It did not. But, I just ran all of those tests with and without this patch, and there were no regressions. So, I think we're good.
Pushed - thanks for the review!
On Thursday, May 29, 2014 07:32:21 PM Chris Forbes wrote:
> Did this test run include forcing ARB_gpu_shader5 on?
>
> On Thu, May 29, 2014 at 7:06 PM, Kenneth Graunke <kenn...@whitecape.org>
wrote:
> > As far as I can tell, Broadwell doesn't need any of the SURFACE_STATE
> > workarounds for textureGather() bugs, so there's no need to emit
> > a second set of identical copies.
> >
> > To keep things simple, just point the gather surface index base to the
> > same place as the texture surface index base.
> >
> > Signed-off-by: Kenneth Graunke <kenn...@whitecape.org>
> > Cc: "10.2" <mesa-sta...@lists.freedesktop.org>
> > ---
> > src/mesa/drivers/dri/i965/brw_shader.cpp | 9 +++++++--
> > src/mesa/drivers/dri/i965/brw_wm_surface_state.c | 14 ++++++++------
> > 2 files changed, 15 insertions(+), 8 deletions(-)
> >
> > No Piglit regressions on Broadwell with -t texture.
> >
> > diff --git a/src/mesa/drivers/dri/i965/brw_shader.cpp
b/src/mesa/drivers/dri/i965/brw_shader.cpp
> > index f4f1334..7fae416 100644
> > --- a/src/mesa/drivers/dri/i965/brw_shader.cpp
> > +++ b/src/mesa/drivers/dri/i965/brw_shader.cpp
> > @@ -740,8 +740,13 @@
backend_visitor::assign_common_binding_table_offsets(uint32_t
next_binding_table
> > }
> >
> > if (prog->UsesGather) {
> > - stage_prog_data->binding_table.gather_texture_start =
next_binding_table_offset;
> > - next_binding_table_offset += num_textures;
> > + if (brw->gen >= 8) {
> > + stage_prog_data->binding_table.gather_texture_start =
> > + stage_prog_data->binding_table.texture_start;
> > + } else {
> > + stage_prog_data->binding_table.gather_texture_start =
next_binding_table_offset;
> > + next_binding_table_offset += num_textures;
> > + }
> > } else {
> > stage_prog_data->binding_table.gather_texture_start = 0xd0d0d0d0;
> > }
> > diff --git a/src/mesa/drivers/dri/i965/brw_wm_surface_state.c
b/src/mesa/drivers/dri/i965/brw_wm_surface_state.c
> > index c9d9614..3279d3b 100644
> > --- a/src/mesa/drivers/dri/i965/brw_wm_surface_state.c
> > +++ b/src/mesa/drivers/dri/i965/brw_wm_surface_state.c
> > @@ -829,12 +829,14 @@ brw_update_texture_surfaces(struct brw_context *brw)
> > /* emit alternate set of surface state for gather. this
> > * allows the surface format to be overriden for only the
> > * gather4 messages. */
> > - if (vs && vs->UsesGather)
> > - update_stage_texture_surfaces(brw, vs, &brw->vs.base, true);
> > - if (gs && gs->UsesGather)
> > - update_stage_texture_surfaces(brw, gs, &brw->gs.base, true);
> > - if (fs && fs->UsesGather)
> > - update_stage_texture_surfaces(brw, fs, &brw->wm.base, true);
> > + if (brw->gen < 8) {
> > + if (vs && vs->UsesGather)
> > + update_stage_texture_surfaces(brw, vs, &brw->vs.base, true);
> > + if (gs && gs->UsesGather)
> > + update_stage_texture_surfaces(brw, gs, &brw->gs.base, true);
> > + if (fs && fs->UsesGather)
> > + update_stage_texture_surfaces(brw, fs, &brw->wm.base, true);
> > + }
> >
> > brw->state.dirty.brw |= BRW_NEW_SURFACES;
> > }
> > --
> > 1.9.1
> >
> > _______________________________________________
> > mesa-stable mailing list
> > mesa-sta...@lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/mesa-stable
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
