* [PATCH urgent 0/3] __restore_processor_state() fixes
@ 2017-12-14 21:19 Andy Lutomirski
2017-12-14 21:19 ` [PATCH urgent 1/3] x86/power/64: Use struct desc_ptr for the IDT in struct saved_context Andy Lutomirski
` (3 more replies)
0 siblings, 4 replies; 13+ messages in thread
From: Andy Lutomirski @ 2017-12-14 21:19 UTC (permalink / raw)
To: X86 ML
Cc: Borislav Petkov, linux-kernel, Pavel Machek, Rafael J. Wysocki,
Zhang Rui, Jarkko Nikula, Linus Torvalds, Josh Poimboeuf,
Andy Lutomirski
__restore_processor_state() was spaghetti code, made no sense, and
had bugs. And I broke resume on 32-bit systems. This series cleans
it up and fixes it (hopefully!).
Andy Lutomirski (3):
x86/power/64: Use struct desc_ptr for the IDT in struct saved_context
x86/power/32: Move SYSENTER MSR restoration to fix_processor_context()
x86/power: Make restore_processor_context() sane
arch/x86/include/asm/suspend_32.h | 8 +++-
arch/x86/include/asm/suspend_64.h | 19 ++++++--
arch/x86/power/cpu.c | 99 ++++++++++++++++++---------------------
3 files changed, 67 insertions(+), 59 deletions(-)
--
2.13.6
^ permalink raw reply [flat|nested] 13+ messages in thread* [PATCH urgent 1/3] x86/power/64: Use struct desc_ptr for the IDT in struct saved_context 2017-12-14 21:19 [PATCH urgent 0/3] __restore_processor_state() fixes Andy Lutomirski @ 2017-12-14 21:19 ` Andy Lutomirski 2017-12-15 14:37 ` [tip:x86/urgent] " tip-bot for Andy Lutomirski 2017-12-14 21:19 ` [PATCH urgent 2/3] x86/power/32: Move SYSENTER MSR restoration to fix_processor_context() Andy Lutomirski ` (2 subsequent siblings) 3 siblings, 1 reply; 13+ messages in thread From: Andy Lutomirski @ 2017-12-14 21:19 UTC (permalink / raw) To: X86 ML Cc: Borislav Petkov, linux-kernel, Pavel Machek, Rafael J. Wysocki, Zhang Rui, Jarkko Nikula, Linus Torvalds, Josh Poimboeuf, Andy Lutomirski x86_64's saved_context nonsensically used separate idt_limit and idt_base fields and then cast &idt_limit to struct desc_ptr *. This was correct (with -fno-strict-aliasing), but it's confusing, served no purpose, and required ifdeffery. Just use struct desc_ptr directly. Tested-by: Jarkko Nikula <jarkko.nikula@linux.intel.com> Signed-off-by: Andy Lutomirski <luto@kernel.org> --- arch/x86/include/asm/suspend_64.h | 3 +-- arch/x86/power/cpu.c | 11 +---------- 2 files changed, 2 insertions(+), 12 deletions(-) diff --git a/arch/x86/include/asm/suspend_64.h b/arch/x86/include/asm/suspend_64.h index 7306e911faee..600e9e0aea51 100644 --- a/arch/x86/include/asm/suspend_64.h +++ b/arch/x86/include/asm/suspend_64.h @@ -30,8 +30,7 @@ struct saved_context { u16 gdt_pad; /* Unused */ struct desc_ptr gdt_desc; u16 idt_pad; - u16 idt_limit; - unsigned long idt_base; + struct desc_ptr idt; u16 ldt; u16 tss; unsigned long tr; diff --git a/arch/x86/power/cpu.c b/arch/x86/power/cpu.c index 5191de14f4df..472bc8c8212b 100644 --- a/arch/x86/power/cpu.c +++ b/arch/x86/power/cpu.c @@ -82,12 +82,8 @@ static void __save_processor_state(struct saved_context *ctxt) /* * descriptor tables */ -#ifdef CONFIG_X86_32 store_idt(&ctxt->idt); -#else -/* CONFIG_X86_64 */ - store_idt((struct desc_ptr *)&ctxt->idt_limit); -#endif + /* * We save it here, but restore it only in the hibernate case. * For ACPI S3 resume, this is loaded via 'early_gdt_desc' in 64-bit @@ -219,12 +215,7 @@ static void notrace __restore_processor_state(struct saved_context *ctxt) * now restore the descriptor tables to their proper values * ltr is done i fix_processor_context(). */ -#ifdef CONFIG_X86_32 load_idt(&ctxt->idt); -#else -/* CONFIG_X86_64 */ - load_idt((const struct desc_ptr *)&ctxt->idt_limit); -#endif #ifdef CONFIG_X86_64 /* -- 2.13.6 ^ permalink raw reply related [flat|nested] 13+ messages in thread
* [tip:x86/urgent] x86/power/64: Use struct desc_ptr for the IDT in struct saved_context 2017-12-14 21:19 ` [PATCH urgent 1/3] x86/power/64: Use struct desc_ptr for the IDT in struct saved_context Andy Lutomirski @ 2017-12-15 14:37 ` tip-bot for Andy Lutomirski 0 siblings, 0 replies; 13+ messages in thread From: tip-bot for Andy Lutomirski @ 2017-12-15 14:37 UTC (permalink / raw) To: linux-tip-commits Cc: linux-kernel, rui.zhang, luto, peterz, pavel, jarkko.nikula, bpetkov, hpa, jpoimboe, rafael.j.wysocki, rjw, tglx, mingo, torvalds Commit-ID: 090edbe23ff57940fca7f57d9165ce57a826bd7a Gitweb: https://git.kernel.org/tip/090edbe23ff57940fca7f57d9165ce57a826bd7a Author: Andy Lutomirski <luto@kernel.org> AuthorDate: Thu, 14 Dec 2017 13:19:05 -0800 Committer: Ingo Molnar <mingo@kernel.org> CommitDate: Fri, 15 Dec 2017 12:18:29 +0100 x86/power/64: Use struct desc_ptr for the IDT in struct saved_context x86_64's saved_context nonsensically used separate idt_limit and idt_base fields and then cast &idt_limit to struct desc_ptr *. This was correct (with -fno-strict-aliasing), but it's confusing, served no purpose, and required #ifdeffery. Simplify this by using struct desc_ptr directly. No change in functionality. Tested-by: Jarkko Nikula <jarkko.nikula@linux.intel.com> Signed-off-by: Andy Lutomirski <luto@kernel.org> Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Acked-by: Thomas Gleixner <tglx@linutronix.de> Cc: Borislav Petkov <bpetkov@suse.de> Cc: Josh Poimboeuf <jpoimboe@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Pavel Machek <pavel@ucw.cz> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Rafael J. Wysocki <rjw@rjwysocki.net> Cc: Zhang Rui <rui.zhang@intel.com> Link: http://lkml.kernel.org/r/967909ce38d341b01d45eff53e278e2728a3a93a.1513286253.git.luto@kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org> --- arch/x86/include/asm/suspend_64.h | 3 +-- arch/x86/power/cpu.c | 11 +---------- 2 files changed, 2 insertions(+), 12 deletions(-) diff --git a/arch/x86/include/asm/suspend_64.h b/arch/x86/include/asm/suspend_64.h index 7306e91..600e9e0 100644 --- a/arch/x86/include/asm/suspend_64.h +++ b/arch/x86/include/asm/suspend_64.h @@ -30,8 +30,7 @@ struct saved_context { u16 gdt_pad; /* Unused */ struct desc_ptr gdt_desc; u16 idt_pad; - u16 idt_limit; - unsigned long idt_base; + struct desc_ptr idt; u16 ldt; u16 tss; unsigned long tr; diff --git a/arch/x86/power/cpu.c b/arch/x86/power/cpu.c index 5191de1..472bc8c 100644 --- a/arch/x86/power/cpu.c +++ b/arch/x86/power/cpu.c @@ -82,12 +82,8 @@ static void __save_processor_state(struct saved_context *ctxt) /* * descriptor tables */ -#ifdef CONFIG_X86_32 store_idt(&ctxt->idt); -#else -/* CONFIG_X86_64 */ - store_idt((struct desc_ptr *)&ctxt->idt_limit); -#endif + /* * We save it here, but restore it only in the hibernate case. * For ACPI S3 resume, this is loaded via 'early_gdt_desc' in 64-bit @@ -219,12 +215,7 @@ static void notrace __restore_processor_state(struct saved_context *ctxt) * now restore the descriptor tables to their proper values * ltr is done i fix_processor_context(). */ -#ifdef CONFIG_X86_32 load_idt(&ctxt->idt); -#else -/* CONFIG_X86_64 */ - load_idt((const struct desc_ptr *)&ctxt->idt_limit); -#endif #ifdef CONFIG_X86_64 /* ^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH urgent 2/3] x86/power/32: Move SYSENTER MSR restoration to fix_processor_context() 2017-12-14 21:19 [PATCH urgent 0/3] __restore_processor_state() fixes Andy Lutomirski 2017-12-14 21:19 ` [PATCH urgent 1/3] x86/power/64: Use struct desc_ptr for the IDT in struct saved_context Andy Lutomirski @ 2017-12-14 21:19 ` Andy Lutomirski 2017-12-15 14:38 ` [tip:x86/urgent] " tip-bot for Andy Lutomirski 2017-12-14 21:19 ` [PATCH urgent 3/3] x86/power: Make restore_processor_context() sane Andy Lutomirski 2017-12-15 1:28 ` [PATCH urgent 0/3] __restore_processor_state() fixes Rafael J. Wysocki 3 siblings, 1 reply; 13+ messages in thread From: Andy Lutomirski @ 2017-12-14 21:19 UTC (permalink / raw) To: X86 ML Cc: Borislav Petkov, linux-kernel, Pavel Machek, Rafael J. Wysocki, Zhang Rui, Jarkko Nikula, Linus Torvalds, Josh Poimboeuf, Andy Lutomirski x86_64 restores system call MSRs in fix_processor_context(), and x86_32 restored them along with segment registers. The 64-bit variant makes more sense, so move the 32-bit code to match the 64-bit code. Tested-by: Jarkko Nikula <jarkko.nikula@linux.intel.com> Signed-off-by: Andy Lutomirski <luto@kernel.org> --- arch/x86/power/cpu.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) diff --git a/arch/x86/power/cpu.c b/arch/x86/power/cpu.c index 472bc8c8212b..033c61e6891b 100644 --- a/arch/x86/power/cpu.c +++ b/arch/x86/power/cpu.c @@ -174,6 +174,9 @@ static void fix_processor_context(void) write_gdt_entry(desc, GDT_ENTRY_TSS, &tss, DESC_TSS); syscall_init(); /* This sets MSR_*STAR and related */ +#else + if (boot_cpu_has(X86_FEATURE_SEP)) + enable_sep_cpu(); #endif load_TR_desc(); /* This does ltr */ load_mm_ldt(current->active_mm); /* This does lldt */ @@ -237,12 +240,6 @@ static void notrace __restore_processor_state(struct saved_context *ctxt) loadsegment(fs, ctxt->fs); loadsegment(gs, ctxt->gs); loadsegment(ss, ctxt->ss); - - /* - * sysenter MSRs - */ - if (boot_cpu_has(X86_FEATURE_SEP)) - enable_sep_cpu(); #else /* CONFIG_X86_64 */ asm volatile ("movw %0, %%ds" :: "r" (ctxt->ds)); -- 2.13.6 ^ permalink raw reply related [flat|nested] 13+ messages in thread
* [tip:x86/urgent] x86/power/32: Move SYSENTER MSR restoration to fix_processor_context() 2017-12-14 21:19 ` [PATCH urgent 2/3] x86/power/32: Move SYSENTER MSR restoration to fix_processor_context() Andy Lutomirski @ 2017-12-15 14:38 ` tip-bot for Andy Lutomirski 0 siblings, 0 replies; 13+ messages in thread From: tip-bot for Andy Lutomirski @ 2017-12-15 14:38 UTC (permalink / raw) To: linux-tip-commits Cc: jpoimboe, mingo, peterz, luto, pavel, jarkko.nikula, torvalds, tglx, rui.zhang, rjw, rafael.j.wysocki, linux-kernel, bpetkov, hpa Commit-ID: 896c80bef4d3b357814a476663158aaf669d0fb3 Gitweb: https://git.kernel.org/tip/896c80bef4d3b357814a476663158aaf669d0fb3 Author: Andy Lutomirski <luto@kernel.org> AuthorDate: Thu, 14 Dec 2017 13:19:06 -0800 Committer: Ingo Molnar <mingo@kernel.org> CommitDate: Fri, 15 Dec 2017 12:18:29 +0100 x86/power/32: Move SYSENTER MSR restoration to fix_processor_context() x86_64 restores system call MSRs in fix_processor_context(), and x86_32 restored them along with segment registers. The 64-bit variant makes more sense, so move the 32-bit code to match the 64-bit code. No side effects are expected to runtime behavior. Tested-by: Jarkko Nikula <jarkko.nikula@linux.intel.com> Signed-off-by: Andy Lutomirski <luto@kernel.org> Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Acked-by: Thomas Gleixner <tglx@linutronix.de> Cc: Borislav Petkov <bpetkov@suse.de> Cc: Josh Poimboeuf <jpoimboe@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Pavel Machek <pavel@ucw.cz> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Rafael J. Wysocki <rjw@rjwysocki.net> Cc: Zhang Rui <rui.zhang@intel.com> Link: http://lkml.kernel.org/r/65158f8d7ee64dd6bbc6c1c83b3b34aaa854e3ae.1513286253.git.luto@kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org> --- arch/x86/power/cpu.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) diff --git a/arch/x86/power/cpu.c b/arch/x86/power/cpu.c index 472bc8c..033c61e 100644 --- a/arch/x86/power/cpu.c +++ b/arch/x86/power/cpu.c @@ -174,6 +174,9 @@ static void fix_processor_context(void) write_gdt_entry(desc, GDT_ENTRY_TSS, &tss, DESC_TSS); syscall_init(); /* This sets MSR_*STAR and related */ +#else + if (boot_cpu_has(X86_FEATURE_SEP)) + enable_sep_cpu(); #endif load_TR_desc(); /* This does ltr */ load_mm_ldt(current->active_mm); /* This does lldt */ @@ -237,12 +240,6 @@ static void notrace __restore_processor_state(struct saved_context *ctxt) loadsegment(fs, ctxt->fs); loadsegment(gs, ctxt->gs); loadsegment(ss, ctxt->ss); - - /* - * sysenter MSRs - */ - if (boot_cpu_has(X86_FEATURE_SEP)) - enable_sep_cpu(); #else /* CONFIG_X86_64 */ asm volatile ("movw %0, %%ds" :: "r" (ctxt->ds)); ^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH urgent 3/3] x86/power: Make restore_processor_context() sane 2017-12-14 21:19 [PATCH urgent 0/3] __restore_processor_state() fixes Andy Lutomirski 2017-12-14 21:19 ` [PATCH urgent 1/3] x86/power/64: Use struct desc_ptr for the IDT in struct saved_context Andy Lutomirski 2017-12-14 21:19 ` [PATCH urgent 2/3] x86/power/32: Move SYSENTER MSR restoration to fix_processor_context() Andy Lutomirski @ 2017-12-14 21:19 ` Andy Lutomirski 2017-12-14 22:32 ` Pavel Machek 2017-12-15 14:38 ` [tip:x86/urgent] " tip-bot for Andy Lutomirski 2017-12-15 1:28 ` [PATCH urgent 0/3] __restore_processor_state() fixes Rafael J. Wysocki 3 siblings, 2 replies; 13+ messages in thread From: Andy Lutomirski @ 2017-12-14 21:19 UTC (permalink / raw) To: X86 ML Cc: Borislav Petkov, linux-kernel, Pavel Machek, Rafael J. Wysocki, Zhang Rui, Jarkko Nikula, Linus Torvalds, Josh Poimboeuf, Andy Lutomirski My previous attempt to fix a couple of bugs in __restore_processor_context() introduced yet another bug. Rather than trying to come up with a minimal fix, let's try to clean it up for real. This patch fixes quite a few things: - The old code saved a nonsensical subset of segment registers. The only registers that need to be saved are those that contain userspace state or those that can't be trivially restored without percpu access working. (On x86_32, we can restore percpu access by writing __KERNEL_PERCPU to %fs. On x86_64, it's easier to save and restore the kernel's GSBASE.) With this patch, we restore hardcoded values to the kernel state where applicable and explicitly restore the user state after fixing all the descriptor tables. - We used to use an unholy mix of inline asm and C helpers for segment register access. Let's get rid of the inline asm. Tested-by: Jarkko Nikula <jarkko.nikula@linux.intel.com> Fixes: 5b06bbcfc2c6 ("x86/power: Fix some ordering bugs in __restore_processor_context()") Signed-off-by: Andy Lutomirski <luto@kernel.org> --- arch/x86/include/asm/suspend_32.h | 8 +++- arch/x86/include/asm/suspend_64.h | 16 +++++++- arch/x86/power/cpu.c | 79 ++++++++++++++++++++------------------- 3 files changed, 62 insertions(+), 41 deletions(-) diff --git a/arch/x86/include/asm/suspend_32.h b/arch/x86/include/asm/suspend_32.h index 982c325dad33..8be6afb58471 100644 --- a/arch/x86/include/asm/suspend_32.h +++ b/arch/x86/include/asm/suspend_32.h @@ -12,7 +12,13 @@ /* image of the saved processor state */ struct saved_context { - u16 es, fs, gs, ss; + /* + * On x86_32, all segment registers, with the possible exception of + * gs, are saved at kernel entry in pt_regs. + */ +#ifdef CONFIG_X86_32_LAZY_GS + u16 gs; +#endif unsigned long cr0, cr2, cr3, cr4; u64 misc_enable; bool misc_enable_saved; diff --git a/arch/x86/include/asm/suspend_64.h b/arch/x86/include/asm/suspend_64.h index 600e9e0aea51..a7af9f53c0cb 100644 --- a/arch/x86/include/asm/suspend_64.h +++ b/arch/x86/include/asm/suspend_64.h @@ -20,8 +20,20 @@ */ struct saved_context { struct pt_regs regs; - u16 ds, es, fs, gs, ss; - unsigned long gs_base, gs_kernel_base, fs_base; + + /* + * User CS and SS are saved in current_pt_regs(). The rest of the + * segment selectors need to be saved and restored here. + */ + u16 ds, es, fs, gs; + + /* + * Usermode FSBASE and GSBASE may not match the fs and gs selectors, + * so we save them separately. We save the kernelmode GSBASE to + * restore percpu access after resume. + */ + unsigned long kernelmode_gs_base, usermode_gs_base, fs_base; + unsigned long cr0, cr2, cr3, cr4, cr8; u64 misc_enable; bool misc_enable_saved; diff --git a/arch/x86/power/cpu.c b/arch/x86/power/cpu.c index 033c61e6891b..36a28eddb435 100644 --- a/arch/x86/power/cpu.c +++ b/arch/x86/power/cpu.c @@ -99,22 +99,18 @@ static void __save_processor_state(struct saved_context *ctxt) /* * segment registers */ -#ifdef CONFIG_X86_32 - savesegment(es, ctxt->es); - savesegment(fs, ctxt->fs); +#ifdef CONFIG_X86_32_LAZY_GS savesegment(gs, ctxt->gs); - savesegment(ss, ctxt->ss); -#else -/* CONFIG_X86_64 */ - asm volatile ("movw %%ds, %0" : "=m" (ctxt->ds)); - asm volatile ("movw %%es, %0" : "=m" (ctxt->es)); - asm volatile ("movw %%fs, %0" : "=m" (ctxt->fs)); - asm volatile ("movw %%gs, %0" : "=m" (ctxt->gs)); - asm volatile ("movw %%ss, %0" : "=m" (ctxt->ss)); +#endif +#ifdef CONFIG_X86_64 + savesegment(gs, ctxt->gs); + savesegment(fs, ctxt->fs); + savesegment(ds, ctxt->ds); + savesegment(es, ctxt->es); rdmsrl(MSR_FS_BASE, ctxt->fs_base); - rdmsrl(MSR_GS_BASE, ctxt->gs_base); - rdmsrl(MSR_KERNEL_GS_BASE, ctxt->gs_kernel_base); + rdmsrl(MSR_GS_BASE, ctxt->kernelmode_gs_base); + rdmsrl(MSR_KERNEL_GS_BASE, ctxt->usermode_gs_base); mtrr_save_fixed_ranges(NULL); rdmsrl(MSR_EFER, ctxt->efer); @@ -189,9 +185,12 @@ static void fix_processor_context(void) } /** - * __restore_processor_state - restore the contents of CPU registers saved - * by __save_processor_state() - * @ctxt - structure to load the registers contents from + * __restore_processor_state - restore the contents of CPU registers saved + * by __save_processor_state() + * @ctxt - structure to load the registers contents from + * + * The asm code that gets us here will have restored a usable GDT, although + * it will be pointing to the wrong alias. */ static void notrace __restore_processor_state(struct saved_context *ctxt) { @@ -214,46 +213,50 @@ static void notrace __restore_processor_state(struct saved_context *ctxt) write_cr2(ctxt->cr2); write_cr0(ctxt->cr0); + /* Restore the IDT. */ + load_idt(&ctxt->idt); + /* - * now restore the descriptor tables to their proper values - * ltr is done i fix_processor_context(). + * Just in case the asm code got us here with the SS, DS, or ES + * out of sync with the GDT, update them. */ - load_idt(&ctxt->idt); + loadsegment(ss, __KERNEL_DS); + loadsegment(ds, __USER_DS); + loadsegment(es, __USER_DS); -#ifdef CONFIG_X86_64 /* - * We need GSBASE restored before percpu access can work. - * percpu access can happen in exception handlers or in complicated - * helpers like load_gs_index(). + * Restore percpu access. Percpu access can happen in exception + * handlers or in complicated helpers like load_gs_index(). */ - wrmsrl(MSR_GS_BASE, ctxt->gs_base); +#ifdef CONFIG_X86_64 + wrmsrl(MSR_GS_BASE, ctxt->kernelmode_gs_base); +#else + loadsegment(fs, __KERNEL_PERCPU); + loadsegment(gs, __KERNEL_STACK_CANARY); #endif + /* Restore the TSS, RO GDT, LDT, and usermode-relevant MSRs. */ fix_processor_context(); /* - * Restore segment registers. This happens after restoring the GDT - * and LDT, which happen in fix_processor_context(). + * Now that we have descriptor tables fully restored and working + * exception handling, restore the usermode segments. */ -#ifdef CONFIG_X86_32 +#ifdef CONFIG_X86_64 + loadsegment(ds, ctxt->es); loadsegment(es, ctxt->es); loadsegment(fs, ctxt->fs); - loadsegment(gs, ctxt->gs); - loadsegment(ss, ctxt->ss); -#else -/* CONFIG_X86_64 */ - asm volatile ("movw %0, %%ds" :: "r" (ctxt->ds)); - asm volatile ("movw %0, %%es" :: "r" (ctxt->es)); - asm volatile ("movw %0, %%fs" :: "r" (ctxt->fs)); load_gs_index(ctxt->gs); - asm volatile ("movw %0, %%ss" :: "r" (ctxt->ss)); /* - * Restore FSBASE and user GSBASE after reloading the respective - * segment selectors. + * Restore FSBASE and GSBASE after restoring the selectors, since + * restoring the selectors clobbers the bases. Keep in mind + * that MSR_KERNEL_GS_BASE is horribly misnamed. */ wrmsrl(MSR_FS_BASE, ctxt->fs_base); - wrmsrl(MSR_KERNEL_GS_BASE, ctxt->gs_kernel_base); + wrmsrl(MSR_KERNEL_GS_BASE, ctxt->usermode_gs_base); +#elif defined(CONFIG_X86_32_LAZY_GS) + loadsegment(gs, ctxt->gs); #endif do_fpu_end(); -- 2.13.6 ^ permalink raw reply related [flat|nested] 13+ messages in thread
* Re: [PATCH urgent 3/3] x86/power: Make restore_processor_context() sane 2017-12-14 21:19 ` [PATCH urgent 3/3] x86/power: Make restore_processor_context() sane Andy Lutomirski @ 2017-12-14 22:32 ` Pavel Machek 2017-12-15 14:38 ` [tip:x86/urgent] " tip-bot for Andy Lutomirski 1 sibling, 0 replies; 13+ messages in thread From: Pavel Machek @ 2017-12-14 22:32 UTC (permalink / raw) To: Andy Lutomirski Cc: X86 ML, Borislav Petkov, linux-kernel, Rafael J. Wysocki, Zhang Rui, Jarkko Nikula, Linus Torvalds, Josh Poimboeuf [-- Attachment #1: Type: text/plain, Size: 1407 bytes --] Hi! > My previous attempt to fix a couple of bugs in > __restore_processor_context() introduced yet another bug. Rather > than trying to come up with a minimal fix, let's try to clean it up > for real. This patch fixes quite a few things: > > - The old code saved a nonsensical subset of segment registers. > The only registers that need to be saved are those that contain > userspace state or those that can't be trivially restored without > percpu access working. (On x86_32, we can restore percpu access > by writing __KERNEL_PERCPU to %fs. On x86_64, it's easier to > save and restore the kernel's GSBASE.) With this patch, we > restore hardcoded values to the kernel state where applicable and > explicitly restore the user state after fixing all the descriptor > tables. > > - We used to use an unholy mix of inline asm and C helpers for > segment register access. Let's get rid of the inline asm. > > Tested-by: Jarkko Nikula <jarkko.nikula@linux.intel.com> > Fixes: 5b06bbcfc2c6 ("x86/power: Fix some ordering bugs in __restore_processor_context()") > Signed-off-by: Andy Lutomirski <luto@kernel.org> Thanks! Tested-by: Pavel Machek <pavel@ucw.cz> Reported-by: Pavel Machek <pavel@ucw.cz> -- (english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html [-- Attachment #2: Digital signature --] [-- Type: application/pgp-signature, Size: 181 bytes --] ^ permalink raw reply [flat|nested] 13+ messages in thread
* [tip:x86/urgent] x86/power: Make restore_processor_context() sane 2017-12-14 21:19 ` [PATCH urgent 3/3] x86/power: Make restore_processor_context() sane Andy Lutomirski 2017-12-14 22:32 ` Pavel Machek @ 2017-12-15 14:38 ` tip-bot for Andy Lutomirski 1 sibling, 0 replies; 13+ messages in thread From: tip-bot for Andy Lutomirski @ 2017-12-15 14:38 UTC (permalink / raw) To: linux-tip-commits Cc: jpoimboe, hpa, bpetkov, torvalds, rafael.j.wysocki, rjw, linux-kernel, pavel, peterz, luto, jarkko.nikula, tglx, mingo, rui.zhang Commit-ID: 7ee18d677989e99635027cee04c878950e0752b9 Gitweb: https://git.kernel.org/tip/7ee18d677989e99635027cee04c878950e0752b9 Author: Andy Lutomirski <luto@kernel.org> AuthorDate: Thu, 14 Dec 2017 13:19:07 -0800 Committer: Ingo Molnar <mingo@kernel.org> CommitDate: Fri, 15 Dec 2017 12:21:38 +0100 x86/power: Make restore_processor_context() sane My previous attempt to fix a couple of bugs in __restore_processor_context(): 5b06bbcfc2c6 ("x86/power: Fix some ordering bugs in __restore_processor_context()") ... introduced yet another bug, breaking suspend-resume. Rather than trying to come up with a minimal fix, let's try to clean it up for real. This patch fixes quite a few things: - The old code saved a nonsensical subset of segment registers. The only registers that need to be saved are those that contain userspace state or those that can't be trivially restored without percpu access working. (On x86_32, we can restore percpu access by writing __KERNEL_PERCPU to %fs. On x86_64, it's easier to save and restore the kernel's GSBASE.) With this patch, we restore hardcoded values to the kernel state where applicable and explicitly restore the user state after fixing all the descriptor tables. - We used to use an unholy mix of inline asm and C helpers for segment register access. Let's get rid of the inline asm. This fixes the reported s2ram hangs and make the code all around more logical. Analyzed-by: Linus Torvalds <torvalds@linux-foundation.org> Reported-by: Jarkko Nikula <jarkko.nikula@linux.intel.com> Reported-by: Pavel Machek <pavel@ucw.cz> Tested-by: Jarkko Nikula <jarkko.nikula@linux.intel.com> Tested-by: Pavel Machek <pavel@ucw.cz> Signed-off-by: Andy Lutomirski <luto@kernel.org> Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Acked-by: Thomas Gleixner <tglx@linutronix.de> Cc: Borislav Petkov <bpetkov@suse.de> Cc: Josh Poimboeuf <jpoimboe@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Rafael J. Wysocki <rjw@rjwysocki.net> Cc: Zhang Rui <rui.zhang@intel.com> Fixes: 5b06bbcfc2c6 ("x86/power: Fix some ordering bugs in __restore_processor_context()") Link: http://lkml.kernel.org/r/398ee68e5c0f766425a7b746becfc810840770ff.1513286253.git.luto@kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org> --- arch/x86/include/asm/suspend_32.h | 8 +++- arch/x86/include/asm/suspend_64.h | 16 +++++++- arch/x86/power/cpu.c | 79 ++++++++++++++++++++------------------- 3 files changed, 62 insertions(+), 41 deletions(-) diff --git a/arch/x86/include/asm/suspend_32.h b/arch/x86/include/asm/suspend_32.h index 982c325..8be6afb 100644 --- a/arch/x86/include/asm/suspend_32.h +++ b/arch/x86/include/asm/suspend_32.h @@ -12,7 +12,13 @@ /* image of the saved processor state */ struct saved_context { - u16 es, fs, gs, ss; + /* + * On x86_32, all segment registers, with the possible exception of + * gs, are saved at kernel entry in pt_regs. + */ +#ifdef CONFIG_X86_32_LAZY_GS + u16 gs; +#endif unsigned long cr0, cr2, cr3, cr4; u64 misc_enable; bool misc_enable_saved; diff --git a/arch/x86/include/asm/suspend_64.h b/arch/x86/include/asm/suspend_64.h index 600e9e0..a7af9f5 100644 --- a/arch/x86/include/asm/suspend_64.h +++ b/arch/x86/include/asm/suspend_64.h @@ -20,8 +20,20 @@ */ struct saved_context { struct pt_regs regs; - u16 ds, es, fs, gs, ss; - unsigned long gs_base, gs_kernel_base, fs_base; + + /* + * User CS and SS are saved in current_pt_regs(). The rest of the + * segment selectors need to be saved and restored here. + */ + u16 ds, es, fs, gs; + + /* + * Usermode FSBASE and GSBASE may not match the fs and gs selectors, + * so we save them separately. We save the kernelmode GSBASE to + * restore percpu access after resume. + */ + unsigned long kernelmode_gs_base, usermode_gs_base, fs_base; + unsigned long cr0, cr2, cr3, cr4, cr8; u64 misc_enable; bool misc_enable_saved; diff --git a/arch/x86/power/cpu.c b/arch/x86/power/cpu.c index 033c61e..36a28ed 100644 --- a/arch/x86/power/cpu.c +++ b/arch/x86/power/cpu.c @@ -99,22 +99,18 @@ static void __save_processor_state(struct saved_context *ctxt) /* * segment registers */ -#ifdef CONFIG_X86_32 - savesegment(es, ctxt->es); - savesegment(fs, ctxt->fs); +#ifdef CONFIG_X86_32_LAZY_GS savesegment(gs, ctxt->gs); - savesegment(ss, ctxt->ss); -#else -/* CONFIG_X86_64 */ - asm volatile ("movw %%ds, %0" : "=m" (ctxt->ds)); - asm volatile ("movw %%es, %0" : "=m" (ctxt->es)); - asm volatile ("movw %%fs, %0" : "=m" (ctxt->fs)); - asm volatile ("movw %%gs, %0" : "=m" (ctxt->gs)); - asm volatile ("movw %%ss, %0" : "=m" (ctxt->ss)); +#endif +#ifdef CONFIG_X86_64 + savesegment(gs, ctxt->gs); + savesegment(fs, ctxt->fs); + savesegment(ds, ctxt->ds); + savesegment(es, ctxt->es); rdmsrl(MSR_FS_BASE, ctxt->fs_base); - rdmsrl(MSR_GS_BASE, ctxt->gs_base); - rdmsrl(MSR_KERNEL_GS_BASE, ctxt->gs_kernel_base); + rdmsrl(MSR_GS_BASE, ctxt->kernelmode_gs_base); + rdmsrl(MSR_KERNEL_GS_BASE, ctxt->usermode_gs_base); mtrr_save_fixed_ranges(NULL); rdmsrl(MSR_EFER, ctxt->efer); @@ -189,9 +185,12 @@ static void fix_processor_context(void) } /** - * __restore_processor_state - restore the contents of CPU registers saved - * by __save_processor_state() - * @ctxt - structure to load the registers contents from + * __restore_processor_state - restore the contents of CPU registers saved + * by __save_processor_state() + * @ctxt - structure to load the registers contents from + * + * The asm code that gets us here will have restored a usable GDT, although + * it will be pointing to the wrong alias. */ static void notrace __restore_processor_state(struct saved_context *ctxt) { @@ -214,46 +213,50 @@ static void notrace __restore_processor_state(struct saved_context *ctxt) write_cr2(ctxt->cr2); write_cr0(ctxt->cr0); + /* Restore the IDT. */ + load_idt(&ctxt->idt); + /* - * now restore the descriptor tables to their proper values - * ltr is done i fix_processor_context(). + * Just in case the asm code got us here with the SS, DS, or ES + * out of sync with the GDT, update them. */ - load_idt(&ctxt->idt); + loadsegment(ss, __KERNEL_DS); + loadsegment(ds, __USER_DS); + loadsegment(es, __USER_DS); -#ifdef CONFIG_X86_64 /* - * We need GSBASE restored before percpu access can work. - * percpu access can happen in exception handlers or in complicated - * helpers like load_gs_index(). + * Restore percpu access. Percpu access can happen in exception + * handlers or in complicated helpers like load_gs_index(). */ - wrmsrl(MSR_GS_BASE, ctxt->gs_base); +#ifdef CONFIG_X86_64 + wrmsrl(MSR_GS_BASE, ctxt->kernelmode_gs_base); +#else + loadsegment(fs, __KERNEL_PERCPU); + loadsegment(gs, __KERNEL_STACK_CANARY); #endif + /* Restore the TSS, RO GDT, LDT, and usermode-relevant MSRs. */ fix_processor_context(); /* - * Restore segment registers. This happens after restoring the GDT - * and LDT, which happen in fix_processor_context(). + * Now that we have descriptor tables fully restored and working + * exception handling, restore the usermode segments. */ -#ifdef CONFIG_X86_32 +#ifdef CONFIG_X86_64 + loadsegment(ds, ctxt->es); loadsegment(es, ctxt->es); loadsegment(fs, ctxt->fs); - loadsegment(gs, ctxt->gs); - loadsegment(ss, ctxt->ss); -#else -/* CONFIG_X86_64 */ - asm volatile ("movw %0, %%ds" :: "r" (ctxt->ds)); - asm volatile ("movw %0, %%es" :: "r" (ctxt->es)); - asm volatile ("movw %0, %%fs" :: "r" (ctxt->fs)); load_gs_index(ctxt->gs); - asm volatile ("movw %0, %%ss" :: "r" (ctxt->ss)); /* - * Restore FSBASE and user GSBASE after reloading the respective - * segment selectors. + * Restore FSBASE and GSBASE after restoring the selectors, since + * restoring the selectors clobbers the bases. Keep in mind + * that MSR_KERNEL_GS_BASE is horribly misnamed. */ wrmsrl(MSR_FS_BASE, ctxt->fs_base); - wrmsrl(MSR_KERNEL_GS_BASE, ctxt->gs_kernel_base); + wrmsrl(MSR_KERNEL_GS_BASE, ctxt->usermode_gs_base); +#elif defined(CONFIG_X86_32_LAZY_GS) + loadsegment(gs, ctxt->gs); #endif do_fpu_end(); ^ permalink raw reply related [flat|nested] 13+ messages in thread
* Re: [PATCH urgent 0/3] __restore_processor_state() fixes 2017-12-14 21:19 [PATCH urgent 0/3] __restore_processor_state() fixes Andy Lutomirski ` (2 preceding siblings ...) 2017-12-14 21:19 ` [PATCH urgent 3/3] x86/power: Make restore_processor_context() sane Andy Lutomirski @ 2017-12-15 1:28 ` Rafael J. Wysocki 2017-12-15 1:35 ` Linus Torvalds 2017-12-15 9:18 ` Jarkko Nikula 3 siblings, 2 replies; 13+ messages in thread From: Rafael J. Wysocki @ 2017-12-15 1:28 UTC (permalink / raw) To: Andy Lutomirski Cc: X86 ML, Borislav Petkov, linux-kernel, Pavel Machek, Zhang Rui, Jarkko Nikula, Linus Torvalds, Josh Poimboeuf On Thursday, December 14, 2017 10:19:04 PM CET Andy Lutomirski wrote: > __restore_processor_state() was spaghetti code, made no sense, and > had bugs. And I broke resume on 32-bit systems. This series cleans > it up and fixes it (hopefully!). > > Andy Lutomirski (3): > x86/power/64: Use struct desc_ptr for the IDT in struct saved_context > x86/power/32: Move SYSENTER MSR restoration to fix_processor_context() > x86/power: Make restore_processor_context() sane For all three: Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Thanks! ^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH urgent 0/3] __restore_processor_state() fixes 2017-12-15 1:28 ` [PATCH urgent 0/3] __restore_processor_state() fixes Rafael J. Wysocki @ 2017-12-15 1:35 ` Linus Torvalds 2017-12-15 11:21 ` Ingo Molnar 2017-12-15 9:18 ` Jarkko Nikula 1 sibling, 1 reply; 13+ messages in thread From: Linus Torvalds @ 2017-12-15 1:35 UTC (permalink / raw) To: Rafael J. Wysocki Cc: Andy Lutomirski, X86 ML, Borislav Petkov, linux-kernel, Pavel Machek, Zhang Rui, Jarkko Nikula, Josh Poimboeuf On Thu, Dec 14, 2017 at 5:28 PM, Rafael J. Wysocki <rjw@rjwysocki.net> wrote: > On Thursday, December 14, 2017 10:19:04 PM CET Andy Lutomirski wrote: >> __restore_processor_state() was spaghetti code, made no sense, and >> had bugs. And I broke resume on 32-bit systems. This series cleans >> it up and fixes it (hopefully!). > > Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> And I'm assuming I'll get these from the x86 -tip tree. If not, somebody please holler, and I'll pick them up directly. Linus ^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH urgent 0/3] __restore_processor_state() fixes 2017-12-15 1:35 ` Linus Torvalds @ 2017-12-15 11:21 ` Ingo Molnar 0 siblings, 0 replies; 13+ messages in thread From: Ingo Molnar @ 2017-12-15 11:21 UTC (permalink / raw) To: Linus Torvalds Cc: Rafael J. Wysocki, Andy Lutomirski, X86 ML, Borislav Petkov, linux-kernel, Pavel Machek, Zhang Rui, Jarkko Nikula, Josh Poimboeuf * Linus Torvalds <torvalds@linux-foundation.org> wrote: > On Thu, Dec 14, 2017 at 5:28 PM, Rafael J. Wysocki <rjw@rjwysocki.net> wrote: > > On Thursday, December 14, 2017 10:19:04 PM CET Andy Lutomirski wrote: > >> __restore_processor_state() was spaghetti code, made no sense, and > >> had bugs. And I broke resume on 32-bit systems. This series cleans > >> it up and fixes it (hopefully!). > > > > Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> > > And I'm assuming I'll get these from the x86 -tip tree. > > If not, somebody please holler, and I'll pick them up directly. I have applied them to tip:x86/urgent with small changes to the changelogs, and will send them to you in a couple of hours. These fixes are really nice, thanks guys for all the effort! Thanks, Ingo ^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH urgent 0/3] __restore_processor_state() fixes 2017-12-15 1:28 ` [PATCH urgent 0/3] __restore_processor_state() fixes Rafael J. Wysocki 2017-12-15 1:35 ` Linus Torvalds @ 2017-12-15 9:18 ` Jarkko Nikula 2017-12-15 11:31 ` Ingo Molnar 1 sibling, 1 reply; 13+ messages in thread From: Jarkko Nikula @ 2017-12-15 9:18 UTC (permalink / raw) To: Rafael J. Wysocki, Andy Lutomirski Cc: X86 ML, Borislav Petkov, linux-kernel, Pavel Machek, Zhang Rui, Linus Torvalds, Josh Poimboeuf On 12/15/2017 03:28 AM, Rafael J. Wysocki wrote: > On Thursday, December 14, 2017 10:19:04 PM CET Andy Lutomirski wrote: >> __restore_processor_state() was spaghetti code, made no sense, and >> had bugs. And I broke resume on 32-bit systems. This series cleans >> it up and fixes it (hopefully!). >> >> Andy Lutomirski (3): >> x86/power/64: Use struct desc_ptr for the IDT in struct saved_context >> x86/power/32: Move SYSENTER MSR restoration to fix_processor_context() >> x86/power: Make restore_processor_context() sane > > For all three: > > Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> > Works still after split up so my tested by remains. Please note Pavel gave his reported/tested-by in 3/3. -- Jarkko ^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH urgent 0/3] __restore_processor_state() fixes 2017-12-15 9:18 ` Jarkko Nikula @ 2017-12-15 11:31 ` Ingo Molnar 0 siblings, 0 replies; 13+ messages in thread From: Ingo Molnar @ 2017-12-15 11:31 UTC (permalink / raw) To: Jarkko Nikula Cc: Rafael J. Wysocki, Andy Lutomirski, X86 ML, Borislav Petkov, linux-kernel, Pavel Machek, Zhang Rui, Linus Torvalds, Josh Poimboeuf * Jarkko Nikula <jarkko.nikula@linux.intel.com> wrote: > On 12/15/2017 03:28 AM, Rafael J. Wysocki wrote: > > On Thursday, December 14, 2017 10:19:04 PM CET Andy Lutomirski wrote: > > > __restore_processor_state() was spaghetti code, made no sense, and > > > had bugs. And I broke resume on 32-bit systems. This series cleans > > > it up and fixes it (hopefully!). > > > > > > Andy Lutomirski (3): > > > x86/power/64: Use struct desc_ptr for the IDT in struct saved_context > > > x86/power/32: Move SYSENTER MSR restoration to fix_processor_context() > > > x86/power: Make restore_processor_context() sane > > > > For all three: > > > > Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> > > > > Works still after split up so my tested by remains. Please note Pavel gave > his reported/tested-by in 3/3. Thanks, I've updated all the changelogs - and I've also added your Reported-by tag to patch #3. Thanks, Ingo ^ permalink raw reply [flat|nested] 13+ messages in thread
end of thread, other threads:[~2017-12-15 14:42 UTC | newest] Thread overview: 13+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2017-12-14 21:19 [PATCH urgent 0/3] __restore_processor_state() fixes Andy Lutomirski 2017-12-14 21:19 ` [PATCH urgent 1/3] x86/power/64: Use struct desc_ptr for the IDT in struct saved_context Andy Lutomirski 2017-12-15 14:37 ` [tip:x86/urgent] " tip-bot for Andy Lutomirski 2017-12-14 21:19 ` [PATCH urgent 2/3] x86/power/32: Move SYSENTER MSR restoration to fix_processor_context() Andy Lutomirski 2017-12-15 14:38 ` [tip:x86/urgent] " tip-bot for Andy Lutomirski 2017-12-14 21:19 ` [PATCH urgent 3/3] x86/power: Make restore_processor_context() sane Andy Lutomirski 2017-12-14 22:32 ` Pavel Machek 2017-12-15 14:38 ` [tip:x86/urgent] " tip-bot for Andy Lutomirski 2017-12-15 1:28 ` [PATCH urgent 0/3] __restore_processor_state() fixes Rafael J. Wysocki 2017-12-15 1:35 ` Linus Torvalds 2017-12-15 11:21 ` Ingo Molnar 2017-12-15 9:18 ` Jarkko Nikula 2017-12-15 11:31 ` Ingo Molnar
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.