Messages in this thread | | | Subject | Re: [PATCH 3/5] perf: New PERF_EVENT_STATE_PAUSED event state | From | Peter Zijlstra <> | Date | Thu, 10 Jun 2010 12:55:17 +0200 |
| |
On Thu, 2010-06-10 at 05:49 +0200, Frederic Weisbecker wrote: > This brings a new PERF_EVENT_STATE_PAUSED state. It means the events > is enabled but we don't want it to run, it must be in the same state > than after a pmu->stop() call. So the event has been reserved and > allocated and it is ready to start after a pmu->start() call. > > It is deemed for hardware events when we want them to be reserved on > the cpu and ready to be started anytime. This is going to be useful > for the new context exclusion that will follow. > > Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> > Cc: Ingo Molnar <mingo@elte.hu> > Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> > Cc: Arnaldo Carvalho de Melo <acme@redhat.com> > Cc: Paul Mackerras <paulus@samba.org> > Cc: Stephane Eranian <eranian@google.com> > Cc: Cyrill Gorcunov <gorcunov@gmail.com> > Cc: Zhang Yanmin <yanmin_zhang@linux.intel.com> > Cc: Steven Rostedt <rostedt@goodmis.org> > --- > arch/x86/kernel/cpu/perf_event.c | 6 ++++-- > include/linux/perf_event.h | 3 ++- > kernel/perf_event.c | 7 ++++--- > 3 files changed, 10 insertions(+), 6 deletions(-) > > diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c > index f2da20f..9b0e52f 100644 > --- a/arch/x86/kernel/cpu/perf_event.c > +++ b/arch/x86/kernel/cpu/perf_event.c > @@ -839,7 +839,8 @@ void hw_perf_enable(void) > match_prev_assignment(hwc, cpuc, i)) > continue; > > - x86_pmu_stop(event); > + if (event->state != PERF_EVENT_STATE_PAUSED) > + x86_pmu_stop(event); > } > > for (i = 0; i < cpuc->n_events; i++) { > @@ -851,7 +852,8 @@ void hw_perf_enable(void) > else if (i < n_running) > continue; > > - x86_pmu_start(event); > + if (event->state != PERF_EVENT_STATE_PAUSED) > + x86_pmu_start(event); > } > cpuc->n_added = 0; > perf_events_lapic_init();
Shouldn't that latter be == PAUSED?
Also, you'll have to audit all struct pmu implementations that stop/disable or disable/disable is good.
Also, I'd rather keep the whole event->state knowledge in the generic code.
| |