle calls vtime_stop_cpu which does the necessary voodoo to get the cpu time accounting right. The udelay code just loads an enabled wait PSW. To correct this rework the vtime_stop_cpu/vtime_start_cpu logic and move the difficult parts to entry[64].S, vtime_stop_cpu can now be called from anywhere and vtime_start_cpu is gone. The correction of the cpu time during wakeup from an enabled wait PSW is done with a critical section in entry[64].S. As vtime_start_cpu is gone, s390_idle_check can be removed as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:27 +07:00
stg %r9,__LC_SYSTEM_TIMER
s390/vtimer: rework virtual timer interface The current virtual timer interface is inherently per-cpu and hard to use. The sole user of the interface is appldata which uses it to execute a function after a specific amount of cputime has been used over all cpus. Rework the virtual timer interface to hook into the cputime accounting. This makes the interface independent from the CPU timer interrupts, and makes the virtual timers global as opposed to per-cpu. Overall the code is greatly simplified. The downside is that the accuracy is not as good as the original implementation, but it is still good enough for appldata. Reviewed-by: Jan Glauber <jang@linux.vnet.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-07-20 16:15:08 +07:00
mvc __LC_LAST_UPDATE_TIMER(8),__TIMER_IDLE_EXIT(%r2)
[S390] rework idle code Whenever the cpu loads an enabled wait PSW it will appear as idle to the underlying host system. The code in default_idle calls vtime_stop_cpu which does the necessary voodoo to get the cpu time accounting right. The udelay code just loads an enabled wait PSW. To correct this rework the vtime_stop_cpu/vtime_start_cpu logic and move the difficult parts to entry[64].S, vtime_stop_cpu can now be called from anywhere and vtime_start_cpu is gone. The correction of the cpu time during wakeup from an enabled wait PSW is done with a critical section in entry[64].S. As vtime_start_cpu is gone, s390_idle_check can be removed as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:27 +07:00
# prepare return psw
s390/time: return with irqs disabled from psw_idle Modify the psw_idle waiting logic in entry[64].S to return with interrupts disabled. This avoids potential issues with udelay and interrupt loops as interrupts are not reenabled after clock comparator interrupts. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-08-23 19:45:58 +07:00
nihh %r8,0xfcfd # clear irq & wait state bits
[S390] rework idle code Whenever the cpu loads an enabled wait PSW it will appear as idle to the underlying host system. The code in default_idle calls vtime_stop_cpu which does the necessary voodoo to get the cpu time accounting right. The udelay code just loads an enabled wait PSW. To correct this rework the vtime_stop_cpu/vtime_start_cpu logic and move the difficult parts to entry[64].S, vtime_stop_cpu can now be called from anywhere and vtime_start_cpu is gone. The correction of the cpu time during wakeup from an enabled wait PSW is done with a critical section in entry[64].S. As vtime_start_cpu is gone, s390_idle_check can be removed as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:27 +07:00
lg %r9,48(%r11) # return from psw_idle
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
BR_R11USE_R14
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.Lcleanup_idle_insn:
.quad .Lpsw_idle_lpsw
[S390] rework idle code Whenever the cpu loads an enabled wait PSW it will appear as idle to the underlying host system. The code in default_idle calls vtime_stop_cpu which does the necessary voodoo to get the cpu time accounting right. The udelay code just loads an enabled wait PSW. To correct this rework the vtime_stop_cpu/vtime_start_cpu logic and move the difficult parts to entry[64].S, vtime_stop_cpu can now be called from anywhere and vtime_start_cpu is gone. The correction of the cpu time during wakeup from an enabled wait PSW is done with a critical section in entry[64].S. As vtime_start_cpu is gone, s390_idle_check can be removed as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-03-11 22:59:27 +07:00
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
.Lcleanup_save_fpu_regs:
s390: fix floating pointer register corruption (again) There is a tricky interaction between the machine check handler and the critical sections of load_fpu_regs and save_fpu_regs functions. If the machine check interrupts one of the two functions the critical section cleanup will complete the function before the machine check handler s390_do_machine_check is called. Trouble is that the machine check handler needs to validate the floating point registers *before* and not *after* the completion of load_fpu_regs/save_fpu_regs. The simplest solution is to rewind the PSW to the start of the load_fpu_regs/save_fpu_regs and retry the function after the return from the machine check handler. Tested-by: Christian Borntraeger <borntraeger@de.ibm.com> Cc: <stable@vger.kernel.org> # 4.3+ Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-03-10 15:52:55 +07:00
larl %r9,save_fpu_regs
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
BR_R11USE_R14
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
.Lcleanup_load_fpu_regs:
s390: fix floating pointer register corruption (again) There is a tricky interaction between the machine check handler and the critical sections of load_fpu_regs and save_fpu_regs functions. If the machine check interrupts one of the two functions the critical section cleanup will complete the function before the machine check handler s390_do_machine_check is called. Trouble is that the machine check handler needs to validate the floating point registers *before* and not *after* the completion of load_fpu_regs/save_fpu_regs. The simplest solution is to rewind the PSW to the start of the load_fpu_regs/save_fpu_regs and retry the function after the return from the machine check handler. Tested-by: Christian Borntraeger <borntraeger@de.ibm.com> Cc: <stable@vger.kernel.org> # 4.3+ Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2016-03-10 15:52:55 +07:00
larl %r9,load_fpu_regs
s390: introduce execute-trampolines for branches Add CONFIG_EXPOLINE to enable the use of the new -mindirect-branch= and -mfunction_return= compiler options to create a kernel fortified against the specte v2 attack. With CONFIG_EXPOLINE=y all indirect branches will be issued with an execute type instruction. For z10 or newer the EXRL instruction will be used, for older machines the EX instruction. The typical indirect call basr %r14,%r1 is replaced with a PC relative call to a new thunk brasl %r14,__s390x_indirect_jump_r1 The thunk contains the EXRL/EX instruction to the indirect branch __s390x_indirect_jump_r1: exrl 0,0f j . 0: br %r1 The detour via the execute type instruction has a performance impact. To get rid of the detour the new kernel parameter "nospectre_v2" and "spectre_v2=[on,off,auto]" can be used. If the parameter is specified the kernel and module code will be patched at runtime. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-01-26 18:46:47 +07:00
BR_R11USE_R14
s390/kernel: lazy restore fpu registers Improve the save and restore behavior of FPU register contents to use the vector extension within the kernel. The kernel does not use floating-point or vector registers and, therefore, saving and restoring the FPU register contents are performed for handling signals or switching processes only. To prepare for using vector instructions and vector registers within the kernel, enhance the save behavior and implement a lazy restore at return to user space from a system call or interrupt. To implement the lazy restore, the save_fpu_regs() sets a CPU information flag, CIF_FPU, to indicate that the FPU registers must be restored. Saving and setting CIF_FPU is performed in an atomic fashion to be interrupt-safe. When the kernel wants to use the vector extension or wants to change the FPU register state for a task during signal handling, the save_fpu_regs() must be called first. The CIF_FPU flag is also set at process switch. At return to user space, the FPU state is restored. In particular, the FPU state includes the floating-point or vector register contents, as well as, vector-enablement and floating-point control. The FPU state restore and clearing CIF_FPU is also performed in an atomic fashion. For KVM, the restore of the FPU register state is performed when restoring the general-purpose guest registers before the SIE instructions is started. Because the path towards the SIE instruction is interruptible, the CIF_FPU flag must be checked again right before going into SIE. If set, the guest registers must be reloaded again by re-entering the outer SIE loop. This is the same behavior as if the SIE critical section is interrupted. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-10 17:53:42 +07:00
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
/*
* Integer constants
*/
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
.align 8
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
.Lcritical_start:
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.quad .L__critical_start
[S390] entry[64].S improvements Another round of cleanup for entry[64].S, in particular the program check handler looks more reasonable now. The code size for the 31 bit kernel has been reduced by 616 byte and by 528 byte for the 64 bit version. Even better the code is a bit faster as well. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-12-27 17:27:15 +07:00
.Lcritical_length:
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.quad .L__critical_end - .L__critical_start
s390: use IS_ENABLED to check if a CONFIG is set to y or m This is shorter and should be used instead of the longer form which checks for both possible config options. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-11-01 16:08:20 +07:00
#if IS_ENABLED(CONFIG_KVM)
s390/kvm: integrate HANDLE_SIE_INTERCEPT into cleanup_critical Currently there are two mechanisms to deal with cleanup work due to interrupts. The HANDLE_SIE_INTERCEPT macro is used to undo the changes required to enter SIE in sie64a. If the SIE instruction causes a program check, or an asynchronous interrupt is received the HANDLE_SIE_INTERCEPT code forwards the program execution to sie_exit. All the other critical sections in entry.S are handled by the code in cleanup_critical that is called by the SWITCH_ASYNC macro. Move the sie64a function to the beginning of the critical section and add the code from HANDLE_SIE_INTERCEPT to cleanup_critical. Add a special case for the sie64a cleanup to the program check handler. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-06-22 22:26:40 +07:00
.Lsie_critical_start:
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.quad .Lsie_gmap
s390/kvm: avoid automatic sie reentry Do not automatically restart the sie instruction in entry64.S after an interrupt, return to the caller with a reason code instead. That allows to deal with RCU and other conditions in C code. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Gleb Natapov <gleb@redhat.com>
2013-05-17 19:41:37 +07:00
.Lsie_critical_length:
s390: use local symbol names in entry[64].S To improve the output of the perf tool hide most of the symbols from entry[64].S by using the '.L' prefix. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2014-12-03 23:00:08 +07:00
.quad .Lsie_done - .Lsie_gmap
s390/nmi: s390: New low level handling for machine check happening in guest Add the logic to check if the machine check happens when the guest is running. If yes, set the exit reason -EINTR in the machine check's interrupt handler. Refactor s390_do_machine_check to avoid panicing the host for some kinds of machine checks which happen when guest is running. Reinject the instruction processing damage's machine checks including Delayed Access Exception instead of damaging the host if it happens in the guest because it could be caused by improper update on TLB entry or other software case and impacts the guest only. Signed-off-by: QingFeng Hao <haoqf@linux.vnet.ibm.com> Acked-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com> Acked-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
2017-06-07 16:30:42 +07:00
.Lsie_crit_mcck_start:
.quad .Lsie_entry
.Lsie_crit_mcck_length:
.quad .Lsie_skip - .Lsie_entry
[S390] move sie code to entry.S The entry to / exit from sie has subtle dependencies to the first level interrupt handler. Move the sie assembler code to entry64.S and replace the SIE_HOOK callback with a test and the new _TIF_SIE bit. In addition this patch fixes several problems in regard to the check for the_TIF_EXIT_SIE bits. The old code checked the TIF bits before executing the interrupt handler and it only modified the instruction address if it pointed directly to the sie instruction. In both cases it could miss a TIF bit that normally would cause an exit from the guest and would reenter the guest context. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2011-07-24 15:48:18 +07:00
#endif
s390: remove 31 bit syscalls Remove the 31 bit syscalls from the syscall table. This is a separate patch just in case I screwed something up so it can be easily reverted. However the conversion was done with a script, so everything should be ok. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-02-13 20:44:29 +07:00
.section .rodata, "a"
#define SYSCALL(esame,emu) .long esame
[S390] ftrace: add system call tracer support System call tracer support for s390. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2009-06-12 15:26:47 +07:00
.globl sys_call_table
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
sys_call_table:
s390/syscalls: use generated syscall_table.h and unistd.h header files Update the uapi/asm/unistd.h to include the generated compat and 64-bit version of the unistd.h and, as well as, the unistd_nr.h header file. Also remove the arch/s390/kernel/syscalls.S file and use the generated system call table, syscall_table.h, instead. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-12-11 20:47:27 +07:00
#include "asm/syscall_table.h"
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#undef SYSCALL
[PATCH] s390: cleanup Kconfig Sanitize some s390 Kconfig options. We have ARCH_S390, ARCH_S390X, ARCH_S390_31, 64BIT, S390_SUPPORT and COMPAT. Replace these 6 options by S390, 64BIT and COMPAT. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-06 15:19:28 +07:00
#ifdef CONFIG_COMPAT
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
s390: remove 31 bit syscalls Remove the 31 bit syscalls from the syscall table. This is a separate patch just in case I screwed something up so it can be easily reverted. However the conversion was done with a script, so everything should be ok. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2015-02-13 20:44:29 +07:00
#define SYSCALL(esame,emu) .long emu
s390: system call path micro optimization Add a pointer to the system call table to the thread_info structure. The TIF_31BIT bit is set or cleared by SET_PERSONALITY exactly once for the lifetime of a process. With the pointer to the correct system call table in thread_info the system call code in entry64.S path can drop the check for TIF_31BIT which saves a couple of instructions. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2013-04-24 17:58:39 +07:00
.globl sys_call_table_emu
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
sys_call_table_emu:
s390/syscalls: use generated syscall_table.h and unistd.h header files Update the uapi/asm/unistd.h to include the generated compat and 64-bit version of the unistd.h and, as well as, the unistd_nr.h header file. Also remove the arch/s390/kernel/syscalls.S file and use the generated system call table, syscall_table.h, instead. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2017-12-11 20:47:27 +07:00
#include "asm/syscall_table.h"
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 05:20:36 +07:00
#undef SYSCALL
#endif
Reference in New Issue Copy Permalink