Commit a2e87d06 authored Apr 06, 2009 by Peter Zijlstra Committed by Ingo Molnar Apr 07, 2009

perf_counter: update mmap() counter read, take 2



Update the userspace read method.

Paul noted that:
 - userspace cannot observe ->lock & 1 on the same cpu.
 - we need a barrier() between reading ->lock and ->index
   to ensure we read them in that prticular order.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
LKML-Reference: <20090406094517.368446033@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

parent 92f22a38

include/linux/perf_counter.h

+10 −14

Original line number	Diff line number	Diff line
		@@ -170,13 +170,10 @@ struct perf_counter_mmap_page {
		* u32 seq;
		* s64 count;
		*
		* again:
		* do {
		* seq = pc->lock;
		* if (unlikely(seq & 1)) {
		* cpu_relax();
		* goto again;
		* }
		*
		* barrier()
		* if (pc->index) {
		* count = pmc_read(pc->index - 1);
		* count += pc->offset;
		@@ -184,8 +181,7 @@ struct perf_counter_mmap_page {
		* goto regular_read;
		*
		* barrier();
		* if (pc->lock != seq)
		* goto again;
		* } while (pc->lock != seq);
		*
		* NOTE: for obvious reason this only works on self-monitoring
		* processes.

Admin message