On 2017/06/06 06:25PM, Peter Zijlstra wrote: > On Tue, Jun 06, 2017 at 08:21:18PM +0530, Naveen N. Rao wrote: > > Many perf sideband events (context switches, namespaces, ...) are useful > > by themselves without the need for subscribing to any overflow events. > > However, it is not possible to subscribe for notifications when such > > records are logged into the ring buffer. Introduce IOC_COUNT_RECORDS as > > a way to request this. > > > > With IOC_COUNT_RECORDS set, IOC_REFRESH takes the number of records > > after which to generate a notification, rather than the number of > > overflow events. > > > > Signed-off-by: Naveen N. Rao <[email protected]> > > --- > > include/linux/perf_event.h | 1 + > > include/uapi/linux/perf_event.h | 1 + > > kernel/events/core.c | 16 +++++++++++++++- > > kernel/events/ring_buffer.c | 9 +++++++++ > > 4 files changed, 26 insertions(+), 1 deletion(-) > > > > diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h > > index 24a635887f28..016f2da2bba7 100644 > > --- a/include/linux/perf_event.h > > +++ b/include/linux/perf_event.h > > @@ -683,6 +683,7 @@ struct perf_event { > > struct irq_work pending; > > > > atomic_t event_limit; > > + bool count_records; > > This is an instant nack ;-) Never, that is _never_ use bool in composite > types.
Ouch! Sorry. I did briefly consider labeling this 'EARLY RFC', but that still wouldn't have done justice :/ Thanks for the review, - Naveen
