On Thu, Dec 05, 2013 at 02:21:08PM +0000, Damien Lespiau wrote: > On Thu, Dec 05, 2013 at 02:09:33PM +0000, Chris Wilson wrote: > > Patches look fine. Just relying on fuse registers being correct is like > > relying on vbt being correct, fraught with anxiety. > > Well, I believe those bits are directly reflecting the fuses/straps > programmed (the sames that will actually disable hw), so it should be > better than VBT. On the other hand, that's the reason why I only check > from IVB on and not from ILK to not take unecessary risks on platforms > with no known fused config. I'm surprised by this. We had a semi-lengthy mail on the matter internally, and ISTR there was no way to actually make this work for all cases. I guess I need to go re-read that. > > In any case, if someone with such a device could test the series :) IMHO the patches can't be merged until it can be verified. If you want to call that a nak-until-then on patch 6, do. I'd much rather have a dynamic solution like this though. > > -- > Damien -- Ben Widawsky, Intel Open Source Technology Center _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx
