Re: [PATCH 02/18] perf: Add persistent event facilities

From: Ingo Molnar
Date: Tue Apr 26 2011 - 08:45:37 EST



* Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:

> On Sat, 2011-04-23 at 18:28 +0200, Borislav Petkov wrote:
> > From: Borislav Petkov <borislav.petkov@xxxxxxx>
> >
> > Add a barebones implementation for registering persistent events with
> > perf. For that, we don't destroy the buffers when they're unmapped and
> > we map them read-only so that multiple agents can access them.
> >
> > Signed-off-by: Borislav Petkov <borislav.petkov@xxxxxxx>
> > ---
> > include/linux/perf_event.h | 22 ++++++++++++++++-
> > kernel/events/Makefile | 2 +-
> > kernel/events/core.c | 30 ++++++++++++++++++----
> > kernel/events/persistent.c | 56 ++++++++++++++++++++++++++++++++++++++++++++
> > 4 files changed, 102 insertions(+), 8 deletions(-)
> > create mode 100644 kernel/events/persistent.c
> >
> > diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h
> > index ee9f1e7..37bfae1 100644
> > --- a/include/linux/perf_event.h
> > +++ b/include/linux/perf_event.h
> > @@ -216,8 +216,9 @@ struct perf_event_attr {
> > precise_ip : 2, /* skid constraint */
> > mmap_data : 1, /* non-exec mmap data */
> > sample_id_all : 1, /* sample_type all events */
> > + persistent : 1, /* event always on */
>
> But how will you find it again?
>
> Persistent yells filesystem to me, and while we didn't use relayfs for
> various reasons, I think we do need something like it at the current stage.

eventfs or sysfs?

We need VFS space for event hierarchy enumeration anyway (and move the current
debugfs bits there). Persistent events could show up in that hiearachy as well.

Thanks,

Ingo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/