Re: [PATCH v2 4/9] drm/i915: Move the hw semaphore initialisation from GEM to the engine

[Joonas Lahtinen <joonas.lahtinen@linux.intel.com>]

On ke, 2016-04-06 at 13:33 +0100, Chris Wilson wrote:
> Since we are setting engine local values that are tied to the hardware,
> move it out of i915_gem_init_seqno() into the intel_ring_init_seqno()
> backend, next to where the other hw semaphore registers are written.
> 
> v2: Make the explanatory comment about always resetting the semaphores to
> 0 irrespective of the value of the reset seqno.
> 
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
> Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
> Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
> ---
>  drivers/gpu/drm/i915/i915_gem.c         |  8 ++------
>  drivers/gpu/drm/i915/intel_ringbuffer.c | 11 +++++++++++
>  2 files changed, 13 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
> index 05711e7f9fb8..7506e850913e 100644
> --- a/drivers/gpu/drm/i915/i915_gem.c
> +++ b/drivers/gpu/drm/i915/i915_gem.c
> @@ -2468,7 +2468,7 @@ i915_gem_init_seqno(struct drm_device *dev, u32 seqno)
>  {
>  	struct drm_i915_private *dev_priv = dev->dev_private;
>  	struct intel_engine_cs *engine;
> -	int ret, j;
> +	int ret;
>  
>  	/* Carefully retire all requests without writing to the rings */
>  	for_each_engine(engine, dev_priv) {
> @@ -2479,13 +2479,9 @@ i915_gem_init_seqno(struct drm_device *dev, u32 seqno)
>  	i915_gem_retire_requests(dev);
>  
>  	/* Finally reset hw state */
> -	for_each_engine(engine, dev_priv) {
> +	for_each_engine(engine, dev_priv)
>  		intel_ring_init_seqno(engine, seqno);
>  
> -		for (j = 0; j < ARRAY_SIZE(engine->semaphore.sync_seqno); j++)
> -			engine->semaphore.sync_seqno[j] = 0;
> -	}
> -
>  	return 0;
>  }
>  
> diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
> index 371f4c1fc33c..69852ac4018d 100644
> --- a/drivers/gpu/drm/i915/intel_ringbuffer.c
> +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
> @@ -2552,14 +2552,25 @@ void intel_ring_init_seqno(struct intel_engine_cs *engine, u32 seqno)
>  {
>  	struct drm_i915_private *dev_priv = to_i915(engine->dev);
>  
> +	/* Our semaphore implementation is strictly monotonic (i.e. we proceed
> +	 * so long as the semaphore value in the register/page is greater
> +	 * than the sync value), so whenever we reset the seqno,
> +	 * so long as we reset the tracking semaphore value to 0, it will
> +	 * always be before the next request's seqno. If we don't reset
> +	 * the semaphore value, then when the seqno moves backwards all
> +	 * future waits will complete instantly (causing rendering corruption).
> +	 */

Yes, this is better comment.

>  	if (INTEL_INFO(dev_priv)->gen == 6 || INTEL_INFO(dev_priv)->gen == 7) {
>  		I915_WRITE(RING_SYNC_0(engine->mmio_base), 0);
>  		I915_WRITE(RING_SYNC_1(engine->mmio_base), 0);
>  		if (HAS_VEBOX(dev_priv))
>  			I915_WRITE(RING_SYNC_2(engine->mmio_base), 0);
>  	}
> +	memset(engine->semaphore.sync_seqno, 0,
> +	       sizeof(engine->semaphore.sync_seqno));
>  
>  	engine->set_seqno(engine, seqno);
> +
>  	engine->hangcheck.seqno = seqno;
>  }
>  
-- 
Joonas Lahtinen
Open Source Technology Center
Intel Corporation
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx