Quoting Lucas De Marchi (2018-08-29 17:01:11) > On Wed, Aug 29, 2018 at 11:32:35AM +0100, Chris Wilson wrote: > > Quoting Lucas De Marchi (2018-08-29 01:35:31) > > > The 2 PCI IDs that are used for the command line overrid mechanism > > > were left defined. > > > > What makes them so special? Why not just match on the override devid? > > because it's a name -> id mapping? It maps a short string like "skl" to > a single specific PCI ID... how useful is that and if we should retain > its behavior, I have dunno. But > i915_pciids.h doesn't have defines for individual PCI IDs, but groups of > them. My bad, I've always used pci-id overrides as a pci-id! > I would either have to create an accessor/iter for gen x in > intel_chipset.c or do some macros to extract the first id from the > i915_pciids.h, just to get an ID that is set in stone and change the > current id used :-/ Having the i915_pciids.h contain the codename (and /rough/ marketing name) was pencilled in to my plans now that it no longer appears to be a freak out. Once we have configurable macros, the extra parameters just disappear when unwanted. -Chris _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
