lkml.org 
[lkml]   [2012]   [Feb]   [21]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
From
Date
SubjectRe: [PATCH v2 3/3] i387: support lazy restore of FPU state
On Mon, Feb 20, 2012 at 6:14 PM, H. Peter Anvin <hpa@zytor.com> wrote:
> On 02/20/2012 06:10 PM, Linus Torvalds wrote:
>>
>> Although I do wonder if we should just make kernel_fpu_begin() be a
>> real function instead of inlining it. I'm not sure it makes sense to
>> inline that thing, and it might be better to export that one instead.
>> Comments?
>
> I would agree with that.

So I have a patch that does that, but it's noticeably bigger.

It uninlines a fair amount of i387.h, and moves it into i387.c. I do
think it's probably the right thing to do, though.

I did a "make allmodconfig" with this on x86-64, but it's quite
possible that x86-32 does additional cases. Does this patch work for
people?

(This is *on*top*of* the quick "let's just get it to work" patch that
just exports the new percpu variable. I already committed that and
pushed it out, since I wanted a quick fix so that people wouldn't be
held up by this)

IOW, if you can try this on top of current -git, that would be lovely.

Linus
arch/x86/include/asm/i387.h | 78 ++--------------------------------------
arch/x86/kernel/cpu/common.c | 2 -
arch/x86/kernel/i387.c | 80 ++++++++++++++++++++++++++++++++++++++++++
3 files changed, 84 insertions(+), 76 deletions(-)

diff --git a/arch/x86/include/asm/i387.h b/arch/x86/include/asm/i387.h
index 247904945d3f..0c1031d354f2 100644
--- a/arch/x86/include/asm/i387.h
+++ b/arch/x86/include/asm/i387.h
@@ -419,70 +419,9 @@ static inline void __clear_fpu(struct task_struct *tsk)
}
}

-/*
- * Were we in an interrupt that interrupted kernel mode?
- *
- * We can do a kernel_fpu_begin/end() pair *ONLY* if that
- * pair does nothing at all: the thread must not have fpu (so
- * that we don't try to save the FPU state), and TS must
- * be set (so that the clts/stts pair does nothing that is
- * visible in the interrupted kernel thread).
- */
-static inline bool interrupted_kernel_fpu_idle(void)
-{
- return !__thread_has_fpu(current) &&
- (read_cr0() & X86_CR0_TS);
-}
-
-/*
- * Were we in user mode (or vm86 mode) when we were
- * interrupted?
- *
- * Doing kernel_fpu_begin/end() is ok if we are running
- * in an interrupt context from user mode - we'll just
- * save the FPU state as required.
- */
-static inline bool interrupted_user_mode(void)
-{
- struct pt_regs *regs = get_irq_regs();
- return regs && user_mode_vm(regs);
-}
-
-/*
- * Can we use the FPU in kernel mode with the
- * whole "kernel_fpu_begin/end()" sequence?
- *
- * It's always ok in process context (ie "not interrupt")
- * but it is sometimes ok even from an irq.
- */
-static inline bool irq_fpu_usable(void)
-{
- return !in_interrupt() ||
- interrupted_user_mode() ||
- interrupted_kernel_fpu_idle();
-}
-
-static inline void kernel_fpu_begin(void)
-{
- struct task_struct *me = current;
-
- WARN_ON_ONCE(!irq_fpu_usable());
- preempt_disable();
- if (__thread_has_fpu(me)) {
- __save_init_fpu(me);
- __thread_clear_has_fpu(me);
- /* We do 'stts()' in kernel_fpu_end() */
- } else {
- percpu_write(fpu_owner_task, NULL);
- clts();
- }
-}
-
-static inline void kernel_fpu_end(void)
-{
- stts();
- preempt_enable();
-}
+extern bool irq_fpu_usable(void);
+extern void kernel_fpu_begin(void);
+extern void kernel_fpu_end(void);

/*
* Some instructions like VIA's padlock instructions generate a spurious
@@ -566,16 +505,7 @@ static inline void save_init_fpu(struct task_struct *tsk)
preempt_enable();
}

-static inline void unlazy_fpu(struct task_struct *tsk)
-{
- preempt_disable();
- if (__thread_has_fpu(tsk)) {
- __save_init_fpu(tsk);
- __thread_fpu_end(tsk);
- } else
- tsk->fpu_counter = 0;
- preempt_enable();
-}
+extern void unlazy_fpu(struct task_struct *tsk);

static inline void clear_fpu(struct task_struct *tsk)
{
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index c0f7d68d318f..cb71b01ab66e 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -1045,7 +1045,6 @@ DEFINE_PER_CPU(char *, irq_stack_ptr) =
DEFINE_PER_CPU(unsigned int, irq_count) = -1;

DEFINE_PER_CPU(struct task_struct *, fpu_owner_task);
-EXPORT_PER_CPU_SYMBOL(fpu_owner_task);

/*
* Special IST stacks which the CPU switches to when it calls
@@ -1115,7 +1114,6 @@ void debug_stack_reset(void)
DEFINE_PER_CPU(struct task_struct *, current_task) = &init_task;
EXPORT_PER_CPU_SYMBOL(current_task);
DEFINE_PER_CPU(struct task_struct *, fpu_owner_task);
-EXPORT_PER_CPU_SYMBOL(fpu_owner_task);

#ifdef CONFIG_CC_STACKPROTECTOR
DEFINE_PER_CPU_ALIGNED(struct stack_canary, stack_canary);
diff --git a/arch/x86/kernel/i387.c b/arch/x86/kernel/i387.c
index 739d8598f789..17b7549c4134 100644
--- a/arch/x86/kernel/i387.c
+++ b/arch/x86/kernel/i387.c
@@ -32,6 +32,86 @@
# define user32_fxsr_struct user_fxsr_struct
#endif

+/*
+ * Were we in an interrupt that interrupted kernel mode?
+ *
+ * We can do a kernel_fpu_begin/end() pair *ONLY* if that
+ * pair does nothing at all: the thread must not have fpu (so
+ * that we don't try to save the FPU state), and TS must
+ * be set (so that the clts/stts pair does nothing that is
+ * visible in the interrupted kernel thread).
+ */
+static inline bool interrupted_kernel_fpu_idle(void)
+{
+ return !__thread_has_fpu(current) &&
+ (read_cr0() & X86_CR0_TS);
+}
+
+/*
+ * Were we in user mode (or vm86 mode) when we were
+ * interrupted?
+ *
+ * Doing kernel_fpu_begin/end() is ok if we are running
+ * in an interrupt context from user mode - we'll just
+ * save the FPU state as required.
+ */
+static inline bool interrupted_user_mode(void)
+{
+ struct pt_regs *regs = get_irq_regs();
+ return regs && user_mode_vm(regs);
+}
+
+/*
+ * Can we use the FPU in kernel mode with the
+ * whole "kernel_fpu_begin/end()" sequence?
+ *
+ * It's always ok in process context (ie "not interrupt")
+ * but it is sometimes ok even from an irq.
+ */
+bool irq_fpu_usable(void)
+{
+ return !in_interrupt() ||
+ interrupted_user_mode() ||
+ interrupted_kernel_fpu_idle();
+}
+EXPORT_SYMBOL(irq_fpu_usable);
+
+void kernel_fpu_begin(void)
+{
+ struct task_struct *me = current;
+
+ WARN_ON_ONCE(!irq_fpu_usable());
+ preempt_disable();
+ if (__thread_has_fpu(me)) {
+ __save_init_fpu(me);
+ __thread_clear_has_fpu(me);
+ /* We do 'stts()' in kernel_fpu_end() */
+ } else {
+ percpu_write(fpu_owner_task, NULL);
+ clts();
+ }
+}
+EXPORT_SYMBOL(kernel_fpu_begin);
+
+void kernel_fpu_end(void)
+{
+ stts();
+ preempt_enable();
+}
+EXPORT_SYMBOL(kernel_fpu_end);
+
+void unlazy_fpu(struct task_struct *tsk)
+{
+ preempt_disable();
+ if (__thread_has_fpu(tsk)) {
+ __save_init_fpu(tsk);
+ __thread_fpu_end(tsk);
+ } else
+ tsk->fpu_counter = 0;
+ preempt_enable();
+}
+EXPORT_SYMBOL(unlazy_fpu);
+
#ifdef CONFIG_MATH_EMULATION
# define HAVE_HWFP (boot_cpu_data.hard_math)
#else
\
 
 \ /
  Last update: 2012-02-21 06:31    [W:0.121 / U:0.140 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site