Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1495644 > unrolled thread

[PATCH 0/9] x86/fpu: remove lazy FPU mode & various FPU cleanups

Back to article view | Back to linux.kernel

Contents

  [PATCH 0/9] x86/fpu: remove lazy FPU mode & various FPU cleanups riel@redhat.com - 2016-10-05 02:40 +0200
    [PATCH 9/9] x86/fpu: split old & new fpu code paths riel@redhat.com - 2016-10-05 02:40 +0200
      [tip:x86/fpu] x86/fpu: Split old & new FPU code paths tip-bot for Rik van Riel <tipbot@zytor.com> - 2016-10-07 12:00 +0200
    [PATCH 1/9] x86/crypto: Remove X86_FEATURE_EAGER_FPU ifdef from the crc32c code riel@redhat.com - 2016-10-05 02:40 +0200
      Re: [PATCH 1/9] x86/crypto: Remove X86_FEATURE_EAGER_FPU ifdef from  the crc32c code Rik van Riel <riel@redhat.com> - 2016-10-05 02:50 +0200
      [tip:x86/fpu] x86/crypto, x86/fpu: Remove X86_FEATURE_EAGER_FPU  #ifdef from the crc32c code tip-bot for Andy Lutomirski <tipbot@zytor.com> - 2016-10-07 11:50 +0200
    [PATCH 7/9] x86/fpu: rename lazy restore functions to "register state valid" riel@redhat.com - 2016-10-05 02:40 +0200
      Re: [PATCH 7/9] x86/fpu: rename lazy restore functions to "register  state valid" Andy Lutomirski <luto@amacapital.net> - 2016-10-05 04:20 +0200
      [tip:x86/fpu] x86/fpu: Rename lazy restore functions to "register  state valid" tip-bot for Rik van Riel <tipbot@zytor.com> - 2016-10-07 11:50 +0200

#1495644 — [PATCH 0/9] x86/fpu: remove lazy FPU mode & various FPU cleanups

This series removes lazy FPU mode, and cleans up various bits
and pieces around the FPU code.

I have run this through a basic floating point test that involves
about 1.5 billion context switches and 45 minutes of swapping at
250MB/second.

This seems to tease out bugs fairly well, though I would not mind
an actual floating point test suite...

[toc] | [next] | [standalone]

#1495645 — [PATCH 9/9] x86/fpu: split old & new fpu code paths

From: Rik van Riel <riel@redhat.com>

Now that CR0.TS is no longer being manipulated, we can simplify
switch_fpu_prepare by no longer nesting the handling of the new
fpu inside the two branches for the old FPU.

Signed-off-by: Rik van Riel <riel@redhat.com>
---
 arch/x86/include/asm/fpu/internal.h | 22 ++++++++--------------
 1 file changed, 8 insertions(+), 14 deletions(-)

diff --git a/arch/x86/include/asm/fpu/internal.h b/arch/x86/include/asm/fpu/internal.h
index aa7a117b43f8..ef52935f8a17 100644
--- a/arch/x86/include/asm/fpu/internal.h
+++ b/arch/x86/include/asm/fpu/internal.h
@@ -581,23 +581,17 @@ switch_fpu_prepare(struct fpu *old_fpu, struct fpu *new_fpu, int cpu)
 		/* But leave fpu_fpregs_owner_ctx! */
 		old_fpu->fpregs_active = 0;
 		trace_x86_fpu_regs_deactivated(old_fpu);
+	} else
+		old_fpu->last_cpu = -1;
 
-		/* Don't change CR0.TS if we just switch! */
-		if (fpu.preload) {
-			fpregs_activate(new_fpu);
-			trace_x86_fpu_regs_activated(new_fpu);
+	if (fpu.preload) {
+		if (fpregs_state_valid(new_fpu, cpu))
+			fpu.preload = 0;
+		else
 			prefetch(&new_fpu->state);
-		}
-	} else {
-		old_fpu->last_cpu = -1;
-		if (fpu.preload) {
-			if (fpregs_state_valid(new_fpu, cpu))
-				fpu.preload = 0;
-			else
-				prefetch(&new_fpu->state);
-			fpregs_activate(new_fpu);
-		}
+		fpregs_activate(new_fpu);
 	}
+
 	return fpu;
 }
 
-- 
2.7.4

[toc] | [prev] | [next] | [standalone]

#1497109 — [tip:x86/fpu] x86/fpu: Split old & new FPU code paths

Commit-ID:  9ad93fe35aff616fca4e2b9581fdeed498605f9e
Gitweb:     http://git.kernel.org/tip/9ad93fe35aff616fca4e2b9581fdeed498605f9e
Author:     Rik van Riel <riel@redhat.com>
AuthorDate: Tue, 4 Oct 2016 20:34:38 -0400
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Fri, 7 Oct 2016 11:14:43 +0200

x86/fpu: Split old & new FPU code paths

Now that CR0.TS is no longer being manipulated, we can simplify
switch_fpu_prepare() by no longer nesting the handling of new_fpu
inside the two branches for the old_fpu.

Signed-off-by: Rik van Riel <riel@redhat.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Quentin Casasnovas <quentin.casasnovas@oracle.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: pbonzini@redhat.com
Link: http://lkml.kernel.org/r/1475627678-20788-10-git-send-email-riel@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/include/asm/fpu/internal.h | 22 ++++++++--------------
 1 file changed, 8 insertions(+), 14 deletions(-)

diff --git a/arch/x86/include/asm/fpu/internal.h b/arch/x86/include/asm/fpu/internal.h
index d0324bc..1dcb29e 100644
--- a/arch/x86/include/asm/fpu/internal.h
+++ b/arch/x86/include/asm/fpu/internal.h
@@ -581,23 +581,17 @@ switch_fpu_prepare(struct fpu *old_fpu, struct fpu *new_fpu, int cpu)
 		/* But leave fpu_fpregs_owner_ctx! */
 		old_fpu->fpregs_active = 0;
 		trace_x86_fpu_regs_deactivated(old_fpu);
+	} else
+		old_fpu->last_cpu = -1;
 
-		/* Don't change CR0.TS if we just switch! */
-		if (fpu.preload) {
-			fpregs_activate(new_fpu);
-			trace_x86_fpu_regs_activated(new_fpu);
+	if (fpu.preload) {
+		if (fpregs_state_valid(new_fpu, cpu))
+			fpu.preload = 0;
+		else
 			prefetch(&new_fpu->state);
-		}
-	} else {
-		old_fpu->last_cpu = -1;
-		if (fpu.preload) {
-			if (fpregs_state_valid(new_fpu, cpu))
-				fpu.preload = 0;
-			else
-				prefetch(&new_fpu->state);
-			fpregs_activate(new_fpu);
-		}
+		fpregs_activate(new_fpu);
 	}
+
 	return fpu;
 }
 

[toc] | [prev] | [next] | [standalone]

#1495646 — [PATCH 1/9] x86/crypto: Remove X86_FEATURE_EAGER_FPU ifdef from the crc32c code

From: Rik van Riel <riel@redhat.com>

From : Andy Lutomirski <luto@kernel.org>

The crypto code was checking both use_eager_fpu() and
defined(X86_FEATURE_EAGER_FPU).  The latter was nonsensical, so
remove it.  This will avoid breakage when we remove
X86_FEATURE_EAGER_FPU.

Signed-off-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Andy Lutomirski <luto@kernel.org>
---
 arch/x86/crypto/crc32c-intel_glue.c | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/arch/x86/crypto/crc32c-intel_glue.c b/arch/x86/crypto/crc32c-intel_glue.c
index 0857b1a1de3b..715399b14ed7 100644
--- a/arch/x86/crypto/crc32c-intel_glue.c
+++ b/arch/x86/crypto/crc32c-intel_glue.c
@@ -58,16 +58,11 @@
 asmlinkage unsigned int crc_pcl(const u8 *buffer, int len,
 				unsigned int crc_init);
 static int crc32c_pcl_breakeven = CRC32C_PCL_BREAKEVEN_EAGERFPU;
-#if defined(X86_FEATURE_EAGER_FPU)
 #define set_pcl_breakeven_point()					\
 do {									\
 	if (!use_eager_fpu())						\
 		crc32c_pcl_breakeven = CRC32C_PCL_BREAKEVEN_NOEAGERFPU;	\
 } while (0)
-#else
-#define set_pcl_breakeven_point()					\
-	(crc32c_pcl_breakeven = CRC32C_PCL_BREAKEVEN_NOEAGERFPU)
-#endif
 #endif /* CONFIG_X86_64 */
 
 static u32 crc32c_intel_le_hw_byte(u32 crc, unsigned char const *data, size_t length)
-- 
2.7.4

[toc] | [prev] | [next] | [standalone]

#1495649 — Re: [PATCH 1/9] x86/crypto: Remove X86_FEATURE_EAGER_FPU ifdef from the crc32c code

[Multipart message — attachments visible in raw view] — view raw

On Tue, 2016-10-04 at 20:34 -0400, riel@redhat.com wrote:
> From: Rik van Riel <riel@redhat.com>
> 
> From : Andy Lutomirski <luto@kernel.org>

OK, I am not sure why "guilt patchbomb" (which invokes
git patchbomb) would do that :(

Obviously these first four patches are Andy's, not my
own.

-- 
All Rights Reversed.

[toc] | [prev] | [next] | [standalone]

#1497107 — [tip:x86/fpu] x86/crypto, x86/fpu: Remove X86_FEATURE_EAGER_FPU #ifdef from the crc32c code

Commit-ID:  02f39b2379fb81557ae864ec8f85421c0250c954
Gitweb:     http://git.kernel.org/tip/02f39b2379fb81557ae864ec8f85421c0250c954
Author:     Andy Lutomirski <luto@kernel.org>
AuthorDate: Tue, 4 Oct 2016 20:34:30 -0400
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Fri, 7 Oct 2016 11:14:06 +0200

x86/crypto, x86/fpu: Remove X86_FEATURE_EAGER_FPU #ifdef from the crc32c code

The crypto code was checking both use_eager_fpu() and
defined(X86_FEATURE_EAGER_FPU).  The latter was nonsensical, so
remove it.  This will avoid breakage when we remove
X86_FEATURE_EAGER_FPU.

Signed-off-by: Andy Lutomirski <luto@kernel.org>
Signed-off-by: Rik van Riel <riel@redhat.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Quentin Casasnovas <quentin.casasnovas@oracle.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: pbonzini@redhat.com
Link: http://lkml.kernel.org/r/1475627678-20788-2-git-send-email-riel@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/crypto/crc32c-intel_glue.c | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/arch/x86/crypto/crc32c-intel_glue.c b/arch/x86/crypto/crc32c-intel_glue.c
index 0857b1a..715399b 100644
--- a/arch/x86/crypto/crc32c-intel_glue.c
+++ b/arch/x86/crypto/crc32c-intel_glue.c
@@ -58,16 +58,11 @@
 asmlinkage unsigned int crc_pcl(const u8 *buffer, int len,
 				unsigned int crc_init);
 static int crc32c_pcl_breakeven = CRC32C_PCL_BREAKEVEN_EAGERFPU;
-#if defined(X86_FEATURE_EAGER_FPU)
 #define set_pcl_breakeven_point()					\
 do {									\
 	if (!use_eager_fpu())						\
 		crc32c_pcl_breakeven = CRC32C_PCL_BREAKEVEN_NOEAGERFPU;	\
 } while (0)
-#else
-#define set_pcl_breakeven_point()					\
-	(crc32c_pcl_breakeven = CRC32C_PCL_BREAKEVEN_NOEAGERFPU)
-#endif
 #endif /* CONFIG_X86_64 */
 
 static u32 crc32c_intel_le_hw_byte(u32 crc, unsigned char const *data, size_t length)

[toc] | [prev] | [next] | [standalone]

#1495648 — [PATCH 7/9] x86/fpu: rename lazy restore functions to "register state valid"

From: Rik van Riel <riel@redhat.com>

Name the functions after the state they track, rather than the function
they currently enable. This should make it more obvious when we use the
fpu_register_state_valid function for something else in the future.

Signed-off-by: Rik van Riel <riel@redhat.com>
---
 arch/x86/include/asm/fpu/internal.h | 26 ++++++++++++++++++++------
 arch/x86/kernel/fpu/core.c          |  4 ++--
 arch/x86/kernel/smpboot.c           |  2 +-
 3 files changed, 23 insertions(+), 9 deletions(-)

diff --git a/arch/x86/include/asm/fpu/internal.h b/arch/x86/include/asm/fpu/internal.h
index 499d6ed0e376..d2cfe16dd9fa 100644
--- a/arch/x86/include/asm/fpu/internal.h
+++ b/arch/x86/include/asm/fpu/internal.h
@@ -479,18 +479,32 @@ extern int copy_fpstate_to_sigframe(void __user *buf, void __user *fp, int size)
 DECLARE_PER_CPU(struct fpu *, fpu_fpregs_owner_ctx);
 
 /*
+ * The in-register FPU state for an FPU context on a CPU is assumed to be
+ * valid if the fpu->last_cpu matches the CPU, and the fpu_fpregs_owner_ctx
+ * matches the FPU.
+ *
+ * If the FPU register state is valid, the kernel can skip restoring the
+ * FPU state from memory.
+ *
+ * Any code that clobbers the FPU registers or updates the in-memory
+ * FPU state for a task MUST let the rest of the kernel know that the
+ * FPU registers are no longer valid for this task. Calling either of
+ * these two invalidate functions is enough, use whichever is convenient.
+ *
  * Must be run with preemption disabled: this clears the fpu_fpregs_owner_ctx,
  * on this CPU.
- *
- * This will disable any lazy FPU state restore of the current FPU state,
- * but if the current thread owns the FPU, it will still be saved by.
  */
-static inline void __cpu_disable_lazy_restore(unsigned int cpu)
+static inline void __cpu_invalidate_fpregs_state(unsigned int cpu)
 {
 	per_cpu(fpu_fpregs_owner_ctx, cpu) = NULL;
 }
 
-static inline int fpu_want_lazy_restore(struct fpu *fpu, unsigned int cpu)
+static inline void __fpu_invalidate_fpregs_state(struct fpu *fpu)
+{
+	fpu->last_cpu = -1;
+}
+
+static inline int fpregs_state_valid(struct fpu *fpu, unsigned int cpu)
 {
 	return fpu == this_cpu_read_stable(fpu_fpregs_owner_ctx) && cpu == fpu->last_cpu;
 }
@@ -588,7 +602,7 @@ switch_fpu_prepare(struct fpu *old_fpu, struct fpu *new_fpu, int cpu)
 	} else {
 		old_fpu->last_cpu = -1;
 		if (fpu.preload) {
-			if (fpu_want_lazy_restore(new_fpu, cpu))
+			if (fpregs_state_valid(new_fpu, cpu))
 				fpu.preload = 0;
 			else
 				prefetch(&new_fpu->state);
diff --git a/arch/x86/kernel/fpu/core.c b/arch/x86/kernel/fpu/core.c
index 6a37d525bdbe..25a45ddfdbcf 100644
--- a/arch/x86/kernel/fpu/core.c
+++ b/arch/x86/kernel/fpu/core.c
@@ -336,7 +336,7 @@ void fpu__activate_fpstate_write(struct fpu *fpu)
 
 	if (fpu->fpstate_active) {
 		/* Invalidate any lazy state: */
-		fpu->last_cpu = -1;
+		__fpu_invalidate_fpregs_state(fpu);
 	} else {
 		fpstate_init(&fpu->state);
 		trace_x86_fpu_init_state(fpu);
@@ -379,7 +379,7 @@ void fpu__current_fpstate_write_begin(void)
 	 * ensures we will not be lazy and skip a XRSTOR in the
 	 * future.
 	 */
-	fpu->last_cpu = -1;
+	__fpu_invalidate_fpregs_state(fpu);
 }
 
 /*
diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
index 42a93621f5b0..ca4c4ca2f6af 100644
--- a/arch/x86/kernel/smpboot.c
+++ b/arch/x86/kernel/smpboot.c
@@ -1111,7 +1111,7 @@ int native_cpu_up(unsigned int cpu, struct task_struct *tidle)
 		return err;
 
 	/* the FPU context is blank, nobody can own it */
-	__cpu_disable_lazy_restore(cpu);
+	__cpu_invalidate_fpregs_state(cpu);
 
 	common_cpu_up(cpu, tidle);
 
-- 
2.7.4

[toc] | [prev] | [next] | [standalone]

#1495669 — Re: [PATCH 7/9] x86/fpu: rename lazy restore functions to "register state valid"

On Tue, Oct 4, 2016 at 5:34 PM,  <riel@redhat.com> wrote:
> From: Rik van Riel <riel@redhat.com>
>
> Name the functions after the state they track, rather than the function
> they currently enable. This should make it more obvious when we use the
> fpu_register_state_valid function for something else in the future.

I like this.

Reviewed-by: Andy Lutomirski <luto@kernel.org>

[toc] | [prev] | [next] | [standalone]

#1497103 — [tip:x86/fpu] x86/fpu: Rename lazy restore functions to "register state valid"

Commit-ID:  25d83b531c1aa4fca5b4e24ed10f493268f162bc
Gitweb:     http://git.kernel.org/tip/25d83b531c1aa4fca5b4e24ed10f493268f162bc
Author:     Rik van Riel <riel@redhat.com>
AuthorDate: Tue, 4 Oct 2016 20:34:36 -0400
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Fri, 7 Oct 2016 11:14:41 +0200

x86/fpu: Rename lazy restore functions to "register state valid"

Name the functions after the state they track, rather than the function
they currently enable. This should make it more obvious when we use the
fpu_register_state_valid() function for something else in the future.

Signed-off-by: Rik van Riel <riel@redhat.com>
Reviewed-by: Andy Lutomirski <luto@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Quentin Casasnovas <quentin.casasnovas@oracle.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: pbonzini@redhat.com
Link: http://lkml.kernel.org/r/1475627678-20788-8-git-send-email-riel@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/include/asm/fpu/internal.h | 26 ++++++++++++++++++++------
 arch/x86/kernel/fpu/core.c          |  4 ++--
 arch/x86/kernel/smpboot.c           |  2 +-
 3 files changed, 23 insertions(+), 9 deletions(-)

diff --git a/arch/x86/include/asm/fpu/internal.h b/arch/x86/include/asm/fpu/internal.h
index 499d6ed..d2cfe16 100644
--- a/arch/x86/include/asm/fpu/internal.h
+++ b/arch/x86/include/asm/fpu/internal.h
@@ -479,18 +479,32 @@ extern int copy_fpstate_to_sigframe(void __user *buf, void __user *fp, int size)
 DECLARE_PER_CPU(struct fpu *, fpu_fpregs_owner_ctx);
 
 /*
+ * The in-register FPU state for an FPU context on a CPU is assumed to be
+ * valid if the fpu->last_cpu matches the CPU, and the fpu_fpregs_owner_ctx
+ * matches the FPU.
+ *
+ * If the FPU register state is valid, the kernel can skip restoring the
+ * FPU state from memory.
+ *
+ * Any code that clobbers the FPU registers or updates the in-memory
+ * FPU state for a task MUST let the rest of the kernel know that the
+ * FPU registers are no longer valid for this task. Calling either of
+ * these two invalidate functions is enough, use whichever is convenient.
+ *
  * Must be run with preemption disabled: this clears the fpu_fpregs_owner_ctx,
  * on this CPU.
- *
- * This will disable any lazy FPU state restore of the current FPU state,
- * but if the current thread owns the FPU, it will still be saved by.
  */
-static inline void __cpu_disable_lazy_restore(unsigned int cpu)
+static inline void __cpu_invalidate_fpregs_state(unsigned int cpu)
 {
 	per_cpu(fpu_fpregs_owner_ctx, cpu) = NULL;
 }
 
-static inline int fpu_want_lazy_restore(struct fpu *fpu, unsigned int cpu)
+static inline void __fpu_invalidate_fpregs_state(struct fpu *fpu)
+{
+	fpu->last_cpu = -1;
+}
+
+static inline int fpregs_state_valid(struct fpu *fpu, unsigned int cpu)
 {
 	return fpu == this_cpu_read_stable(fpu_fpregs_owner_ctx) && cpu == fpu->last_cpu;
 }
@@ -588,7 +602,7 @@ switch_fpu_prepare(struct fpu *old_fpu, struct fpu *new_fpu, int cpu)
 	} else {
 		old_fpu->last_cpu = -1;
 		if (fpu.preload) {
-			if (fpu_want_lazy_restore(new_fpu, cpu))
+			if (fpregs_state_valid(new_fpu, cpu))
 				fpu.preload = 0;
 			else
 				prefetch(&new_fpu->state);
diff --git a/arch/x86/kernel/fpu/core.c b/arch/x86/kernel/fpu/core.c
index 6a37d52..25a45dd 100644
--- a/arch/x86/kernel/fpu/core.c
+++ b/arch/x86/kernel/fpu/core.c
@@ -336,7 +336,7 @@ void fpu__activate_fpstate_write(struct fpu *fpu)
 
 	if (fpu->fpstate_active) {
 		/* Invalidate any lazy state: */
-		fpu->last_cpu = -1;
+		__fpu_invalidate_fpregs_state(fpu);
 	} else {
 		fpstate_init(&fpu->state);
 		trace_x86_fpu_init_state(fpu);
@@ -379,7 +379,7 @@ void fpu__current_fpstate_write_begin(void)
 	 * ensures we will not be lazy and skip a XRSTOR in the
 	 * future.
 	 */
-	fpu->last_cpu = -1;
+	__fpu_invalidate_fpregs_state(fpu);
 }
 
 /*
diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
index 42a9362..ca4c4ca 100644
--- a/arch/x86/kernel/smpboot.c
+++ b/arch/x86/kernel/smpboot.c
@@ -1111,7 +1111,7 @@ int native_cpu_up(unsigned int cpu, struct task_struct *tidle)
 		return err;
 
 	/* the FPU context is blank, nobody can own it */
-	__cpu_disable_lazy_restore(cpu);
+	__cpu_invalidate_fpregs_state(cpu);
 
 	common_cpu_up(cpu, tidle);
 

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web