Hi Chris,
> Now that we are careful to always force-restore contexts upon rewinding
> (where necessary), we can restore our optimisation to skip over
> completed active execlists when dequeuing.
>
> Referenecs: 35f3fd8182ba ("drm/i915/execlists: Workaround switching back to a
> completed context")
> References: 8ab3a3812aa9 ("drm/i915/gt: Incrementally check for rewinding")
> Signed-off-by: Chris Wilson <[email protected]>
> Cc: Mika Kuoppala <[email protected]>
> ---
> .../drm/i915/gt/intel_execlists_submission.c | 34 +++++++++----------
> 1 file changed, 16 insertions(+), 18 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
> b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
> index 524c8b54d220..ac1be7a632d3 100644
> --- a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
> +++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
> @@ -1224,12 +1224,20 @@ static void set_preempt_timeout(struct
> intel_engine_cs *engine,
> active_preempt_timeout(engine, rq));
> }
>
> +static bool completed(const struct i915_request *rq)
> +{
> + if (i915_request_has_sentinel(rq))
> + return false;
> +
> + return __i915_request_is_complete(rq);
> +}
> +
> static void execlists_dequeue(struct intel_engine_cs *engine)
> {
> struct intel_engine_execlists * const execlists = &engine->execlists;
> struct i915_request **port = execlists->pending;
> struct i915_request ** const last_port = port + execlists->port_mask;
> - struct i915_request *last = *execlists->active;
> + struct i915_request *last, * const *active;
> struct virtual_engine *ve;
> struct rb_node *rb;
> bool submit = false;
> @@ -1266,21 +1274,13 @@ static void execlists_dequeue(struct intel_engine_cs
> *engine)
> * i.e. we will retrigger preemption following the ack in case
> * of trouble.
> *
> - * In theory we can skip over completed contexts that have not
> - * yet been processed by events (as those events are in flight):
> - *
> - * while ((last = *active) && i915_request_completed(last))
> - * active++;
> - *
> - * However, the GPU cannot handle this as it will ultimately
> - * find itself trying to jump back into a context it has just
> - * completed and barf.
> */
> + active = execlists->active;
> + while ((last = *active) && completed(last))
> + active++;
looks reasonable to me...
Reviewed-by: Andi Shyti <[email protected]>
Thanks,
Andi
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
