2007-08-09 14:07:41

by Chris Snook

[permalink] [raw]
Subject: [PATCH 16/24] make atomic_read() behave consistently on s390

From: Chris Snook <[email protected]>

Make atomic[64]_read() volatile on s390, to ensure memory is actually read
each time.

Signed-off-by: Chris Snook <[email protected]>

--- linux-2.6.23-rc2-orig/include/asm-s390/atomic.h 2007-08-08 17:48:53.000000000 -0400
+++ linux-2.6.23-rc2/include/asm-s390/atomic.h 2007-08-09 07:18:56.000000000 -0400
@@ -67,7 +67,11 @@ typedef struct {

#endif /* __GNUC__ */

-#define atomic_read(v) ((v)->counter)
+/*
+ * Casting to volatile here minimizes the need for barriers,
+ * without having to declare the type itself as volatile.
+ */
+#define atomic_read(v) (*(volatile int *)&(v)->counter)
#define atomic_set(v,i) (((v)->counter) = (i))

static __inline__ int atomic_add_return(int i, atomic_t * v)
@@ -182,7 +186,7 @@ typedef struct {

#endif /* __GNUC__ */

-#define atomic64_read(v) ((v)->counter)
+#define atomic64_read(v) (*(volatile long long *)&(v)->counter)
#define atomic64_set(v,i) (((v)->counter) = (i))

static __inline__ long long atomic64_add_return(long long i, atomic64_t * v)


2007-08-09 14:33:58

by Martin Schwidefsky

[permalink] [raw]
Subject: Re: [PATCH 16/24] make atomic_read() behave consistently on s390

On Thu, 2007-08-09 at 10:06 -0400, Chris Snook wrote:
> From: Chris Snook <[email protected]>
>
> Make atomic[64]_read() volatile on s390, to ensure memory is actually read
> each time.
>
> Signed-off-by: Chris Snook <[email protected]>

Acked-by: Martin Schwidefsky <[email protected]>

--
blue skies,
Martin.

"Reality continues to ruin my life." - Calvin.