Hi Andrzej, On Fri, Jul 21, 2023 at 11:41:22AM +0200, Andrzej Hajda wrote: > On 20.07.2023 23:07, Andi Shyti wrote: > > We always assumed that a device might either have AUX or FLAT > > CCS, but this is an approximation that is not always true as it > > requires some further per device checks. > > > > Add the "has_aux_ccs" flag in the intel_device_info structure in > > order to have a per device flag indicating of the AUX CCS. > > As Matt mentioned in v6, aux_ccs is present also in older platforms. > This is about presence/necessity (?) of aux_ccs table invalidation. > Maybe has_aux_ccs_inv, dunno? yes, true! It's aux_ccs_inv. > Moreover you define flag per device, but this is rather per engine, > theoretically could be also: > MTL: > .aux_ccs_inv_mask = BIT(RCS0) | BIT(BCS0) | ... > Others: > .aux_ccs_inv_mask = BIT(RCS0) | ... there is already some engine discrimination that is mandatory later in the series. Doing it here it's a bit replicated. > looks overkill, > maybe helper function would be simpler, up to you. Yes, a helper function that checks on the platform and returns true or false... it looks hardcoded to me, while the info structure is hardcoded by definition and looks easier to maintain by just toggling on/off a single flag in that structure. That's why I decided to place it there. But, because there are already two people suggesting it, then I will try it in v8. Thanks, Andi
