diff options
author | David S. Miller <davem@davemloft.net> | 2012-10-16 13:05:25 -0700 |
---|---|---|
committer | David S. Miller <davem@davemloft.net> | 2012-10-16 13:05:25 -0700 |
commit | e793d8c6740f8fe704fa216e95685f4d92c4c4b9 (patch) | |
tree | f35eaea2baa8b83f4ffed9c12f55dbd2e32abb25 /arch/sparc | |
parent | 916ca14aaf12a7191118adb51bb95e3c7866380d (diff) | |
download | lwn-e793d8c6740f8fe704fa216e95685f4d92c4c4b9.tar.gz lwn-e793d8c6740f8fe704fa216e95685f4d92c4c4b9.zip |
sparc64: Fix bit twiddling in sparc_pmu_enable_event().
There was a serious disconnect in the logic happening in
sparc_pmu_disable_event() vs. sparc_pmu_enable_event().
Event disable is implemented by programming a NOP event into the PCR.
However, event enable was not reversing this operation. Instead, it
was setting the User/Priv/Hypervisor trace enable bits.
That's not sparc_pmu_enable_event()'s job, that's what
sparc_pmu_enable() and sparc_pmu_disable() do .
The intent of sparc_pmu_enable_event() is clear, since it first clear
out the event type encoding field. So fix this by OR'ing in the event
encoding rather than the trace enable bits.
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'arch/sparc')
-rw-r--r-- | arch/sparc/kernel/perf_event.c | 6 |
1 files changed, 4 insertions, 2 deletions
diff --git a/arch/sparc/kernel/perf_event.c b/arch/sparc/kernel/perf_event.c index 9e96f849a744..885a8af74064 100644 --- a/arch/sparc/kernel/perf_event.c +++ b/arch/sparc/kernel/perf_event.c @@ -817,15 +817,17 @@ static u64 nop_for_index(int idx) static inline void sparc_pmu_enable_event(struct cpu_hw_events *cpuc, struct hw_perf_event *hwc, int idx) { - u64 val, mask = mask_for_index(idx); + u64 enc, val, mask = mask_for_index(idx); int pcr_index = 0; if (sparc_pmu->num_pcrs > 1) pcr_index = idx; + enc = perf_event_get_enc(cpuc->events[idx]); + val = cpuc->pcr[pcr_index]; val &= ~mask; - val |= hwc->config; + val |= event_encoding(enc, idx); cpuc->pcr[pcr_index] = val; pcr_ops->write_pcr(pcr_index, cpuc->pcr[pcr_index]); |