Re: [PATCH RFC 2/3] perf/x86/intel/pt: Add support for pause_resume()

[James Clark <james.clark@arm.com>]

On 23/11/2023 12:18, Adrian Hunter wrote:
> Prevent tracing to start if aux_paused.
> 
> Implement pause_resume() callback. When aux_paused, stop tracing. When
> not aux_paused, only start tracing if it isn't currently meant to be
> stopped.
> 
> Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
> ---
>  arch/x86/events/intel/pt.c | 12 ++++++++++++
>  1 file changed, 12 insertions(+)
> 
> diff --git a/arch/x86/events/intel/pt.c b/arch/x86/events/intel/pt.c
> index 42a55794004a..aa883b64814a 100644
> --- a/arch/x86/events/intel/pt.c
> +++ b/arch/x86/events/intel/pt.c
> @@ -418,6 +418,9 @@ static void pt_config_start(struct perf_event *event)
>  	struct pt *pt = this_cpu_ptr(&pt_ctx);
>  	u64 ctl = event->hw.config;
>  
> +	if (event->aux_paused)
> +		return;
> +
>  	ctl |= RTIT_CTL_TRACEEN;
>  	if (READ_ONCE(pt->vmx_on))
>  		perf_aux_output_flag(&pt->handle, PERF_AUX_FLAG_PARTIAL);
> @@ -1563,6 +1566,14 @@ EXPORT_SYMBOL_GPL(intel_pt_handle_vmx);
>   * PMU callbacks
>   */
>  
> +static void pt_event_pause_resume(struct perf_event *event)
> +{
> +	if (event->aux_paused)
> +		pt_config_stop(event);
> +	else if (!event->hw.state)
> +		pt_config_start(event);
> +}

It seems like having a single pause/resume callback rather than separate
pause and resume ones pushes some of the event state management into the
individual drivers and would be prone to code duplication and divergent
behavior.

Would it be possible to move the conditions from here into the core code
and call separate functions instead?

> +
>  static void pt_event_start(struct perf_event *event, int mode)
>  {
>  	struct hw_perf_event *hwc = &event->hw;
> @@ -1798,6 +1809,7 @@ static __init int pt_init(void)
>  	pt_pmu.pmu.del			 = pt_event_del;
>  	pt_pmu.pmu.start		 = pt_event_start;
>  	pt_pmu.pmu.stop			 = pt_event_stop;
> +	pt_pmu.pmu.pause_resume		 = pt_event_pause_resume;

The general idea seems ok to me. Is there a reason to not use the
existing start() stop() callbacks, rather than adding a new one?

I assume it's intended to be something like an optimisation where you
can turn it on and off without having to do the full setup, teardown and
emit an AUX record because you know the process being traced never gets
switched out?

Could you make it so that it works out of the box, with the option of
later optimisation if you do something like this (not here but something
like this in events/core.c):

  /* Use specialised pause/resume if it exists, otherwise use more
   * expensive start/stop.
   */
  if (pmu->pause_resume)
	pmu->pause_resume(...)
  else
        pmu->stop(...)


>  	pt_pmu.pmu.snapshot_aux		 = pt_event_snapshot_aux;
>  	pt_pmu.pmu.read			 = pt_event_read;
>  	pt_pmu.pmu.setup_aux		 = pt_buffer_setup_aux;