[Intel-gfx] [PATCH 1/2] drm/i915: Restore GTT domain tracking of unbound objects upon resume

[Daniel Vetter]

On Wed, Feb 06, 2013 at 09:31:20AM +0000, Chris Wilson wrote:
> In order for the objects to be coherent in uncached memory (as they are
> presumed to be on the unbound list) we need to clflush them because
> experience dictates that the CPU caches will be dirtied across
> hibernation.
> 
> Cc: stable at vger.kernel.org
> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>

I think this change will freak out Jesse, since it'll kill resume time
when we clflush a few GB of memory. So could you throw another patch on
top of this series (maybe after the unbound dropping) which calls
restore_gtt only when thawing from hibernate, but not when resuming?

> ---
>  drivers/gpu/drm/i915/i915_gem_gtt.c |    3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
> index bdaca3f..368c821 100644
> --- a/drivers/gpu/drm/i915/i915_gem_gtt.c
> +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
> @@ -404,6 +404,9 @@ void i915_gem_restore_gtt_mappings(struct drm_device *dev)
>  		i915_gem_gtt_bind_object(obj, obj->cache_level);
>  	}
>  
> +	list_for_each_entry(obj, &dev_priv->mm.unbound_list, gtt_list)
> +		i915_gem_clflush_object(obj);
> +
>  	i915_gem_chipset_flush(dev);
>  }
>  
> -- 
> 1.7.10.4
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx

-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch