[Intel-gfx] [PATCH v2] drm/i915/execlists: Track begin/end of execlists submission sequences

Mika Kuoppala mika.kuoppala at linux.intel.com
Tue Apr 3 11:17:57 UTC 2018


Chris Wilson <chris at chris-wilson.co.uk> writes:

> We would like to start doing some bookkeeping at the beginning, between
> contexts and at the end of execlists submission. We already mark the
> beginning and end using EXECLISTS_ACTIVE_USER, to provide an indication
> when the HW is idle. This give us a pair of sequence points we can then
> expand on for further bookkeeping.
>
> v2: Refactor guc submission to share the same begin/end.
>
> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> Cc: Mika Kuoppala <mika.kuoppala at linux.intel.com>
> Cc: Francisco Jerez <currojerez at riseup.net>
> Reviewed-by: Francisco Jerez <currojerez at riseup.net> #v1
> ---
>  drivers/gpu/drm/i915/intel_guc_submission.c | 17 ++++++----
>  drivers/gpu/drm/i915/intel_lrc.c            | 50 ++++++++++++++++++++++-------
>  drivers/gpu/drm/i915/intel_ringbuffer.h     | 15 ++++++++-
>  3 files changed, 63 insertions(+), 19 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_guc_submission.c b/drivers/gpu/drm/i915/intel_guc_submission.c
> index 207cda062626..749f27916a02 100644
> --- a/drivers/gpu/drm/i915/intel_guc_submission.c
> +++ b/drivers/gpu/drm/i915/intel_guc_submission.c
> @@ -728,7 +728,7 @@ static void guc_dequeue(struct intel_engine_cs *engine)
>  	execlists->first = rb;
>  	if (submit) {
>  		port_assign(port, last);
> -		execlists_set_active(execlists, EXECLISTS_ACTIVE_USER);
> +		execlists_user_begin(execlists, execlists->port);
>  		guc_submit(engine);
>  	}
>  
> @@ -748,17 +748,20 @@ static void guc_submission_tasklet(unsigned long data)
>  	struct execlist_port *port = execlists->port;
>  	struct i915_request *rq;
>  
> -	rq = port_request(&port[0]);
> +	rq = port_request(port);
>  	while (rq && i915_request_completed(rq)) {
>  		trace_i915_request_out(rq);
>  		i915_request_put(rq);
>  
> -		execlists_port_complete(execlists, port);
> -
> -		rq = port_request(&port[0]);
> +		port = execlists_port_complete(execlists, port);
> +		if (port_isset(port)) {
> +			execlists_user_begin(execlists, port);
> +			rq = port_request(port);
> +		} else {
> +			execlists_user_end(execlists);
> +			rq = NULL;

I did ponder doing the user_begin/end pair in complete
but better not to hide it in there as we might end up
doing completes without notifying users.

Reviewed-by: Mika Kuoppala <mika.kuoppala at linux.intel.com>

> +		}
>  	}
> -	if (!rq)
> -		execlists_clear_active(execlists, EXECLISTS_ACTIVE_USER);
>  
>  	if (execlists_is_active(execlists, EXECLISTS_ACTIVE_PREEMPT) &&
>  	    intel_read_status_page(engine, I915_GEM_HWS_PREEMPT_INDEX) ==
> diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
> index f60b61bf8b3b..4d08875422b6 100644
> --- a/drivers/gpu/drm/i915/intel_lrc.c
> +++ b/drivers/gpu/drm/i915/intel_lrc.c
> @@ -374,6 +374,19 @@ execlists_context_status_change(struct i915_request *rq, unsigned long status)
>  				   status, rq);
>  }
>  
> +inline void
> +execlists_user_begin(struct intel_engine_execlists *execlists,
> +		     const struct execlist_port *port)
> +{
> +	execlists_set_active_once(execlists, EXECLISTS_ACTIVE_USER);
> +}
> +
> +inline void
> +execlists_user_end(struct intel_engine_execlists *execlists)
> +{
> +	execlists_clear_active(execlists, EXECLISTS_ACTIVE_USER);
> +}
> +
>  static inline void
>  execlists_context_schedule_in(struct i915_request *rq)
>  {
> @@ -711,7 +724,7 @@ static void execlists_dequeue(struct intel_engine_cs *engine)
>  	spin_unlock_irq(&engine->timeline->lock);
>  
>  	if (submit) {
> -		execlists_set_active(execlists, EXECLISTS_ACTIVE_USER);
> +		execlists_user_begin(execlists, execlists->port);
>  		execlists_submit_ports(engine);
>  	}
>  
> @@ -742,7 +755,7 @@ execlists_cancel_port_requests(struct intel_engine_execlists * const execlists)
>  		port++;
>  	}
>  
> -	execlists_clear_active(execlists, EXECLISTS_ACTIVE_USER);
> +	execlists_user_end(execlists);
>  }
>  
>  static void clear_gtiir(struct intel_engine_cs *engine)
> @@ -873,7 +886,7 @@ static void execlists_submission_tasklet(unsigned long data)
>  {
>  	struct intel_engine_cs * const engine = (struct intel_engine_cs *)data;
>  	struct intel_engine_execlists * const execlists = &engine->execlists;
> -	struct execlist_port * const port = execlists->port;
> +	struct execlist_port *port = execlists->port;
>  	struct drm_i915_private *dev_priv = engine->i915;
>  	bool fw = false;
>  
> @@ -1012,10 +1025,28 @@ static void execlists_submission_tasklet(unsigned long data)
>  
>  			GEM_BUG_ON(count == 0);
>  			if (--count == 0) {
> +				/*
> +				 * On the final event corresponding to the
> +				 * submission of this context, we expect either
> +				 * an element-switch event or a completion
> +				 * event (and on completion, the active-idle
> +				 * marker). No more preemptions, lite-restore
> +				 * or otherwise.
> +				 */
>  				GEM_BUG_ON(status & GEN8_CTX_STATUS_PREEMPTED);
>  				GEM_BUG_ON(port_isset(&port[1]) &&
>  					   !(status & GEN8_CTX_STATUS_ELEMENT_SWITCH));
> +				GEM_BUG_ON(!port_isset(&port[1]) &&
> +					   !(status & GEN8_CTX_STATUS_ACTIVE_IDLE));
> +
> +				/*
> +				 * We rely on the hardware being strongly
> +				 * ordered, that the breadcrumb write is
> +				 * coherent (visible from the CPU) before the
> +				 * user interrupt and CSB is processed.
> +				 */
>  				GEM_BUG_ON(!i915_request_completed(rq));
> +
>  				execlists_context_schedule_out(rq);
>  				trace_i915_request_out(rq);
>  				i915_request_put(rq);
> @@ -1023,17 +1054,14 @@ static void execlists_submission_tasklet(unsigned long data)
>  				GEM_TRACE("%s completed ctx=%d\n",
>  					  engine->name, port->context_id);
>  
> -				execlists_port_complete(execlists, port);
> +				port = execlists_port_complete(execlists, port);
> +				if (port_isset(port))
> +					execlists_user_begin(execlists, port);
> +				else
> +					execlists_user_end(execlists);
>  			} else {
>  				port_set(port, port_pack(rq, count));
>  			}
> -
> -			/* After the final element, the hw should be idle */
> -			GEM_BUG_ON(port_count(port) == 0 &&
> -				   !(status & GEN8_CTX_STATUS_ACTIVE_IDLE));
> -			if (port_count(port) == 0)
> -				execlists_clear_active(execlists,
> -						       EXECLISTS_ACTIVE_USER);
>  		}
>  
>  		if (head != execlists->csb_head) {
> diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
> index a02c7b3b9d55..40461e29cdab 100644
> --- a/drivers/gpu/drm/i915/intel_ringbuffer.h
> +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
> @@ -638,6 +638,13 @@ execlists_set_active(struct intel_engine_execlists *execlists,
>  	__set_bit(bit, (unsigned long *)&execlists->active);
>  }
>  
> +static inline bool
> +execlists_set_active_once(struct intel_engine_execlists *execlists,
> +			  unsigned int bit)
> +{
> +	return !__test_and_set_bit(bit, (unsigned long *)&execlists->active);
> +}
> +
>  static inline void
>  execlists_clear_active(struct intel_engine_execlists *execlists,
>  		       unsigned int bit)
> @@ -652,6 +659,10 @@ execlists_is_active(const struct intel_engine_execlists *execlists,
>  	return test_bit(bit, (unsigned long *)&execlists->active);
>  }
>  
> +void execlists_user_begin(struct intel_engine_execlists *execlists,
> +			  const struct execlist_port *port);
> +void execlists_user_end(struct intel_engine_execlists *execlists);
> +
>  void
>  execlists_cancel_port_requests(struct intel_engine_execlists * const execlists);
>  
> @@ -664,7 +675,7 @@ execlists_num_ports(const struct intel_engine_execlists * const execlists)
>  	return execlists->port_mask + 1;
>  }
>  
> -static inline void
> +static inline struct execlist_port *
>  execlists_port_complete(struct intel_engine_execlists * const execlists,
>  			struct execlist_port * const port)
>  {
> @@ -675,6 +686,8 @@ execlists_port_complete(struct intel_engine_execlists * const execlists,
>  
>  	memmove(port, port + 1, m * sizeof(struct execlist_port));
>  	memset(port + m, 0, sizeof(struct execlist_port));
> +
> +	return port;
>  }
>  
>  static inline unsigned int
> -- 
> 2.16.3


More information about the Intel-gfx mailing list