On ti, 2016-02-23 at 14:55 +0000, Chris Wilson wrote: > On Tue, Feb 23, 2016 at 04:47:17PM +0200, Imre Deak wrote: > > The device needs to be in D0 state whenever we call these > > functions, so > > add the corresponding assert checks. > > No. In quite a few of those cases we are calling iowrite to non- > device > memory, even normal pages. Hm right didn't think about that. I guess the only case we care about then are accesses through the GTT. > How's the separation of struct_mutex from rpm going so that we can > forgo > adding assertions and use explicit power management instead? It's planned to be done, but no one is working on that yet. This is something we could still have regardless, similarly to other helpers accessing the device. > -Chris > _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
