i387: do not preload FPU state at task switch time

Yes, taking the trap to re-load the FPU/MMX state is expensive, but so
is spending several days looking for a bug in the state save/restore
code.  And the preload code has some rather subtle interactions with
both paravirtualization support and segment state restore, so it's not
nearly as simple as it should be.

Also, now that we no longer necessarily depend on a single bit (ie
TS_USEDFPU) for keeping track of the state of the FPU, we migth be able
to do better.  If we are really switching between two processes that
keep touching the FP state, save/restore is inevitable, but in the case
of having one process that does most of the FPU usage, we may actually
be able to do much better than the preloading.

In particular, we may be able to keep track of which CPU the process ran
on last, and also per CPU keep track of which process' FP state that CPU
has.  For modern CPU's that don't destroy the FPU contents on save time,
that would allow us to do a lazy restore by just re-enabling the
existing FPU state - with no restore cost at all!
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

arch/x86/include/asm/i387.h

@@ -30,7 +30,6 @@ extern void fpu_init(void);
 extern void mxcsr_feature_mask_init(void);
 extern int init_fpu(struct task_struct *child);
 extern void math_state_restore(void);
-extern void __math_state_restore(void);
 extern int dump_fpu(struct pt_regs *, struct user_i387_struct *);
 
 extern user_regset_active_fn fpregs_active, xfpregs_active;

arch/x86/kernel/process_32.c

@@ -299,23 +299,11 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
 				 *next = &next_p->thread;
 	int cpu = smp_processor_id();
 	struct tss_struct *tss = &per_cpu(init_tss, cpu);
-	bool preload_fpu;
 
 	/* never put a printk in __switch_to... printk() calls wake_up*() indirectly */
 
-	/*
-	 * If the task has used fpu the last 5 timeslices, just do a full
-	 * restore of the math state immediately to avoid the trap; the
-	 * chances of needing FPU soon are obviously high now
-	 */
-	preload_fpu = tsk_used_math(next_p) && next_p->fpu_counter > 5;
-
 	__unlazy_fpu(prev_p);
 
-	/* we're going to use this soon, after a few expensive things */
-	if (preload_fpu)
-		prefetch(next->fpu.state);
-
 	/*
 	 * Reload esp0.
 	 */
@@ -354,11 +342,6 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
 		     task_thread_info(next_p)->flags & _TIF_WORK_CTXSW_NEXT))
 		__switch_to_xtra(prev_p, next_p, tss);
 
-	/* If we're going to preload the fpu context, make sure clts
-	   is run while we're batching the cpu state updates. */
-	if (preload_fpu)
-		clts();
-
 	/*
 	 * Leave lazy mode, flushing any hypercalls made here.
 	 * This must be done before restoring TLS segments so
@@ -368,9 +351,6 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
 	 */
 	arch_end_context_switch(next_p);
 
-	if (preload_fpu)
-		__math_state_restore();
-
 	/*
 	 * Restore %gs if needed (which is common)
 	 */

arch/x86/kernel/process_64.c

@@ -386,18 +386,6 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
 	int cpu = smp_processor_id();
 	struct tss_struct *tss = &per_cpu(init_tss, cpu);
 	unsigned fsindex, gsindex;
-	bool preload_fpu;
-
-	/*
-	 * If the task has used fpu the last 5 timeslices, just do a full
-	 * restore of the math state immediately to avoid the trap; the
-	 * chances of needing FPU soon are obviously high now
-	 */
-	preload_fpu = tsk_used_math(next_p) && next_p->fpu_counter > 5;
-
-	/* we're going to use this soon, after a few expensive things */
-	if (preload_fpu)
-		prefetch(next->fpu.state);
 
 	/*
 	 * Reload esp0, LDT and the page table pointer:
@@ -430,10 +418,6 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
 	/* Must be after DS reload */
 	__unlazy_fpu(prev_p);
 
-	/* Make sure cpu is ready for new context */
-	if (preload_fpu)
-		clts();
-
 	/*
 	 * Leave lazy mode, flushing any hypercalls made here.
 	 * This must be done before restoring TLS segments so
@@ -492,13 +476,6 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
 		     task_thread_info(prev_p)->flags & _TIF_WORK_CTXSW_PREV))
 		__switch_to_xtra(prev_p, next_p, tss);
 
-	/*
-	 * Preload the FPU context, now that we've determined that the
-	 * task is likely to be using it. 
-	 */
-	if (preload_fpu)
-		__math_state_restore();
-
 	return prev_p;
 }
 

arch/x86/kernel/traps.c

@@ -570,28 +570,6 @@ asmlinkage void __attribute__((weak)) smp_threshold_interrupt(void)
 {
 }
 
-/*
- * __math_state_restore assumes that cr0.TS is already clear and the
- * fpu state is all ready for use.  Used during context switch.
- */
-void __math_state_restore(void)
-{
-	struct thread_info *thread = current_thread_info();
-	struct task_struct *tsk = thread->task;
-
-	/*
-	 * Paranoid restore. send a SIGSEGV if we fail to restore the state.
-	 */
-	if (unlikely(restore_fpu_checking(tsk))) {
-		stts();
-		force_sig(SIGSEGV, tsk);
-		return;
-	}
-
-	__thread_set_has_fpu(thread);	/* clts in caller! */
-	tsk->fpu_counter++;
-}
-
 /*
  * 'math_state_restore()' saves the current math information in the
  * old math state array, and gets the new ones from the current task
@@ -622,9 +600,18 @@ void math_state_restore(void)
 		local_irq_disable();
 	}
 
-	clts();				/* Allow maths ops (or we recurse) */
+	__thread_fpu_begin(thread);
 
-	__math_state_restore();
+	/*
+	 * Paranoid restore. send a SIGSEGV if we fail to restore the state.
+	 */
+	if (unlikely(restore_fpu_checking(tsk))) {
+		__thread_fpu_end(thread);
+		force_sig(SIGSEGV, tsk);
+		return;
+	}
+
+	tsk->fpu_counter++;
 }
 EXPORT_SYMBOL_GPL(math_state_restore);