[PATCH] remove get_cpu_ptr()

Ingo points out that it's unusable anyway, because with some configs the
get_cpu() is evaluated and with others it is not.

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

1 files changed, 0 insertions, 22 deletions

diff --git a/include/linux/percpu.h b/include/linux/percpu.h
index 6aedd71cc02c..a493e372201e 100644
--- a/include/linux/percpu.h
+++ b/include/linux/percpu.h
@@ -58,26 +58,4 @@ static inline void free_percpu(const void *ptr)
 #define alloc_percpu(type) \
 	((type *)(__alloc_percpu(sizeof(type), __alignof__(type))))
 
-/* 
- * Use these with alloc_percpu. If
- * 1. You want to operate on memory allocated by alloc_percpu (dereference
- *    and read/modify/write)  AND 
- * 2. You want "this cpu's version" of the object AND 
- * 3. You want to do this safely since:
- *    a. On multiprocessors, you don't want to switch between cpus after 
- *    you've read the current processor id due to preemption -- this would 
- *    take away the implicit  advantage to not have any kind of traditional 
- *    serialization for per-cpu data
- *    b. On uniprocessors, you don't want another kernel thread messing
- *    up with the same per-cpu data due to preemption
- *    
- * So, Use get_cpu_ptr to disable preemption and get pointer to the 
- * local cpu version of the per-cpu object. Use put_cpu_ptr to enable
- * preemption.  Operations on per-cpu data between get_ and put_ is
- * then considered to be safe. And ofcourse, "Thou shalt not sleep between 
- * get_cpu_ptr and put_cpu_ptr"
- */
-#define get_cpu_ptr(ptr) per_cpu_ptr(ptr, get_cpu())
-#define put_cpu_ptr(ptr) put_cpu()
-
 #endif /* __LINUX_PERCPU_H */