From 0c44c2d0f459cd7e275242b72f500137c4fa834d Mon Sep 17 00:00:00 2001
From: Peter Zijlstra <peterz@infradead.org>
Date: Wed, 11 Sep 2013 15:19:24 +0200
Subject: x86: Use asm goto to implement better modify_and_test() functions

Linus suggested using asm goto to get rid of the typical SETcc + TEST
instruction pair -- which also clobbers an extra register -- for our
typical modify_and_test() functions.

Because asm goto doesn't allow output fields it has to include an
unconditinal memory clobber when it changes a memory variable to force
a reload.

Luckily all atomic ops already imply a compiler barrier to go along
with their memory barrier semantics.

Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-0mtn9siwbeo1d33bap1422se@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/include/asm/atomic.h      | 29 +++++----------------------
 arch/x86/include/asm/atomic64_64.h | 28 ++++----------------------
 arch/x86/include/asm/bitops.h      | 24 ++++------------------
 arch/x86/include/asm/local.h       | 28 ++++----------------------
 arch/x86/include/asm/rmwcc.h       | 41 ++++++++++++++++++++++++++++++++++++++
 5 files changed, 58 insertions(+), 92 deletions(-)
 create mode 100644 arch/x86/include/asm/rmwcc.h

(limited to 'arch/x86')

diff --git a/arch/x86/include/asm/atomic.h b/arch/x86/include/asm/atomic.h
index 722aa3b0462..da31c8b8a92 100644
--- a/arch/x86/include/asm/atomic.h
+++ b/arch/x86/include/asm/atomic.h
@@ -6,6 +6,7 @@
 #include <asm/processor.h>
 #include <asm/alternative.h>
 #include <asm/cmpxchg.h>
+#include <asm/rmwcc.h>
 
 /*
  * Atomic operations that C can't guarantee us.  Useful for
@@ -76,12 +77,7 @@ static inline void atomic_sub(int i, atomic_t *v)
  */
 static inline int atomic_sub_and_test(int i, atomic_t *v)
 {
-	unsigned char c;
-
-	asm volatile(LOCK_PREFIX "subl %2,%0; sete %1"
-		     : "+m" (v->counter), "=qm" (c)
-		     : "ir" (i) : "memory");
-	return c;
+	GEN_BINARY_RMWcc(LOCK_PREFIX "subl", v->counter, i, "%0", "e");
 }
 
 /**
@@ -118,12 +114,7 @@ static inline void atomic_dec(atomic_t *v)
  */
 static inline int atomic_dec_and_test(atomic_t *v)
 {
-	unsigned char c;
-
-	asm volatile(LOCK_PREFIX "decl %0; sete %1"
-		     : "+m" (v->counter), "=qm" (c)
-		     : : "memory");
-	return c != 0;
+	GEN_UNARY_RMWcc(LOCK_PREFIX "decl", v->counter, "%0", "e");
 }
 
 /**
@@ -136,12 +127,7 @@ static inline int atomic_dec_and_test(atomic_t *v)
  */
 static inline int atomic_inc_and_test(atomic_t *v)
 {
-	unsigned char c;
-
-	asm volatile(LOCK_PREFIX "incl %0; sete %1"
-		     : "+m" (v->counter), "=qm" (c)
-		     : : "memory");
-	return c != 0;
+	GEN_UNARY_RMWcc(LOCK_PREFIX "incl", v->counter, "%0", "e");
 }
 
 /**
@@ -155,12 +141,7 @@ static inline int atomic_inc_and_test(atomic_t *v)
  */
 static inline int atomic_add_negative(int i, atomic_t *v)
 {
-	unsigned char c;
-
-	asm volatile(LOCK_PREFIX "addl %2,%0; sets %1"
-		     : "+m" (v->counter), "=qm" (c)
-		     : "ir" (i) : "memory");
-	return c;
+	GEN_BINARY_RMWcc(LOCK_PREFIX "addl", v->counter, i, "%0", "s");
 }
 
 /**
diff --git a/arch/x86/include/asm/atomic64_64.h b/arch/x86/include/asm/atomic64_64.h
index 0e1cbfc8ee0..3f065c985ae 100644
--- a/arch/x86/include/asm/atomic64_64.h
+++ b/arch/x86/include/asm/atomic64_64.h
@@ -72,12 +72,7 @@ static inline void atomic64_sub(long i, atomic64_t *v)
  */
 static inline int atomic64_sub_and_test(long i, atomic64_t *v)
 {
-	unsigned char c;
-
-	asm volatile(LOCK_PREFIX "subq %2,%0; sete %1"
-		     : "=m" (v->counter), "=qm" (c)
-		     : "er" (i), "m" (v->counter) : "memory");
-	return c;
+	GEN_BINARY_RMWcc(LOCK_PREFIX "subq", v->counter, i, "%0", "e");
 }
 
 /**
@@ -116,12 +111,7 @@ static inline void atomic64_dec(atomic64_t *v)
  */
 static inline int atomic64_dec_and_test(atomic64_t *v)
 {
-	unsigned char c;
-
-	asm volatile(LOCK_PREFIX "decq %0; sete %1"
-		     : "=m" (v->counter), "=qm" (c)
-		     : "m" (v->counter) : "memory");
-	return c != 0;
+	GEN_UNARY_RMWcc(LOCK_PREFIX "decq", v->counter, "%0", "e");
 }
 
 /**
@@ -134,12 +124,7 @@ static inline int atomic64_dec_and_test(atomic64_t *v)
  */
 static inline int atomic64_inc_and_test(atomic64_t *v)
 {
-	unsigned char c;
-
-	asm volatile(LOCK_PREFIX "incq %0; sete %1"
-		     : "=m" (v->counter), "=qm" (c)
-		     : "m" (v->counter) : "memory");
-	return c != 0;
+	GEN_UNARY_RMWcc(LOCK_PREFIX "incq", v->counter, "%0", "e");
 }
 
 /**
@@ -153,12 +138,7 @@ static inline int atomic64_inc_and_test(atomic64_t *v)
  */
 static inline int atomic64_add_negative(long i, atomic64_t *v)
 {
-	unsigned char c;
-
-	asm volatile(LOCK_PREFIX "addq %2,%0; sets %1"
-		     : "=m" (v->counter), "=qm" (c)
-		     : "er" (i), "m" (v->counter) : "memory");
-	return c;
+	GEN_BINARY_RMWcc(LOCK_PREFIX "addq", v->counter, i, "%0", "s");
 }
 
 /**
diff --git a/arch/x86/include/asm/bitops.h b/arch/x86/include/asm/bitops.h
index 41639ce8fd6..6d76d093598 100644
--- a/arch/x86/include/asm/bitops.h
+++ b/arch/x86/include/asm/bitops.h
@@ -14,6 +14,7 @@
 
 #include <linux/compiler.h>
 #include <asm/alternative.h>
+#include <asm/rmwcc.h>
 
 #if BITS_PER_LONG == 32
 # define _BITOPS_LONG_SHIFT 5
@@ -204,12 +205,7 @@ static inline void change_bit(long nr, volatile unsigned long *addr)
  */
 static inline int test_and_set_bit(long nr, volatile unsigned long *addr)
 {
-	int oldbit;
-
-	asm volatile(LOCK_PREFIX "bts %2,%1\n\t"
-		     "sbb %0,%0" : "=r" (oldbit), ADDR : "Ir" (nr) : "memory");
-
-	return oldbit;
+	GEN_BINARY_RMWcc(LOCK_PREFIX "bts", *addr, nr, "%0", "c");
 }
 
 /**
@@ -255,13 +251,7 @@ static inline int __test_and_set_bit(long nr, volatile unsigned long *addr)
  */
 static inline int test_and_clear_bit(long nr, volatile unsigned long *addr)
 {
-	int oldbit;
-
-	asm volatile(LOCK_PREFIX "btr %2,%1\n\t"
-		     "sbb %0,%0"
-		     : "=r" (oldbit), ADDR : "Ir" (nr) : "memory");
-
-	return oldbit;
+	GEN_BINARY_RMWcc(LOCK_PREFIX "btr", *addr, nr, "%0", "c");
 }
 
 /**
@@ -314,13 +304,7 @@ static inline int __test_and_change_bit(long nr, volatile unsigned long *addr)
  */
 static inline int test_and_change_bit(long nr, volatile unsigned long *addr)
 {
-	int oldbit;
-
-	asm volatile(LOCK_PREFIX "btc %2,%1\n\t"
-		     "sbb %0,%0"
-		     : "=r" (oldbit), ADDR : "Ir" (nr) : "memory");
-
-	return oldbit;
+	GEN_BINARY_RMWcc(LOCK_PREFIX "btc", *addr, nr, "%0", "c");
 }
 
 static __always_inline int constant_test_bit(long nr, const volatile unsigned long *addr)
diff --git a/arch/x86/include/asm/local.h b/arch/x86/include/asm/local.h
index 2d89e3980cb..5b23e605e70 100644
--- a/arch/x86/include/asm/local.h
+++ b/arch/x86/include/asm/local.h
@@ -52,12 +52,7 @@ static inline void local_sub(long i, local_t *l)
  */
 static inline int local_sub_and_test(long i, local_t *l)
 {
-	unsigned char c;
-
-	asm volatile(_ASM_SUB "%2,%0; sete %1"
-		     : "+m" (l->a.counter), "=qm" (c)
-		     : "ir" (i) : "memory");
-	return c;
+	GEN_BINARY_RMWcc(_ASM_SUB, l->a.counter, i, "%0", "e");
 }
 
 /**
@@ -70,12 +65,7 @@ static inline int local_sub_and_test(long i, local_t *l)
  */
 static inline int local_dec_and_test(local_t *l)
 {
-	unsigned char c;
-
-	asm volatile(_ASM_DEC "%0; sete %1"
-		     : "+m" (l->a.counter), "=qm" (c)
-		     : : "memory");
-	return c != 0;
+	GEN_UNARY_RMWcc(_ASM_DEC, l->a.counter, "%0", "e");
 }
 
 /**
@@ -88,12 +78,7 @@ static inline int local_dec_and_test(local_t *l)
  */
 static inline int local_inc_and_test(local_t *l)
 {
-	unsigned char c;
-
-	asm volatile(_ASM_INC "%0; sete %1"
-		     : "+m" (l->a.counter), "=qm" (c)
-		     : : "memory");
-	return c != 0;
+	GEN_UNARY_RMWcc(_ASM_INC, l->a.counter, "%0", "e");
 }
 
 /**
@@ -107,12 +92,7 @@ static inline int local_inc_and_test(local_t *l)
  */
 static inline int local_add_negative(long i, local_t *l)
 {
-	unsigned char c;
-
-	asm volatile(_ASM_ADD "%2,%0; sets %1"
-		     : "+m" (l->a.counter), "=qm" (c)
-		     : "ir" (i) : "memory");
-	return c;
+	GEN_BINARY_RMWcc(_ASM_ADD, l->a.counter, i, "%0", "s");
 }
 
 /**
diff --git a/arch/x86/include/asm/rmwcc.h b/arch/x86/include/asm/rmwcc.h
new file mode 100644
index 00000000000..735f1849795
--- /dev/null
+++ b/arch/x86/include/asm/rmwcc.h
@@ -0,0 +1,41 @@
+#ifndef _ASM_X86_RMWcc
+#define _ASM_X86_RMWcc
+
+#ifdef CC_HAVE_ASM_GOTO
+
+#define __GEN_RMWcc(fullop, var, cc, ...)				\
+do {									\
+	asm volatile goto (fullop "; j" cc " %l[cc_label]"		\
+			: : "m" (var), ## __VA_ARGS__ 			\
+			: "memory" : cc_label);				\
+	return 0;							\
+cc_label:								\
+	return 1;							\
+} while (0)
+
+#define GEN_UNARY_RMWcc(op, var, arg0, cc) 				\
+	__GEN_RMWcc(op " " arg0, var, cc)
+
+#define GEN_BINARY_RMWcc(op, var, val, arg0, cc)			\
+	__GEN_RMWcc(op " %1, " arg0, var, cc, "er" (val))
+
+#else /* !CC_HAVE_ASM_GOTO */
+
+#define __GEN_RMWcc(fullop, var, cc, ...)				\
+do {									\
+	char c;								\
+	asm volatile (fullop "; set" cc " %1"				\
+			: "+m" (var), "=qm" (c)				\
+			: __VA_ARGS__ : "memory");			\
+	return c != 0;							\
+} while (0)
+
+#define GEN_UNARY_RMWcc(op, var, arg0, cc)				\
+	__GEN_RMWcc(op " " arg0, var, cc)
+
+#define GEN_BINARY_RMWcc(op, var, val, arg0, cc)			\
+	__GEN_RMWcc(op " %2, " arg0, var, cc, "er" (val))
+
+#endif /* CC_HAVE_ASM_GOTO */
+
+#endif /* _ASM_X86_RMWcc */
-- 
cgit v1.2.3-70-g09d2


From ea8117478918a4734586d35ff530721b682425be Mon Sep 17 00:00:00 2001
From: Peter Zijlstra <peterz@infradead.org>
Date: Wed, 11 Sep 2013 12:43:13 +0200
Subject: sched, idle: Fix the idle polling state logic

Mike reported that commit 7d1a9417 ("x86: Use generic idle loop")
regressed several workloads and caused excessive reschedule
interrupts.

The patch in question failed to notice that the x86 code had an
inverted sense of the polling state versus the new generic code (x86:
default polling, generic: default !polling).

Fix the two prominent x86 mwait based idle drivers and introduce a few
new generic polling helpers (fixing the wrong smp_mb__after_clear_bit
usage).

Also switch the idle routines to using tif_need_resched() which is an
immediate TIF_NEED_RESCHED test as opposed to need_resched which will
end up being slightly different.

Reported-by: Mike Galbraith <bitbucket@online.de>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: lenb@kernel.org
Cc: tglx@linutronix.de
Link: http://lkml.kernel.org/n/tip-nc03imb0etuefmzybzj7sprf@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/kernel/process.c     |  6 ++--
 drivers/acpi/processor_idle.c | 46 ++++++-------------------
 drivers/idle/intel_idle.c     |  2 +-
 include/linux/sched.h         | 78 +++++++++++++++++++++++++++++++++++++++----
 include/linux/thread_info.h   |  2 ++
 kernel/cpu/idle.c             |  9 +++--
 6 files changed, 91 insertions(+), 52 deletions(-)

(limited to 'arch/x86')

diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c
index c83516be105..3fb8d95ab8b 100644
--- a/arch/x86/kernel/process.c
+++ b/arch/x86/kernel/process.c
@@ -391,9 +391,9 @@ static void amd_e400_idle(void)
 		 * The switch back from broadcast mode needs to be
 		 * called with interrupts disabled.
 		 */
-		 local_irq_disable();
-		 clockevents_notify(CLOCK_EVT_NOTIFY_BROADCAST_EXIT, &cpu);
-		 local_irq_enable();
+		local_irq_disable();
+		clockevents_notify(CLOCK_EVT_NOTIFY_BROADCAST_EXIT, &cpu);
+		local_irq_enable();
 	} else
 		default_idle();
 }
diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c
index f98dd00b51a..c7414a545a4 100644
--- a/drivers/acpi/processor_idle.c
+++ b/drivers/acpi/processor_idle.c
@@ -119,17 +119,10 @@ static struct dmi_system_id processor_power_dmi_table[] = {
  */
 static void acpi_safe_halt(void)
 {
-	current_thread_info()->status &= ~TS_POLLING;
-	/*
-	 * TS_POLLING-cleared state must be visible before we
-	 * test NEED_RESCHED:
-	 */
-	smp_mb();
-	if (!need_resched()) {
+	if (!tif_need_resched()) {
 		safe_halt();
 		local_irq_disable();
 	}
-	current_thread_info()->status |= TS_POLLING;
 }
 
 #ifdef ARCH_APICTIMER_STOPS_ON_C3
@@ -737,6 +730,11 @@ static int acpi_idle_enter_c1(struct cpuidle_device *dev,
 	if (unlikely(!pr))
 		return -EINVAL;
 
+	if (cx->entry_method == ACPI_CSTATE_FFH) {
+		if (current_set_polling_and_test())
+			return -EINVAL;
+	}
+
 	lapic_timer_state_broadcast(pr, cx, 1);
 	acpi_idle_do_entry(cx);
 
@@ -790,18 +788,9 @@ static int acpi_idle_enter_simple(struct cpuidle_device *dev,
 	if (unlikely(!pr))
 		return -EINVAL;
 
-	if (cx->entry_method != ACPI_CSTATE_FFH) {
-		current_thread_info()->status &= ~TS_POLLING;
-		/*
-		 * TS_POLLING-cleared state must be visible before we test
-		 * NEED_RESCHED:
-		 */
-		smp_mb();
-
-		if (unlikely(need_resched())) {
-			current_thread_info()->status |= TS_POLLING;
+	if (cx->entry_method == ACPI_CSTATE_FFH) {
+		if (current_set_polling_and_test())
 			return -EINVAL;
-		}
 	}
 
 	/*
@@ -819,9 +808,6 @@ static int acpi_idle_enter_simple(struct cpuidle_device *dev,
 
 	sched_clock_idle_wakeup_event(0);
 
-	if (cx->entry_method != ACPI_CSTATE_FFH)
-		current_thread_info()->status |= TS_POLLING;
-
 	lapic_timer_state_broadcast(pr, cx, 0);
 	return index;
 }
@@ -858,18 +844,9 @@ static int acpi_idle_enter_bm(struct cpuidle_device *dev,
 		}
 	}
 
-	if (cx->entry_method != ACPI_CSTATE_FFH) {
-		current_thread_info()->status &= ~TS_POLLING;
-		/*
-		 * TS_POLLING-cleared state must be visible before we test
-		 * NEED_RESCHED:
-		 */
-		smp_mb();
-
-		if (unlikely(need_resched())) {
-			current_thread_info()->status |= TS_POLLING;
+	if (cx->entry_method == ACPI_CSTATE_FFH) {
+		if (current_set_polling_and_test())
 			return -EINVAL;
-		}
 	}
 
 	acpi_unlazy_tlb(smp_processor_id());
@@ -915,9 +892,6 @@ static int acpi_idle_enter_bm(struct cpuidle_device *dev,
 
 	sched_clock_idle_wakeup_event(0);
 
-	if (cx->entry_method != ACPI_CSTATE_FFH)
-		current_thread_info()->status |= TS_POLLING;
-
 	lapic_timer_state_broadcast(pr, cx, 0);
 	return index;
 }
diff --git a/drivers/idle/intel_idle.c b/drivers/idle/intel_idle.c
index fa6964d8681..f116d664b47 100644
--- a/drivers/idle/intel_idle.c
+++ b/drivers/idle/intel_idle.c
@@ -359,7 +359,7 @@ static int intel_idle(struct cpuidle_device *dev,
 	if (!(lapic_timer_reliable_states & (1 << (cstate))))
 		clockevents_notify(CLOCK_EVT_NOTIFY_BROADCAST_ENTER, &cpu);
 
-	if (!need_resched()) {
+	if (!current_set_polling_and_test()) {
 
 		__monitor((void *)&current_thread_info()->flags, 0, 0);
 		smp_mb();
diff --git a/include/linux/sched.h b/include/linux/sched.h
index b5344de1658..e783ec52295 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -2479,34 +2479,98 @@ static inline int tsk_is_polling(struct task_struct *p)
 {
 	return task_thread_info(p)->status & TS_POLLING;
 }
-static inline void current_set_polling(void)
+static inline void __current_set_polling(void)
 {
 	current_thread_info()->status |= TS_POLLING;
 }
 
-static inline void current_clr_polling(void)
+static inline bool __must_check current_set_polling_and_test(void)
+{
+	__current_set_polling();
+
+	/*
+	 * Polling state must be visible before we test NEED_RESCHED,
+	 * paired by resched_task()
+	 */
+	smp_mb();
+
+	return unlikely(tif_need_resched());
+}
+
+static inline void __current_clr_polling(void)
 {
 	current_thread_info()->status &= ~TS_POLLING;
-	smp_mb__after_clear_bit();
+}
+
+static inline bool __must_check current_clr_polling_and_test(void)
+{
+	__current_clr_polling();
+
+	/*
+	 * Polling state must be visible before we test NEED_RESCHED,
+	 * paired by resched_task()
+	 */
+	smp_mb();
+
+	return unlikely(tif_need_resched());
 }
 #elif defined(TIF_POLLING_NRFLAG)
 static inline int tsk_is_polling(struct task_struct *p)
 {
 	return test_tsk_thread_flag(p, TIF_POLLING_NRFLAG);
 }
-static inline void current_set_polling(void)
+
+static inline void __current_set_polling(void)
 {
 	set_thread_flag(TIF_POLLING_NRFLAG);
 }
 
-static inline void current_clr_polling(void)
+static inline bool __must_check current_set_polling_and_test(void)
+{
+	__current_set_polling();
+
+	/*
+	 * Polling state must be visible before we test NEED_RESCHED,
+	 * paired by resched_task()
+	 *
+	 * XXX: assumes set/clear bit are identical barrier wise.
+	 */
+	smp_mb__after_clear_bit();
+
+	return unlikely(tif_need_resched());
+}
+
+static inline void __current_clr_polling(void)
 {
 	clear_thread_flag(TIF_POLLING_NRFLAG);
 }
+
+static inline bool __must_check current_clr_polling_and_test(void)
+{
+	__current_clr_polling();
+
+	/*
+	 * Polling state must be visible before we test NEED_RESCHED,
+	 * paired by resched_task()
+	 */
+	smp_mb__after_clear_bit();
+
+	return unlikely(tif_need_resched());
+}
+
 #else
 static inline int tsk_is_polling(struct task_struct *p) { return 0; }
-static inline void current_set_polling(void) { }
-static inline void current_clr_polling(void) { }
+static inline void __current_set_polling(void) { }
+static inline void __current_clr_polling(void) { }
+
+static inline bool __must_check current_set_polling_and_test(void)
+{
+	return unlikely(tif_need_resched());
+}
+static inline bool __must_check current_clr_polling_and_test(void)
+{
+	return unlikely(tif_need_resched());
+}
 #endif
 
 /*
diff --git a/include/linux/thread_info.h b/include/linux/thread_info.h
index a629e4b2321..fddbe2023a5 100644
--- a/include/linux/thread_info.h
+++ b/include/linux/thread_info.h
@@ -118,6 +118,8 @@ static inline __deprecated void set_need_resched(void)
 	 */
 }
 
+#define tif_need_resched() test_thread_flag(TIF_NEED_RESCHED)
+
 #if defined TIF_RESTORE_SIGMASK && !defined HAVE_SET_RESTORE_SIGMASK
 /*
  * An arch can define its own version of set_restore_sigmask() to get the
diff --git a/kernel/cpu/idle.c b/kernel/cpu/idle.c
index e695c0a0bcb..c261409500e 100644
--- a/kernel/cpu/idle.c
+++ b/kernel/cpu/idle.c
@@ -44,7 +44,7 @@ static inline int cpu_idle_poll(void)
 	rcu_idle_enter();
 	trace_cpu_idle_rcuidle(0, smp_processor_id());
 	local_irq_enable();
-	while (!need_resched())
+	while (!tif_need_resched())
 		cpu_relax();
 	trace_cpu_idle_rcuidle(PWR_EVENT_EXIT, smp_processor_id());
 	rcu_idle_exit();
@@ -92,8 +92,7 @@ static void cpu_idle_loop(void)
 			if (cpu_idle_force_poll || tick_check_broadcast_expired()) {
 				cpu_idle_poll();
 			} else {
-				current_clr_polling();
-				if (!need_resched()) {
+				if (!current_clr_polling_and_test()) {
 					stop_critical_timings();
 					rcu_idle_enter();
 					arch_cpu_idle();
@@ -103,7 +102,7 @@ static void cpu_idle_loop(void)
 				} else {
 					local_irq_enable();
 				}
-				current_set_polling();
+				__current_set_polling();
 			}
 			arch_cpu_idle_exit();
 		}
@@ -129,7 +128,7 @@ void cpu_startup_entry(enum cpuhp_state state)
 	 */
 	boot_init_stack_canary();
 #endif
-	current_set_polling();
+	__current_set_polling();
 	arch_cpu_idle_prepare();
 	cpu_idle_loop();
 }
-- 
cgit v1.2.3-70-g09d2


From a787870924dbd6f321661e06d4ec1c7a408c9ccf Mon Sep 17 00:00:00 2001
From: Peter Zijlstra <peterz@infradead.org>
Date: Wed, 14 Aug 2013 14:55:40 +0200
Subject: sched, arch: Create asm/preempt.h

In order to prepare to per-arch implementations of preempt_count move
the required bits into an asm-generic header and use this for all
archs.

Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-h5j0c1r3e3fk015m30h8f1zx@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/alpha/include/asm/Kbuild      |  1 +
 arch/arc/include/asm/Kbuild        |  1 +
 arch/arm/include/asm/Kbuild        |  1 +
 arch/arm64/include/asm/Kbuild      |  1 +
 arch/avr32/include/asm/Kbuild      |  1 +
 arch/blackfin/include/asm/Kbuild   |  1 +
 arch/c6x/include/asm/Kbuild        |  1 +
 arch/cris/include/asm/Kbuild       |  1 +
 arch/frv/include/asm/Kbuild        |  1 +
 arch/h8300/include/asm/Kbuild      |  1 +
 arch/hexagon/include/asm/Kbuild    |  1 +
 arch/ia64/include/asm/Kbuild       |  1 +
 arch/m32r/include/asm/Kbuild       |  1 +
 arch/m68k/include/asm/Kbuild       |  1 +
 arch/metag/include/asm/Kbuild      |  1 +
 arch/microblaze/include/asm/Kbuild |  1 +
 arch/mips/include/asm/Kbuild       |  1 +
 arch/mn10300/include/asm/Kbuild    |  1 +
 arch/openrisc/include/asm/Kbuild   |  1 +
 arch/parisc/include/asm/Kbuild     |  1 +
 arch/powerpc/include/asm/Kbuild    |  1 +
 arch/s390/include/asm/Kbuild       |  1 +
 arch/score/include/asm/Kbuild      |  1 +
 arch/sh/include/asm/Kbuild         |  1 +
 arch/sparc/include/asm/Kbuild      |  1 +
 arch/tile/include/asm/Kbuild       |  1 +
 arch/um/include/asm/Kbuild         |  1 +
 arch/unicore32/include/asm/Kbuild  |  1 +
 arch/x86/include/asm/Kbuild        |  1 +
 arch/xtensa/include/asm/Kbuild     |  1 +
 include/asm-generic/preempt.h      | 54 ++++++++++++++++++++++++++++++++++++++
 include/linux/preempt.h            | 49 +---------------------------------
 32 files changed, 85 insertions(+), 48 deletions(-)
 create mode 100644 include/asm-generic/preempt.h

(limited to 'arch/x86')

diff --git a/arch/alpha/include/asm/Kbuild b/arch/alpha/include/asm/Kbuild
index a6e85f448c1..f01fb505ad5 100644
--- a/arch/alpha/include/asm/Kbuild
+++ b/arch/alpha/include/asm/Kbuild
@@ -3,3 +3,4 @@ generic-y += clkdev.h
 
 generic-y += exec.h
 generic-y += trace_clock.h
+generic-y += preempt.h
diff --git a/arch/arc/include/asm/Kbuild b/arch/arc/include/asm/Kbuild
index d8dd660898b..5943f7f9d32 100644
--- a/arch/arc/include/asm/Kbuild
+++ b/arch/arc/include/asm/Kbuild
@@ -46,3 +46,4 @@ generic-y += ucontext.h
 generic-y += user.h
 generic-y += vga.h
 generic-y += xor.h
+generic-y += preempt.h
diff --git a/arch/arm/include/asm/Kbuild b/arch/arm/include/asm/Kbuild
index d3db39860b9..4e6838d4ddf 100644
--- a/arch/arm/include/asm/Kbuild
+++ b/arch/arm/include/asm/Kbuild
@@ -33,3 +33,4 @@ generic-y += timex.h
 generic-y += trace_clock.h
 generic-y += types.h
 generic-y += unaligned.h
+generic-y += preempt.h
diff --git a/arch/arm64/include/asm/Kbuild b/arch/arm64/include/asm/Kbuild
index 79a642d199f..519f89f5b6a 100644
--- a/arch/arm64/include/asm/Kbuild
+++ b/arch/arm64/include/asm/Kbuild
@@ -50,3 +50,4 @@ generic-y += unaligned.h
 generic-y += user.h
 generic-y += vga.h
 generic-y += xor.h
+generic-y += preempt.h
diff --git a/arch/avr32/include/asm/Kbuild b/arch/avr32/include/asm/Kbuild
index d22af851f3f..b946080ee8b 100644
--- a/arch/avr32/include/asm/Kbuild
+++ b/arch/avr32/include/asm/Kbuild
@@ -3,3 +3,4 @@ generic-y	+= clkdev.h
 generic-y	+= exec.h
 generic-y	+= trace_clock.h
 generic-y	+= param.h
+generic-y += preempt.h
diff --git a/arch/blackfin/include/asm/Kbuild b/arch/blackfin/include/asm/Kbuild
index 127826f8a37..f2b43474b0e 100644
--- a/arch/blackfin/include/asm/Kbuild
+++ b/arch/blackfin/include/asm/Kbuild
@@ -44,3 +44,4 @@ generic-y += ucontext.h
 generic-y += unaligned.h
 generic-y += user.h
 generic-y += xor.h
+generic-y += preempt.h
diff --git a/arch/c6x/include/asm/Kbuild b/arch/c6x/include/asm/Kbuild
index e49f918531a..fc0b3c35602 100644
--- a/arch/c6x/include/asm/Kbuild
+++ b/arch/c6x/include/asm/Kbuild
@@ -56,3 +56,4 @@ generic-y += ucontext.h
 generic-y += user.h
 generic-y += vga.h
 generic-y += xor.h
+generic-y += preempt.h
diff --git a/arch/cris/include/asm/Kbuild b/arch/cris/include/asm/Kbuild
index c8325455520..b06caf649a9 100644
--- a/arch/cris/include/asm/Kbuild
+++ b/arch/cris/include/asm/Kbuild
@@ -11,3 +11,4 @@ generic-y += module.h
 generic-y += trace_clock.h
 generic-y += vga.h
 generic-y += xor.h
+generic-y += preempt.h
diff --git a/arch/frv/include/asm/Kbuild b/arch/frv/include/asm/Kbuild
index c5d76702830..74742dc6a3d 100644
--- a/arch/frv/include/asm/Kbuild
+++ b/arch/frv/include/asm/Kbuild
@@ -2,3 +2,4 @@
 generic-y += clkdev.h
 generic-y += exec.h
 generic-y += trace_clock.h
+generic-y += preempt.h
diff --git a/arch/h8300/include/asm/Kbuild b/arch/h8300/include/asm/Kbuild
index 8ada3cf0c98..7e0e7213a48 100644
--- a/arch/h8300/include/asm/Kbuild
+++ b/arch/h8300/include/asm/Kbuild
@@ -6,3 +6,4 @@ generic-y += mmu.h
 generic-y += module.h
 generic-y += trace_clock.h
 generic-y += xor.h
+generic-y += preempt.h
diff --git a/arch/hexagon/include/asm/Kbuild b/arch/hexagon/include/asm/Kbuild
index 1da17caac23..67c3450309b 100644
--- a/arch/hexagon/include/asm/Kbuild
+++ b/arch/hexagon/include/asm/Kbuild
@@ -53,3 +53,4 @@ generic-y += types.h
 generic-y += ucontext.h
 generic-y += unaligned.h
 generic-y += xor.h
+generic-y += preempt.h
diff --git a/arch/ia64/include/asm/Kbuild b/arch/ia64/include/asm/Kbuild
index a3456f34f67..f93ee087e8f 100644
--- a/arch/ia64/include/asm/Kbuild
+++ b/arch/ia64/include/asm/Kbuild
@@ -3,4 +3,5 @@ generic-y += clkdev.h
 generic-y += exec.h
 generic-y += kvm_para.h
 generic-y += trace_clock.h
+generic-y += preempt.h
 generic-y += vtime.h
\ No newline at end of file
diff --git a/arch/m32r/include/asm/Kbuild b/arch/m32r/include/asm/Kbuild
index bebdc36ebb0..2b58c5f0bc3 100644
--- a/arch/m32r/include/asm/Kbuild
+++ b/arch/m32r/include/asm/Kbuild
@@ -3,3 +3,4 @@ generic-y += clkdev.h
 generic-y += exec.h
 generic-y += module.h
 generic-y += trace_clock.h
+generic-y += preempt.h
diff --git a/arch/m68k/include/asm/Kbuild b/arch/m68k/include/asm/Kbuild
index 09d77a862da..a5d27f272a5 100644
--- a/arch/m68k/include/asm/Kbuild
+++ b/arch/m68k/include/asm/Kbuild
@@ -31,3 +31,4 @@ generic-y += trace_clock.h
 generic-y += types.h
 generic-y += word-at-a-time.h
 generic-y += xor.h
+generic-y += preempt.h
diff --git a/arch/metag/include/asm/Kbuild b/arch/metag/include/asm/Kbuild
index 6ae0ccb632c..84d0c1d6b9b 100644
--- a/arch/metag/include/asm/Kbuild
+++ b/arch/metag/include/asm/Kbuild
@@ -52,3 +52,4 @@ generic-y += unaligned.h
 generic-y += user.h
 generic-y += vga.h
 generic-y += xor.h
+generic-y += preempt.h
diff --git a/arch/microblaze/include/asm/Kbuild b/arch/microblaze/include/asm/Kbuild
index d3c51a6a601..ce0bbf8f564 100644
--- a/arch/microblaze/include/asm/Kbuild
+++ b/arch/microblaze/include/asm/Kbuild
@@ -3,3 +3,4 @@ generic-y += clkdev.h
 generic-y += exec.h
 generic-y += trace_clock.h
 generic-y += syscalls.h
+generic-y += preempt.h
diff --git a/arch/mips/include/asm/Kbuild b/arch/mips/include/asm/Kbuild
index 454ddf9bb76..1acbb8b77a7 100644
--- a/arch/mips/include/asm/Kbuild
+++ b/arch/mips/include/asm/Kbuild
@@ -11,5 +11,6 @@ generic-y += sections.h
 generic-y += segment.h
 generic-y += serial.h
 generic-y += trace_clock.h
+generic-y += preempt.h
 generic-y += ucontext.h
 generic-y += xor.h
diff --git a/arch/mn10300/include/asm/Kbuild b/arch/mn10300/include/asm/Kbuild
index c5d76702830..74742dc6a3d 100644
--- a/arch/mn10300/include/asm/Kbuild
+++ b/arch/mn10300/include/asm/Kbuild
@@ -2,3 +2,4 @@
 generic-y += clkdev.h
 generic-y += exec.h
 generic-y += trace_clock.h
+generic-y += preempt.h
diff --git a/arch/openrisc/include/asm/Kbuild b/arch/openrisc/include/asm/Kbuild
index 195653e851d..78405625e79 100644
--- a/arch/openrisc/include/asm/Kbuild
+++ b/arch/openrisc/include/asm/Kbuild
@@ -67,3 +67,4 @@ generic-y += ucontext.h
 generic-y += user.h
 generic-y += word-at-a-time.h
 generic-y += xor.h
+generic-y += preempt.h
diff --git a/arch/parisc/include/asm/Kbuild b/arch/parisc/include/asm/Kbuild
index ff4c9faed54..a603b9ebe54 100644
--- a/arch/parisc/include/asm/Kbuild
+++ b/arch/parisc/include/asm/Kbuild
@@ -4,3 +4,4 @@ generic-y += word-at-a-time.h auxvec.h user.h cputime.h emergency-restart.h \
 	  div64.h irq_regs.h kdebug.h kvm_para.h local64.h local.h param.h \
 	  poll.h xor.h clkdev.h exec.h
 generic-y += trace_clock.h
+generic-y += preempt.h
diff --git a/arch/powerpc/include/asm/Kbuild b/arch/powerpc/include/asm/Kbuild
index 704e6f10ae8..d8f9d2f18a2 100644
--- a/arch/powerpc/include/asm/Kbuild
+++ b/arch/powerpc/include/asm/Kbuild
@@ -2,4 +2,5 @@
 generic-y += clkdev.h
 generic-y += rwsem.h
 generic-y += trace_clock.h
+generic-y += preempt.h
 generic-y += vtime.h
\ No newline at end of file
diff --git a/arch/s390/include/asm/Kbuild b/arch/s390/include/asm/Kbuild
index f313f9cbcf4..7a5288f3479 100644
--- a/arch/s390/include/asm/Kbuild
+++ b/arch/s390/include/asm/Kbuild
@@ -2,3 +2,4 @@
 
 generic-y += clkdev.h
 generic-y += trace_clock.h
+generic-y += preempt.h
diff --git a/arch/score/include/asm/Kbuild b/arch/score/include/asm/Kbuild
index e1c7bb999b0..f3414ade77a 100644
--- a/arch/score/include/asm/Kbuild
+++ b/arch/score/include/asm/Kbuild
@@ -4,3 +4,4 @@ header-y +=
 generic-y += clkdev.h
 generic-y += trace_clock.h
 generic-y += xor.h
+generic-y += preempt.h
diff --git a/arch/sh/include/asm/Kbuild b/arch/sh/include/asm/Kbuild
index 280bea9e5e2..231efbb6810 100644
--- a/arch/sh/include/asm/Kbuild
+++ b/arch/sh/include/asm/Kbuild
@@ -34,3 +34,4 @@ generic-y += termios.h
 generic-y += trace_clock.h
 generic-y += ucontext.h
 generic-y += xor.h
+generic-y += preempt.h
diff --git a/arch/sparc/include/asm/Kbuild b/arch/sparc/include/asm/Kbuild
index 7e4a97fbded..bf390667657 100644
--- a/arch/sparc/include/asm/Kbuild
+++ b/arch/sparc/include/asm/Kbuild
@@ -16,3 +16,4 @@ generic-y += serial.h
 generic-y += trace_clock.h
 generic-y += types.h
 generic-y += word-at-a-time.h
+generic-y += preempt.h
diff --git a/arch/tile/include/asm/Kbuild b/arch/tile/include/asm/Kbuild
index 664d6ad23f8..22f3bd147fa 100644
--- a/arch/tile/include/asm/Kbuild
+++ b/arch/tile/include/asm/Kbuild
@@ -38,3 +38,4 @@ generic-y += termios.h
 generic-y += trace_clock.h
 generic-y += types.h
 generic-y += xor.h
+generic-y += preempt.h
diff --git a/arch/um/include/asm/Kbuild b/arch/um/include/asm/Kbuild
index b30f34a7988..fdde187e608 100644
--- a/arch/um/include/asm/Kbuild
+++ b/arch/um/include/asm/Kbuild
@@ -3,3 +3,4 @@ generic-y += hw_irq.h irq_regs.h kdebug.h percpu.h sections.h topology.h xor.h
 generic-y += ftrace.h pci.h io.h param.h delay.h mutex.h current.h exec.h
 generic-y += switch_to.h clkdev.h
 generic-y += trace_clock.h
+generic-y += preempt.h
diff --git a/arch/unicore32/include/asm/Kbuild b/arch/unicore32/include/asm/Kbuild
index 89d8b6c4e39..00045cbe5c6 100644
--- a/arch/unicore32/include/asm/Kbuild
+++ b/arch/unicore32/include/asm/Kbuild
@@ -60,3 +60,4 @@ generic-y += unaligned.h
 generic-y += user.h
 generic-y += vga.h
 generic-y += xor.h
+generic-y += preempt.h
diff --git a/arch/x86/include/asm/Kbuild b/arch/x86/include/asm/Kbuild
index 7f669853317..eca20286a91 100644
--- a/arch/x86/include/asm/Kbuild
+++ b/arch/x86/include/asm/Kbuild
@@ -5,3 +5,4 @@ genhdr-y += unistd_64.h
 genhdr-y += unistd_x32.h
 
 generic-y += clkdev.h
+generic-y += preempt.h
diff --git a/arch/xtensa/include/asm/Kbuild b/arch/xtensa/include/asm/Kbuild
index 1b982641ec3..228d6aee3a1 100644
--- a/arch/xtensa/include/asm/Kbuild
+++ b/arch/xtensa/include/asm/Kbuild
@@ -28,3 +28,4 @@ generic-y += termios.h
 generic-y += topology.h
 generic-y += trace_clock.h
 generic-y += xor.h
+generic-y += preempt.h
diff --git a/include/asm-generic/preempt.h b/include/asm-generic/preempt.h
new file mode 100644
index 00000000000..a1fc6590a74
--- /dev/null
+++ b/include/asm-generic/preempt.h
@@ -0,0 +1,54 @@
+#ifndef __ASM_PREEMPT_H
+#define __ASM_PREEMPT_H
+
+#include <linux/thread_info.h>
+
+/*
+ * We mask the PREEMPT_NEED_RESCHED bit so as not to confuse all current users
+ * that think a non-zero value indicates we cannot preempt.
+ */
+static __always_inline int preempt_count(void)
+{
+	return current_thread_info()->preempt_count & ~PREEMPT_NEED_RESCHED;
+}
+
+static __always_inline int *preempt_count_ptr(void)
+{
+	return &current_thread_info()->preempt_count;
+}
+
+/*
+ * We now loose PREEMPT_NEED_RESCHED and cause an extra reschedule; however the
+ * alternative is loosing a reschedule. Better schedule too often -- also this
+ * should be a very rare operation.
+ */
+static __always_inline void preempt_count_set(int pc)
+{
+	*preempt_count_ptr() = pc;
+}
+
+/*
+ * We fold the NEED_RESCHED bit into the preempt count such that
+ * preempt_enable() can decrement and test for needing to reschedule with a
+ * single instruction.
+ *
+ * We invert the actual bit, so that when the decrement hits 0 we know we both
+ * need to resched (the bit is cleared) and can resched (no preempt count).
+ */
+
+static __always_inline void set_preempt_need_resched(void)
+{
+	*preempt_count_ptr() &= ~PREEMPT_NEED_RESCHED;
+}
+
+static __always_inline void clear_preempt_need_resched(void)
+{
+	*preempt_count_ptr() |= PREEMPT_NEED_RESCHED;
+}
+
+static __always_inline bool test_preempt_need_resched(void)
+{
+	return !(*preempt_count_ptr() & PREEMPT_NEED_RESCHED);
+}
+
+#endif /* __ASM_PREEMPT_H */
diff --git a/include/linux/preempt.h b/include/linux/preempt.h
index 92e341853e4..df8e245e872 100644
--- a/include/linux/preempt.h
+++ b/include/linux/preempt.h
@@ -6,7 +6,6 @@
  * preempt_count (used for kernel preemption, interrupt count, etc.)
  */
 
-#include <linux/thread_info.h>
 #include <linux/linkage.h>
 #include <linux/list.h>
 
@@ -16,53 +15,7 @@
  */
 #define PREEMPT_NEED_RESCHED	0x80000000
 
-/*
- * We mask the PREEMPT_NEED_RESCHED bit so as not to confuse all current users
- * that think a non-zero value indicates we cannot preempt.
- */
-static __always_inline int preempt_count(void)
-{
-	return current_thread_info()->preempt_count & ~PREEMPT_NEED_RESCHED;
-}
-
-static __always_inline int *preempt_count_ptr(void)
-{
-	return &current_thread_info()->preempt_count;
-}
-
-/*
- * We now loose PREEMPT_NEED_RESCHED and cause an extra reschedule; however the
- * alternative is loosing a reschedule. Better schedule too often -- also this
- * should be a very rare operation.
- */
-static __always_inline void preempt_count_set(int pc)
-{
-	*preempt_count_ptr() = pc;
-}
-
-/*
- * We fold the NEED_RESCHED bit into the preempt count such that
- * preempt_enable() can decrement and test for needing to reschedule with a
- * single instruction.
- *
- * We invert the actual bit, so that when the decrement hits 0 we know we both
- * need to resched (the bit is cleared) and can resched (no preempt count).
- */
-
-static __always_inline void set_preempt_need_resched(void)
-{
-	*preempt_count_ptr() &= ~PREEMPT_NEED_RESCHED;
-}
-
-static __always_inline void clear_preempt_need_resched(void)
-{
-	*preempt_count_ptr() |= PREEMPT_NEED_RESCHED;
-}
-
-static __always_inline bool test_preempt_need_resched(void)
-{
-	return !(*preempt_count_ptr() & PREEMPT_NEED_RESCHED);
-}
+#include <asm/preempt.h>
 
 #if defined(CONFIG_DEBUG_PREEMPT) || defined(CONFIG_PREEMPT_TRACER)
   extern void add_preempt_count(int val);
-- 
cgit v1.2.3-70-g09d2


From bdb43806589096ac4272fe1307e789846ac08d7c Mon Sep 17 00:00:00 2001
From: Peter Zijlstra <peterz@infradead.org>
Date: Tue, 10 Sep 2013 12:15:23 +0200
Subject: sched: Extract the basic add/sub preempt_count modifiers

Rewrite the preempt_count macros in order to extract the 3 basic
preempt_count value modifiers:

  __preempt_count_add()
  __preempt_count_sub()

and the new:

  __preempt_count_dec_and_test()

And since we're at it anyway, replace the unconventional
$op_preempt_count names with the more conventional preempt_count_$op.

Since these basic operators are equivalent to the previous _notrace()
variants, do away with the _notrace() versions.

Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-ewbpdbupy9xpsjhg960zwbv8@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/mips/mm/init.c           |   5 +-
 arch/x86/kernel/traps.c       |   4 +-
 include/asm-generic/preempt.h |  35 ++++++++++++++
 include/linux/hardirq.h       |   8 ++--
 include/linux/preempt.h       | 106 +++++++++++++++++++-----------------------
 include/linux/sched.h         |   5 --
 include/linux/uaccess.h       |   8 +---
 kernel/context_tracking.c     |   2 +-
 kernel/sched/core.c           |  29 +++++-------
 kernel/softirq.c              |  14 +++---
 10 files changed, 113 insertions(+), 103 deletions(-)

(limited to 'arch/x86')

diff --git a/arch/mips/mm/init.c b/arch/mips/mm/init.c
index e205ef598e9..12156176c7c 100644
--- a/arch/mips/mm/init.c
+++ b/arch/mips/mm/init.c
@@ -124,7 +124,7 @@ void *kmap_coherent(struct page *page, unsigned long addr)
 
 	BUG_ON(Page_dcache_dirty(page));
 
-	inc_preempt_count();
+	pagefault_disable();
 	idx = (addr >> PAGE_SHIFT) & (FIX_N_COLOURS - 1);
 #ifdef CONFIG_MIPS_MT_SMTC
 	idx += FIX_N_COLOURS * smp_processor_id() +
@@ -193,8 +193,7 @@ void kunmap_coherent(void)
 	write_c0_entryhi(old_ctx);
 	EXIT_CRITICAL(flags);
 #endif
-	dec_preempt_count();
-	preempt_check_resched();
+	pagefault_enable();
 }
 
 void copy_user_highpage(struct page *to, struct page *from,
diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
index 8c8093b146c..729aa779ff7 100644
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -88,7 +88,7 @@ static inline void conditional_sti(struct pt_regs *regs)
 
 static inline void preempt_conditional_sti(struct pt_regs *regs)
 {
-	inc_preempt_count();
+	preempt_count_inc();
 	if (regs->flags & X86_EFLAGS_IF)
 		local_irq_enable();
 }
@@ -103,7 +103,7 @@ static inline void preempt_conditional_cli(struct pt_regs *regs)
 {
 	if (regs->flags & X86_EFLAGS_IF)
 		local_irq_disable();
-	dec_preempt_count();
+	preempt_count_dec();
 }
 
 static int __kprobes
diff --git a/include/asm-generic/preempt.h b/include/asm-generic/preempt.h
index 8100b1ec171..82d958fc382 100644
--- a/include/asm-generic/preempt.h
+++ b/include/asm-generic/preempt.h
@@ -65,4 +65,39 @@ static __always_inline bool test_preempt_need_resched(void)
 	return !(*preempt_count_ptr() & PREEMPT_NEED_RESCHED);
 }
 
+/*
+ * The various preempt_count add/sub methods
+ */
+
+static __always_inline void __preempt_count_add(int val)
+{
+	*preempt_count_ptr() += val;
+}
+
+static __always_inline void __preempt_count_sub(int val)
+{
+	*preempt_count_ptr() -= val;
+}
+
+static __always_inline bool __preempt_count_dec_and_test(void)
+{
+	return !--*preempt_count_ptr();
+}
+
+/*
+ * Returns true when we need to resched -- even if we can not.
+ */
+static __always_inline bool need_resched(void)
+{
+	return unlikely(test_preempt_need_resched());
+}
+
+/*
+ * Returns true when we need to resched and can (barring IRQ state).
+ */
+static __always_inline bool should_resched(void)
+{
+	return unlikely(!*preempt_count_ptr());
+}
+
 #endif /* __ASM_PREEMPT_H */
diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h
index 1e041063b22..d9cf963ac83 100644
--- a/include/linux/hardirq.h
+++ b/include/linux/hardirq.h
@@ -33,7 +33,7 @@ extern void rcu_nmi_exit(void);
 #define __irq_enter()					\
 	do {						\
 		account_irq_enter_time(current);	\
-		add_preempt_count(HARDIRQ_OFFSET);	\
+		preempt_count_add(HARDIRQ_OFFSET);	\
 		trace_hardirq_enter();			\
 	} while (0)
 
@@ -49,7 +49,7 @@ extern void irq_enter(void);
 	do {						\
 		trace_hardirq_exit();			\
 		account_irq_exit_time(current);		\
-		sub_preempt_count(HARDIRQ_OFFSET);	\
+		preempt_count_sub(HARDIRQ_OFFSET);	\
 	} while (0)
 
 /*
@@ -62,7 +62,7 @@ extern void irq_exit(void);
 		lockdep_off();					\
 		ftrace_nmi_enter();				\
 		BUG_ON(in_nmi());				\
-		add_preempt_count(NMI_OFFSET + HARDIRQ_OFFSET);	\
+		preempt_count_add(NMI_OFFSET + HARDIRQ_OFFSET);	\
 		rcu_nmi_enter();				\
 		trace_hardirq_enter();				\
 	} while (0)
@@ -72,7 +72,7 @@ extern void irq_exit(void);
 		trace_hardirq_exit();				\
 		rcu_nmi_exit();					\
 		BUG_ON(!in_nmi());				\
-		sub_preempt_count(NMI_OFFSET + HARDIRQ_OFFSET);	\
+		preempt_count_sub(NMI_OFFSET + HARDIRQ_OFFSET);	\
 		ftrace_nmi_exit();				\
 		lockdep_on();					\
 	} while (0)
diff --git a/include/linux/preempt.h b/include/linux/preempt.h
index df8e245e872..2343d871529 100644
--- a/include/linux/preempt.h
+++ b/include/linux/preempt.h
@@ -18,97 +18,86 @@
 #include <asm/preempt.h>
 
 #if defined(CONFIG_DEBUG_PREEMPT) || defined(CONFIG_PREEMPT_TRACER)
-  extern void add_preempt_count(int val);
-  extern void sub_preempt_count(int val);
+extern void preempt_count_add(int val);
+extern void preempt_count_sub(int val);
+#define preempt_count_dec_and_test() ({ preempt_count_sub(1); should_resched(); })
 #else
-# define add_preempt_count(val)	do { *preempt_count_ptr() += (val); } while (0)
-# define sub_preempt_count(val)	do { *preempt_count_ptr() -= (val); } while (0)
+#define preempt_count_add(val)	__preempt_count_add(val)
+#define preempt_count_sub(val)	__preempt_count_sub(val)
+#define preempt_count_dec_and_test() __preempt_count_dec_and_test()
 #endif
 
-#define inc_preempt_count() add_preempt_count(1)
-#define dec_preempt_count() sub_preempt_count(1)
-
-#ifdef CONFIG_PREEMPT
-
-asmlinkage void preempt_schedule(void);
-
-#define preempt_check_resched() \
-do { \
-	if (unlikely(!*preempt_count_ptr())) \
-		preempt_schedule(); \
-} while (0)
-
-#ifdef CONFIG_CONTEXT_TRACKING
-
-void preempt_schedule_context(void);
-
-#define preempt_check_resched_context() \
-do { \
-	if (unlikely(!*preempt_count_ptr())) \
-		preempt_schedule_context(); \
-} while (0)
-#else
-
-#define preempt_check_resched_context() preempt_check_resched()
-
-#endif /* CONFIG_CONTEXT_TRACKING */
-
-#else /* !CONFIG_PREEMPT */
-
-#define preempt_check_resched()		do { } while (0)
-#define preempt_check_resched_context()	do { } while (0)
-
-#endif /* CONFIG_PREEMPT */
+#define __preempt_count_inc() __preempt_count_add(1)
+#define __preempt_count_dec() __preempt_count_sub(1)
 
+#define preempt_count_inc() preempt_count_add(1)
+#define preempt_count_dec() preempt_count_sub(1)
 
 #ifdef CONFIG_PREEMPT_COUNT
 
 #define preempt_disable() \
 do { \
-	inc_preempt_count(); \
+	preempt_count_inc(); \
 	barrier(); \
 } while (0)
 
 #define sched_preempt_enable_no_resched() \
 do { \
 	barrier(); \
-	dec_preempt_count(); \
+	preempt_count_dec(); \
 } while (0)
 
-#define preempt_enable_no_resched()	sched_preempt_enable_no_resched()
+#define preempt_enable_no_resched() sched_preempt_enable_no_resched()
 
+#ifdef CONFIG_PREEMPT
+asmlinkage void preempt_schedule(void);
 #define preempt_enable() \
 do { \
-	preempt_enable_no_resched(); \
-	preempt_check_resched(); \
+	barrier(); \
+	if (unlikely(preempt_count_dec_and_test())) \
+		preempt_schedule(); \
 } while (0)
 
-/* For debugging and tracer internals only! */
-#define add_preempt_count_notrace(val)			\
-	do { *preempt_count_ptr() += (val); } while (0)
-#define sub_preempt_count_notrace(val)			\
-	do { *preempt_count_ptr() -= (val); } while (0)
-#define inc_preempt_count_notrace() add_preempt_count_notrace(1)
-#define dec_preempt_count_notrace() sub_preempt_count_notrace(1)
+#define preempt_check_resched() \
+do { \
+	if (should_resched()) \
+		preempt_schedule(); \
+} while (0)
+
+#else
+#define preempt_enable() preempt_enable_no_resched()
+#define preempt_check_resched() do { } while (0)
+#endif
 
 #define preempt_disable_notrace() \
 do { \
-	inc_preempt_count_notrace(); \
+	__preempt_count_inc(); \
 	barrier(); \
 } while (0)
 
 #define preempt_enable_no_resched_notrace() \
 do { \
 	barrier(); \
-	dec_preempt_count_notrace(); \
+	__preempt_count_dec(); \
 } while (0)
 
-/* preempt_check_resched is OK to trace */
+#ifdef CONFIG_PREEMPT
+
+#ifdef CONFIG_CONTEXT_TRACKING
+asmlinkage void preempt_schedule_context(void);
+#else
+#define preempt_schedule_context() preempt_schedule()
+#endif
+
 #define preempt_enable_notrace() \
 do { \
-	preempt_enable_no_resched_notrace(); \
-	preempt_check_resched_context(); \
+	barrier(); \
+	if (unlikely(__preempt_count_dec_and_test())) \
+		preempt_schedule_context(); \
 } while (0)
+#else
+#define preempt_enable_notrace() preempt_enable_no_resched_notrace()
+#endif
 
 #else /* !CONFIG_PREEMPT_COUNT */
 
@@ -118,10 +107,11 @@ do { \
  * that can cause faults and scheduling migrate into our preempt-protected
  * region.
  */
-#define preempt_disable()		barrier()
+#define preempt_disable()			barrier()
 #define sched_preempt_enable_no_resched()	barrier()
-#define preempt_enable_no_resched()	barrier()
-#define preempt_enable()		barrier()
+#define preempt_enable_no_resched()		barrier()
+#define preempt_enable()			barrier()
+#define preempt_check_resched()			do { } while (0)
 
 #define preempt_disable_notrace()		barrier()
 #define preempt_enable_no_resched_notrace()	barrier()
diff --git a/include/linux/sched.h b/include/linux/sched.h
index 9fa151fb968..06ac17c7e63 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -2409,11 +2409,6 @@ static inline int signal_pending_state(long state, struct task_struct *p)
 	return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
 }
 
-static inline int need_resched(void)
-{
-	return unlikely(test_preempt_need_resched());
-}
-
 /*
  * cond_resched() and cond_resched_lock(): latency reduction via
  * explicit rescheduling in places that are safe. The return
diff --git a/include/linux/uaccess.h b/include/linux/uaccess.h
index 5ca0951e185..9d8cf056e66 100644
--- a/include/linux/uaccess.h
+++ b/include/linux/uaccess.h
@@ -15,7 +15,7 @@
  */
 static inline void pagefault_disable(void)
 {
-	inc_preempt_count();
+	preempt_count_inc();
 	/*
 	 * make sure to have issued the store before a pagefault
 	 * can hit.
@@ -30,11 +30,7 @@ static inline void pagefault_enable(void)
 	 * the pagefault handler again.
 	 */
 	barrier();
-	dec_preempt_count();
-	/*
-	 * make sure we do..
-	 */
-	barrier();
+	preempt_count_dec();
 	preempt_check_resched();
 }
 
diff --git a/kernel/context_tracking.c b/kernel/context_tracking.c
index 247091bf058..013161f1c80 100644
--- a/kernel/context_tracking.c
+++ b/kernel/context_tracking.c
@@ -111,7 +111,7 @@ void context_tracking_user_enter(void)
  * instead of preempt_schedule() to exit user context if needed before
  * calling the scheduler.
  */
-void __sched notrace preempt_schedule_context(void)
+asmlinkage void __sched notrace preempt_schedule_context(void)
 {
 	enum ctx_state prev_ctx;
 
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 0ba4e419239..9c84a9ab189 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -2219,7 +2219,7 @@ notrace unsigned long get_parent_ip(unsigned long addr)
 #if defined(CONFIG_PREEMPT) && (defined(CONFIG_DEBUG_PREEMPT) || \
 				defined(CONFIG_PREEMPT_TRACER))
 
-void __kprobes add_preempt_count(int val)
+void __kprobes preempt_count_add(int val)
 {
 #ifdef CONFIG_DEBUG_PREEMPT
 	/*
@@ -2228,7 +2228,7 @@ void __kprobes add_preempt_count(int val)
 	if (DEBUG_LOCKS_WARN_ON((preempt_count() < 0)))
 		return;
 #endif
-	add_preempt_count_notrace(val);
+	__preempt_count_add(val);
 #ifdef CONFIG_DEBUG_PREEMPT
 	/*
 	 * Spinlock count overflowing soon?
@@ -2239,9 +2239,9 @@ void __kprobes add_preempt_count(int val)
 	if (preempt_count() == val)
 		trace_preempt_off(CALLER_ADDR0, get_parent_ip(CALLER_ADDR1));
 }
-EXPORT_SYMBOL(add_preempt_count);
+EXPORT_SYMBOL(preempt_count_add);
 
-void __kprobes sub_preempt_count(int val)
+void __kprobes preempt_count_sub(int val)
 {
 #ifdef CONFIG_DEBUG_PREEMPT
 	/*
@@ -2259,9 +2259,9 @@ void __kprobes sub_preempt_count(int val)
 
 	if (preempt_count() == val)
 		trace_preempt_on(CALLER_ADDR0, get_parent_ip(CALLER_ADDR1));
-	sub_preempt_count_notrace(val);
+	__preempt_count_sub(val);
 }
-EXPORT_SYMBOL(sub_preempt_count);
+EXPORT_SYMBOL(preempt_count_sub);
 
 #endif
 
@@ -2525,9 +2525,9 @@ asmlinkage void __sched notrace preempt_schedule(void)
 		return;
 
 	do {
-		add_preempt_count_notrace(PREEMPT_ACTIVE);
+		__preempt_count_add(PREEMPT_ACTIVE);
 		__schedule();
-		sub_preempt_count_notrace(PREEMPT_ACTIVE);
+		__preempt_count_sub(PREEMPT_ACTIVE);
 
 		/*
 		 * Check again in case we missed a preemption opportunity
@@ -2554,11 +2554,11 @@ asmlinkage void __sched preempt_schedule_irq(void)
 	prev_state = exception_enter();
 
 	do {
-		add_preempt_count(PREEMPT_ACTIVE);
+		__preempt_count_add(PREEMPT_ACTIVE);
 		local_irq_enable();
 		__schedule();
 		local_irq_disable();
-		sub_preempt_count(PREEMPT_ACTIVE);
+		__preempt_count_sub(PREEMPT_ACTIVE);
 
 		/*
 		 * Check again in case we missed a preemption opportunity
@@ -3798,16 +3798,11 @@ SYSCALL_DEFINE0(sched_yield)
 	return 0;
 }
 
-static inline int should_resched(void)
-{
-	return need_resched() && !(preempt_count() & PREEMPT_ACTIVE);
-}
-
 static void __cond_resched(void)
 {
-	add_preempt_count(PREEMPT_ACTIVE);
+	__preempt_count_add(PREEMPT_ACTIVE);
 	__schedule();
-	sub_preempt_count(PREEMPT_ACTIVE);
+	__preempt_count_sub(PREEMPT_ACTIVE);
 }
 
 int __sched _cond_resched(void)
diff --git a/kernel/softirq.c b/kernel/softirq.c
index a90de70cf1f..3e88612fc87 100644
--- a/kernel/softirq.c
+++ b/kernel/softirq.c
@@ -100,13 +100,13 @@ static void __local_bh_disable(unsigned long ip, unsigned int cnt)
 
 	raw_local_irq_save(flags);
 	/*
-	 * The preempt tracer hooks into add_preempt_count and will break
+	 * The preempt tracer hooks into preempt_count_add and will break
 	 * lockdep because it calls back into lockdep after SOFTIRQ_OFFSET
 	 * is set and before current->softirq_enabled is cleared.
 	 * We must manually increment preempt_count here and manually
 	 * call the trace_preempt_off later.
 	 */
-	add_preempt_count_notrace(cnt);
+	__preempt_count_add(cnt);
 	/*
 	 * Were softirqs turned off above:
 	 */
@@ -120,7 +120,7 @@ static void __local_bh_disable(unsigned long ip, unsigned int cnt)
 #else /* !CONFIG_TRACE_IRQFLAGS */
 static inline void __local_bh_disable(unsigned long ip, unsigned int cnt)
 {
-	add_preempt_count(cnt);
+	preempt_count_add(cnt);
 	barrier();
 }
 #endif /* CONFIG_TRACE_IRQFLAGS */
@@ -139,7 +139,7 @@ static void __local_bh_enable(unsigned int cnt)
 
 	if (softirq_count() == cnt)
 		trace_softirqs_on(_RET_IP_);
-	sub_preempt_count(cnt);
+	preempt_count_sub(cnt);
 }
 
 /*
@@ -169,12 +169,12 @@ static inline void _local_bh_enable_ip(unsigned long ip)
 	 * Keep preemption disabled until we are done with
 	 * softirq processing:
  	 */
-	sub_preempt_count(SOFTIRQ_DISABLE_OFFSET - 1);
+	preempt_count_sub(SOFTIRQ_DISABLE_OFFSET - 1);
 
 	if (unlikely(!in_interrupt() && local_softirq_pending()))
 		do_softirq();
 
-	dec_preempt_count();
+	preempt_count_dec();
 #ifdef CONFIG_TRACE_IRQFLAGS
 	local_irq_enable();
 #endif
@@ -360,7 +360,7 @@ void irq_exit(void)
 
 	account_irq_exit_time(current);
 	trace_hardirq_exit();
-	sub_preempt_count(HARDIRQ_OFFSET);
+	preempt_count_sub(HARDIRQ_OFFSET);
 	if (!in_interrupt() && local_softirq_pending())
 		invoke_softirq();
 
-- 
cgit v1.2.3-70-g09d2


From c2daa3bed53a81171cf8c1a36db798e82b91afe8 Mon Sep 17 00:00:00 2001
From: Peter Zijlstra <peterz@infradead.org>
Date: Wed, 14 Aug 2013 14:51:00 +0200
Subject: sched, x86: Provide a per-cpu preempt_count implementation

Convert x86 to use a per-cpu preemption count. The reason for doing so
is that accessing per-cpu variables is a lot cheaper than accessing
thread_info variables.

We still need to save/restore the actual preemption count due to
PREEMPT_ACTIVE so we place the per-cpu __preempt_count variable in the
same cache-line as the other hot __switch_to() variables such as
current_task.

NOTE: this save/restore is required even for !PREEMPT kernels as
cond_resched() also relies on preempt_count's PREEMPT_ACTIVE to ignore
task_struct::state.

Also rename thread_info::preempt_count to ensure nobody is
'accidentally' still poking at it.

Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-gzn5rfsf8trgjoqx8hyayy3q@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/include/asm/Kbuild        |  1 -
 arch/x86/include/asm/preempt.h     | 98 ++++++++++++++++++++++++++++++++++++++
 arch/x86/include/asm/thread_info.h |  5 +-
 arch/x86/kernel/asm-offsets.c      |  1 -
 arch/x86/kernel/cpu/common.c       |  5 ++
 arch/x86/kernel/entry_32.S         |  7 +--
 arch/x86/kernel/entry_64.S         |  4 +-
 arch/x86/kernel/irq_32.c           |  4 --
 arch/x86/kernel/process_32.c       |  8 ++++
 arch/x86/kernel/process_64.c       |  8 ++++
 10 files changed, 124 insertions(+), 17 deletions(-)
 create mode 100644 arch/x86/include/asm/preempt.h

(limited to 'arch/x86')

diff --git a/arch/x86/include/asm/Kbuild b/arch/x86/include/asm/Kbuild
index eca20286a91..7f669853317 100644
--- a/arch/x86/include/asm/Kbuild
+++ b/arch/x86/include/asm/Kbuild
@@ -5,4 +5,3 @@ genhdr-y += unistd_64.h
 genhdr-y += unistd_x32.h
 
 generic-y += clkdev.h
-generic-y += preempt.h
diff --git a/arch/x86/include/asm/preempt.h b/arch/x86/include/asm/preempt.h
new file mode 100644
index 00000000000..1309942b95e
--- /dev/null
+++ b/arch/x86/include/asm/preempt.h
@@ -0,0 +1,98 @@
+#ifndef __ASM_PREEMPT_H
+#define __ASM_PREEMPT_H
+
+#include <asm/rmwcc.h>
+#include <asm/percpu.h>
+#include <linux/thread_info.h>
+
+DECLARE_PER_CPU(int, __preempt_count);
+
+/*
+ * We mask the PREEMPT_NEED_RESCHED bit so as not to confuse all current users
+ * that think a non-zero value indicates we cannot preempt.
+ */
+static __always_inline int preempt_count(void)
+{
+	return __this_cpu_read_4(__preempt_count) & ~PREEMPT_NEED_RESCHED;
+}
+
+static __always_inline void preempt_count_set(int pc)
+{
+	__this_cpu_write_4(__preempt_count, pc);
+}
+
+/*
+ * must be macros to avoid header recursion hell
+ */
+#define task_preempt_count(p) \
+	(task_thread_info(p)->saved_preempt_count & ~PREEMPT_NEED_RESCHED)
+
+#define init_task_preempt_count(p) do { \
+	task_thread_info(p)->saved_preempt_count = PREEMPT_DISABLED; \
+} while (0)
+
+#define init_idle_preempt_count(p, cpu) do { \
+	task_thread_info(p)->saved_preempt_count = PREEMPT_ENABLED; \
+	per_cpu(__preempt_count, (cpu)) = PREEMPT_ENABLED; \
+} while (0)
+
+/*
+ * We fold the NEED_RESCHED bit into the preempt count such that
+ * preempt_enable() can decrement and test for needing to reschedule with a
+ * single instruction.
+ *
+ * We invert the actual bit, so that when the decrement hits 0 we know we both
+ * need to resched (the bit is cleared) and can resched (no preempt count).
+ */
+
+static __always_inline void set_preempt_need_resched(void)
+{
+	__this_cpu_and_4(__preempt_count, ~PREEMPT_NEED_RESCHED);
+}
+
+static __always_inline void clear_preempt_need_resched(void)
+{
+	__this_cpu_or_4(__preempt_count, PREEMPT_NEED_RESCHED);
+}
+
+static __always_inline bool test_preempt_need_resched(void)
+{
+	return !(__this_cpu_read_4(__preempt_count) & PREEMPT_NEED_RESCHED);
+}
+
+/*
+ * The various preempt_count add/sub methods
+ */
+
+static __always_inline void __preempt_count_add(int val)
+{
+	__this_cpu_add_4(__preempt_count, val);
+}
+
+static __always_inline void __preempt_count_sub(int val)
+{
+	__this_cpu_add_4(__preempt_count, -val);
+}
+
+static __always_inline bool __preempt_count_dec_and_test(void)
+{
+	GEN_UNARY_RMWcc("decl", __preempt_count, __percpu_arg(0), "e");
+}
+
+/*
+ * Returns true when we need to resched -- even if we can not.
+ */
+static __always_inline bool need_resched(void)
+{
+	return unlikely(test_preempt_need_resched());
+}
+
+/*
+ * Returns true when we need to resched and can (barring IRQ state).
+ */
+static __always_inline bool should_resched(void)
+{
+	return unlikely(!__this_cpu_read_4(__preempt_count));
+}
+
+#endif /* __ASM_PREEMPT_H */
diff --git a/arch/x86/include/asm/thread_info.h b/arch/x86/include/asm/thread_info.h
index 27811190cbd..c46a46be1ec 100644
--- a/arch/x86/include/asm/thread_info.h
+++ b/arch/x86/include/asm/thread_info.h
@@ -28,8 +28,7 @@ struct thread_info {
 	__u32			flags;		/* low level flags */
 	__u32			status;		/* thread synchronous flags */
 	__u32			cpu;		/* current CPU */
-	int			preempt_count;	/* 0 => preemptable,
-						   <0 => BUG */
+	int			saved_preempt_count;
 	mm_segment_t		addr_limit;
 	struct restart_block    restart_block;
 	void __user		*sysenter_return;
@@ -49,7 +48,7 @@ struct thread_info {
 	.exec_domain	= &default_exec_domain,	\
 	.flags		= 0,			\
 	.cpu		= 0,			\
-	.preempt_count	= INIT_PREEMPT_COUNT,	\
+	.saved_preempt_count = INIT_PREEMPT_COUNT,	\
 	.addr_limit	= KERNEL_DS,		\
 	.restart_block = {			\
 		.fn = do_no_restart_syscall,	\
diff --git a/arch/x86/kernel/asm-offsets.c b/arch/x86/kernel/asm-offsets.c
index 28610822fb3..9f6b9341950 100644
--- a/arch/x86/kernel/asm-offsets.c
+++ b/arch/x86/kernel/asm-offsets.c
@@ -32,7 +32,6 @@ void common(void) {
 	OFFSET(TI_flags, thread_info, flags);
 	OFFSET(TI_status, thread_info, status);
 	OFFSET(TI_addr_limit, thread_info, addr_limit);
-	OFFSET(TI_preempt_count, thread_info, preempt_count);
 
 	BLANK();
 	OFFSET(crypto_tfm_ctx_offset, crypto_tfm, __crt_ctx);
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index 2793d1f095a..5223fe6dec7 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -1095,6 +1095,9 @@ DEFINE_PER_CPU(char *, irq_stack_ptr) =
 
 DEFINE_PER_CPU(unsigned int, irq_count) __visible = -1;
 
+DEFINE_PER_CPU(int, __preempt_count) = INIT_PREEMPT_COUNT;
+EXPORT_PER_CPU_SYMBOL(__preempt_count);
+
 DEFINE_PER_CPU(struct task_struct *, fpu_owner_task);
 
 /*
@@ -1169,6 +1172,8 @@ void debug_stack_reset(void)
 
 DEFINE_PER_CPU(struct task_struct *, current_task) = &init_task;
 EXPORT_PER_CPU_SYMBOL(current_task);
+DEFINE_PER_CPU(int, __preempt_count) = INIT_PREEMPT_COUNT;
+EXPORT_PER_CPU_SYMBOL(__preempt_count);
 DEFINE_PER_CPU(struct task_struct *, fpu_owner_task);
 
 #ifdef CONFIG_CC_STACKPROTECTOR
diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S
index f0dcb0ceb6a..fd1bc1b15e6 100644
--- a/arch/x86/kernel/entry_32.S
+++ b/arch/x86/kernel/entry_32.S
@@ -362,12 +362,9 @@ END(ret_from_exception)
 #ifdef CONFIG_PREEMPT
 ENTRY(resume_kernel)
 	DISABLE_INTERRUPTS(CLBR_ANY)
-	cmpl $0,TI_preempt_count(%ebp)	# non-zero preempt_count ?
-	jnz restore_all
 need_resched:
-	movl TI_flags(%ebp), %ecx	# need_resched set ?
-	testb $_TIF_NEED_RESCHED, %cl
-	jz restore_all
+	cmpl $0,PER_CPU_VAR(__preempt_count)
+	jnz restore_all
 	testl $X86_EFLAGS_IF,PT_EFLAGS(%esp)	# interrupts off (exception path) ?
 	jz restore_all
 	call preempt_schedule_irq
diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S
index 1b69951a81e..6a43e7d29fe 100644
--- a/arch/x86/kernel/entry_64.S
+++ b/arch/x86/kernel/entry_64.S
@@ -1118,10 +1118,8 @@ retint_signal:
 	/* Returning to kernel space. Check if we need preemption */
 	/* rcx:	 threadinfo. interrupts off. */
 ENTRY(retint_kernel)
-	cmpl $0,TI_preempt_count(%rcx)
+	cmpl $0,PER_CPU_VAR(__preempt_count)
 	jnz  retint_restore_args
-	bt  $TIF_NEED_RESCHED,TI_flags(%rcx)
-	jnc  retint_restore_args
 	bt   $9,EFLAGS-ARGOFFSET(%rsp)	/* interrupts off? */
 	jnc  retint_restore_args
 	call preempt_schedule_irq
diff --git a/arch/x86/kernel/irq_32.c b/arch/x86/kernel/irq_32.c
index 4186755f1d7..3fe066359ac 100644
--- a/arch/x86/kernel/irq_32.c
+++ b/arch/x86/kernel/irq_32.c
@@ -100,9 +100,6 @@ execute_on_irq_stack(int overflow, struct irq_desc *desc, int irq)
 	irqctx->tinfo.task = curctx->tinfo.task;
 	irqctx->tinfo.previous_esp = current_stack_pointer;
 
-	/* Copy the preempt_count so that the [soft]irq checks work. */
-	irqctx->tinfo.preempt_count = curctx->tinfo.preempt_count;
-
 	if (unlikely(overflow))
 		call_on_stack(print_stack_overflow, isp);
 
@@ -131,7 +128,6 @@ void irq_ctx_init(int cpu)
 					       THREAD_SIZE_ORDER));
 	memset(&irqctx->tinfo, 0, sizeof(struct thread_info));
 	irqctx->tinfo.cpu		= cpu;
-	irqctx->tinfo.preempt_count	= HARDIRQ_OFFSET;
 	irqctx->tinfo.addr_limit	= MAKE_MM_SEG(0);
 
 	per_cpu(hardirq_ctx, cpu) = irqctx;
diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c
index 884f98f6935..c2ec1aa6d45 100644
--- a/arch/x86/kernel/process_32.c
+++ b/arch/x86/kernel/process_32.c
@@ -291,6 +291,14 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
 	if (get_kernel_rpl() && unlikely(prev->iopl != next->iopl))
 		set_iopl_mask(next->iopl);
 
+	/*
+	 * If it were not for PREEMPT_ACTIVE we could guarantee that the
+	 * preempt_count of all tasks was equal here and this would not be
+	 * needed.
+	 */
+	task_thread_info(prev_p)->saved_preempt_count = this_cpu_read(__preempt_count);
+	this_cpu_write(__preempt_count, task_thread_info(next_p)->saved_preempt_count);
+
 	/*
 	 * Now maybe handle debug registers and/or IO bitmaps
 	 */
diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
index bb1dc51bab0..45ab4d6fc8a 100644
--- a/arch/x86/kernel/process_64.c
+++ b/arch/x86/kernel/process_64.c
@@ -363,6 +363,14 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
 	this_cpu_write(old_rsp, next->usersp);
 	this_cpu_write(current_task, next_p);
 
+	/*
+	 * If it were not for PREEMPT_ACTIVE we could guarantee that the
+	 * preempt_count of all tasks was equal here and this would not be
+	 * needed.
+	 */
+	task_thread_info(prev_p)->saved_preempt_count = this_cpu_read(__preempt_count);
+	this_cpu_write(__preempt_count, task_thread_info(next_p)->saved_preempt_count);
+
 	this_cpu_write(kernel_stack,
 		  (unsigned long)task_stack_page(next_p) +
 		  THREAD_SIZE - KERNEL_STACK_OFFSET);
-- 
cgit v1.2.3-70-g09d2


From 1a338ac32ca630f67df25b4a16436cccc314e997 Mon Sep 17 00:00:00 2001
From: Peter Zijlstra <peterz@infradead.org>
Date: Wed, 14 Aug 2013 14:51:00 +0200
Subject: sched, x86: Optimize the preempt_schedule() call

Remove the bloat of the C calling convention out of the
preempt_enable() sites by creating an ASM wrapper which allows us to
do an asm("call ___preempt_schedule") instead.

calling.h bits by Andi Kleen

Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-tk7xdi1cvvxewixzke8t8le1@git.kernel.org
[ Fixed build error. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/include/asm/calling.h   | 50 ++++++++++++++++++++++++++++++++++++++++
 arch/x86/include/asm/preempt.h   | 10 ++++++++
 arch/x86/kernel/Makefile         |  2 ++
 arch/x86/kernel/i386_ksyms_32.c  |  7 ++++++
 arch/x86/kernel/preempt.S        | 25 ++++++++++++++++++++
 arch/x86/kernel/x8664_ksyms_64.c |  7 ++++++
 include/asm-generic/preempt.h    | 10 ++++++++
 include/linux/preempt.h          | 13 ++++-------
 8 files changed, 116 insertions(+), 8 deletions(-)
 create mode 100644 arch/x86/kernel/preempt.S

(limited to 'arch/x86')

diff --git a/arch/x86/include/asm/calling.h b/arch/x86/include/asm/calling.h
index 0fa67503391..cb4c73bfeb4 100644
--- a/arch/x86/include/asm/calling.h
+++ b/arch/x86/include/asm/calling.h
@@ -48,6 +48,8 @@ For 32-bit we have the following conventions - kernel is built with
 
 #include <asm/dwarf2.h>
 
+#ifdef CONFIG_X86_64
+
 /*
  * 64-bit system call stack frame layout defines and helpers,
  * for assembly code:
@@ -192,3 +194,51 @@ For 32-bit we have the following conventions - kernel is built with
 	.macro icebp
 	.byte 0xf1
 	.endm
+
+#else /* CONFIG_X86_64 */
+
+/*
+ * For 32bit only simplified versions of SAVE_ALL/RESTORE_ALL. These
+ * are different from the entry_32.S versions in not changing the segment
+ * registers. So only suitable for in kernel use, not when transitioning
+ * from or to user space. The resulting stack frame is not a standard
+ * pt_regs frame. The main use case is calling C code from assembler
+ * when all the registers need to be preserved.
+ */
+
+	.macro SAVE_ALL
+	pushl_cfi %eax
+	CFI_REL_OFFSET eax, 0
+	pushl_cfi %ebp
+	CFI_REL_OFFSET ebp, 0
+	pushl_cfi %edi
+	CFI_REL_OFFSET edi, 0
+	pushl_cfi %esi
+	CFI_REL_OFFSET esi, 0
+	pushl_cfi %edx
+	CFI_REL_OFFSET edx, 0
+	pushl_cfi %ecx
+	CFI_REL_OFFSET ecx, 0
+	pushl_cfi %ebx
+	CFI_REL_OFFSET ebx, 0
+	.endm
+
+	.macro RESTORE_ALL
+	popl_cfi %ebx
+	CFI_RESTORE ebx
+	popl_cfi %ecx
+	CFI_RESTORE ecx
+	popl_cfi %edx
+	CFI_RESTORE edx
+	popl_cfi %esi
+	CFI_RESTORE esi
+	popl_cfi %edi
+	CFI_RESTORE edi
+	popl_cfi %ebp
+	CFI_RESTORE ebp
+	popl_cfi %eax
+	CFI_RESTORE eax
+	.endm
+
+#endif /* CONFIG_X86_64 */
+
diff --git a/arch/x86/include/asm/preempt.h b/arch/x86/include/asm/preempt.h
index 1309942b95e..1de41690ff9 100644
--- a/arch/x86/include/asm/preempt.h
+++ b/arch/x86/include/asm/preempt.h
@@ -95,4 +95,14 @@ static __always_inline bool should_resched(void)
 	return unlikely(!__this_cpu_read_4(__preempt_count));
 }
 
+#ifdef CONFIG_PREEMPT
+  extern asmlinkage void ___preempt_schedule(void);
+# define __preempt_schedule() asm ("call ___preempt_schedule")
+  extern asmlinkage void preempt_schedule(void);
+# ifdef CONFIG_CONTEXT_TRACKING
+    extern asmlinkage void ___preempt_schedule_context(void);
+#   define __preempt_schedule_context() asm ("call ___preempt_schedule_context")
+# endif
+#endif
+
 #endif /* __ASM_PREEMPT_H */
diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile
index a5408b965c9..9b0a34e2cd7 100644
--- a/arch/x86/kernel/Makefile
+++ b/arch/x86/kernel/Makefile
@@ -36,6 +36,8 @@ obj-y			+= tsc.o io_delay.o rtc.o
 obj-y			+= pci-iommu_table.o
 obj-y			+= resource.o
 
+obj-$(CONFIG_PREEMPT)	+= preempt.o
+
 obj-y				+= process.o
 obj-y				+= i387.o xsave.o
 obj-y				+= ptrace.o
diff --git a/arch/x86/kernel/i386_ksyms_32.c b/arch/x86/kernel/i386_ksyms_32.c
index 0fa69127209..05fd74f537d 100644
--- a/arch/x86/kernel/i386_ksyms_32.c
+++ b/arch/x86/kernel/i386_ksyms_32.c
@@ -37,3 +37,10 @@ EXPORT_SYMBOL(strstr);
 
 EXPORT_SYMBOL(csum_partial);
 EXPORT_SYMBOL(empty_zero_page);
+
+#ifdef CONFIG_PREEMPT
+EXPORT_SYMBOL(___preempt_schedule);
+#ifdef CONFIG_CONTEXT_TRACKING
+EXPORT_SYMBOL(___preempt_schedule_context);
+#endif
+#endif
diff --git a/arch/x86/kernel/preempt.S b/arch/x86/kernel/preempt.S
new file mode 100644
index 00000000000..ca7f0d58a87
--- /dev/null
+++ b/arch/x86/kernel/preempt.S
@@ -0,0 +1,25 @@
+
+#include <linux/linkage.h>
+#include <asm/dwarf2.h>
+#include <asm/asm.h>
+#include <asm/calling.h>
+
+ENTRY(___preempt_schedule)
+	CFI_STARTPROC
+	SAVE_ALL
+	call preempt_schedule
+	RESTORE_ALL
+	ret
+	CFI_ENDPROC
+
+#ifdef CONFIG_CONTEXT_TRACKING
+
+ENTRY(___preempt_schedule_context)
+	CFI_STARTPROC
+	SAVE_ALL
+	call preempt_schedule_context
+	RESTORE_ALL
+	ret
+	CFI_ENDPROC
+
+#endif
diff --git a/arch/x86/kernel/x8664_ksyms_64.c b/arch/x86/kernel/x8664_ksyms_64.c
index b014d9414d0..040681928e9 100644
--- a/arch/x86/kernel/x8664_ksyms_64.c
+++ b/arch/x86/kernel/x8664_ksyms_64.c
@@ -66,3 +66,10 @@ EXPORT_SYMBOL(empty_zero_page);
 #ifndef CONFIG_PARAVIRT
 EXPORT_SYMBOL(native_load_gs_index);
 #endif
+
+#ifdef CONFIG_PREEMPT
+EXPORT_SYMBOL(___preempt_schedule);
+#ifdef CONFIG_CONTEXT_TRACKING
+EXPORT_SYMBOL(___preempt_schedule_context);
+#endif
+#endif
diff --git a/include/asm-generic/preempt.h b/include/asm-generic/preempt.h
index 82d958fc382..5dc14ed3791 100644
--- a/include/asm-generic/preempt.h
+++ b/include/asm-generic/preempt.h
@@ -100,4 +100,14 @@ static __always_inline bool should_resched(void)
 	return unlikely(!*preempt_count_ptr());
 }
 
+#ifdef CONFIG_PREEMPT
+extern asmlinkage void preempt_schedule(void);
+#define __preempt_schedule() preempt_schedule()
+
+#ifdef CONFIG_CONTEXT_TRACKING
+extern asmlinkage void preempt_schedule_context(void);
+#define __preempt_schedule_context() preempt_schedule_context()
+#endif
+#endif /* CONFIG_PREEMPT */
+
 #endif /* __ASM_PREEMPT_H */
diff --git a/include/linux/preempt.h b/include/linux/preempt.h
index 2343d871529..a3d9dc8c2c0 100644
--- a/include/linux/preempt.h
+++ b/include/linux/preempt.h
@@ -50,18 +50,17 @@ do { \
 #define preempt_enable_no_resched() sched_preempt_enable_no_resched()
 
 #ifdef CONFIG_PREEMPT
-asmlinkage void preempt_schedule(void);
 #define preempt_enable() \
 do { \
 	barrier(); \
 	if (unlikely(preempt_count_dec_and_test())) \
-		preempt_schedule(); \
+		__preempt_schedule(); \
 } while (0)
 
 #define preempt_check_resched() \
 do { \
 	if (should_resched()) \
-		preempt_schedule(); \
+		__preempt_schedule(); \
 } while (0)
 
 #else
@@ -83,17 +82,15 @@ do { \
 
 #ifdef CONFIG_PREEMPT
 
-#ifdef CONFIG_CONTEXT_TRACKING
-asmlinkage void preempt_schedule_context(void);
-#else
-#define preempt_schedule_context() preempt_schedule()
+#ifndef CONFIG_CONTEXT_TRACKING
+#define __preempt_schedule_context() __preempt_schedule()
 #endif
 
 #define preempt_enable_notrace() \
 do { \
 	barrier(); \
 	if (unlikely(__preempt_count_dec_and_test())) \
-		preempt_schedule_context(); \
+		__preempt_schedule_context(); \
 } while (0)
 #else
 #define preempt_enable_notrace() preempt_enable_no_resched_notrace()
-- 
cgit v1.2.3-70-g09d2


From 75f93fed50c2abadbab6ef546b265f51ca975b27 Mon Sep 17 00:00:00 2001
From: Peter Zijlstra <peterz@infradead.org>
Date: Fri, 27 Sep 2013 17:30:03 +0200
Subject: sched: Revert need_resched() to look at TIF_NEED_RESCHED

Yuanhan reported a serious throughput regression in his pigz
benchmark. Using the ftrace patch I found that several idle
paths need more TLC before we can switch the generic
need_resched() over to preempt_need_resched.

The preemption paths benefit most from preempt_need_resched and
do indeed use it; all other need_resched() users don't really
care that much so reverting need_resched() back to
tif_need_resched() is the simple and safe solution.

Reported-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: Fengguang Wu <fengguang.wu@intel.com>
Cc: Huang Ying <ying.huang@intel.com>
Cc: lkp@linux.intel.com
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/20130927153003.GF15690@laptop.programming.kicks-ass.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/include/asm/preempt.h | 8 --------
 include/asm-generic/preempt.h  | 8 --------
 include/linux/sched.h          | 5 +++++
 3 files changed, 5 insertions(+), 16 deletions(-)

(limited to 'arch/x86')

diff --git a/arch/x86/include/asm/preempt.h b/arch/x86/include/asm/preempt.h
index 1de41690ff9..8729723636f 100644
--- a/arch/x86/include/asm/preempt.h
+++ b/arch/x86/include/asm/preempt.h
@@ -79,14 +79,6 @@ static __always_inline bool __preempt_count_dec_and_test(void)
 	GEN_UNARY_RMWcc("decl", __preempt_count, __percpu_arg(0), "e");
 }
 
-/*
- * Returns true when we need to resched -- even if we can not.
- */
-static __always_inline bool need_resched(void)
-{
-	return unlikely(test_preempt_need_resched());
-}
-
 /*
  * Returns true when we need to resched and can (barring IRQ state).
  */
diff --git a/include/asm-generic/preempt.h b/include/asm-generic/preempt.h
index 5dc14ed3791..ddf2b420ac8 100644
--- a/include/asm-generic/preempt.h
+++ b/include/asm-generic/preempt.h
@@ -84,14 +84,6 @@ static __always_inline bool __preempt_count_dec_and_test(void)
 	return !--*preempt_count_ptr();
 }
 
-/*
- * Returns true when we need to resched -- even if we can not.
- */
-static __always_inline bool need_resched(void)
-{
-	return unlikely(test_preempt_need_resched());
-}
-
 /*
  * Returns true when we need to resched and can (barring IRQ state).
  */
diff --git a/include/linux/sched.h b/include/linux/sched.h
index b09798b672f..2ac5285db43 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -2577,6 +2577,11 @@ static inline bool __must_check current_clr_polling_and_test(void)
 }
 #endif
 
+static __always_inline bool need_resched(void)
+{
+	return unlikely(tif_need_resched());
+}
+
 /*
  * Thread group CPU time accounting.
  */
-- 
cgit v1.2.3-70-g09d2


From 88f182dd779b9d350b4774c12d16633a5b60f50c Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@kernel.org>
Date: Thu, 10 Oct 2013 10:16:30 +0200
Subject: x86: Apply the asm_volatile_goto() compiler quirk

Apply the asm_volatile_goto() compiler quirk to the new rmwcc.h
file as well, introduced in:

   c2daa3bed53a sched, x86: Provide a per-cpu preempt_count implementation

Reported-and-tested-by: Fengguang Wu <fengguang.wu@intel.com>
Reported-by: Oleg Nesterov <oleg@redhat.com>
Reported-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Suggested-by: Jakub Jelinek <jakub@redhat.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/include/asm/rmwcc.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'arch/x86')

diff --git a/arch/x86/include/asm/rmwcc.h b/arch/x86/include/asm/rmwcc.h
index 735f1849795..1ff990f1de8 100644
--- a/arch/x86/include/asm/rmwcc.h
+++ b/arch/x86/include/asm/rmwcc.h
@@ -5,7 +5,7 @@
 
 #define __GEN_RMWcc(fullop, var, cc, ...)				\
 do {									\
-	asm volatile goto (fullop "; j" cc " %l[cc_label]"		\
+	asm_volatile_goto (fullop "; j" cc " %l[cc_label]"		\
 			: : "m" (var), ## __VA_ARGS__ 			\
 			: "memory" : cc_label);				\
 	return 0;							\
-- 
cgit v1.2.3-70-g09d2