Quoting Mika Kuoppala (2020-04-20 13:54:21) > Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> writes: > > > Quoting Mika Kuoppala (2020-04-20 11:54:38) > >> Further, intel_runtime_pm_get is missing. > > > > For what? We acquire the wakeref via the request on the engine. > > > > We don't talk to intel_runtime_pm directly, everything we should be > > doing is engine specific, which knows which gt and the power management > > for that. > > I was worried about the read_cagf(). > > But as it is implied that the request will be running > and we have total control of it during reading th cagf, > it will work like this. Yes, we are only looking at the HW inside either the request or explicit pm_gt. And our mmio debug is good enough to warn if we do attempt to read cagf without the rpm wakeref. -Chris _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
