On Tue, Apr 26, 2016 at 04:41:08PM -0700, Kenneth Graunke wrote: > My intention was to have the render ring prelude occur at the first > point where we start emiting render commands into a batch. > > Gen4-5 have a single ring that handles both BLT and 3D commands, so it's > possible to have a BLT -> RENDER transition in the middle of a batch > (not just at the start when prev_ring == UNKNOWN). We want to call the > prelude here as well. > > Today, this means we'll miss a OA bookend snapshot. Topi's also > thinking of moving some code here that absolutely must run before RENDER > commands happen, at which point fixing this will be more critical.
I dropped those plans in the end but I think this is worth fixing anyway: Reviewed-by: Topi Pohjolainen <topi.pohjolai...@intel.com> > > Signed-off-by: Kenneth Graunke <kenn...@whitecape.org> > --- > src/mesa/drivers/dri/i965/intel_batchbuffer.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/src/mesa/drivers/dri/i965/intel_batchbuffer.c > b/src/mesa/drivers/dri/i965/intel_batchbuffer.c > index e41f927..47a2864 100644 > --- a/src/mesa/drivers/dri/i965/intel_batchbuffer.c > +++ b/src/mesa/drivers/dri/i965/intel_batchbuffer.c > @@ -128,7 +128,7 @@ intel_batchbuffer_require_space(struct brw_context *brw, > GLuint sz, > */ > brw->batch.ring = ring; > > - if (unlikely(prev_ring == UNKNOWN_RING && ring == RENDER_RING)) > + if (unlikely(prev_ring != RENDER_RING && ring == RENDER_RING)) > intel_batchbuffer_emit_render_ring_prelude(brw); > } > > -- > 2.8.0 > _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
