On Thu, Apr 21, 2016 at 09:48:36AM +0300, Joonas Lahtinen wrote: > On ke, 2016-04-20 at 19:42 +0100, Chris Wilson wrote: > > The code to switch_mm() is already handled by i915_switch_context(), the > > only difference required to setup the aliasing ppgtt is that we need to > > emit te switch_mm() on the first context, i.e. when transitioning from > > engine->last_context == NULL. This allows us to defer the > > initialisation of the GPU from early device initialisation to first use, > > which should marginally speed up both. The caveat is that we then defer > > the context initialisation until first use - i.e. we cannot assume that > > the GPU engines are initialised. For example, this means that power > > contexts for rc6 (Ironlake) need to explicitly loaded, as they are. > > > > Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> > > Cc: Mika Kuoppala <mika.kuoppala@xxxxxxxxx> > > Some comments below. > > Reviewed-by: Joonas Lahtinen <joonas.lahtinen@xxxxxxxxxxxxxxx> > > > diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c > > index aafcb4942acf..14c9b29294c5 100644 > > --- a/drivers/gpu/drm/i915/i915_gem.c > > +++ b/drivers/gpu/drm/i915/i915_gem.c > > @@ -4910,34 +4910,6 @@ i915_gem_init_hw(struct drm_device *dev) > > if (ret) > > goto out; > > This whole if (ret) goto out can be removed as out is right after the > removed code block. Oh, this piece of code is earmarked to be removed and now I even have a bugzilla for it! https://bugs.freedesktop.org/show_bug.cgi?id=95023 is an example where we are doing the wrong thing with the current i915_gem_init_seqno. -Chris -- Chris Wilson, Intel Open Source Technology Centre _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
