* x86 fpu cleanups
@ 2019-06-04 7:15 Christoph Hellwig
2019-06-04 7:15 ` [PATCH 1/3] x86/fpu: Simplify kernel_fpu_end Christoph Hellwig
` (3 more replies)
0 siblings, 4 replies; 16+ messages in thread
From: Christoph Hellwig @ 2019-06-04 7:15 UTC (permalink / raw)
To: x86; +Cc: Sebastian Andrzej Siewior, linux-kernel
Hi all,
this series finishes off some cleanups that are possible now that
Sebastians fpu rework is finished.
^ permalink raw reply [flat|nested] 16+ messages in thread* [PATCH 1/3] x86/fpu: Simplify kernel_fpu_end 2019-06-04 7:15 x86 fpu cleanups Christoph Hellwig @ 2019-06-04 7:15 ` Christoph Hellwig 2019-06-17 10:31 ` [tip:x86/fpu] x86/fpu: Simplify kernel_fpu_end() tip-bot for Christoph Hellwig 2019-06-04 7:15 ` [PATCH 2/3] x86/fpu: Simplify kernel_fpu_begin Christoph Hellwig ` (2 subsequent siblings) 3 siblings, 1 reply; 16+ messages in thread From: Christoph Hellwig @ 2019-06-04 7:15 UTC (permalink / raw) To: x86; +Cc: Sebastian Andrzej Siewior, linux-kernel Remove two little helpers and merge them into kernel_fpu_end to streamline the function. Signed-off-by: Christoph Hellwig <hch@lst.de> --- arch/x86/kernel/fpu/core.c | 15 +++------------ 1 file changed, 3 insertions(+), 12 deletions(-) diff --git a/arch/x86/kernel/fpu/core.c b/arch/x86/kernel/fpu/core.c index 466fca686fb9..1d09af1158e1 100644 --- a/arch/x86/kernel/fpu/core.c +++ b/arch/x86/kernel/fpu/core.c @@ -49,12 +49,6 @@ static void kernel_fpu_disable(void) this_cpu_write(in_kernel_fpu, true); } -static void kernel_fpu_enable(void) -{ - WARN_ON_FPU(!this_cpu_read(in_kernel_fpu)); - this_cpu_write(in_kernel_fpu, false); -} - static bool kernel_fpu_disabled(void) { return this_cpu_read(in_kernel_fpu); @@ -115,11 +109,6 @@ static void __kernel_fpu_begin(void) __cpu_invalidate_fpregs_state(); } -static void __kernel_fpu_end(void) -{ - kernel_fpu_enable(); -} - void kernel_fpu_begin(void) { preempt_disable(); @@ -129,7 +118,9 @@ EXPORT_SYMBOL_GPL(kernel_fpu_begin); void kernel_fpu_end(void) { - __kernel_fpu_end(); + WARN_ON_FPU(!this_cpu_read(in_kernel_fpu)); + + this_cpu_write(in_kernel_fpu, false); preempt_enable(); } EXPORT_SYMBOL_GPL(kernel_fpu_end); -- 2.20.1 ^ permalink raw reply related [flat|nested] 16+ messages in thread
* [tip:x86/fpu] x86/fpu: Simplify kernel_fpu_end() 2019-06-04 7:15 ` [PATCH 1/3] x86/fpu: Simplify kernel_fpu_end Christoph Hellwig @ 2019-06-17 10:31 ` tip-bot for Christoph Hellwig 0 siblings, 0 replies; 16+ messages in thread From: tip-bot for Christoph Hellwig @ 2019-06-17 10:31 UTC (permalink / raw) To: linux-tip-commits Cc: tglx, bigeasy, hch, mingo, x86, linux-kernel, mingo, nstange, dave.hansen, bp, hpa, riel Commit-ID: b78ea19ac22fd7b32d7828066cce3d8f2db5226a Gitweb: https://git.kernel.org/tip/b78ea19ac22fd7b32d7828066cce3d8f2db5226a Author: Christoph Hellwig <hch@lst.de> AuthorDate: Tue, 4 Jun 2019 09:15:22 +0200 Committer: Borislav Petkov <bp@suse.de> CommitDate: Mon, 17 Jun 2019 10:43:43 +0200 x86/fpu: Simplify kernel_fpu_end() Remove two little helpers and merge them into kernel_fpu_end() to streamline the function. Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Borislav Petkov <bp@suse.de> Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Cc: Dave Hansen <dave.hansen@intel.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Nicolai Stange <nstange@suse.de> Cc: Rik van Riel <riel@surriel.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: x86-ml <x86@kernel.org> Link: https://lkml.kernel.org/r/20190604071524.12835-2-hch@lst.de --- arch/x86/kernel/fpu/core.c | 15 +++------------ 1 file changed, 3 insertions(+), 12 deletions(-) diff --git a/arch/x86/kernel/fpu/core.c b/arch/x86/kernel/fpu/core.c index 649fbc3fcf9f..8e046068d20f 100644 --- a/arch/x86/kernel/fpu/core.c +++ b/arch/x86/kernel/fpu/core.c @@ -49,12 +49,6 @@ static void kernel_fpu_disable(void) this_cpu_write(in_kernel_fpu, true); } -static void kernel_fpu_enable(void) -{ - WARN_ON_FPU(!this_cpu_read(in_kernel_fpu)); - this_cpu_write(in_kernel_fpu, false); -} - static bool kernel_fpu_disabled(void) { return this_cpu_read(in_kernel_fpu); @@ -115,11 +109,6 @@ static void __kernel_fpu_begin(void) __cpu_invalidate_fpregs_state(); } -static void __kernel_fpu_end(void) -{ - kernel_fpu_enable(); -} - void kernel_fpu_begin(void) { preempt_disable(); @@ -129,7 +118,9 @@ EXPORT_SYMBOL_GPL(kernel_fpu_begin); void kernel_fpu_end(void) { - __kernel_fpu_end(); + WARN_ON_FPU(!this_cpu_read(in_kernel_fpu)); + + this_cpu_write(in_kernel_fpu, false); preempt_enable(); } EXPORT_SYMBOL_GPL(kernel_fpu_end); ^ permalink raw reply related [flat|nested] 16+ messages in thread
* [PATCH 2/3] x86/fpu: Simplify kernel_fpu_begin 2019-06-04 7:15 x86 fpu cleanups Christoph Hellwig 2019-06-04 7:15 ` [PATCH 1/3] x86/fpu: Simplify kernel_fpu_end Christoph Hellwig @ 2019-06-04 7:15 ` Christoph Hellwig 2019-06-04 11:47 ` Peter Zijlstra 2019-06-17 10:31 ` [tip:x86/fpu] x86/fpu: Simplify kernel_fpu_begin() tip-bot for Christoph Hellwig 2019-06-04 7:15 ` [PATCH 3/3] x86/fpu: remove the fpu__save export Christoph Hellwig 2019-06-04 17:54 ` [PATCH 4/3] x86/fpu: don't use current->mm to check for a kthread Christoph Hellwig 3 siblings, 2 replies; 16+ messages in thread From: Christoph Hellwig @ 2019-06-04 7:15 UTC (permalink / raw) To: x86; +Cc: Sebastian Andrzej Siewior, linux-kernel Merge two helpers into the main function, remove a pointless local variable and flatten a conditional. Signed-off-by: Christoph Hellwig <hch@lst.de> --- arch/x86/kernel/fpu/core.c | 35 +++++++++++------------------------ 1 file changed, 11 insertions(+), 24 deletions(-) diff --git a/arch/x86/kernel/fpu/core.c b/arch/x86/kernel/fpu/core.c index 1d09af1158e1..03c2d306e6f2 100644 --- a/arch/x86/kernel/fpu/core.c +++ b/arch/x86/kernel/fpu/core.c @@ -43,12 +43,6 @@ static DEFINE_PER_CPU(bool, in_kernel_fpu); */ DEFINE_PER_CPU(struct fpu *, fpu_fpregs_owner_ctx); -static void kernel_fpu_disable(void) -{ - WARN_ON_FPU(this_cpu_read(in_kernel_fpu)); - this_cpu_write(in_kernel_fpu, true); -} - static bool kernel_fpu_disabled(void) { return this_cpu_read(in_kernel_fpu); @@ -88,32 +82,25 @@ bool irq_fpu_usable(void) } EXPORT_SYMBOL(irq_fpu_usable); -static void __kernel_fpu_begin(void) +void kernel_fpu_begin(void) { - struct fpu *fpu = ¤t->thread.fpu; + preempt_disable(); WARN_ON_FPU(!irq_fpu_usable()); + WARN_ON_FPU(this_cpu_read(in_kernel_fpu)); - kernel_fpu_disable(); + this_cpu_write(in_kernel_fpu, true); - if (current->mm) { - if (!test_thread_flag(TIF_NEED_FPU_LOAD)) { - set_thread_flag(TIF_NEED_FPU_LOAD); - /* - * Ignore return value -- we don't care if reg state - * is clobbered. - */ - copy_fpregs_to_fpstate(fpu); - } + if (current->mm && !test_thread_flag(TIF_NEED_FPU_LOAD)) { + set_thread_flag(TIF_NEED_FPU_LOAD); + /* + * Ignore return value -- we don't care if reg state + * is clobbered. + */ + copy_fpregs_to_fpstate(¤t->thread.fpu); } __cpu_invalidate_fpregs_state(); } - -void kernel_fpu_begin(void) -{ - preempt_disable(); - __kernel_fpu_begin(); -} EXPORT_SYMBOL_GPL(kernel_fpu_begin); void kernel_fpu_end(void) -- 2.20.1 ^ permalink raw reply related [flat|nested] 16+ messages in thread
* Re: [PATCH 2/3] x86/fpu: Simplify kernel_fpu_begin 2019-06-04 7:15 ` [PATCH 2/3] x86/fpu: Simplify kernel_fpu_begin Christoph Hellwig @ 2019-06-04 11:47 ` Peter Zijlstra 2019-06-04 13:11 ` Christoph Hellwig 2019-06-17 10:31 ` [tip:x86/fpu] x86/fpu: Simplify kernel_fpu_begin() tip-bot for Christoph Hellwig 1 sibling, 1 reply; 16+ messages in thread From: Peter Zijlstra @ 2019-06-04 11:47 UTC (permalink / raw) To: Christoph Hellwig; +Cc: x86, Sebastian Andrzej Siewior, linux-kernel On Tue, Jun 04, 2019 at 09:15:23AM +0200, Christoph Hellwig wrote: > +void kernel_fpu_begin(void) > { > + preempt_disable(); > > WARN_ON_FPU(!irq_fpu_usable()); > + WARN_ON_FPU(this_cpu_read(in_kernel_fpu)); > > + this_cpu_write(in_kernel_fpu, true); > > + if (current->mm && !test_thread_flag(TIF_NEED_FPU_LOAD)) { Did that want to be: !(current->flags & PF_KTHREAD), instead? Because I'm thinking that kernel_fpu_begin() on a kthread that has use_mm() employed shouldn't be doing this.. > + set_thread_flag(TIF_NEED_FPU_LOAD); > + /* > + * Ignore return value -- we don't care if reg state > + * is clobbered. > + */ > + copy_fpregs_to_fpstate(¤t->thread.fpu); > } > __cpu_invalidate_fpregs_state(); > } ^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH 2/3] x86/fpu: Simplify kernel_fpu_begin 2019-06-04 11:47 ` Peter Zijlstra @ 2019-06-04 13:11 ` Christoph Hellwig 2019-06-04 13:34 ` Peter Zijlstra 0 siblings, 1 reply; 16+ messages in thread From: Christoph Hellwig @ 2019-06-04 13:11 UTC (permalink / raw) To: Peter Zijlstra Cc: Christoph Hellwig, x86, Sebastian Andrzej Siewior, linux-kernel On Tue, Jun 04, 2019 at 01:47:01PM +0200, Peter Zijlstra wrote: > > + if (current->mm && !test_thread_flag(TIF_NEED_FPU_LOAD)) { > > Did that want to be: !(current->flags & PF_KTHREAD), instead? > > Because I'm thinking that kernel_fpu_begin() on a kthread that has > use_mm() employed shouldn't be doing this.. Well, this is intended to not change semantics. If we should fix this is should be a separate patch before or after this series. ^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH 2/3] x86/fpu: Simplify kernel_fpu_begin 2019-06-04 13:11 ` Christoph Hellwig @ 2019-06-04 13:34 ` Peter Zijlstra 2019-06-04 14:34 ` Christoph Hellwig 0 siblings, 1 reply; 16+ messages in thread From: Peter Zijlstra @ 2019-06-04 13:34 UTC (permalink / raw) To: Christoph Hellwig; +Cc: x86, Sebastian Andrzej Siewior, linux-kernel On Tue, Jun 04, 2019 at 03:11:38PM +0200, Christoph Hellwig wrote: > On Tue, Jun 04, 2019 at 01:47:01PM +0200, Peter Zijlstra wrote: > > > + if (current->mm && !test_thread_flag(TIF_NEED_FPU_LOAD)) { > > > > Did that want to be: !(current->flags & PF_KTHREAD), instead? > > > > Because I'm thinking that kernel_fpu_begin() on a kthread that has > > use_mm() employed shouldn't be doing this.. > > Well, this is intended to not change semantics. If we should fix > this is should be a separate patch before or after this series. Sure; it's just that I just noticed it. We've recently ran into a similar issue elsewhere. ^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH 2/3] x86/fpu: Simplify kernel_fpu_begin 2019-06-04 13:34 ` Peter Zijlstra @ 2019-06-04 14:34 ` Christoph Hellwig 0 siblings, 0 replies; 16+ messages in thread From: Christoph Hellwig @ 2019-06-04 14:34 UTC (permalink / raw) To: Peter Zijlstra Cc: Christoph Hellwig, x86, Sebastian Andrzej Siewior, linux-kernel On Tue, Jun 04, 2019 at 03:34:17PM +0200, Peter Zijlstra wrote: > > Well, this is intended to not change semantics. If we should fix > > this is should be a separate patch before or after this series. > > Sure; it's just that I just noticed it. We've recently ran into a > similar issue elsewhere. Ok, I'll send a follow on. ^ permalink raw reply [flat|nested] 16+ messages in thread
* [tip:x86/fpu] x86/fpu: Simplify kernel_fpu_begin() 2019-06-04 7:15 ` [PATCH 2/3] x86/fpu: Simplify kernel_fpu_begin Christoph Hellwig 2019-06-04 11:47 ` Peter Zijlstra @ 2019-06-17 10:31 ` tip-bot for Christoph Hellwig 1 sibling, 0 replies; 16+ messages in thread From: tip-bot for Christoph Hellwig @ 2019-06-17 10:31 UTC (permalink / raw) To: linux-tip-commits Cc: mingo, hpa, hch, bp, bigeasy, x86, riel, nstange, tglx, mingo, dave.hansen, linux-kernel Commit-ID: 6d79d86f9600510e08ad45c72b9d7e664e439e62 Gitweb: https://git.kernel.org/tip/6d79d86f9600510e08ad45c72b9d7e664e439e62 Author: Christoph Hellwig <hch@lst.de> AuthorDate: Tue, 4 Jun 2019 09:15:23 +0200 Committer: Borislav Petkov <bp@suse.de> CommitDate: Mon, 17 Jun 2019 12:19:49 +0200 x86/fpu: Simplify kernel_fpu_begin() Merge two helpers into the main function, remove a pointless local variable and flatten a conditional. Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Borislav Petkov <bp@suse.de> Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Cc: Dave Hansen <dave.hansen@intel.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Nicolai Stange <nstange@suse.de> Cc: Rik van Riel <riel@surriel.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: x86-ml <x86@kernel.org> Link: https://lkml.kernel.org/r/20190604071524.12835-3-hch@lst.de --- arch/x86/kernel/fpu/core.c | 36 ++++++++++++------------------------ 1 file changed, 12 insertions(+), 24 deletions(-) diff --git a/arch/x86/kernel/fpu/core.c b/arch/x86/kernel/fpu/core.c index 8e046068d20f..3f92cfad88f0 100644 --- a/arch/x86/kernel/fpu/core.c +++ b/arch/x86/kernel/fpu/core.c @@ -43,12 +43,6 @@ static DEFINE_PER_CPU(bool, in_kernel_fpu); */ DEFINE_PER_CPU(struct fpu *, fpu_fpregs_owner_ctx); -static void kernel_fpu_disable(void) -{ - WARN_ON_FPU(this_cpu_read(in_kernel_fpu)); - this_cpu_write(in_kernel_fpu, true); -} - static bool kernel_fpu_disabled(void) { return this_cpu_read(in_kernel_fpu); @@ -88,32 +82,26 @@ bool irq_fpu_usable(void) } EXPORT_SYMBOL(irq_fpu_usable); -static void __kernel_fpu_begin(void) +void kernel_fpu_begin(void) { - struct fpu *fpu = ¤t->thread.fpu; + preempt_disable(); WARN_ON_FPU(!irq_fpu_usable()); + WARN_ON_FPU(this_cpu_read(in_kernel_fpu)); - kernel_fpu_disable(); + this_cpu_write(in_kernel_fpu, true); - if (!(current->flags & PF_KTHREAD)) { - if (!test_thread_flag(TIF_NEED_FPU_LOAD)) { - set_thread_flag(TIF_NEED_FPU_LOAD); - /* - * Ignore return value -- we don't care if reg state - * is clobbered. - */ - copy_fpregs_to_fpstate(fpu); - } + if (!(current->flags & PF_KTHREAD) && + !test_thread_flag(TIF_NEED_FPU_LOAD)) { + set_thread_flag(TIF_NEED_FPU_LOAD); + /* + * Ignore return value -- we don't care if reg state + * is clobbered. + */ + copy_fpregs_to_fpstate(¤t->thread.fpu); } __cpu_invalidate_fpregs_state(); } - -void kernel_fpu_begin(void) -{ - preempt_disable(); - __kernel_fpu_begin(); -} EXPORT_SYMBOL_GPL(kernel_fpu_begin); void kernel_fpu_end(void) ^ permalink raw reply related [flat|nested] 16+ messages in thread
* [PATCH 3/3] x86/fpu: remove the fpu__save export 2019-06-04 7:15 x86 fpu cleanups Christoph Hellwig 2019-06-04 7:15 ` [PATCH 1/3] x86/fpu: Simplify kernel_fpu_end Christoph Hellwig 2019-06-04 7:15 ` [PATCH 2/3] x86/fpu: Simplify kernel_fpu_begin Christoph Hellwig @ 2019-06-04 7:15 ` Christoph Hellwig 2019-06-17 10:32 ` [tip:x86/fpu] x86/fpu: Remove the fpu__save() export tip-bot for Christoph Hellwig 2019-06-04 17:54 ` [PATCH 4/3] x86/fpu: don't use current->mm to check for a kthread Christoph Hellwig 3 siblings, 1 reply; 16+ messages in thread From: Christoph Hellwig @ 2019-06-04 7:15 UTC (permalink / raw) To: x86; +Cc: Sebastian Andrzej Siewior, linux-kernel This function is only use by the core fpu code. Signed-off-by: Christoph Hellwig <hch@lst.de> --- arch/x86/kernel/fpu/core.c | 1 - 1 file changed, 1 deletion(-) diff --git a/arch/x86/kernel/fpu/core.c b/arch/x86/kernel/fpu/core.c index 03c2d306e6f2..5e0240d029fd 100644 --- a/arch/x86/kernel/fpu/core.c +++ b/arch/x86/kernel/fpu/core.c @@ -133,7 +133,6 @@ void fpu__save(struct fpu *fpu) trace_x86_fpu_after_save(fpu); fpregs_unlock(); } -EXPORT_SYMBOL_GPL(fpu__save); /* * Legacy x87 fpstate state init: -- 2.20.1 ^ permalink raw reply related [flat|nested] 16+ messages in thread
* [tip:x86/fpu] x86/fpu: Remove the fpu__save() export 2019-06-04 7:15 ` [PATCH 3/3] x86/fpu: remove the fpu__save export Christoph Hellwig @ 2019-06-17 10:32 ` tip-bot for Christoph Hellwig 0 siblings, 0 replies; 16+ messages in thread From: tip-bot for Christoph Hellwig @ 2019-06-17 10:32 UTC (permalink / raw) To: linux-tip-commits Cc: hpa, dave.hansen, mingo, bp, nstange, x86, linux-kernel, bigeasy, tglx, mingo, hch, riel Commit-ID: 466329bf407cc5143c3211620faa2c132b9d9a06 Gitweb: https://git.kernel.org/tip/466329bf407cc5143c3211620faa2c132b9d9a06 Author: Christoph Hellwig <hch@lst.de> AuthorDate: Tue, 4 Jun 2019 09:15:24 +0200 Committer: Borislav Petkov <bp@suse.de> CommitDate: Mon, 17 Jun 2019 12:21:26 +0200 x86/fpu: Remove the fpu__save() export This function is only use by the core FPU code. Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Borislav Petkov <bp@suse.de> Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Cc: Dave Hansen <dave.hansen@intel.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Nicolai Stange <nstange@suse.de> Cc: Rik van Riel <riel@surriel.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: x86-ml <x86@kernel.org> Link: https://lkml.kernel.org/r/20190604071524.12835-4-hch@lst.de --- arch/x86/kernel/fpu/core.c | 1 - 1 file changed, 1 deletion(-) diff --git a/arch/x86/kernel/fpu/core.c b/arch/x86/kernel/fpu/core.c index 3f92cfad88f0..12c70840980e 100644 --- a/arch/x86/kernel/fpu/core.c +++ b/arch/x86/kernel/fpu/core.c @@ -134,7 +134,6 @@ void fpu__save(struct fpu *fpu) trace_x86_fpu_after_save(fpu); fpregs_unlock(); } -EXPORT_SYMBOL_GPL(fpu__save); /* * Legacy x87 fpstate state init: ^ permalink raw reply related [flat|nested] 16+ messages in thread
* [PATCH 4/3] x86/fpu: don't use current->mm to check for a kthread 2019-06-04 7:15 x86 fpu cleanups Christoph Hellwig ` (2 preceding siblings ...) 2019-06-04 7:15 ` [PATCH 3/3] x86/fpu: remove the fpu__save export Christoph Hellwig @ 2019-06-04 17:54 ` Christoph Hellwig 2019-06-11 7:52 ` Sebastian Andrzej Siewior ` (2 more replies) 3 siblings, 3 replies; 16+ messages in thread From: Christoph Hellwig @ 2019-06-04 17:54 UTC (permalink / raw) To: x86; +Cc: Sebastian Andrzej Siewior, linux-kernel current->mm can be non-NULL if a kthread calls use_mm. Check for PF_KTHREAD instead to decide when to store user mode FP state. Fixes: 2722146eb784 ("x86/fpu: Remove fpu->initialized") Reported-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Christoph Hellwig <hch@lst.de> --- arch/x86/include/asm/fpu/internal.h | 6 +++--- arch/x86/kernel/fpu/core.c | 3 ++- 2 files changed, 5 insertions(+), 4 deletions(-) diff --git a/arch/x86/include/asm/fpu/internal.h b/arch/x86/include/asm/fpu/internal.h index 9e27fa05a7ae..4c95c365058a 100644 --- a/arch/x86/include/asm/fpu/internal.h +++ b/arch/x86/include/asm/fpu/internal.h @@ -536,7 +536,7 @@ static inline void __fpregs_load_activate(void) struct fpu *fpu = ¤t->thread.fpu; int cpu = smp_processor_id(); - if (WARN_ON_ONCE(current->mm == NULL)) + if (WARN_ON_ONCE(current->flags & PF_KTHREAD)) return; if (!fpregs_state_valid(fpu, cpu)) { @@ -567,11 +567,11 @@ static inline void __fpregs_load_activate(void) * otherwise. * * The FPU context is only stored/restored for a user task and - * ->mm is used to distinguish between kernel and user threads. + * PF_KTHREAD is used to distinguish between kernel and user threads. */ static inline void switch_fpu_prepare(struct fpu *old_fpu, int cpu) { - if (static_cpu_has(X86_FEATURE_FPU) && current->mm) { + if (static_cpu_has(X86_FEATURE_FPU) && !(current->flags & PF_KTHREAD)) { if (!copy_fpregs_to_fpstate(old_fpu)) old_fpu->last_cpu = -1; else diff --git a/arch/x86/kernel/fpu/core.c b/arch/x86/kernel/fpu/core.c index 5e0240d029fd..12c70840980e 100644 --- a/arch/x86/kernel/fpu/core.c +++ b/arch/x86/kernel/fpu/core.c @@ -91,7 +91,8 @@ void kernel_fpu_begin(void) this_cpu_write(in_kernel_fpu, true); - if (current->mm && !test_thread_flag(TIF_NEED_FPU_LOAD)) { + if (!(current->flags & PF_KTHREAD) && + !test_thread_flag(TIF_NEED_FPU_LOAD)) { set_thread_flag(TIF_NEED_FPU_LOAD); /* * Ignore return value -- we don't care if reg state -- 2.20.1 ^ permalink raw reply related [flat|nested] 16+ messages in thread
* Re: [PATCH 4/3] x86/fpu: don't use current->mm to check for a kthread 2019-06-04 17:54 ` [PATCH 4/3] x86/fpu: don't use current->mm to check for a kthread Christoph Hellwig @ 2019-06-11 7:52 ` Sebastian Andrzej Siewior 2019-06-13 19:03 ` Borislav Petkov 2019-06-13 19:07 ` [tip:x86/urgent] x86/fpu: Don't " tip-bot for Christoph Hellwig 2 siblings, 0 replies; 16+ messages in thread From: Sebastian Andrzej Siewior @ 2019-06-11 7:52 UTC (permalink / raw) To: Christoph Hellwig; +Cc: x86, linux-kernel On 2019-06-04 19:54:12 [+0200], Christoph Hellwig wrote: > current->mm can be non-NULL if a kthread calls use_mm. Check for > PF_KTHREAD instead to decide when to store user mode FP state. > > Fixes: 2722146eb784 ("x86/fpu: Remove fpu->initialized") > Reported-by: Peter Zijlstra <peterz@infradead.org> > Signed-off-by: Christoph Hellwig <hch@lst.de> Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> For the whole series. Sebastian ^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH 4/3] x86/fpu: don't use current->mm to check for a kthread 2019-06-04 17:54 ` [PATCH 4/3] x86/fpu: don't use current->mm to check for a kthread Christoph Hellwig 2019-06-11 7:52 ` Sebastian Andrzej Siewior @ 2019-06-13 19:03 ` Borislav Petkov 2019-06-13 19:07 ` [tip:x86/urgent] x86/fpu: Don't " tip-bot for Christoph Hellwig 2 siblings, 0 replies; 16+ messages in thread From: Borislav Petkov @ 2019-06-13 19:03 UTC (permalink / raw) To: Christoph Hellwig; +Cc: x86, Sebastian Andrzej Siewior, linux-kernel On Tue, Jun 04, 2019 at 07:54:12PM +0200, Christoph Hellwig wrote: > current->mm can be non-NULL if a kthread calls use_mm. Check for > PF_KTHREAD instead to decide when to store user mode FP state. > > Fixes: 2722146eb784 ("x86/fpu: Remove fpu->initialized") > Reported-by: Peter Zijlstra <peterz@infradead.org> > Signed-off-by: Christoph Hellwig <hch@lst.de> > --- > arch/x86/include/asm/fpu/internal.h | 6 +++--- > arch/x86/kernel/fpu/core.c | 3 ++- > 2 files changed, 5 insertions(+), 4 deletions(-) I had to take this one first because of the Fixes: tag and expedite it through tip:x86/urgent. Check the tip-bot notification mail whether it is still ok. I'll redo the other ones ontop, or you can if you feel bored. :) Thx. -- Regards/Gruss, Boris. Good mailing practices for 400: avoid top-posting and trim the reply. ^ permalink raw reply [flat|nested] 16+ messages in thread
* [tip:x86/urgent] x86/fpu: Don't use current->mm to check for a kthread 2019-06-04 17:54 ` [PATCH 4/3] x86/fpu: don't use current->mm to check for a kthread Christoph Hellwig 2019-06-11 7:52 ` Sebastian Andrzej Siewior 2019-06-13 19:03 ` Borislav Petkov @ 2019-06-13 19:07 ` tip-bot for Christoph Hellwig 2 siblings, 0 replies; 16+ messages in thread From: tip-bot for Christoph Hellwig @ 2019-06-13 19:07 UTC (permalink / raw) To: linux-tip-commits Cc: bigeasy, hch, jannh, bp, mingo, aubrey.li, dave.hansen, linux-kernel, hpa, x86, tglx, riel, mingo, nstange, peterz Commit-ID: 8d3289f2fa1e0c7e2f72c7352f1efb75d2ad7c76 Gitweb: https://git.kernel.org/tip/8d3289f2fa1e0c7e2f72c7352f1efb75d2ad7c76 Author: Christoph Hellwig <hch@lst.de> AuthorDate: Tue, 4 Jun 2019 19:54:12 +0200 Committer: Borislav Petkov <bp@suse.de> CommitDate: Thu, 13 Jun 2019 20:57:49 +0200 x86/fpu: Don't use current->mm to check for a kthread current->mm can be non-NULL if a kthread calls use_mm(). Check for PF_KTHREAD instead to decide when to store user mode FP state. Fixes: 2722146eb784 ("x86/fpu: Remove fpu->initialized") Reported-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Borislav Petkov <bp@suse.de> Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Cc: Aubrey Li <aubrey.li@intel.com> Cc: Dave Hansen <dave.hansen@intel.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Jann Horn <jannh@google.com> Cc: Nicolai Stange <nstange@suse.de> Cc: Rik van Riel <riel@surriel.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: x86-ml <x86@kernel.org> Link: https://lkml.kernel.org/r/20190604175411.GA27477@lst.de --- arch/x86/include/asm/fpu/internal.h | 6 +++--- arch/x86/kernel/fpu/core.c | 2 +- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/arch/x86/include/asm/fpu/internal.h b/arch/x86/include/asm/fpu/internal.h index 9e27fa05a7ae..4c95c365058a 100644 --- a/arch/x86/include/asm/fpu/internal.h +++ b/arch/x86/include/asm/fpu/internal.h @@ -536,7 +536,7 @@ static inline void __fpregs_load_activate(void) struct fpu *fpu = ¤t->thread.fpu; int cpu = smp_processor_id(); - if (WARN_ON_ONCE(current->mm == NULL)) + if (WARN_ON_ONCE(current->flags & PF_KTHREAD)) return; if (!fpregs_state_valid(fpu, cpu)) { @@ -567,11 +567,11 @@ static inline void __fpregs_load_activate(void) * otherwise. * * The FPU context is only stored/restored for a user task and - * ->mm is used to distinguish between kernel and user threads. + * PF_KTHREAD is used to distinguish between kernel and user threads. */ static inline void switch_fpu_prepare(struct fpu *old_fpu, int cpu) { - if (static_cpu_has(X86_FEATURE_FPU) && current->mm) { + if (static_cpu_has(X86_FEATURE_FPU) && !(current->flags & PF_KTHREAD)) { if (!copy_fpregs_to_fpstate(old_fpu)) old_fpu->last_cpu = -1; else diff --git a/arch/x86/kernel/fpu/core.c b/arch/x86/kernel/fpu/core.c index 466fca686fb9..649fbc3fcf9f 100644 --- a/arch/x86/kernel/fpu/core.c +++ b/arch/x86/kernel/fpu/core.c @@ -102,7 +102,7 @@ static void __kernel_fpu_begin(void) kernel_fpu_disable(); - if (current->mm) { + if (!(current->flags & PF_KTHREAD)) { if (!test_thread_flag(TIF_NEED_FPU_LOAD)) { set_thread_flag(TIF_NEED_FPU_LOAD); /* ^ permalink raw reply related [flat|nested] 16+ messages in thread
* x86: FPU cleanups @ 2010-08-28 16:04 Brian Gerst 0 siblings, 0 replies; 16+ messages in thread From: Brian Gerst @ 2010-08-28 16:04 UTC (permalink / raw) To: hpa; +Cc: x86, linux-kernel [PATCH 01/11] x86: Use correct type for %cr4 [PATCH 02/11] x86: Merge fpu_init() [PATCH 03/11] x86: Merge tolerant_fwait() [PATCH 04/11] x86: Merge __save_init_fpu() [PATCH 05/11] x86-64: Disable preemption when using TS_USEDFPU [PATCH 06/11] x86-64: Fix %cs value in convert_from_fxsr() [PATCH 07/11] x86-64: Simplify constraints for fxsave/fxtstor [PATCH 08/11] x86-32: Remove math_emulate stub [PATCH 09/11] x86: Merge fpu_save_init() [PATCH 10/11] x86: Remove unnecessary ifdefs from i387 code. [PATCH 11/11] x86: Remove PSHUFB_XMM5_* macros arch/x86/include/asm/i387.h | 188 ++++++++++---------------------------- arch/x86/include/asm/processor.h | 4 +- arch/x86/kernel/cpu/common.c | 7 -- arch/x86/kernel/i387.c | 49 ++++------ arch/x86/kernel/process_64.c | 2 +- arch/x86/kernel/traps.c | 35 +------ 6 files changed, 78 insertions(+), 207 deletions(-) ^ permalink raw reply [flat|nested] 16+ messages in thread
end of thread, other threads:[~2019-06-17 10:33 UTC | newest] Thread overview: 16+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2019-06-04 7:15 x86 fpu cleanups Christoph Hellwig 2019-06-04 7:15 ` [PATCH 1/3] x86/fpu: Simplify kernel_fpu_end Christoph Hellwig 2019-06-17 10:31 ` [tip:x86/fpu] x86/fpu: Simplify kernel_fpu_end() tip-bot for Christoph Hellwig 2019-06-04 7:15 ` [PATCH 2/3] x86/fpu: Simplify kernel_fpu_begin Christoph Hellwig 2019-06-04 11:47 ` Peter Zijlstra 2019-06-04 13:11 ` Christoph Hellwig 2019-06-04 13:34 ` Peter Zijlstra 2019-06-04 14:34 ` Christoph Hellwig 2019-06-17 10:31 ` [tip:x86/fpu] x86/fpu: Simplify kernel_fpu_begin() tip-bot for Christoph Hellwig 2019-06-04 7:15 ` [PATCH 3/3] x86/fpu: remove the fpu__save export Christoph Hellwig 2019-06-17 10:32 ` [tip:x86/fpu] x86/fpu: Remove the fpu__save() export tip-bot for Christoph Hellwig 2019-06-04 17:54 ` [PATCH 4/3] x86/fpu: don't use current->mm to check for a kthread Christoph Hellwig 2019-06-11 7:52 ` Sebastian Andrzej Siewior 2019-06-13 19:03 ` Borislav Petkov 2019-06-13 19:07 ` [tip:x86/urgent] x86/fpu: Don't " tip-bot for Christoph Hellwig -- strict thread matches above, loose matches on Subject: below -- 2010-08-28 16:04 x86: FPU cleanups Brian Gerst
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).