[PATCH v8 7/9] drm/i915/gt: Ensure memory quiesced before invalidation for all engines

Andrzej Hajda andrzej.hajda at intel.com
Mon Jul 24 09:17:55 UTC 2023



On 24.07.2023 11:14, Andi Shyti wrote:
> Hi Andrzej,
>
>>>    	intel_engine_mask_t aux_inv = 0;
>>> -	u32 cmd, *cs;
>>> +	u32 cmd_flush = 0;
>>> +	u32 cmd = 4;
>>> +	u32 *cs;
>>> -	cmd = 4;
>>> -	if (mode & EMIT_INVALIDATE) {
>>> +	if (mode & EMIT_INVALIDATE)
>>>    		cmd += 2;
>>> -		if (gen12_needs_ccs_aux_inv(rq->engine) &&
>>> -		    (rq->engine->class == VIDEO_DECODE_CLASS ||
>>> -		     rq->engine->class == VIDEO_ENHANCEMENT_CLASS)) {
>>> -			aux_inv = rq->engine->mask &
>>> -				~GENMASK(_BCS(I915_MAX_BCS - 1), BCS0);
>>> -			if (aux_inv)
>>> -				cmd += 4;
>>> -		}
>>> +	if (gen12_needs_ccs_aux_inv(rq->engine))
>>> +		aux_inv = rq->engine->mask &
>>> +			  ~GENMASK(_BCS(I915_MAX_BCS - 1), BCS0);
>> Shouldn't we remove BCS check for MTL? And move it inside
>> gen12_needs_ccs_aux_inv?
>> Btw aux_inv is used as bool, make better is to make it bool.
> Both the cleanups come in patch 9. I wanted to move it initially
> before, but per engine check come later in the series.
>
> I think would need to re-architecture all the patch structure if
> I want to remove it :)
>
> Are you strong with this change?

Nope, if it finally arrives then OK for me.

Reviewed-by: Andrzej Hajda <andrzej.hajda at intel.com>

Regards
Andrzej



>
> Andi



More information about the dri-devel mailing list