[Mesa-dev] [PATCH v3 13/34] i965/blorp: Add a generic ISL-based surface state emit path

Pohjolainen, Topi topi.pohjolainen at intel.com
Wed Jul 6 06:13:21 UTC 2016


On Tue, Jul 05, 2016 at 07:12:49AM -0700, Jason Ekstrand wrote:
>    On Jul 5, 2016 6:32 AM, "Pohjolainen, Topi"
>    <[1]topi.pohjolainen at intel.com> wrote:
>    >
>    > On Wed, Jun 29, 2016 at 05:14:51PM -0700, Jason Ekstrand wrote:
>    > > Reviewed-by: Topi Pohjolainen <[2]topi.pohjolainen at intel.com>
>    > >
>    > > v3: Stomp more surface state fields
>    > > ---
>    > >  src/mesa/drivers/dri/i965/brw_blorp.c | 163
>    ++++++++++++++++++++++++++++++++++
>    > >  src/mesa/drivers/dri/i965/brw_blorp.h |   6 ++
>    > >  2 files changed, 169 insertions(+)
>    > >
>    > > diff --git a/src/mesa/drivers/dri/i965/brw_blorp.c
>    b/src/mesa/drivers/dri/i965/brw_blorp.c
>    > > index 9590968..275f10c 100644
>    > > --- a/src/mesa/drivers/dri/i965/brw_blorp.c
>    > > +++ b/src/mesa/drivers/dri/i965/brw_blorp.c
>    > > @@ -240,6 +240,169 @@ brw_blorp_compile_nir_shader(struct
>    brw_context *brw, struct nir_shader *nir,
>    > >     return program;
>    > >  }
>    > >
>    > > +static enum isl_msaa_layout
>    > > +get_isl_msaa_layout(enum intel_msaa_layout layout)
>    > > +{
>    > > +   switch (layout) {
>    > > +   case INTEL_MSAA_LAYOUT_NONE:
>    > > +      return ISL_MSAA_LAYOUT_NONE;
>    > > +   case INTEL_MSAA_LAYOUT_IMS:
>    > > +      return ISL_MSAA_LAYOUT_INTERLEAVED;
>    > > +   case INTEL_MSAA_LAYOUT_UMS:
>    > > +   case INTEL_MSAA_LAYOUT_CMS:
>    > > +      return ISL_MSAA_LAYOUT_ARRAY;
>    > > +   default:
>    > > +      unreachable("Invalid MSAA layout");
>    > > +   }
>    > > +}
>    > > +
>    > > +struct surface_state_info {
>    > > +   unsigned num_dwords;
>    > > +   unsigned ss_align; /* Required alignment of
>    RENDER_SURFACE_STATE in bytes */
>    > > +   unsigned reloc_dw;
>    > > +   unsigned aux_reloc_dw;
>    > > +   unsigned tex_mocs;
>    > > +   unsigned rb_mocs;
>    > > +};
>    > > +
>    > > +static const struct surface_state_info surface_state_infos[] = {
>    > > +   [6] = {6,  32, 1,  0},
>    > > +   [7] = {8,  32, 1,  6,  GEN7_MOCS_L3, GEN7_MOCS_L3},
>    > > +   [8] = {13, 64, 8,  10, BDW_MOCS_WB,  BDW_MOCS_PTE},
>    > > +   [9] = {16, 64, 8,  10, SKL_MOCS_WB,  SKL_MOCS_PTE},
>    > > +};
>    > > +
>    > > +uint32_t
>    > > +brw_blorp_emit_surface_state(struct brw_context *brw,
>    > > +                             const struct brw_blorp_surface_info
>    *surface,
>    > > +                             uint32_t read_domains, uint32_t
>    write_domain,
>    > > +                             bool is_render_target)
>    > > +{
>    > > +   /* TODO: This should go in the context */
>    > > +   struct isl_device isl_dev;
>    > > +   isl_device_init(&isl_dev, brw->intelScreen->devinfo,
>    brw->has_swizzling);
>    > > +
>    > > +   const struct surface_state_info ss_info =
>    surface_state_infos[brw->gen];
>    > > +
>    > > +   struct isl_surf surf;
>    > > +   intel_miptree_get_isl_surf(brw, surface->mt, &surf);
>    > > +
>    > > +   /* Stomp surface dimensions and tiling (if needed) with info
>    from blorp */
>    > > +   surf.dim = ISL_SURF_DIM_2D;
>    > > +   surf.dim_layout = ISL_DIM_LAYOUT_GEN4_2D;
>    > > +   surf.msaa_layout = get_isl_msaa_layout(surface->msaa_layout);
>    > > +   surf.logical_level0_px.width = surface->width;
>    > > +   surf.logical_level0_px.height = surface->height;
>    > > +   surf.logical_level0_px.depth = 1;
>    > > +   surf.logical_level0_px.array_len = 1;
>    >
>    > You added these two lines which make sense. You didn't need them
>    before, did
>    > anything change or just assertions firing?
> 
>    The newly added X/YOffset assertions require this.
> 
>    > > +   surf.levels = 1;
>    > > +   surf.samples = MAX2(surface->num_samples, 1);
>    > > +
>    > > +   /* Alignment doesn't matter since we have 1 miplevel and 1
>    array slice so
>    > > +    * just pick something that works for everybody.
>    > > +    */
>    > > +   surf.image_alignment_el = isl_extent3d(4, 4, 1);
>    >
>    > Same here, looks sane but makes me womnder what made you add this?
> 
>    There are two cases where we can end up with an alignment which is
>    invalid for 3D surfaces: gen6 stencil and gen9 1D. In
>    Intel_mipmap_tree, gen9 1D surfaces have an invalid halign value and we
>    just special-case gen9 1D to use 64 everywhere.  When I started using
>    ISL for offset calculations in blorp, I needed image_alignment_el to
>    have the correct value of (64, 1, 1) so I had to add that to
>    miptree_get_isl_surf which meant we now had to undo it here.

But this patch would still work without this, or am I misunderstanding? We
still call brw_blorp_compute_tile_offsets() down below. In fact, I don't
really understand why setting image_alignment_el in miptree_get_isl_surf()
would help isl-based offset calculation either. That is going to replace
the blorp specific call below, and we are overriding the calculation
of miptree_get_isl_surf() here.

I suppose there is just too many things happening the same time for me to keep
track. Some preparing for future patches and some needed locally...

> 
>    > > +
>    > > +   if (brw->gen == 6 && surface->num_samples > 1) {
>    > > +      /* Since gen6 uses INTEL_MSAA_LAYOUT_IMS, width and height
>    are measured
>    > > +       * in samples.  But SURFACE_STATE wants them in pixels, so
>    we need to
>    > > +       * divide them each by 2.
>    > > +       */
>    > > +      surf.logical_level0_px.width /= 2;
>    > > +      surf.logical_level0_px.height /= 2;
>    > > +   }
>    > > +
>    > > +   if (brw->gen == 6 && surf.image_alignment_el.height > 4) {
>    > > +      /* This can happen on stencil buffers on Sandy Bridge due to
>    the
>    > > +       * single-LOD work-around.  It's fairly harmless as long as
>    we don't
>    > > +       * pass a bogus value into isl_surf_fill_state().
>    > > +       */
>    > > +      surf.image_alignment_el = isl_extent3d(4, 2, 1);
>    >
>    > Earlier you set this unconditionally to isl_extent3d(4, 4, 1), isn't
>    that
>    > valid?
> 
>    Yes, I believe this can go now.
> 
>    > > +   }
>    > > +
>    > > +   if (surface->map_stencil_as_y_tiled) {
>    > > +      /* We need to fake W-tiling with Y-tiling */
>    > > +      surf.tiling = ISL_TILING_Y0;
>    > > +      surf.row_pitch *= 2;
>    > > +   }
>    > > +
>    > > +   union isl_color_value clear_color = { .u32 = { 0, 0, 0, 0 } };
>    > > +
>    > > +   struct isl_surf *aux_surf = NULL, aux_surf_s;
>    > > +   uint64_t aux_offset = 0;
>    > > +   enum isl_aux_usage aux_usage = ISL_AUX_USAGE_NONE;
>    > > +   if (surface->mt->mcs_mt) {
>    > > +      /* We should probably to similar stomping to above but most
>    of the aux
>    > > +       * surf gets ignored when we fill out the surface state
>    anyway so
>    > > +       * there's no point.
>    > > +       */
>    > > +      intel_miptree_get_aux_isl_surf(brw, surface->mt,
>    &aux_surf_s, &aux_usage);
>    > > +      aux_surf = &aux_surf_s;
>    > > +      assert(surface->mt->mcs_mt->offset == 0);
>    > > +      aux_offset = surface->mt->mcs_mt->bo->offset64;
>    > > +
>    > > +      /* We only really need a clear color if we also have an
>    auxiliary
>    > > +       * surface.  Without one, it does nothing.
>    > > +       */
>    > > +      clear_color = intel_miptree_get_isl_clear_color(brw,
>    surface->mt);
>    > > +   }
>    > > +
>    > > +   struct isl_view view = {
>    > > +      .format = surface->brw_surfaceformat,
>    > > +      .base_level = 0,
>    > > +      .levels = 1,
>    > > +      .base_array_layer = 0,
>    > > +      .array_len = 1,
>    > > +      .channel_select = {
>    > > +         ISL_CHANNEL_SELECT_RED,
>    > > +         ISL_CHANNEL_SELECT_GREEN,
>    > > +         ISL_CHANNEL_SELECT_BLUE,
>    > > +         ISL_CHANNEL_SELECT_ALPHA,
>    > > +      },
>    > > +      .usage = is_render_target ? ISL_SURF_USAGE_RENDER_TARGET_BIT
>    :
>    > > +                                  ISL_SURF_USAGE_TEXTURE_BIT,
>    > > +   };
>    > > +
>    > > +   uint32_t offset, tile_x, tile_y;
>    > > +   offset = brw_blorp_compute_tile_offsets(surface, &tile_x,
>    &tile_y);
>    > > +
>    > > +   uint32_t surf_offset;
>    > > +   uint32_t *dw = brw_state_batch(brw, AUB_TRACE_SURFACE_STATE,
>    > > +                                  ss_info.num_dwords * 4,
>    ss_info.ss_align,
>    > > +                                  &surf_offset);
>    > > +
>    > > +   const uint32_t mocs = is_render_target ? ss_info.rb_mocs :
>    ss_info.tex_mocs;
>    > > +
>    > > +   isl_surf_fill_state(&isl_dev, dw, .surf = &surf, .view = &view,
>    > > +                       .address = surface->mt->bo->offset64 +
>    offset,
>    > > +                       .aux_surf = aux_surf, .aux_usage =
>    aux_usage,
>    > > +                       .aux_address = aux_offset,
>    > > +                       .mocs = mocs, .clear_color = clear_color,
>    > > +                       .x_offset_sa = tile_x, .y_offset_sa =
>    tile_y);
>    > > +
>    > > +   /* Emit relocation to surface contents */
>    > > +   drm_intel_bo_emit_reloc(brw->[3]batch.bo,
>    > > +                           surf_offset + ss_info.reloc_dw * 4,
>    > > +                           surface->mt->bo,
>    > > +                           dw[ss_info.reloc_dw] -
>    surface->mt->bo->offset64,
>    > > +                           read_domains, write_domain);
>    > > +
>    > > +   if (aux_surf) {
>    > > +      /* On gen7 and prior, the bottom 12 bits of the MCS base
>    address are
>    > > +       * used to store other information.  This should be ok,
>    however, because
>    > > +       * surface buffer addresses are always 4K page alinged.
>    > > +       */
>    > > +      assert((aux_offset & 0xfff) == 0);
>    > > +      drm_intel_bo_emit_reloc(brw->[4]batch.bo,
>    > > +                              surf_offset + ss_info.aux_reloc_dw *
>    4,
>    > > +                              surface->mt->mcs_mt->bo,
>    > > +                              dw[ss_info.aux_reloc_dw] & 0xfff,
>    > > +                              read_domains, write_domain);
>    > > +   }
>    > > +
>    > > +   return surf_offset;
>    > > +}
>    > > +
>    > >  /**
>    > >   * Perform a HiZ or depth resolve operation.
>    > >   *
>    > > diff --git a/src/mesa/drivers/dri/i965/brw_blorp.h
>    b/src/mesa/drivers/dri/i965/brw_blorp.h
>    > > index 7ec5875..bacdecb 100644
>    > > --- a/src/mesa/drivers/dri/i965/brw_blorp.h
>    > > +++ b/src/mesa/drivers/dri/i965/brw_blorp.h
>    > > @@ -372,6 +372,12 @@ brw_blorp_compile_nir_shader(struct
>    brw_context *brw, struct nir_shader *nir,
>    > >                               struct brw_blorp_prog_data
>    *prog_data,
>    > >                               unsigned *program_size);
>    > >
>    > > +uint32_t
>    > > +brw_blorp_emit_surface_state(struct brw_context *brw,
>    > > +                             const struct brw_blorp_surface_info
>    *surface,
>    > > +                             uint32_t read_domains, uint32_t
>    write_domain,
>    > > +                             bool is_render_target);
>    > > +
>    > >  void
>    > >  gen6_blorp_init(struct brw_context *brw);
>    > >
>    > > --
>    > > 2.5.0.400.gff86faf
>    > >
>    > > _______________________________________________
>    > > mesa-dev mailing list
>    > > [5]mesa-dev at lists.freedesktop.org
>    > > [6]https://lists.freedesktop.org/mailman/listinfo/mesa-dev
> 
> References
> 
>    1. mailto:topi.pohjolainen at intel.com
>    2. mailto:topi.pohjolainen at intel.com
>    3. http://batch.bo/
>    4. http://batch.bo/
>    5. mailto:mesa-dev at lists.freedesktop.org
>    6. https://lists.freedesktop.org/mailman/listinfo/mesa-dev


More information about the mesa-dev mailing list