lkml.org 
[lkml]   [2017]   [Aug]   [4]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [PATCH v6 2/3]: perf/core: use context tstamp_data for skipped events on mux interrupt
From
Date
On 04.08.2017 15:51, Peter Zijlstra wrote:
> On Fri, Aug 04, 2017 at 02:35:34PM +0200, Peter Zijlstra wrote:
>> Something like:
>>
>> __update_state_and_time(event, new_state)
>> {
>> u64 delta, now = perf_event_time(event);
>> int old_state = event->state;
>>
>> event->tstamp = now;
>> event->state = new_state;
>>
>> delta = now - event->tstamp;
> Obv should go above the tstamp assignment
>
>> switch (state) {
>> case STATE_ACTIVE:
>> WARN_ON_ONCE(old_state != STATE_INACTIVE);
>> event->total_time_enabled += delta;
>> break;
>>
>> case STATE_INACTIVE:
>> switch (old_state) {
>> case STATE_OFF:
>> /* ignore the OFF -> INACTIVE period */
>> break;
>>
>> case STATE_ACTIVE:
>> event->total_time_enabled += delta;
>> event->total_time_running += delta;
>> break;
>>
>> default:
>> WARN_ONCE();
>> }
>> break;
>>
>> case STATE_OFF:
>> WARN_ON_ONCE(old_state != STATE_INACTIVE)
>> event->total_time_enabled += delta;
>> break;
>> }
>> }
>
> So that's a straight fwd state machine that deals with:
>
> OFF <-> INACTIVE <-> ACTIVE
>
> but I think something like:
>
> __update_state_and_time(event, new_state)
> {
> u64 delta, new = perf_event_time(event);
> int old_state = event->state;
>
> delta = now - event->tstamp;
> event->tstamp = now;
> event->state = new_state;
>
> if (old_state == STATE_OFF)
> return;
>
> event->total_time_enabled += delta;
>
> if (old_state == STATE_ACTIVE)
> event->total_time_running += delta;
> }
>
> is equivalent and generates smaller code.. but again, double check (also
> it doesn't validate the state transitions).

Accepted.

>

\
 
 \ /
  Last update: 2017-08-04 16:27    [W:0.111 / U:2.756 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site