Skip to contentSkip to conversation

perf_counter: x86: More accurate counter update

tip:perfcounters/core tip-bot for Peter Zijlstra 1 participant 1 message May 15, 2009 x86

lore
Share & tools
lore
#0 lore raw
Commit-ID:  ec3232bdf8518bea8410f0027f870b24d3aa8753
Gitweb:     http://git.kernel.org/tip/ec3232bdf8518bea8410f0027f870b24d3aa8753
Author:     Peter Zijlstra <[email protected]>
AuthorDate: Wed, 13 May 2009 09:45:19 +0200
Committer:  Ingo Molnar <[email protected]>
CommitDate: Fri, 15 May 2009 09:46:54 +0200

perf_counter: x86: More accurate counter update

Take the counter width into account instead of assuming 32 bits.

In particular Nehalem has 44 bit wide counters, and all arithmetics should happen on a 44-bit signed integer basis.

[ Impact: fix rare event imprecision, warning message on Nehalem ]

Cc Paul Mackerras <[email protected]>
Cc Corey Ashford <[email protected]>

LKML-Reference: <new-submission>

---
 arch/x86/kernel/cpu/perf_counter.c |    9 ++++++---
 1 files changed, 6 insertions(+), 3 deletions(-)
arch/x86/kernel/cpu/perf_counter.c +6 -3
index f7772ff..3a92a2b 100644
@@ -138,7 +138,9 @@ static u64
 x86_perf_counter_update(struct perf_counter *counter,
 			struct hw_perf_counter *hwc, int idx)
 {
-	u64 prev_raw_count, new_raw_count, delta;
+	int shift = 64 - x86_pmu.counter_bits;
+	u64 prev_raw_count, new_raw_count;
+	s64 delta;
 
 	/*
 	 * Careful: an NMI might modify the previous counter value.
@@ -161,9 +163,10 @@ again:
 	 * (counter-)time and add that to the generic counter.
 	 *
 	 * Careful, not all hw sign-extends above the physical width
-	 * of the count, so we do that by clipping the delta to 32 bits:
+	 * of the count.
 	 */
-	delta = (u64)(u32)((s32)new_raw_count - (s32)prev_raw_count);
+	delta = (new_raw_count << shift) - (prev_raw_count << shift);
+	delta >>= shift;
 
 	atomic64_add(delta, &counter->count);
 	atomic64_sub(delta, &hwc->period_left);