Skip to contentSkip to conversation

add this_cpu_{add_return|[cmp]xchg}() operations

PATCH In review Tejun Heo 2 participants 8 messages Dec 18, 2010 – Dec 20, 2010

Subsystems (2)
lore
Share & tools
lore

Tejun Heo

author maintainer
#0 lore raw

PATCH 1/1 add this_cpu_{add_return|[cmp]xchg}() operations

Hello,

Here are the patches that I applied to percpu#this_cpu_ops branch. This branch won't be rebased and contains only the commits which implement the new this_cpu_* ops, so that it can be pulled into other trees without patches which make use of the operations which modify other subsystems.

Christoph, I rearranged things a bit to make it conform better to the overall organization. It would be great if you can go over it once just in case.

There are two git branches.

git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu.git this_cpu_ops
git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu.git for-2.6.38

Both branches won't be rebased. this_cpu_ops contains only patches from this series. for-2.6.38 contains other percpu changes mostly composed of patches which apply this_cpu_*() ops.

Thank you.

--
tejun

author maintainer
#1 lore raw

PATCH 1/6 percpu: Generic support for this_cpu_add, sub, dec, inc_return

From: Christoph Lameter <[email protected]>

Introduce generic support for this_cpu_add_return etc.

The fallback is to realize these operations with simpler __this_cpu_ops.

tj: - Reformatted __cpu_size_call_return2() to make it more consistent
with its neighbors.
- Dropped unnecessary temp variable ret__ from __this_cpu_generic_add_return().

---
 include/linux/percpu.h |   71 ++++++++++++++++++++++++++++++++++++++++++++++++
 1 files changed, 71 insertions(+), 0 deletions(-)
include/linux/percpu.h +71 -0
index 5095b83..4d593de 100644
@@ -240,6 +240,21 @@ extern void __bad_size_call_parameter(void);
 	pscr_ret__;							\
 })
 
+#define __pcpu_size_call_return2(stem, variable, ...)			\
+({									\
+	typeof(variable) pscr2_ret__;					\
+	__verify_pcpu_ptr(&(variable));					\
+	switch(sizeof(variable)) {					\
+	case 1: pscr2_ret__ = stem##1(variable, __VA_ARGS__); break;	\
+	case 2: pscr2_ret__ = stem##2(variable, __VA_ARGS__); break;	\
+	case 4: pscr2_ret__ = stem##4(variable, __VA_ARGS__); break;	\
+	case 8: pscr2_ret__ = stem##8(variable, __VA_ARGS__); break;	\
+	default:							\
+		__bad_size_call_parameter(); break;			\
+	}								\
+	pscr2_ret__;							\
+})
+
 #define __pcpu_size_call(stem, variable, ...)				\
 do {									\
 	__verify_pcpu_ptr(&(variable));					\
@@ -529,6 +544,62 @@ do {									\
 # define __this_cpu_xor(pcp, val)	__pcpu_size_call(__this_cpu_xor_, (pcp), (val))
 #endif
 
+#define _this_cpu_generic_add_return(pcp, val)				\
+({									\
+	typeof(pcp) ret__;						\
+	preempt_disable();						\
+	__this_cpu_add(pcp, val);					\
+	ret__ = __this_cpu_read(pcp);					\
+	preempt_enable();						\
+	ret__;								\
+})
+
+#ifndef this_cpu_add_return
+# ifndef this_cpu_add_return_1
+#  define this_cpu_add_return_1(pcp, val)	_this_cpu_generic_add_return(pcp, val)
+# endif
+# ifndef this_cpu_add_return_2
+#  define this_cpu_add_return_2(pcp, val)	_this_cpu_generic_add_return(pcp, val)
+# endif
+# ifndef this_cpu_add_return_4
+#  define this_cpu_add_return_4(pcp, val)	_this_cpu_generic_add_return(pcp, val)
+# endif
+# ifndef this_cpu_add_return_8
+#  define this_cpu_add_return_8(pcp, val)	_this_cpu_generic_add_return(pcp, val)
+# endif
+# define this_cpu_add_return(pcp, val)	__pcpu_size_call_return2(this_cpu_add_return_, pcp, val)
+#endif
+
+#define this_cpu_sub_return(pcp, val)	this_cpu_add_return(pcp, -(val))
+#define this_cpu_inc_return(pcp)	this_cpu_add_return(pcp, 1)
+#define this_cpu_dec_return(pcp)	this_cpu_add_return(pcp, -1)
+
+#define __this_cpu_generic_add_return(pcp, val)				\
+({									\
+	__this_cpu_add(pcp, val);					\
+	__this_cpu_read(pcp);						\
+})
+
+#ifndef __this_cpu_add_return
+# ifndef __this_cpu_add_return_1
+#  define __this_cpu_add_return_1(pcp, val)	__this_cpu_generic_add_return(pcp, val)
+# endif
+# ifndef __this_cpu_add_return_2
+#  define __this_cpu_add_return_2(pcp, val)	__this_cpu_generic_add_return(pcp, val)
+# endif
+# ifndef __this_cpu_add_return_4
+#  define __this_cpu_add_return_4(pcp, val)	__this_cpu_generic_add_return(pcp, val)
+# endif
+# ifndef __this_cpu_add_return_8
+#  define __this_cpu_add_return_8(pcp, val)	__this_cpu_generic_add_return(pcp, val)
+# endif
+# define __this_cpu_add_return(pcp, val)	__pcpu_size_call_return2(this_cpu_add_return_, pcp, val)
+#endif
+
+#define __this_cpu_sub_return(pcp, val)	this_cpu_add_return(pcp, -(val))
+#define __this_cpu_inc_return(pcp)	this_cpu_add_return(pcp, 1)
+#define __this_cpu_dec_return(pcp)	this_cpu_add_return(pcp, -1)
+
 /*
  * IRQ safe versions of the per cpu RMW operations. Note that these operations
  * are *not* safe against modification of the same variable from another
Signature · 1 line
1.7.1
author maintainer
#2 lore raw

PATCH 2/6 x86: Support for this_cpu_add, sub, dec, inc_return

In reply to Tejun Heo

From: Christoph Lameter <[email protected]>

Supply an implementation for x86 in order to generate more efficient code.

V2->V3:
- Cleanup
- Remove strange type checking from percpu_add_return_op.

tj: - Dropped unused typedef from percpu_add_return_op().
- Renamed ret__ to paro_ret__ in percpu_add_return_op().
- Minor indentation adjustments.

---
 arch/x86/include/asm/percpu.h |   43 +++++++++++++++++++++++++++++++++++++++++
 1 files changed, 43 insertions(+), 0 deletions(-)
arch/x86/include/asm/percpu.h +43 -0
index f899e01..38f9e96 100644
@@ -177,6 +177,39 @@ do {									\
 	}								\
 } while (0)
 
+/*
+ * Add return operation
+ */
+#define percpu_add_return_op(var, val)					\
+({									\
+	typeof(var) paro_ret__ = val;					\
+	switch (sizeof(var)) {						\
+	case 1:								\
+		asm("xaddb %0, "__percpu_arg(1)				\
+			    : "+q" (paro_ret__), "+m" (var)		\
+			    : : "memory");				\
+		break;							\
+	case 2:								\
+		asm("xaddw %0, "__percpu_arg(1)				\
+			    : "+r" (paro_ret__), "+m" (var)		\
+			    : : "memory");				\
+		break;							\
+	case 4:								\
+		asm("xaddl %0, "__percpu_arg(1)				\
+			    : "+r" (paro_ret__), "+m" (var)		\
+			    : : "memory");				\
+		break;							\
+	case 8:								\
+		asm("xaddq %0, "__percpu_arg(1)				\
+			    : "+re" (paro_ret__), "+m" (var)		\
+			    : : "memory");				\
+		break;							\
+	default: __bad_percpu_size();					\
+	}								\
+	paro_ret__ += val;						\
+	paro_ret__;							\
+})
+
 #define percpu_from_op(op, var, constraint)		\
 ({							\
 	typeof(var) pfo_ret__;				\
@@ -300,6 +333,14 @@ do {									\
 #define irqsafe_cpu_xor_2(pcp, val)	percpu_to_op("xor", (pcp), val)
 #define irqsafe_cpu_xor_4(pcp, val)	percpu_to_op("xor", (pcp), val)
 
+#ifndef CONFIG_M386
+#define __this_cpu_add_return_1(pcp, val) percpu_add_return_op(pcp, val)
+#define __this_cpu_add_return_2(pcp, val) percpu_add_return_op(pcp, val)
+#define __this_cpu_add_return_4(pcp, val) percpu_add_return_op(pcp, val)
+#define this_cpu_add_return_1(pcp, val)	percpu_add_return_op(pcp, val)
+#define this_cpu_add_return_2(pcp, val)	percpu_add_return_op(pcp, val)
+#define this_cpu_add_return_4(pcp, val)	percpu_add_return_op(pcp, val)
+#endif
 /*
  * Per cpu atomic 64 bit operations are only available under 64 bit.
  * 32 bit must fall back to generic operations.
@@ -324,6 +365,8 @@ do {									\
 #define irqsafe_cpu_or_8(pcp, val)	percpu_to_op("or", (pcp), val)
 #define irqsafe_cpu_xor_8(pcp, val)	percpu_to_op("xor", (pcp), val)
 
+#define __this_cpu_add_return_8(pcp, val) percpu_add_return_op(pcp, val)
+#define this_cpu_add_return_8(pcp, val)	percpu_add_return_op(pcp, val)
 #endif
 
 /* This is not atomic against other CPUs -- CPU preemption needs to be off */
Signature · 1 line
1.7.1
author maintainer
#3 lore raw

PATCH 3/6 percpu,x86: relocate this_cpu_add_return() and friends

In reply to Tejun Heo

- include/linux/percpu.h: this_cpu_add_return() and friends were located next to __this_cpu_add_return(). However, the overall organization is to first group by preemption safeness. Relocate this_cpu_add_return() and friends to preemption-safe area.

- arch/x86/include/asm/percpu.h: Relocate percpu_add_return_op() after other more basic operations. Relocate [__]this_cpu_add_return_8() so that they're first grouped by preemption safeness.

Cc Christoph Lameter <[email protected]>
---
 arch/x86/include/asm/percpu.h |   71 ++++++++++++++++++++---------------------
 include/linux/percpu.h        |   60 +++++++++++++++++-----------------
 2 files changed, 65 insertions(+), 66 deletions(-)
arch/x86/include/asm/percpu.h +35 -36
index 38f9e96..dd0cd4b 100644
@@ -177,39 +177,6 @@ do {									\
 	}								\
 } while (0)
 
-/*
- * Add return operation
- */
-#define percpu_add_return_op(var, val)					\
-({									\
-	typeof(var) paro_ret__ = val;					\
-	switch (sizeof(var)) {						\
-	case 1:								\
-		asm("xaddb %0, "__percpu_arg(1)				\
-			    : "+q" (paro_ret__), "+m" (var)		\
-			    : : "memory");				\
-		break;							\
-	case 2:								\
-		asm("xaddw %0, "__percpu_arg(1)				\
-			    : "+r" (paro_ret__), "+m" (var)		\
-			    : : "memory");				\
-		break;							\
-	case 4:								\
-		asm("xaddl %0, "__percpu_arg(1)				\
-			    : "+r" (paro_ret__), "+m" (var)		\
-			    : : "memory");				\
-		break;							\
-	case 8:								\
-		asm("xaddq %0, "__percpu_arg(1)				\
-			    : "+re" (paro_ret__), "+m" (var)		\
-			    : : "memory");				\
-		break;							\
-	default: __bad_percpu_size();					\
-	}								\
-	paro_ret__ += val;						\
-	paro_ret__;							\
-})
-
 #define percpu_from_op(op, var, constraint)		\
 ({							\
 	typeof(var) pfo_ret__;				\
@@ -263,6 +230,39 @@ do {									\
 })
 
 /*
+ * Add return operation
+ */
+#define percpu_add_return_op(var, val)					\
+({									\
+	typeof(var) paro_ret__ = val;					\
+	switch (sizeof(var)) {						\
+	case 1:								\
+		asm("xaddb %0, "__percpu_arg(1)				\
+			    : "+q" (paro_ret__), "+m" (var)		\
+			    : : "memory");				\
+		break;							\
+	case 2:								\
+		asm("xaddw %0, "__percpu_arg(1)				\
+			    : "+r" (paro_ret__), "+m" (var)		\
+			    : : "memory");				\
+		break;							\
+	case 4:								\
+		asm("xaddl %0, "__percpu_arg(1)				\
+			    : "+r" (paro_ret__), "+m" (var)		\
+			    : : "memory");				\
+		break;							\
+	case 8:								\
+		asm("xaddq %0, "__percpu_arg(1)				\
+			    : "+re" (paro_ret__), "+m" (var)		\
+			    : : "memory");				\
+		break;							\
+	default: __bad_percpu_size();					\
+	}								\
+	paro_ret__ += val;						\
+	paro_ret__;							\
+})
+
+/*
  * percpu_read() makes gcc load the percpu variable every time it is
  * accessed while percpu_read_stable() allows the value to be cached.
  * percpu_read_stable() is more efficient and can be used if its value
@@ -352,6 +352,7 @@ do {									\
 #define __this_cpu_and_8(pcp, val)	percpu_to_op("and", (pcp), val)
 #define __this_cpu_or_8(pcp, val)	percpu_to_op("or", (pcp), val)
 #define __this_cpu_xor_8(pcp, val)	percpu_to_op("xor", (pcp), val)
+#define __this_cpu_add_return_8(pcp, val) percpu_add_return_op(pcp, val)
 
 #define this_cpu_read_8(pcp)		percpu_from_op("mov", (pcp), "m"(pcp))
 #define this_cpu_write_8(pcp, val)	percpu_to_op("mov", (pcp), val)
@@ -359,14 +360,12 @@ do {									\
 #define this_cpu_and_8(pcp, val)	percpu_to_op("and", (pcp), val)
 #define this_cpu_or_8(pcp, val)		percpu_to_op("or", (pcp), val)
 #define this_cpu_xor_8(pcp, val)	percpu_to_op("xor", (pcp), val)
+#define this_cpu_add_return_8(pcp, val)	percpu_add_return_op(pcp, val)
 
 #define irqsafe_cpu_add_8(pcp, val)	percpu_add_op((pcp), val)
 #define irqsafe_cpu_and_8(pcp, val)	percpu_to_op("and", (pcp), val)
 #define irqsafe_cpu_or_8(pcp, val)	percpu_to_op("or", (pcp), val)
 #define irqsafe_cpu_xor_8(pcp, val)	percpu_to_op("xor", (pcp), val)
-
-#define __this_cpu_add_return_8(pcp, val) percpu_add_return_op(pcp, val)
-#define this_cpu_add_return_8(pcp, val)	percpu_add_return_op(pcp, val)
 #endif
 
 /* This is not atomic against other CPUs -- CPU preemption needs to be off */
include/linux/percpu.h +30 -30

72 lines not on this page · raw message

Signature · 1 line
1.7.1
author maintainer
#4 lore raw

PATCH 4/6 percpu: Generic this_cpu_cmpxchg() and this_cpu_xchg support

In reply to Tejun Heo

From: Christoph Lameter <[email protected]>

Generic code to provide new per cpu atomic features

this_cpu_cmpxchg
this_cpu_xchg

Fallback occurs to functions using interrupts disable/enable to ensure correct per cpu atomicity.

Fallback to regular cmpxchg and xchg is not possible since per cpu atomic semantics include the guarantee that the current cpus per cpu data is accessed atomically. Use of regular cmpxchg and xchg requires the determination of the address of the per cpu data before regular cmpxchg or xchg which therefore cannot be atomically included in an xchg or cmpxchg without segment override.

tj: - Relocated new ops to conform better to the general organization.
- This patch contains a trivial comment fix.

---
 include/linux/percpu.h |  134 +++++++++++++++++++++++++++++++++++++++++++++++-
 1 files changed, 133 insertions(+), 1 deletions(-)
include/linux/percpu.h +133 -1
index 3484e88..27c3c6f 100644
@@ -447,6 +447,59 @@ do {									\
 #define this_cpu_inc_return(pcp)	this_cpu_add_return(pcp, 1)
 #define this_cpu_dec_return(pcp)	this_cpu_add_return(pcp, -1)
 
+#define _this_cpu_generic_xchg(pcp, nval)				\
+({	typeof(pcp) ret__;						\
+	preempt_disable();						\
+	ret__ = __this_cpu_read(pcp);					\
+	__this_cpu_write(pcp, nval);					\
+	preempt_enable();						\
+	ret__;								\
+})
+
+#ifndef this_cpu_xchg
+# ifndef this_cpu_xchg_1
+#  define this_cpu_xchg_1(pcp, nval)	_this_cpu_generic_xchg(pcp, nval)
+# endif
+# ifndef this_cpu_xchg_2
+#  define this_cpu_xchg_2(pcp, nval)	_this_cpu_generic_xchg(pcp, nval)
+# endif
+# ifndef this_cpu_xchg_4
+#  define this_cpu_xchg_4(pcp, nval)	_this_cpu_generic_xchg(pcp, nval)
+# endif
+# ifndef this_cpu_xchg_8
+#  define this_cpu_xchg_8(pcp, nval)	_this_cpu_generic_xchg(pcp, nval)
+# endif
+# define this_cpu_xchg(pcp, nval)	\
+	__pcpu_size_call_return2(this_cpu_xchg_, (pcp), nval)
+#endif
+
+#define _this_cpu_generic_cmpxchg(pcp, oval, nval)			\
+({	typeof(pcp) ret__;						\
+	preempt_disable();						\
+	ret__ = __this_cpu_read(pcp);					\
+	if (ret__ == (oval))						\
+		__this_cpu_write(pcp, nval);				\
+	preempt_enable();						\
+	ret__;								\
+})
+
+#ifndef this_cpu_cmpxchg
+# ifndef this_cpu_cmpxchg_1
+#  define this_cpu_cmpxchg_1(pcp, oval, nval)	_this_cpu_generic_cmpxchg(pcp, oval, nval)
+# endif
+# ifndef this_cpu_cmpxchg_2
+#  define this_cpu_cmpxchg_2(pcp, oval, nval)	_this_cpu_generic_cmpxchg(pcp, oval, nval)
+# endif
+# ifndef this_cpu_cmpxchg_4
+#  define this_cpu_cmpxchg_4(pcp, oval, nval)	_this_cpu_generic_cmpxchg(pcp, oval, nval)
+# endif
+# ifndef this_cpu_cmpxchg_8
+#  define this_cpu_cmpxchg_8(pcp, oval, nval)	_this_cpu_generic_cmpxchg(pcp, oval, nval)
+# endif
+# define this_cpu_cmpxchg(pcp, oval, nval)	\
+	__pcpu_size_call_return2(this_cpu_cmpxchg_, pcp, oval, nval)
+#endif
+
 /*
  * Generic percpu operations that do not require preemption handling.
  * Either we do not care about races or the caller has the
@@ -600,11 +653,61 @@ do {									\
 #define __this_cpu_inc_return(pcp)	this_cpu_add_return(pcp, 1)
 #define __this_cpu_dec_return(pcp)	this_cpu_add_return(pcp, -1)
 
+#define __this_cpu_generic_xchg(pcp, nval)				\
+({	typeof(pcp) ret__;						\
+	ret__ = __this_cpu_read(pcp);					\
+	__this_cpu_write(pcp, nval);					\
+	ret__;								\
+})
+
+#ifndef __this_cpu_xchg
+# ifndef __this_cpu_xchg_1
+#  define __this_cpu_xchg_1(pcp, nval)	__this_cpu_generic_xchg(pcp, nval)
+# endif
+# ifndef __this_cpu_xchg_2
+#  define __this_cpu_xchg_2(pcp, nval)	__this_cpu_generic_xchg(pcp, nval)
+# endif
+# ifndef __this_cpu_xchg_4
+#  define __this_cpu_xchg_4(pcp, nval)	__this_cpu_generic_xchg(pcp, nval)
+# endif
+# ifndef __this_cpu_xchg_8
+#  define __this_cpu_xchg_8(pcp, nval)	__this_cpu_generic_xchg(pcp, nval)
+# endif
+# define __this_cpu_xchg(pcp, nval)	\
+	__pcpu_size_call_return2(__this_cpu_xchg_, (pcp), nval)
+#endif
+
+#define __this_cpu_generic_cmpxchg(pcp, oval, nval)			\
+({									\
+	typeof(pcp) ret__;						\
+	ret__ = __this_cpu_read(pcp);					\
+	if (ret__ == (oval))						\
+		__this_cpu_write(pcp, nval);				\
+	ret__;								\
+})
+
+#ifndef __this_cpu_cmpxchg
+# ifndef __this_cpu_cmpxchg_1
+#  define __this_cpu_cmpxchg_1(pcp, oval, nval)	__this_cpu_generic_cmpxchg(pcp, oval, nval)
+# endif
+# ifndef __this_cpu_cmpxchg_2
+#  define __this_cpu_cmpxchg_2(pcp, oval, nval)	__this_cpu_generic_cmpxchg(pcp, oval, nval)
+# endif
+# ifndef __this_cpu_cmpxchg_4
+#  define __this_cpu_cmpxchg_4(pcp, oval, nval)	__this_cpu_generic_cmpxchg(pcp, oval, nval)
+# endif
+# ifndef __this_cpu_cmpxchg_8
+#  define __this_cpu_cmpxchg_8(pcp, oval, nval)	__this_cpu_generic_cmpxchg(pcp, oval, nval)
+# endif
+# define __this_cpu_cmpxchg(pcp, oval, nval)	\
+	__pcpu_size_call_return2(__this_cpu_cmpxchg_, pcp, oval, nval)
+#endif
+
 /*
  * IRQ safe versions of the per cpu RMW operations. Note that these operations
  * are *not* safe against modification of the same variable from another
  * processors (which one gets when using regular atomic operations)
- . They are guaranteed to be atomic vs. local interrupts and
+ * They are guaranteed to be atomic vs. local interrupts and
  * preemption only.
  */
 #define irqsafe_cpu_generic_to_op(pcp, val, op)				\
@@ -691,4 +794,33 @@ do {									\
 # define irqsafe_cpu_xor(pcp, val) __pcpu_size_call(irqsafe_cpu_xor_, (val))
 #endif
 
+#define irqsafe_cpu_generic_cmpxchg(pcp, oval, nval)			\
+({									\
+	typeof(pcp) ret__;						\
+	unsigned long flags;						\
+	local_irq_save(flags);						\
+	ret__ = __this_cpu_read(pcp);					\
+	if (ret__ == (oval))						\
+		__this_cpu_write(pcp, nval);				\
+	local_irq_restore(flags);					\
+	ret__;								\
+})
+
+#ifndef irqsafe_cpu_cmpxchg
+# ifndef irqsafe_cpu_cmpxchg_1
+#  define irqsafe_cpu_cmpxchg_1(pcp, oval, nval)	irqsafe_cpu_generic_cmpxchg(pcp, oval, nval)
+# endif
+# ifndef irqsafe_cpu_cmpxchg_2
+#  define irqsafe_cpu_cmpxchg_2(pcp, oval, nval)	irqsafe_cpu_generic_cmpxchg(pcp, oval, nval)
+# endif
+# ifndef irqsafe_cpu_cmpxchg_4
+#  define irqsafe_cpu_cmpxchg_4(pcp, oval, nval)	irqsafe_cpu_generic_cmpxchg(pcp, oval, nval)
+# endif
+# ifndef irqsafe_cpu_cmpxchg_8
+#  define irqsafe_cpu_cmpxchg_8(pcp, oval, nval)	irqsafe_cpu_generic_cmpxchg(pcp, oval, nval)
+# endif
+# define irqsafe_cpu_cmpxchg(pcp, oval, nval)		\
+	__pcpu_size_call_return2(irqsafe_cpu_cmpxchg_, (pcp), oval, nval)
+#endif
+
 #endif /* __LINUX_PERCPU_H */
Signature · 1 line
1.7.1
author maintainer
#5 lore raw

PATCH 5/6 x86: this_cpu_cmpxchg and this_cpu_xchg operations

In reply to Tejun Heo

From: Christoph Lameter <[email protected]>

Provide support as far as the hardware capabilities of the x86 cpus allow.

Define CONFIG_CMPXCHG_LOCAL in Kconfig.cpu to allow core code to test for fast cpuops implementations.

V1->V2:
- Take out the definition for this_cpu_cmpxchg_8 and move it into a separate patch.

tj: - Reordered ops to better follow this_cpu_* organization.
- Renamed macro temp variables similar to their existing neighbours.

---
 arch/x86/Kconfig.cpu          |    3 +
 arch/x86/include/asm/percpu.h |  107 ++++++++++++++++++++++++++++++++++++++++-
 2 files changed, 109 insertions(+), 1 deletions(-)
arch/x86/Kconfig.cpu +3 -0
index 2ac9069..15588a0 100644
@@ -310,6 +310,9 @@ config X86_INTERNODE_CACHE_SHIFT
 config X86_CMPXCHG
 	def_bool X86_64 || (X86_32 && !M386)
 
+config CMPXCHG_LOCAL
+	def_bool X86_64 || (X86_32 && !M386)
+
 config X86_L1_CACHE_SHIFT
 	int
 	default "7" if MPENTIUM4 || MPSC
arch/x86/include/asm/percpu.h +106 -1

139 lines not on this page · raw message

Signature · 1 line
1.7.1
author maintainer
#6 lore raw

PATCH 6/6 cpuops: Use cmpxchg for xchg to avoid lock semantics

In reply to Tejun Heo

From: Christoph Lameter <[email protected]>

Use cmpxchg instead of xchg to realize this_cpu_xchg.

xchg will cause LOCK overhead since LOCK is always implied but cmpxchg will not.

Baselines:

xchg() = 18 cycles (no segment prefix, LOCK semantics)
__this_cpu_xchg = 1 cycle

(simulated using this_cpu_read/write, two prefixes. Looks like the cpu can use loop optimization to get rid of most of the overhead)

Cycles before:

this_cpu_xchg = 37 cycles (segment prefix and LOCK (implied by xchg))

After:

this_cpu_xchg = 11 cycle (using cmpxchg without lock semantics)

---
 arch/x86/include/asm/percpu.h |   21 +++++++++++++++------
 1 files changed, 15 insertions(+), 6 deletions(-)
arch/x86/include/asm/percpu.h +15 -6
index b85ade5..8ee4516 100644
@@ -263,8 +263,9 @@ do {									\
 })
 
 /*
- * Beware: xchg on x86 has an implied lock prefix. There will be the cost of
- * full lock semantics even though they are not needed.
+ * xchg is implemented using cmpxchg without a lock prefix. xchg is
+ * expensive due to the implied lock prefix.  The processor cannot prefetch
+ * cachelines if xchg is used.
  */
 #define percpu_xchg_op(var, nval)					\
 ({									\
@@ -272,25 +273,33 @@ do {									\
 	typeof(var) pxo_new__ = (nval);					\
 	switch (sizeof(var)) {						\
 	case 1:								\
-		asm("xchgb %2, "__percpu_arg(1)				\
+		asm("\n1:mov "__percpu_arg(1)",%%al"			\
+		    "\n\tcmpxchgb %2, "__percpu_arg(1)			\
+		    "\n\tjnz 1b"					\
 			    : "=a" (pxo_ret__), "+m" (var)		\
 			    : "q" (pxo_new__)				\
 			    : "memory");				\
 		break;							\
 	case 2:								\
-		asm("xchgw %2, "__percpu_arg(1)				\
+		asm("\n1:mov "__percpu_arg(1)",%%ax"			\
+		    "\n\tcmpxchgw %2, "__percpu_arg(1)			\
+		    "\n\tjnz 1b"					\
 			    : "=a" (pxo_ret__), "+m" (var)		\
 			    : "r" (pxo_new__)				\
 			    : "memory");				\
 		break;							\
 	case 4:								\
-		asm("xchgl %2, "__percpu_arg(1)				\
+		asm("\n1:mov "__percpu_arg(1)",%%eax"			\
+		    "\n\tcmpxchgl %2, "__percpu_arg(1)			\
+		    "\n\tjnz 1b"					\
 			    : "=a" (pxo_ret__), "+m" (var)		\
 			    : "r" (pxo_new__)				\
 			    : "memory");				\
 		break;							\
 	case 8:								\
-		asm("xchgq %2, "__percpu_arg(1)				\
+		asm("\n1:mov "__percpu_arg(1)",%%rax"			\
+		    "\n\tcmpxchgq %2, "__percpu_arg(1)			\
+		    "\n\tjnz 1b"					\
 			    : "=a" (pxo_ret__), "+m" (var)		\
 			    : "r" (pxo_new__)				\
 			    : "memory");				\
Signature · 1 line
1.7.1
#7 lore raw
In reply to Tejun Heo
Quoted 12 lines · Tejun Heo

Hello,

Here are the patches that I applied to percpu#this_cpu_ops branch. This branch won't be rebased and contains only the commits which implement the new this_cpu_* ops, so that it can be pulled into other trees without patches which make use of the operations which modify other subsystems.

Christoph, I rearranged things a bit to make it conform better to the overall organization. It would be great if you can go over it once just in case.

Quoted 8 lines

There are two git branches.

git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu.git this_cpu_ops
git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu.git for-2.6.38

Both branches won't be rebased. this_cpu_ops contains only patches from this series. for-2.6.38 contains other percpu changes mostly composed of patches which apply this_cpu_*() ops.

Thanks. Christoph, mind sending me the slub patches? There's been so many I've lost track which one is the latest.

Pekka