oprofile: don't request cache line alignment for cpu_buffer
Eric Dumazet [Wed, 14 May 2008 23:05:31 +0000 (16:05 -0700)]
Alignment was previously requested because cpu_buffer was an [NR_CPUS]
array, to avoid cache line sharing between CPUS.

After commit 608dfddd845da5ab6accef70154c8910529699f7 (oprofile: change
cpu_buffer from array to per_cpu variable ), we dont need to force an
alignement anymore since cpu_buffer sits in per_cpu zone.

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Cc: Mike Travis <travis@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

drivers/oprofile/cpu_buffer.c
drivers/oprofile/cpu_buffer.h

index efcbf4b..2450b3a 100644 (file)
@@ -27,7 +27,7 @@
 #include "buffer_sync.h"
 #include "oprof.h"
 
-DEFINE_PER_CPU_SHARED_ALIGNED(struct oprofile_cpu_buffer, cpu_buffer);
+DEFINE_PER_CPU(struct oprofile_cpu_buffer, cpu_buffer);
 
 static void wq_sync_buffer(struct work_struct *work);
 
index 1358817..c3e366b 100644 (file)
@@ -46,7 +46,7 @@ struct oprofile_cpu_buffer {
        unsigned long sample_invalid_eip;
        int cpu;
        struct delayed_work work;
-} ____cacheline_aligned;
+};
 
 DECLARE_PER_CPU(struct oprofile_cpu_buffer, cpu_buffer);