[Intel-gfx] [PATCH 08/18] drm/i915/execlists: Wait for ELSP submission on restart

Mika Kuoppala mika.kuoppala at linux.intel.com
Fri May 25 12:37:14 UTC 2018


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

> After a reset, we will ensure that there is at least one request
> submitted to HW to ensure that a context is loaded for powersaving.
> Let's wait for this submission via a tasklet to complete before we drop
> our forcewake, ensuring the system is ready for rc6 before we let it
> possible sleep.

s/possible/possibly ?

>
> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> ---
>  drivers/gpu/drm/i915/i915_gem.h  |  6 ++++++
>  drivers/gpu/drm/i915/intel_lrc.c | 15 ++++++++++++++-
>  2 files changed, 20 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/i915/i915_gem.h b/drivers/gpu/drm/i915/i915_gem.h
> index 62ee4e385365..261da577829a 100644
> --- a/drivers/gpu/drm/i915/i915_gem.h
> +++ b/drivers/gpu/drm/i915/i915_gem.h
> @@ -82,4 +82,10 @@ static inline void __tasklet_disable_sync_once(struct tasklet_struct *t)
>  		tasklet_unlock_wait(t);
>  }
>  
> +static inline void __tasklet_enable_sync_once(struct tasklet_struct *t)
> +{
> +	if (atomic_dec_return(&t->count) == 0)
> +		tasklet_kill(t);
> +}
> +
>  #endif /* __I915_GEM_H__ */
> diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
> index b3a5cfb7689f..f2fb48b1a7b7 100644
> --- a/drivers/gpu/drm/i915/intel_lrc.c
> +++ b/drivers/gpu/drm/i915/intel_lrc.c
> @@ -1877,6 +1877,8 @@ execlists_reset_prepare(struct intel_engine_cs *engine)
>  
>  	GEM_TRACE("%s\n", engine->name);
>  
> +	intel_uncore_forcewake_get(engine->i915, FORCEWAKE_ALL);
> +
>  	/*
>  	 * Prevent request submission to the hardware until we have
>  	 * completed the reset in i915_gem_reset_finish(). If a request
> @@ -2007,7 +2009,18 @@ static void execlists_reset(struct intel_engine_cs *engine,
>  
>  static void execlists_reset_finish(struct intel_engine_cs *engine)
>  {
> -	tasklet_enable(&engine->execlists.tasklet);
> +	/*
> +	 * Flush the tasklet while we still have the forcewake to be sure
> +	 * that it is not allowed to sleep before we restart and reload a
> +	 * context.
> +	 *
> +	 * As before (with execlists_reset_prepare) we rely on the caller
> +	 * serialising mutiple attempts to reset so that we know that we

s/mutiple/multiple

> +	 * are the only one manipulating tasklet state.
> +	 */
> +	__tasklet_enable_sync_once(&engine->execlists.tasklet);
> +
> +	intel_uncore_forcewake_put(engine->i915, FORCEWAKE_ALL);

as discussed in irc, we already hold forcewakes on behalf of
prepare_engine() above us.


With forcewake_put/get removed,

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

>  
>  	GEM_TRACE("%s\n", engine->name);
>  }
> -- 
> 2.17.0
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx


More information about the Intel-gfx mailing list