linux_dsm_epyc7002/arch/x86/power/hibernate_asm_64.S

148 lines
3.8 KiB
ArmAsm
Raw Normal View History

/*
* Hibernation support for x86-64
*
* Distribute under GPLv2.
*
* Copyright 2007 Rafael J. Wysocki <rjw@sisk.pl>
* Copyright 2005 Andi Kleen <ak@suse.de>
* Copyright 2004 Pavel Machek <pavel@suse.cz>
*
Hibernation: Arbitrary boot kernel support on x86_64 Make it possible to restore a hibernation image on x86_64 with the help of a kernel different from the one in the image. The idea is to split the core restoration code into two separate parts and to place each of them in a different page.  The first part belongs to the boot kernel and is executed as the last step of the image kernel's memory restoration procedure.  Before being executed, it is relocated to a safe page that won't be overwritten while copying the image kernel pages. The final operation performed by it is a jump to the second part of the core restoration code that belongs to the image kernel and has just been restored. This code makes the CPU switch to the image kernel's page tables and restores the state of general purpose registers (including the stack pointer) from before the hibernation. The main issue with this idea is that in order to jump to the second part of the core restoration code the boot kernel needs to know its address.  However, this address may be passed to it in the image header.  Namely, the part of the image header previously used for checking if the version of the image kernel is correct can be replaced with some architecture specific data that will allow the boot kernel to jump to the right address within the image kernel.  These data should also be used for checking if the image kernel is compatible with the boot kernel (as far as the memory restroration procedure is concerned). It can be done, for example, with the help of a "magic" value that has to be equal in both kernels, so that they can be regarded as compatible. Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Pavel Machek <pavel@ucw.cz> Cc: Andi Kleen <ak@suse.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-18 17:04:53 +07:00
* swsusp_arch_resume must not use any stack or any nonlocal variables while
* copying pages:
*
* Its rewriting one kernel image with another. What is stack in "old"
* image could very well be data page in "new" image, and overwriting
* your own stack under you is bad idea.
*/
.text
#include <linux/linkage.h>
#include <asm/segment.h>
#include <asm/page_types.h>
#include <asm/asm-offsets.h>
#include <asm/processor-flags.h>
ENTRY(swsusp_arch_suspend)
movq $saved_context, %rax
movq %rsp, pt_regs_sp(%rax)
movq %rbp, pt_regs_bp(%rax)
movq %rsi, pt_regs_si(%rax)
movq %rdi, pt_regs_di(%rax)
movq %rbx, pt_regs_bx(%rax)
movq %rcx, pt_regs_cx(%rax)
movq %rdx, pt_regs_dx(%rax)
movq %r8, pt_regs_r8(%rax)
movq %r9, pt_regs_r9(%rax)
movq %r10, pt_regs_r10(%rax)
movq %r11, pt_regs_r11(%rax)
movq %r12, pt_regs_r12(%rax)
movq %r13, pt_regs_r13(%rax)
movq %r14, pt_regs_r14(%rax)
movq %r15, pt_regs_r15(%rax)
pushfq
popq pt_regs_flags(%rax)
Hibernation: Arbitrary boot kernel support on x86_64 Make it possible to restore a hibernation image on x86_64 with the help of a kernel different from the one in the image. The idea is to split the core restoration code into two separate parts and to place each of them in a different page.  The first part belongs to the boot kernel and is executed as the last step of the image kernel's memory restoration procedure.  Before being executed, it is relocated to a safe page that won't be overwritten while copying the image kernel pages. The final operation performed by it is a jump to the second part of the core restoration code that belongs to the image kernel and has just been restored. This code makes the CPU switch to the image kernel's page tables and restores the state of general purpose registers (including the stack pointer) from before the hibernation. The main issue with this idea is that in order to jump to the second part of the core restoration code the boot kernel needs to know its address.  However, this address may be passed to it in the image header.  Namely, the part of the image header previously used for checking if the version of the image kernel is correct can be replaced with some architecture specific data that will allow the boot kernel to jump to the right address within the image kernel.  These data should also be used for checking if the image kernel is compatible with the boot kernel (as far as the memory restroration procedure is concerned). It can be done, for example, with the help of a "magic" value that has to be equal in both kernels, so that they can be regarded as compatible. Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Pavel Machek <pavel@ucw.cz> Cc: Andi Kleen <ak@suse.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-18 17:04:53 +07:00
/* save the address of restore_registers */
movq $restore_registers, %rax
movq %rax, restore_jump_address(%rip)
/* save cr3 */
movq %cr3, %rax
movq %rax, restore_cr3(%rip)
Hibernation: Arbitrary boot kernel support on x86_64 Make it possible to restore a hibernation image on x86_64 with the help of a kernel different from the one in the image. The idea is to split the core restoration code into two separate parts and to place each of them in a different page.  The first part belongs to the boot kernel and is executed as the last step of the image kernel's memory restoration procedure.  Before being executed, it is relocated to a safe page that won't be overwritten while copying the image kernel pages. The final operation performed by it is a jump to the second part of the core restoration code that belongs to the image kernel and has just been restored. This code makes the CPU switch to the image kernel's page tables and restores the state of general purpose registers (including the stack pointer) from before the hibernation. The main issue with this idea is that in order to jump to the second part of the core restoration code the boot kernel needs to know its address.  However, this address may be passed to it in the image header.  Namely, the part of the image header previously used for checking if the version of the image kernel is correct can be replaced with some architecture specific data that will allow the boot kernel to jump to the right address within the image kernel.  These data should also be used for checking if the image kernel is compatible with the boot kernel (as far as the memory restroration procedure is concerned). It can be done, for example, with the help of a "magic" value that has to be equal in both kernels, so that they can be regarded as compatible. Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Pavel Machek <pavel@ucw.cz> Cc: Andi Kleen <ak@suse.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-18 17:04:53 +07:00
call swsusp_save
ret
[PATCH] x86_64: Set up safe page tables during resume The following patch makes swsusp avoid the possible temporary corruption of page translation tables during resume on x86-64. This is achieved by creating a copy of the relevant page tables that will not be modified by swsusp and can be safely used by it on resume. The problem is that during resume on x86-64 swsusp may temporarily corrupt the page tables used for the direct mapping of RAM. If that happens, a page fault occurs and cannot be handled properly, which leads to the solid hang of the affected system. This leads to the loss of the system's state from before suspend and may result in the loss of data or the corruption of filesystems, so it is a serious issue. Also, it appears to happen quite often (for me, as often as 50% of the time). The problem is related to the fact that (at least) one of the PMD entries used in the direct memory mapping (starting at PAGE_OFFSET) points to a page table the physical address of which is much greater than the physical address of the PMD entry itself. Moreover, unfortunately, the physical address of the page table before suspend (i.e. the one stored in the suspend image) happens to be different to the physical address of the corresponding page table used during resume (i.e. the one that is valid right before swsusp_arch_resume() in arch/x86_64/kernel/suspend_asm.S is executed). Thus while the image is restored, the "offending" PMD entry gets overwritten, so it does not point to the right physical address any more (i.e. there's no page table at the address pointed to by it, because it points to the address the page table has been at during suspend). Consequently, if the PMD entry is used later on, and it _is_ used in the process of copying the image pages, a page fault occurs, but it cannot be handled in the normal way and the system hangs. In principle we can call create_resume_mapping() from swsusp_arch_resume() (ie. from suspend_asm.S), but then the memory allocations in create_resume_mapping(), resume_pud_mapping(), and resume_pmd_mapping() must be made carefully so that we use _only_ NosaveFree pages in them (the other pages are overwritten by the loop in swsusp_arch_resume()). Additionally, we are in atomic context at that time, so we cannot use GFP_KERNEL. Moreover, if one of the allocations fails, we should free all of the allocated pages, so we need to trace them somehow. All of this is done in the appended patch, except that the functions populating the page tables are located in arch/x86_64/kernel/suspend.c rather than in init.c. It may be done in a more elegan way in the future, with the help of some swsusp patches that are in the works now. [AK: move some externs into headers, renamed a function] Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Signed-off-by: Andi Kleen <ak@suse.de> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-10 02:19:40 +07:00
ENTRY(restore_image)
/* switch to temporary page tables */
movq $__PAGE_OFFSET, %rdx
movq temp_level4_pgt(%rip), %rax
subq %rdx, %rax
movq %rax, %cr3
/* Flush TLB */
movq mmu_cr4_features(%rip), %rax
movq %rax, %rdx
andq $~(X86_CR4_PGE), %rdx
movq %rdx, %cr4; # turn off PGE
movq %cr3, %rcx; # flush TLB
movq %rcx, %cr3;
movq %rax, %cr4; # turn PGE back on
Hibernation: Arbitrary boot kernel support on x86_64 Make it possible to restore a hibernation image on x86_64 with the help of a kernel different from the one in the image. The idea is to split the core restoration code into two separate parts and to place each of them in a different page.  The first part belongs to the boot kernel and is executed as the last step of the image kernel's memory restoration procedure.  Before being executed, it is relocated to a safe page that won't be overwritten while copying the image kernel pages. The final operation performed by it is a jump to the second part of the core restoration code that belongs to the image kernel and has just been restored. This code makes the CPU switch to the image kernel's page tables and restores the state of general purpose registers (including the stack pointer) from before the hibernation. The main issue with this idea is that in order to jump to the second part of the core restoration code the boot kernel needs to know its address.  However, this address may be passed to it in the image header.  Namely, the part of the image header previously used for checking if the version of the image kernel is correct can be replaced with some architecture specific data that will allow the boot kernel to jump to the right address within the image kernel.  These data should also be used for checking if the image kernel is compatible with the boot kernel (as far as the memory restroration procedure is concerned). It can be done, for example, with the help of a "magic" value that has to be equal in both kernels, so that they can be regarded as compatible. Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Pavel Machek <pavel@ucw.cz> Cc: Andi Kleen <ak@suse.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-18 17:04:53 +07:00
/* prepare to jump to the image kernel */
movq restore_jump_address(%rip), %rax
movq restore_cr3(%rip), %rbx
Hibernation: Arbitrary boot kernel support on x86_64 Make it possible to restore a hibernation image on x86_64 with the help of a kernel different from the one in the image. The idea is to split the core restoration code into two separate parts and to place each of them in a different page.  The first part belongs to the boot kernel and is executed as the last step of the image kernel's memory restoration procedure.  Before being executed, it is relocated to a safe page that won't be overwritten while copying the image kernel pages. The final operation performed by it is a jump to the second part of the core restoration code that belongs to the image kernel and has just been restored. This code makes the CPU switch to the image kernel's page tables and restores the state of general purpose registers (including the stack pointer) from before the hibernation. The main issue with this idea is that in order to jump to the second part of the core restoration code the boot kernel needs to know its address.  However, this address may be passed to it in the image header.  Namely, the part of the image header previously used for checking if the version of the image kernel is correct can be replaced with some architecture specific data that will allow the boot kernel to jump to the right address within the image kernel.  These data should also be used for checking if the image kernel is compatible with the boot kernel (as far as the memory restroration procedure is concerned). It can be done, for example, with the help of a "magic" value that has to be equal in both kernels, so that they can be regarded as compatible. Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Pavel Machek <pavel@ucw.cz> Cc: Andi Kleen <ak@suse.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-18 17:04:53 +07:00
/* prepare to copy image data to their original locations */
movq restore_pblist(%rip), %rdx
Hibernation: Arbitrary boot kernel support on x86_64 Make it possible to restore a hibernation image on x86_64 with the help of a kernel different from the one in the image. The idea is to split the core restoration code into two separate parts and to place each of them in a different page.  The first part belongs to the boot kernel and is executed as the last step of the image kernel's memory restoration procedure.  Before being executed, it is relocated to a safe page that won't be overwritten while copying the image kernel pages. The final operation performed by it is a jump to the second part of the core restoration code that belongs to the image kernel and has just been restored. This code makes the CPU switch to the image kernel's page tables and restores the state of general purpose registers (including the stack pointer) from before the hibernation. The main issue with this idea is that in order to jump to the second part of the core restoration code the boot kernel needs to know its address.  However, this address may be passed to it in the image header.  Namely, the part of the image header previously used for checking if the version of the image kernel is correct can be replaced with some architecture specific data that will allow the boot kernel to jump to the right address within the image kernel.  These data should also be used for checking if the image kernel is compatible with the boot kernel (as far as the memory restroration procedure is concerned). It can be done, for example, with the help of a "magic" value that has to be equal in both kernels, so that they can be regarded as compatible. Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Pavel Machek <pavel@ucw.cz> Cc: Andi Kleen <ak@suse.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-18 17:04:53 +07:00
movq relocated_restore_code(%rip), %rcx
jmpq *%rcx
/* code below has been relocated to a safe page */
ENTRY(core_restore_code)
loop:
testq %rdx, %rdx
jz done
/* get addresses from the pbe and copy the page */
movq pbe_address(%rdx), %rsi
movq pbe_orig_address(%rdx), %rdi
Hibernation: Arbitrary boot kernel support on x86_64 Make it possible to restore a hibernation image on x86_64 with the help of a kernel different from the one in the image. The idea is to split the core restoration code into two separate parts and to place each of them in a different page.  The first part belongs to the boot kernel and is executed as the last step of the image kernel's memory restoration procedure.  Before being executed, it is relocated to a safe page that won't be overwritten while copying the image kernel pages. The final operation performed by it is a jump to the second part of the core restoration code that belongs to the image kernel and has just been restored. This code makes the CPU switch to the image kernel's page tables and restores the state of general purpose registers (including the stack pointer) from before the hibernation. The main issue with this idea is that in order to jump to the second part of the core restoration code the boot kernel needs to know its address.  However, this address may be passed to it in the image header.  Namely, the part of the image header previously used for checking if the version of the image kernel is correct can be replaced with some architecture specific data that will allow the boot kernel to jump to the right address within the image kernel.  These data should also be used for checking if the image kernel is compatible with the boot kernel (as far as the memory restroration procedure is concerned). It can be done, for example, with the help of a "magic" value that has to be equal in both kernels, so that they can be regarded as compatible. Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Pavel Machek <pavel@ucw.cz> Cc: Andi Kleen <ak@suse.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-18 17:04:53 +07:00
movq $(PAGE_SIZE >> 3), %rcx
rep
movsq
/* progress to the next pbe */
movq pbe_next(%rdx), %rdx
jmp loop
done:
Hibernation: Arbitrary boot kernel support on x86_64 Make it possible to restore a hibernation image on x86_64 with the help of a kernel different from the one in the image. The idea is to split the core restoration code into two separate parts and to place each of them in a different page.  The first part belongs to the boot kernel and is executed as the last step of the image kernel's memory restoration procedure.  Before being executed, it is relocated to a safe page that won't be overwritten while copying the image kernel pages. The final operation performed by it is a jump to the second part of the core restoration code that belongs to the image kernel and has just been restored. This code makes the CPU switch to the image kernel's page tables and restores the state of general purpose registers (including the stack pointer) from before the hibernation. The main issue with this idea is that in order to jump to the second part of the core restoration code the boot kernel needs to know its address.  However, this address may be passed to it in the image header.  Namely, the part of the image header previously used for checking if the version of the image kernel is correct can be replaced with some architecture specific data that will allow the boot kernel to jump to the right address within the image kernel.  These data should also be used for checking if the image kernel is compatible with the boot kernel (as far as the memory restroration procedure is concerned). It can be done, for example, with the help of a "magic" value that has to be equal in both kernels, so that they can be regarded as compatible. Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Pavel Machek <pavel@ucw.cz> Cc: Andi Kleen <ak@suse.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-18 17:04:53 +07:00
/* jump to the restore_registers address from the image header */
jmpq *%rax
/*
* NOTE: This assumes that the boot kernel's text mapping covers the
* image kernel's page containing restore_registers and the address of
* this page is the same as in the image kernel's text mapping (it
* should always be true, because the text mapping is linear, starting
* from 0, and is supposed to cover the entire kernel text for every
* kernel).
*
* code below belongs to the image kernel
*/
ENTRY(restore_registers)
[PATCH] x86_64: Set up safe page tables during resume The following patch makes swsusp avoid the possible temporary corruption of page translation tables during resume on x86-64. This is achieved by creating a copy of the relevant page tables that will not be modified by swsusp and can be safely used by it on resume. The problem is that during resume on x86-64 swsusp may temporarily corrupt the page tables used for the direct mapping of RAM. If that happens, a page fault occurs and cannot be handled properly, which leads to the solid hang of the affected system. This leads to the loss of the system's state from before suspend and may result in the loss of data or the corruption of filesystems, so it is a serious issue. Also, it appears to happen quite often (for me, as often as 50% of the time). The problem is related to the fact that (at least) one of the PMD entries used in the direct memory mapping (starting at PAGE_OFFSET) points to a page table the physical address of which is much greater than the physical address of the PMD entry itself. Moreover, unfortunately, the physical address of the page table before suspend (i.e. the one stored in the suspend image) happens to be different to the physical address of the corresponding page table used during resume (i.e. the one that is valid right before swsusp_arch_resume() in arch/x86_64/kernel/suspend_asm.S is executed). Thus while the image is restored, the "offending" PMD entry gets overwritten, so it does not point to the right physical address any more (i.e. there's no page table at the address pointed to by it, because it points to the address the page table has been at during suspend). Consequently, if the PMD entry is used later on, and it _is_ used in the process of copying the image pages, a page fault occurs, but it cannot be handled in the normal way and the system hangs. In principle we can call create_resume_mapping() from swsusp_arch_resume() (ie. from suspend_asm.S), but then the memory allocations in create_resume_mapping(), resume_pud_mapping(), and resume_pmd_mapping() must be made carefully so that we use _only_ NosaveFree pages in them (the other pages are overwritten by the loop in swsusp_arch_resume()). Additionally, we are in atomic context at that time, so we cannot use GFP_KERNEL. Moreover, if one of the allocations fails, we should free all of the allocated pages, so we need to trace them somehow. All of this is done in the appended patch, except that the functions populating the page tables are located in arch/x86_64/kernel/suspend.c rather than in init.c. It may be done in a more elegan way in the future, with the help of some swsusp patches that are in the works now. [AK: move some externs into headers, renamed a function] Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Signed-off-by: Andi Kleen <ak@suse.de> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-10 02:19:40 +07:00
/* go back to the original page tables */
movq %rbx, %cr3
/* Flush TLB, including "global" things (vmalloc) */
movq mmu_cr4_features(%rip), %rax
movq %rax, %rdx
andq $~(X86_CR4_PGE), %rdx
movq %rdx, %cr4; # turn off PGE
movq %cr3, %rcx; # flush TLB
movq %rcx, %cr3
movq %rax, %cr4; # turn PGE back on
/* We don't restore %rax, it must be 0 anyway */
movq $saved_context, %rax
movq pt_regs_sp(%rax), %rsp
movq pt_regs_bp(%rax), %rbp
movq pt_regs_si(%rax), %rsi
movq pt_regs_di(%rax), %rdi
movq pt_regs_bx(%rax), %rbx
movq pt_regs_cx(%rax), %rcx
movq pt_regs_dx(%rax), %rdx
movq pt_regs_r8(%rax), %r8
movq pt_regs_r9(%rax), %r9
movq pt_regs_r10(%rax), %r10
movq pt_regs_r11(%rax), %r11
movq pt_regs_r12(%rax), %r12
movq pt_regs_r13(%rax), %r13
movq pt_regs_r14(%rax), %r14
movq pt_regs_r15(%rax), %r15
pushq pt_regs_flags(%rax)
popfq
xorq %rax, %rax
Hibernation: Arbitrary boot kernel support on x86_64 Make it possible to restore a hibernation image on x86_64 with the help of a kernel different from the one in the image. The idea is to split the core restoration code into two separate parts and to place each of them in a different page.  The first part belongs to the boot kernel and is executed as the last step of the image kernel's memory restoration procedure.  Before being executed, it is relocated to a safe page that won't be overwritten while copying the image kernel pages. The final operation performed by it is a jump to the second part of the core restoration code that belongs to the image kernel and has just been restored. This code makes the CPU switch to the image kernel's page tables and restores the state of general purpose registers (including the stack pointer) from before the hibernation. The main issue with this idea is that in order to jump to the second part of the core restoration code the boot kernel needs to know its address.  However, this address may be passed to it in the image header.  Namely, the part of the image header previously used for checking if the version of the image kernel is correct can be replaced with some architecture specific data that will allow the boot kernel to jump to the right address within the image kernel.  These data should also be used for checking if the image kernel is compatible with the boot kernel (as far as the memory restroration procedure is concerned). It can be done, for example, with the help of a "magic" value that has to be equal in both kernels, so that they can be regarded as compatible. Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Pavel Machek <pavel@ucw.cz> Cc: Andi Kleen <ak@suse.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-18 17:04:53 +07:00
/* tell the hibernation core that we've just restored the memory */
movq %rax, in_suspend(%rip)
ret