Quoting Tvrtko Ursulin (2018-05-01 16:41:15)
>
> On 01/05/2018 16:26, Tvrtko Ursulin wrote:
> >
> > On 01/05/2018 16:13, Chris Wilson wrote:
> >> Include a bool to show whether the request is complete in every
> >> tracepoint. This especially helps when tracing the flow of requests
> >> through the HW.
> >>
> >> v2: Limit to in/out HW events, not all general request tracepoints.
> >>
> >> Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
> >> ---
> >> drivers/gpu/drm/i915/i915_trace.h | 17 +++++++++++------
> >> drivers/gpu/drm/i915/intel_guc_submission.c | 2 +-
> >> drivers/gpu/drm/i915/intel_lrc.c | 2 +-
> >> 3 files changed, 13 insertions(+), 8 deletions(-)
> >>
> >> diff --git a/drivers/gpu/drm/i915/i915_trace.h
> >> b/drivers/gpu/drm/i915/i915_trace.h
> >> index 408827bf5d96..66956af25da3 100644
> >> --- a/drivers/gpu/drm/i915/i915_trace.h
> >> +++ b/drivers/gpu/drm/i915/i915_trace.h
> >> @@ -691,6 +691,8 @@ DECLARE_EVENT_CLASS(i915_request_hw,
> >> __field(u32, seqno)
> >> __field(u32, global_seqno)
> >> __field(u32, port)
> >> + __field(u32, prio)
> >> + __field(u32, completed)
> >> ),
> >> TP_fast_assign(
> >> @@ -701,12 +703,15 @@ DECLARE_EVENT_CLASS(i915_request_hw,
> >> __entry->seqno = rq->fence.seqno;
> >> __entry->global_seqno = rq->global_seqno;
> >> __entry->port = port;
> >> + __entry->prio = rq->sched.attr.priority;
> >> + __entry->completed = i915_request_completed(rq);
> >> ),
> >> - TP_printk("dev=%u, hw_id=%u, ring=%u, ctx=%u, seqno=%u,
> >> global=%u, port=%u",
> >> + TP_printk("dev=%u, hw_id=%u, ring=%u, ctx=%u, seqno=%u,
> >> prio=%u, global=%u, port=%u, completed?=%d",
> >> __entry->dev, __entry->hw_id, __entry->ring,
> >> - __entry->ctx, __entry->seqno,
> >> - __entry->global_seqno, __entry->port)
> >> + __entry->ctx, __entry->seqno, __entry->prio,
> >> + __entry->global_seqno, __entry->port,
> >> + __entry->completed)
> >> );
> >> DEFINE_EVENT(i915_request_hw, i915_request_in,
> >> @@ -714,9 +719,9 @@ DEFINE_EVENT(i915_request_hw, i915_request_in,
> >> TP_ARGS(rq, port)
> >> );
> >> -DEFINE_EVENT(i915_request, i915_request_out,
> >> - TP_PROTO(struct i915_request *rq),
> >> - TP_ARGS(rq)
> >> +DEFINE_EVENT(i915_request_hw, i915_request_out,
> >> + TP_PROTO(struct i915_request *rq, unsigned int port),
> >> + TP_ARGS(rq, port)
> >> );
> >> #else
> >> #if !defined(TRACE_HEADER_MULTI_READ)
> >> diff --git a/drivers/gpu/drm/i915/intel_guc_submission.c
> >> b/drivers/gpu/drm/i915/intel_guc_submission.c
> >> index c6bb5bebddfc..bd440687892a 100644
> >> --- a/drivers/gpu/drm/i915/intel_guc_submission.c
> >> +++ b/drivers/gpu/drm/i915/intel_guc_submission.c
> >> @@ -762,7 +762,7 @@ static void guc_submission_tasklet(unsigned long
> >> data)
> >> rq = port_request(port);
> >> while (rq && i915_request_completed(rq)) {
> >> - trace_i915_request_out(rq);
> >> + trace_i915_request_out(rq, 0);
> >> i915_request_put(rq);
> >> port = execlists_port_complete(execlists, port);
> >> diff --git a/drivers/gpu/drm/i915/intel_lrc.c
> >> b/drivers/gpu/drm/i915/intel_lrc.c
> >> index ffc74f562f77..396d9596d0aa 100644
> >> --- a/drivers/gpu/drm/i915/intel_lrc.c
> >> +++ b/drivers/gpu/drm/i915/intel_lrc.c
> >> @@ -402,7 +402,7 @@ execlists_context_schedule_out(struct i915_request
> >> *rq, unsigned long status)
> >> {
> >> intel_engine_context_out(rq->engine);
> >> execlists_context_status_change(rq, status);
> >> - trace_i915_request_out(rq);
> >> + trace_i915_request_out(rq, 0);
> >> }
> >> static void
> >>
> >
> > Another one! :) For simplicity:
> >
> > Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@xxxxxxxxx>
>
> Well actually on a second thought I think it would be tidier to define
> in and out separately so not to pollute the logs with pointless fields.
> If you are busy I am happy to do that.
If your fingers are on the keyboard already, go for it.
-Chris
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
