* [PATCH 1/6] percpu: Generic support for this_cpu_add, sub, dec, inc_return
2010-12-18 15:24 [PATCHSET percpu#this_cpu_ops] add this_cpu_{add_return|[cmp]xchg}() operations Tejun Heo
@ 2010-12-18 15:24 ` Tejun Heo
2010-12-18 15:24 ` [PATCH 2/6] x86: Support " Tejun Heo
` (5 subsequent siblings)
6 siblings, 0 replies; 8+ messages in thread
From: Tejun Heo @ 2010-12-18 15:24 UTC (permalink / raw)
To: penberg, akpm, linux-kernel, eric.dumazet, hpa, mathieu.desnoyers
Cc: Christoph Lameter, Tejun Heo
From: Christoph Lameter <cl@linux.com>
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().
Reviewed-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Acked-by: H. Peter Anvin <hpa@zytor.com>
Signed-off-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
---
include/linux/percpu.h | 71 ++++++++++++++++++++++++++++++++++++++++++++++++
1 files changed, 71 insertions(+), 0 deletions(-)
diff --git a/include/linux/percpu.h b/include/linux/percpu.h
index 5095b83..4d593de 100644
--- a/include/linux/percpu.h
+++ b/include/linux/percpu.h
@@ -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
--
1.7.1
^ permalink raw reply related [flat|nested] 8+ messages in thread* [PATCH 2/6] x86: Support for this_cpu_add, sub, dec, inc_return
2010-12-18 15:24 [PATCHSET percpu#this_cpu_ops] add this_cpu_{add_return|[cmp]xchg}() operations Tejun Heo
2010-12-18 15:24 ` [PATCH 1/6] percpu: Generic support for this_cpu_add, sub, dec, inc_return Tejun Heo
@ 2010-12-18 15:24 ` Tejun Heo
2010-12-18 15:24 ` [PATCH 3/6] percpu,x86: relocate this_cpu_add_return() and friends Tejun Heo
` (4 subsequent siblings)
6 siblings, 0 replies; 8+ messages in thread
From: Tejun Heo @ 2010-12-18 15:24 UTC (permalink / raw)
To: penberg, akpm, linux-kernel, eric.dumazet, hpa, mathieu.desnoyers
Cc: Christoph Lameter, Tejun Heo
From: Christoph Lameter <cl@linux.com>
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.
Acked-by: H. Peter Anvin <hpa@zytor.com>
Signed-off-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
---
arch/x86/include/asm/percpu.h | 43 +++++++++++++++++++++++++++++++++++++++++
1 files changed, 43 insertions(+), 0 deletions(-)
diff --git a/arch/x86/include/asm/percpu.h b/arch/x86/include/asm/percpu.h
index f899e01..38f9e96 100644
--- a/arch/x86/include/asm/percpu.h
+++ b/arch/x86/include/asm/percpu.h
@@ -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 */
--
1.7.1
^ permalink raw reply related [flat|nested] 8+ messages in thread* [PATCH 3/6] percpu,x86: relocate this_cpu_add_return() and friends
2010-12-18 15:24 [PATCHSET percpu#this_cpu_ops] add this_cpu_{add_return|[cmp]xchg}() operations Tejun Heo
2010-12-18 15:24 ` [PATCH 1/6] percpu: Generic support for this_cpu_add, sub, dec, inc_return Tejun Heo
2010-12-18 15:24 ` [PATCH 2/6] x86: Support " Tejun Heo
@ 2010-12-18 15:24 ` Tejun Heo
2010-12-18 15:24 ` [PATCH 4/6] percpu: Generic this_cpu_cmpxchg() and this_cpu_xchg support Tejun Heo
` (3 subsequent siblings)
6 siblings, 0 replies; 8+ messages in thread
From: Tejun Heo @ 2010-12-18 15:24 UTC (permalink / raw)
To: penberg, akpm, linux-kernel, eric.dumazet, hpa, mathieu.desnoyers
Cc: Tejun Heo, Christoph Lameter
- 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.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Christoph Lameter <cl@linux.com>
---
arch/x86/include/asm/percpu.h | 71 ++++++++++++++++++++---------------------
include/linux/percpu.h | 60 +++++++++++++++++-----------------
2 files changed, 65 insertions(+), 66 deletions(-)
diff --git a/arch/x86/include/asm/percpu.h b/arch/x86/include/asm/percpu.h
index 38f9e96..dd0cd4b 100644
--- a/arch/x86/include/asm/percpu.h
+++ b/arch/x86/include/asm/percpu.h
@@ -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 */
diff --git a/include/linux/percpu.h b/include/linux/percpu.h
index 4d593de..3484e88 100644
--- a/include/linux/percpu.h
+++ b/include/linux/percpu.h
@@ -417,6 +417,36 @@ do { \
# define this_cpu_xor(pcp, val) __pcpu_size_call(this_cpu_or_, (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)
+
/*
* Generic percpu operations that do not require preemption handling.
* Either we do not care about races or the caller has the
@@ -544,36 +574,6 @@ 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); \
--
1.7.1
^ permalink raw reply related [flat|nested] 8+ messages in thread* [PATCH 4/6] percpu: Generic this_cpu_cmpxchg() and this_cpu_xchg support
2010-12-18 15:24 [PATCHSET percpu#this_cpu_ops] add this_cpu_{add_return|[cmp]xchg}() operations Tejun Heo
` (2 preceding siblings ...)
2010-12-18 15:24 ` [PATCH 3/6] percpu,x86: relocate this_cpu_add_return() and friends Tejun Heo
@ 2010-12-18 15:24 ` Tejun Heo
2010-12-18 15:24 ` [PATCH 5/6] x86: this_cpu_cmpxchg and this_cpu_xchg operations Tejun Heo
` (2 subsequent siblings)
6 siblings, 0 replies; 8+ messages in thread
From: Tejun Heo @ 2010-12-18 15:24 UTC (permalink / raw)
To: penberg, akpm, linux-kernel, eric.dumazet, hpa, mathieu.desnoyers
Cc: Christoph Lameter, Tejun Heo
From: Christoph Lameter <cl@linux.com>
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.
Signed-off-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
---
include/linux/percpu.h | 134 +++++++++++++++++++++++++++++++++++++++++++++++-
1 files changed, 133 insertions(+), 1 deletions(-)
diff --git a/include/linux/percpu.h b/include/linux/percpu.h
index 3484e88..27c3c6f 100644
--- a/include/linux/percpu.h
+++ b/include/linux/percpu.h
@@ -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 */
--
1.7.1
^ permalink raw reply related [flat|nested] 8+ messages in thread* [PATCH 5/6] x86: this_cpu_cmpxchg and this_cpu_xchg operations
2010-12-18 15:24 [PATCHSET percpu#this_cpu_ops] add this_cpu_{add_return|[cmp]xchg}() operations Tejun Heo
` (3 preceding siblings ...)
2010-12-18 15:24 ` [PATCH 4/6] percpu: Generic this_cpu_cmpxchg() and this_cpu_xchg support Tejun Heo
@ 2010-12-18 15:24 ` Tejun Heo
2010-12-18 15:24 ` [PATCH 6/6] cpuops: Use cmpxchg for xchg to avoid lock semantics Tejun Heo
2010-12-20 15:53 ` [PATCHSET percpu#this_cpu_ops] add this_cpu_{add_return|[cmp]xchg}() operations Pekka Enberg
6 siblings, 0 replies; 8+ messages in thread
From: Tejun Heo @ 2010-12-18 15:24 UTC (permalink / raw)
To: penberg, akpm, linux-kernel, eric.dumazet, hpa, mathieu.desnoyers
Cc: Christoph Lameter, Tejun Heo
From: Christoph Lameter <cl@linux.com>
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.
Signed-off-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
---
arch/x86/Kconfig.cpu | 3 +
arch/x86/include/asm/percpu.h | 107 ++++++++++++++++++++++++++++++++++++++++-
2 files changed, 109 insertions(+), 1 deletions(-)
diff --git a/arch/x86/Kconfig.cpu b/arch/x86/Kconfig.cpu
index 2ac9069..15588a0 100644
--- a/arch/x86/Kconfig.cpu
+++ b/arch/x86/Kconfig.cpu
@@ -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
diff --git a/arch/x86/include/asm/percpu.h b/arch/x86/include/asm/percpu.h
index dd0cd4b..b85ade5 100644
--- a/arch/x86/include/asm/percpu.h
+++ b/arch/x86/include/asm/percpu.h
@@ -263,6 +263,83 @@ 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.
+ */
+#define percpu_xchg_op(var, nval) \
+({ \
+ typeof(var) pxo_ret__; \
+ typeof(var) pxo_new__ = (nval); \
+ switch (sizeof(var)) { \
+ case 1: \
+ asm("xchgb %2, "__percpu_arg(1) \
+ : "=a" (pxo_ret__), "+m" (var) \
+ : "q" (pxo_new__) \
+ : "memory"); \
+ break; \
+ case 2: \
+ asm("xchgw %2, "__percpu_arg(1) \
+ : "=a" (pxo_ret__), "+m" (var) \
+ : "r" (pxo_new__) \
+ : "memory"); \
+ break; \
+ case 4: \
+ asm("xchgl %2, "__percpu_arg(1) \
+ : "=a" (pxo_ret__), "+m" (var) \
+ : "r" (pxo_new__) \
+ : "memory"); \
+ break; \
+ case 8: \
+ asm("xchgq %2, "__percpu_arg(1) \
+ : "=a" (pxo_ret__), "+m" (var) \
+ : "r" (pxo_new__) \
+ : "memory"); \
+ break; \
+ default: __bad_percpu_size(); \
+ } \
+ pxo_ret__; \
+})
+
+/*
+ * cmpxchg has no such implied lock semantics as a result it is much
+ * more efficient for cpu local operations.
+ */
+#define percpu_cmpxchg_op(var, oval, nval) \
+({ \
+ typeof(var) pco_ret__; \
+ typeof(var) pco_old__ = (oval); \
+ typeof(var) pco_new__ = (nval); \
+ switch (sizeof(var)) { \
+ case 1: \
+ asm("cmpxchgb %2, "__percpu_arg(1) \
+ : "=a" (pco_ret__), "+m" (var) \
+ : "q" (pco_new__), "0" (pco_old__) \
+ : "memory"); \
+ break; \
+ case 2: \
+ asm("cmpxchgw %2, "__percpu_arg(1) \
+ : "=a" (pco_ret__), "+m" (var) \
+ : "r" (pco_new__), "0" (pco_old__) \
+ : "memory"); \
+ break; \
+ case 4: \
+ asm("cmpxchgl %2, "__percpu_arg(1) \
+ : "=a" (pco_ret__), "+m" (var) \
+ : "r" (pco_new__), "0" (pco_old__) \
+ : "memory"); \
+ break; \
+ case 8: \
+ asm("cmpxchgq %2, "__percpu_arg(1) \
+ : "=a" (pco_ret__), "+m" (var) \
+ : "r" (pco_new__), "0" (pco_old__) \
+ : "memory"); \
+ break; \
+ default: __bad_percpu_size(); \
+ } \
+ pco_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
@@ -300,6 +377,12 @@ do { \
#define __this_cpu_xor_1(pcp, val) percpu_to_op("xor", (pcp), val)
#define __this_cpu_xor_2(pcp, val) percpu_to_op("xor", (pcp), val)
#define __this_cpu_xor_4(pcp, val) percpu_to_op("xor", (pcp), val)
+/*
+ * Generic fallback operations for __this_cpu_xchg_[1-4] are okay and much
+ * faster than an xchg with forced lock semantics.
+ */
+#define __this_cpu_xchg_8(pcp, nval) percpu_xchg_op(pcp, nval)
+#define __this_cpu_cmpxchg_8(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval)
#define this_cpu_read_1(pcp) percpu_from_op("mov", (pcp), "m"(pcp))
#define this_cpu_read_2(pcp) percpu_from_op("mov", (pcp), "m"(pcp))
@@ -319,6 +402,11 @@ do { \
#define this_cpu_xor_1(pcp, val) percpu_to_op("xor", (pcp), val)
#define this_cpu_xor_2(pcp, val) percpu_to_op("xor", (pcp), val)
#define this_cpu_xor_4(pcp, val) percpu_to_op("xor", (pcp), val)
+#define this_cpu_xchg_1(pcp, nval) percpu_xchg_op(pcp, nval)
+#define this_cpu_xchg_2(pcp, nval) percpu_xchg_op(pcp, nval)
+#define this_cpu_xchg_4(pcp, nval) percpu_xchg_op(pcp, nval)
+#define this_cpu_xchg_8(pcp, nval) percpu_xchg_op(pcp, nval)
+#define this_cpu_cmpxchg_8(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval)
#define irqsafe_cpu_add_1(pcp, val) percpu_add_op((pcp), val)
#define irqsafe_cpu_add_2(pcp, val) percpu_add_op((pcp), val)
@@ -332,15 +420,32 @@ do { \
#define irqsafe_cpu_xor_1(pcp, val) percpu_to_op("xor", (pcp), val)
#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)
+#define irqsafe_cpu_xchg_1(pcp, nval) percpu_xchg_op(pcp, nval)
+#define irqsafe_cpu_xchg_2(pcp, nval) percpu_xchg_op(pcp, nval)
+#define irqsafe_cpu_xchg_4(pcp, nval) percpu_xchg_op(pcp, nval)
+#define irqsafe_cpu_xchg_8(pcp, nval) percpu_xchg_op(pcp, nval)
+#define irqsafe_cpu_cmpxchg_8(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval)
#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_cmpxchg_1(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval)
+#define __this_cpu_cmpxchg_2(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval)
+#define __this_cpu_cmpxchg_4(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval)
+
#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
+#define this_cpu_cmpxchg_1(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval)
+#define this_cpu_cmpxchg_2(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval)
+#define this_cpu_cmpxchg_4(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval)
+
+#define irqsafe_cpu_cmpxchg_1(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval)
+#define irqsafe_cpu_cmpxchg_2(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval)
+#define irqsafe_cpu_cmpxchg_4(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval)
+#endif /* !CONFIG_M386 */
+
/*
* Per cpu atomic 64 bit operations are only available under 64 bit.
* 32 bit must fall back to generic operations.
--
1.7.1
^ permalink raw reply related [flat|nested] 8+ messages in thread* [PATCH 6/6] cpuops: Use cmpxchg for xchg to avoid lock semantics
2010-12-18 15:24 [PATCHSET percpu#this_cpu_ops] add this_cpu_{add_return|[cmp]xchg}() operations Tejun Heo
` (4 preceding siblings ...)
2010-12-18 15:24 ` [PATCH 5/6] x86: this_cpu_cmpxchg and this_cpu_xchg operations Tejun Heo
@ 2010-12-18 15:24 ` Tejun Heo
2010-12-20 15:53 ` [PATCHSET percpu#this_cpu_ops] add this_cpu_{add_return|[cmp]xchg}() operations Pekka Enberg
6 siblings, 0 replies; 8+ messages in thread
From: Tejun Heo @ 2010-12-18 15:24 UTC (permalink / raw)
To: penberg, akpm, linux-kernel, eric.dumazet, hpa, mathieu.desnoyers
Cc: Christoph Lameter, Tejun Heo
From: Christoph Lameter <cl@linux.com>
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)
Signed-off-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
---
arch/x86/include/asm/percpu.h | 21 +++++++++++++++------
1 files changed, 15 insertions(+), 6 deletions(-)
diff --git a/arch/x86/include/asm/percpu.h b/arch/x86/include/asm/percpu.h
index b85ade5..8ee4516 100644
--- a/arch/x86/include/asm/percpu.h
+++ b/arch/x86/include/asm/percpu.h
@@ -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"); \
--
1.7.1
^ permalink raw reply related [flat|nested] 8+ messages in thread* Re: [PATCHSET percpu#this_cpu_ops] add this_cpu_{add_return|[cmp]xchg}() operations
2010-12-18 15:24 [PATCHSET percpu#this_cpu_ops] add this_cpu_{add_return|[cmp]xchg}() operations Tejun Heo
` (5 preceding siblings ...)
2010-12-18 15:24 ` [PATCH 6/6] cpuops: Use cmpxchg for xchg to avoid lock semantics Tejun Heo
@ 2010-12-20 15:53 ` Pekka Enberg
6 siblings, 0 replies; 8+ messages in thread
From: Pekka Enberg @ 2010-12-20 15:53 UTC (permalink / raw)
To: Tejun Heo; +Cc: akpm, linux-kernel, eric.dumazet, hpa, mathieu.desnoyers
On Sat, 2010-12-18 at 16:24 +0100, Tejun Heo wrote:
> 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.
Thanks. Christoph, mind sending me the slub patches? There's been so
many I've lost track which one is the latest.
Pekka
^ permalink raw reply [flat|nested] 8+ messages in thread