On 2018-09-29 13:44:49, Steinar H. Gunderson wrote: > On Sat, Sep 29, 2018 at 01:37:40PM +0200, Sebastian Ramacher wrote: > > That might be. I tried multiple times to get feedback from upstream what we need > > to disable after removing the non-free shaders, but they simply ignore the > > issue. So please, especially since you have the hardware to test it (I only > > tried video decoding on Skylake - I don't have a Haswell machine), I'm happy to > > take patches. > > Sure, I'll be happy to test, but I'm unsure exactly how? I mean, keeping .set_vpp = 1 > _seems_ to work just fine for me, but I have no idea what it would break, if > anything. As long as the code doesn't hit another VA_STATUS_ERROR_UNIMPLEMENTED, I think it should be fine. All the missing shaders should trigger that error code if code relying on the shaders is called. Cheers > > > FWIW all of that should be easier once we switch to the new driver currently > > under development. It has an explicit switch to disable all non-free parts. > > That sounds like good news, although I assume it won't reach buster. > > /* Steinar */ > -- > Homepage: https://www.sesse.net/ > -- Sebastian Ramacher
Attachment:
signature.asc
Description: PGP signature