1 files changed, 26 insertions, 14 deletions

diff --git a/include/asm-i386/local.h b/include/asm-i386/local.h
index e67fa08260f..12060e22f7e 100644
--- a/include/asm-i386/local.h
+++ b/include/asm-i386/local.h
@@ -17,32 +17,30 @@ static __inline__ void local_inc(local_t *v)
 {
 	__asm__ __volatile__(
 		"incl %0"
-		:"=m" (v->counter)
-		:"m" (v->counter));
+		:"+m" (v->counter));
 }
 
 static __inline__ void local_dec(local_t *v)
 {
 	__asm__ __volatile__(
 		"decl %0"
-		:"=m" (v->counter)
-		:"m" (v->counter));
+		:"+m" (v->counter));
 }
 
 static __inline__ void local_add(long i, local_t *v)
 {
 	__asm__ __volatile__(
 		"addl %1,%0"
-		:"=m" (v->counter)
-		:"ir" (i), "m" (v->counter));
+		:"+m" (v->counter)
+		:"ir" (i));
 }
 
 static __inline__ void local_sub(long i, local_t *v)
 {
 	__asm__ __volatile__(
 		"subl %1,%0"
-		:"=m" (v->counter)
-		:"ir" (i), "m" (v->counter));
+		:"+m" (v->counter)
+		:"ir" (i));
 }
 
 /* On x86, these are no better than the atomic variants. */
@@ -55,12 +53,26 @@ static __inline__ void local_sub(long i, local_t *v)
  * much more efficient than these naive implementations.  Note they take
  * a variable, not an address.
  */
-#define cpu_local_read(v)	local_read(&__get_cpu_var(v))
-#define cpu_local_set(v, i)	local_set(&__get_cpu_var(v), (i))
-#define cpu_local_inc(v)	local_inc(&__get_cpu_var(v))
-#define cpu_local_dec(v)	local_dec(&__get_cpu_var(v))
-#define cpu_local_add(i, v)	local_add((i), &__get_cpu_var(v))
-#define cpu_local_sub(i, v)	local_sub((i), &__get_cpu_var(v))
+
+/* Need to disable preemption for the cpu local counters otherwise we could
+   still access a variable of a previous CPU in a non atomic way. */
+#define cpu_local_wrap_v(v)	 	\
+	({ local_t res__;		\
+	   preempt_disable(); 		\
+	   res__ = (v);			\
+	   preempt_enable();		\
+	   res__; })
+#define cpu_local_wrap(v)		\
+	({ preempt_disable();		\
+	   v;				\
+	   preempt_enable(); })		\
+
+#define cpu_local_read(v)    cpu_local_wrap_v(local_read(&__get_cpu_var(v)))
+#define cpu_local_set(v, i)  cpu_local_wrap(local_set(&__get_cpu_var(v), (i)))
+#define cpu_local_inc(v)     cpu_local_wrap(local_inc(&__get_cpu_var(v)))
+#define cpu_local_dec(v)     cpu_local_wrap(local_dec(&__get_cpu_var(v)))
+#define cpu_local_add(i, v)  cpu_local_wrap(local_add((i), &__get_cpu_var(v)))
+#define cpu_local_sub(i, v)  cpu_local_wrap(local_sub((i), &__get_cpu_var(v)))
 
 #define __cpu_local_inc(v)	cpu_local_inc(v)
 #define __cpu_local_dec(v)	cpu_local_dec(v)