On Tue, Nov 11, 2014 at 11:34:58AM +0000, Michel Thierry wrote: > On 11/10/2014 4:38 PM, Mika Kuoppala wrote: > >>- int (*init_context)(struct intel_engine_cs *ring); > >>+ int (*init_context)(struct intel_ringbuffer *ringbuf); > >What is the rationale of this? This seems odd. > > > >Why not: > > > >int (*init_context)(struct intel_engine_cs *ring, > > struct intel_context *ctx); > > > >-Mika > > We want to have the emit function for both legacy and execlist submission > modes (intel_ring_workarounds_emit & intel_logical_ring_workarounds_emit). > It's easy to get the ring from a ringbuf, but not the other way around.
ctx already has the ringbuf pointer for execlist that you are looking for. And for legacy contexts there's only one, per-engine. So from that quick look Mika's interface proposal does make more sense to me. -Daneil -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
