[Mesa-dev] [PATCH 10/22] intel/blorp: Add an entrypoint for clearing depth and stencil
Pohjolainen, Topi
topi.pohjolainen at gmail.com
Tue Oct 11 18:54:09 UTC 2016
On Mon, Oct 10, 2016 at 10:24:58AM -0700, Jason Ekstrand wrote:
> Topi,
> There are some patches in this branch:
> [1]https://cgit.freedesktop.org/~jekstrand/mesa/log/?h=wip/i965-blorp-d
> s
> That I intend to send or squash in as part of a v2. Feel free to look
> at them and/or review preemptively.
Patches
intel/blorp: Emit a NULL render target for depth/stencil-only operations
i965: Use blorp for depth/stencil clears on gen6+
look good to me and are:
Reviewed-by: Topi Pohjolainen <topi.pohjolainen at intel.com>
> --Jason
> On Sun, Oct 9, 2016 at 11:00 PM, Pohjolainen, Topi
> <[2]topi.pohjolainen at gmail.com> wrote:
>
> On Sun, Oct 09, 2016 at 10:52:52PM -0700, Jason Ekstrand wrote:
> > On Oct 9, 2016 10:48 PM, "Pohjolainen, Topi"
> > <[1][3]topi.pohjolainen at gmail.com> wrote:
> > >
> > > On Fri, Oct 07, 2016 at 09:41:08PM -0700, Jason Ekstrand
> wrote:
> > > > Signed-off-by: Jason Ekstrand <[2][4]jason at jlekstrand.net>
>
> > > > ---
> > > > src/intel/blorp/blorp.h | 10 ++++++++
> > > > src/intel/blorp/blorp_clear.c | 58
> > +++++++++++++++++++++++++++++++++++++++++++
> > > > 2 files changed, 68 insertions(+)
> > > >
> > > > diff --git a/src/intel/blorp/blorp.h b/src/intel/blorp/blorp.h
> > > > index 480f1bb..263d194 100644
> > > > --- a/src/intel/blorp/blorp.h
> > > > +++ b/src/intel/blorp/blorp.h
> > > > @@ -134,6 +134,16 @@ blorp_clear(struct blorp_batch *batch,
> > > > const bool color_write_disable[4]);
> > > >
> > > > void
> > > > +blorp_clear_depth_stencil(struct blorp_batch *batch,
> > > > + const struct blorp_surf *depth,
> > > > + const struct blorp_surf *stencil,
> > > > + uint32_t level, uint32_t
> start_layer,
> > > > + uint32_t num_layers,
> > > > + uint32_t x0, uint32_t y0, uint32_t
> x1,
> > uint32_t y1,
> > > > + bool clear_depth, float
> depth_value,
> > > > + bool clear_stencil, uint8_t
> > stencil_value);
> > > > +
> > > > +void
> > > > blorp_ccs_resolve(struct blorp_batch *batch,
> > > > struct blorp_surf *surf, enum isl_format
> > format);
> > > >
> > > > diff --git a/src/intel/blorp/blorp_clear.c
> > b/src/intel/blorp/blorp_clear.c
> > > > index a8f29fd..1d6bf1c 100644
> > > > --- a/src/intel/blorp/blorp_clear.c
> > > > +++ b/src/intel/blorp/blorp_clear.c
> > > > @@ -306,6 +306,64 @@ blorp_clear(struct blorp_batch *batch,
> > > > }
> > > >
> > > > void
> > > > +blorp_clear_depth_stencil(struct blorp_batch *batch,
> > > > + const struct blorp_surf *depth,
> > > > + const struct blorp_surf *stencil,
> > > > + uint32_t level, uint32_t
> start_layer,
> > > > + uint32_t num_layers,
> > > > + uint32_t x0, uint32_t y0, uint32_t
> x1,
> > uint32_t y1,
> > > > + bool clear_depth, float
> depth_value,
> > > > + bool clear_stencil, uint8_t
> > stencil_value)
> > > > +{
> > > > + struct blorp_params params;
> > > > + blorp_params_init(¶ms);
> > > > +
> > > > + params.x0 = x0;
> > > > + params.y0 = y0;
> > > > + params.x1 = x1;
> > > > + params.y1 = y1;
> > > > +
> > > > + while (num_layers > 0) {
> > > > + params.num_layers = num_layers;
> > > > +
> > > > + if (clear_stencil) {
> > > > + brw_blorp_surface_info_init(batch->blorp,
> > ¶ms.stencil, stencil,
> > > > + level, start_layer,
> > > > + ISL_FORMAT_UNSUPPORTED,
> > true);
> > > > + params.stencil_ref = stencil_value;
> > > > +
> > > > + params.dst.surf.samples =
> params.stencil.surf.samples;
> > > > + params.dst.surf.logical_level0_px =
> > > > + params.stencil.surf.logical_level0_px;
> > > > +
> > > > + if (params.stencil.view.array_len <
> params.num_layers)
> > > > + params.num_layers =
> params.stencil.view.array_len;
> > > > + }
> > > > +
> > > > + if (clear_depth) {
> > > > + brw_blorp_surface_info_init(batch->blorp,
> ¶ms.depth,
> > depth,
> > > > + level, start_layer,
> > > > + ISL_FORMAT_UNSUPPORTED,
> > true);
> > > > + params.z = depth_value;
> > > > + params.depth_format =
> > > > + isl_format_get_depth_format(depth->surf->format,
> > false);
> > > > +
> > > > + params.dst.surf.samples = params.depth.surf.samples;
> > > > + params.dst.surf.logical_level0_px =
> > > > + params.depth.surf.logical_level0_px;
> > > > +
> > > > + if (params.depth.view.array_len < params.num_layers)
> > > > + params.num_layers = params.depth.view.array_len;
> > >
> > > Stencil already does the same, could we add an assert here:
> > >
> > > assert(!clear_stencil ||
> > > params.depth.view.array_len ==
> > > params.stencil.view.array_len)
> >
> > Sure.
> >
> > >
> > > Moreover, I thought that start_layer + num_layers <=
> view.array_len
> > should
> > > always apply. You seem to explicitly prepare for that not to
> hold.
> > Could you
> > > give an example?
> >
> > I had to do this recently with the other clears as well. Sandy
> Bridge
> > has a max of 512 layers for rendering. In the surface_info_init
> > function, we clamp as needed and this ensures that 16k-slice
> textures
> > get 32 draw calls.
>
> Ah, right, that makes sense. How would you feel we added that as
> comment?
> /* Sandy Bridge has a max of 512 layers for rendering. In the
> * surface_info_init function, we clamp as needed and this
> ensures that
> * 16k-slice textures get 32 draw calls.
> */
> assert(start_layer + num_layers <= view.array_len || gen == 6);
> Otherwise:
> Reviewed-by: Topi Pohjolainen <[5]topi.pohjolainen at intel.com>
> >
> > > > + }
> > > > +
> > > > + batch->blorp->exec(batch, ¶ms);
> > > > +
> > > > + start_layer += params.num_layers;
> > > > + num_layers -= params.num_layers;
> > > > + }
> > > > +}
> > > > +
> > > > +void
> > > > blorp_ccs_resolve(struct blorp_batch *batch,
> > > > struct blorp_surf *surf, enum isl_format
> format)
> > > > {
> > > > --
> > > > 2.5.0.400.gff86faf
> > > >
> > > > _______________________________________________
> > > > mesa-dev mailing list
> > > > [3][6]mesa-dev at lists.freedesktop.org
> > > > [4][7]https://lists.freedesktop.
> org/mailman/listinfo/mesa-dev
> >
> > References
> >
> > 1. mailto:[8]topi.pohjolainen at gmail.com
> > 2. mailto:[9]jason at jlekstrand.net
> > 3. mailto:[10]mesa-dev at lists.freedesktop.org
> > 4. [11]https://lists.freedesktop.org/mailman/listinfo/mesa-dev
>
> References
>
> 1. https://cgit.freedesktop.org/~jekstrand/mesa/log/?h=wip/i965-blorp-ds
> 2. mailto:topi.pohjolainen at gmail.com
> 3. mailto:topi.pohjolainen at gmail.com
> 4. mailto:jason at jlekstrand.net
> 5. mailto:topi.pohjolainen at intel.com
> 6. mailto:mesa-dev at lists.freedesktop.org
> 7. https://lists.freedesktop.org/mailman/listinfo/mesa-dev
> 8. mailto:topi.pohjolainen at gmail.com
> 9. mailto:jason at jlekstrand.net
> 10. mailto:mesa-dev at lists.freedesktop.org
> 11. https://lists.freedesktop.org/mailman/listinfo/mesa-dev
More information about the mesa-dev
mailing list