From: Olof Johansson <[EMAIL PROTECTED]>
For prefetches of NULL (as when walking a short linked list), PPC64 will in
some cases take a performance hit. The hardware needs to do the TLB walk,
and said walk will always miss, which means (up to) two L2 misses as
penalty. This seems to hurt overall performance, so for NULL pointers skip
the prefetch alltogether.
Signed-off-by: Olof Johansson <[EMAIL PROTECTED]>
Signed-off-by: Andrew Morton <[EMAIL PROTECTED]>
---
25-akpm/include/asm-ppc64/processor.h | 6 ++++++
1 files changed, 6 insertions(+)
diff -puN include/asm-ppc64/processor.h~ppc64-no-prefetch-for-null-pointers
include/asm-ppc64/processor.h
--- 25/include/asm-ppc64/processor.h~ppc64-no-prefetch-for-null-pointers
2005-04-12 03:21:16.364649096 -0700
+++ 25-akpm/include/asm-ppc64/processor.h 2005-04-12 03:21:16.367648640
-0700
@@ -642,11 +642,17 @@ static inline unsigned long __pack_fe01(
static inline void prefetch(const void *x)
{
+ if (unlikely(!x))
+ return;
+
__asm__ __volatile__ ("dcbt 0,%0" : : "r" (x));
}
static inline void prefetchw(const void *x)
{
+ if (unlikely(!x))
+ return;
+
__asm__ __volatile__ ("dcbtst 0,%0" : : "r" (x));
}
_
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
