[Intel-gfx] [PATCH 1/6] drm/i915: Refactor export_fence() after i915_vma_move_to_active()

Tvrtko Ursulin tvrtko.ursulin at linux.intel.com
Mon Jul 2 12:29:07 UTC 2018


On 02/07/2018 12:44, Chris Wilson wrote:
> Quoting Tvrtko Ursulin (2018-07-02 12:34:51)
>>> @@ -1921,6 +1914,8 @@ void i915_vma_move_to_active(struct i915_vma *vma,
>>>    
>>>        if (flags & EXEC_OBJECT_NEEDS_FENCE)
>>>                i915_gem_active_set(&vma->last_fence, rq);
>>> +
>>> +     export_fence(vma, rq, flags);
>>
>> Don't care about new extra reservation object operations from the other
>> i915_move_to_active_callers?
> 
> No. Not filling the reservation_object is an underhand shortcut. It
> should never be not correct to track the fences for an object. Doing it
> in one place correctly, should prevent cheating.

Fair enough.

Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin at intel.com>

Regards,

Tvrtko


More information about the Intel-gfx mailing list