diff options
| author | Will Deacon <will.deacon@arm.com> | 2011-10-03 21:30:53 +0400 |
|---|---|---|
| committer | Russell King <rmk+kernel@arm.linux.org.uk> | 2011-10-15 14:04:22 +0400 |
| commit | 29a541f6c1f6e4a85628bb86071b9e72c9f8be2c (patch) | |
| tree | 9f132fd63c08266901f3427d624891ac7e2da7b7 /arch/arm | |
| parent | 002ea9eefec98dada56fd5f8e432a4e8570c2a26 (diff) | |
| download | linux-29a541f6c1f6e4a85628bb86071b9e72c9f8be2c.tar.xz | |
ARM: 7117/1: perf: fix HW_CACHE_* events on Cortex-A9
Using COHERENT_LINE_{MISS,HIT} for cache misses and references
respectively is completely wrong. Instead, use the L1D events which
are a better and more useful approximation despite ignoring instruction
traffic.
Reported-by: Alasdair Grant <alasdair.grant@arm.com>
Reported-by: Matt Horsnell <matt.horsnell@arm.com>
Reported-by: Michael Williams <michael.williams@arm.com>
Cc: stable@kernel.org
Cc: Jean Pihet <j-pihet@ti.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Diffstat (limited to 'arch/arm')
| -rw-r--r-- | arch/arm/kernel/perf_event_v7.c | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/arch/arm/kernel/perf_event_v7.c b/arch/arm/kernel/perf_event_v7.c index 4c851834f68e..6be3e2e4d838 100644 --- a/arch/arm/kernel/perf_event_v7.c +++ b/arch/arm/kernel/perf_event_v7.c @@ -321,8 +321,8 @@ static const unsigned armv7_a9_perf_map[PERF_COUNT_HW_MAX] = { [PERF_COUNT_HW_CPU_CYCLES] = ARMV7_PERFCTR_CPU_CYCLES, [PERF_COUNT_HW_INSTRUCTIONS] = ARMV7_PERFCTR_INST_OUT_OF_RENAME_STAGE, - [PERF_COUNT_HW_CACHE_REFERENCES] = ARMV7_PERFCTR_COHERENT_LINE_HIT, - [PERF_COUNT_HW_CACHE_MISSES] = ARMV7_PERFCTR_COHERENT_LINE_MISS, + [PERF_COUNT_HW_CACHE_REFERENCES] = ARMV7_PERFCTR_DCACHE_ACCESS, + [PERF_COUNT_HW_CACHE_MISSES] = ARMV7_PERFCTR_DCACHE_REFILL, [PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = ARMV7_PERFCTR_PC_WRITE, [PERF_COUNT_HW_BRANCH_MISSES] = ARMV7_PERFCTR_PC_BRANCH_MIS_PRED, [PERF_COUNT_HW_BUS_CYCLES] = ARMV7_PERFCTR_CLOCK_CYCLES, |