[Freedreno] [PATCH 2/2] freedreno/ir3: add a pass to lower tg4 to txl, enable gather on a4xx

Rob Clark robdclark at gmail.com
Mon Nov 20 17:47:02 UTC 2017


On Sun, Nov 19, 2017 at 2:54 PM, Ilia Mirkin <imirkin at alum.mit.edu> wrote:
> Unfortunately Adreno A4xx hardware returns incorrect results with the
> GATHER4 opcodes. As a result, we have to lower to 4 individual texture
> calls (txl since we have to force lod to 0). We achieve this using
> offsets, including on cube maps which normally never have offsets.
>
> Signed-off-by: Ilia Mirkin <imirkin at alum.mit.edu>
> ---
>
> This pass relies on the hw doing the "right thing", working with nonconst
> offsets, and not having the usual limits (since the gather offset will in
> effect get offset by another 1).
>
> It fails two tests out of all the gather ones:
>
> bin/zero-tex-coord textureGather
> tests/spec/arb_gpu_shader5/execution/built-in-functions/fs-textureGatherOffset-uniform-array-offset.shader_test
>
> We haven't fully investigated why yet, but this is a good start.
>
> Note that the blob does this differently - they modify the source coordinate.
> However this seems unnecessary given that the hw can be made to use the
> offsets.
>
> Also please note that my knowledge of nir is minimal. Please carefully check
> that I used the right helpers/etc. This was largely a result of seeing what
> doesn't result in assertions.
>
>  docs/features.txt                                  |   4 +-
>  src/gallium/drivers/freedreno/Makefile.sources     |   1 +
>  src/gallium/drivers/freedreno/freedreno_screen.c   |   2 +-
>  .../drivers/freedreno/ir3/ir3_compiler_nir.c       |   7 +-
>  src/gallium/drivers/freedreno/ir3/ir3_nir.c        |   2 +
>  src/gallium/drivers/freedreno/ir3/ir3_nir.h        |   1 +
>  .../freedreno/ir3/ir3_nir_lower_tg4_to_tex.c       | 139 +++++++++++++++++++++
>  src/gallium/drivers/freedreno/meson.build          |   1 +
>  8 files changed, 152 insertions(+), 5 deletions(-)
>  create mode 100644 src/gallium/drivers/freedreno/ir3/ir3_nir_lower_tg4_to_tex.c
>
> diff --git a/docs/features.txt b/docs/features.txt
> index 633d2593738..99fb1715e0b 100644
> --- a/docs/features.txt
> +++ b/docs/features.txt
> @@ -130,7 +130,7 @@ GL 4.0, GLSL 4.00 --- all DONE: i965/gen7+, nvc0, r600, radeonsi
>    GL_ARB_tessellation_shader                            DONE (i965/gen7+)
>    GL_ARB_texture_buffer_object_rgb32                    DONE (freedreno, i965/gen6+, llvmpipe, softpipe, swr)
>    GL_ARB_texture_cube_map_array                         DONE (i965/gen6+, nv50, llvmpipe, softpipe)
> -  GL_ARB_texture_gather                                 DONE (freedreno/a5xx, i965/gen6+, nv50, llvmpipe, softpipe, swr)
> +  GL_ARB_texture_gather                                 DONE (freedreno, i965/gen6+, nv50, llvmpipe, softpipe, swr)
>    GL_ARB_texture_query_lod                              DONE (freedreno, i965, nv50, llvmpipe, softpipe)
>    GL_ARB_transform_feedback2                            DONE (i965/gen6+, nv50, llvmpipe, softpipe, swr)
>    GL_ARB_transform_feedback3                            DONE (i965/gen7+, llvmpipe, softpipe, swr)
> @@ -256,7 +256,7 @@ GLES3.1, GLSL ES 3.1 -- all DONE: i965/hsw+, nvc0, radeonsi
>    GL_ARB_texture_multisample (Multisample textures)     DONE (i965/gen7+, nv50, r600, llvmpipe, softpipe)
>    GL_ARB_texture_storage_multisample                    DONE (all drivers that support GL_ARB_texture_multisample)
>    GL_ARB_vertex_attrib_binding                          DONE (all drivers)
> -  GS5 Enhanced textureGather                            DONE (i965/gen7+, r600)
> +  GS5 Enhanced textureGather                            DONE (freedreno, i965/gen7+, r600)
>    GS5 Packing/bitfield/conversion functions             DONE (i965/gen6+, r600)
>    GL_EXT_shader_integer_mix                             DONE (all drivers that support GLSL)
>
> diff --git a/src/gallium/drivers/freedreno/Makefile.sources b/src/gallium/drivers/freedreno/Makefile.sources
> index b109a5a7a21..40c2eff0455 100644
> --- a/src/gallium/drivers/freedreno/Makefile.sources
> +++ b/src/gallium/drivers/freedreno/Makefile.sources
> @@ -168,6 +168,7 @@ ir3_SOURCES := \
>         ir3/ir3_nir.c \
>         ir3/ir3_nir.h \
>         ir3/ir3_nir_lower_if_else.c \
> +       ir3/ir3_nir_lower_tg4_to_tex.c \
>         ir3/ir3_print.c \
>         ir3/ir3_ra.c \
>         ir3/ir3_sched.c \
> diff --git a/src/gallium/drivers/freedreno/freedreno_screen.c b/src/gallium/drivers/freedreno/freedreno_screen.c
> index e61344fd104..62e4a574b90 100644
> --- a/src/gallium/drivers/freedreno/freedreno_screen.c
> +++ b/src/gallium/drivers/freedreno/freedreno_screen.c
> @@ -264,7 +264,7 @@ fd_screen_get_param(struct pipe_screen *pscreen, enum pipe_cap param)
>                 return 0;
>
>         case PIPE_CAP_MAX_TEXTURE_GATHER_COMPONENTS:
> -               if (is_a5xx(screen))
> +               if (is_a4xx(screen) || is_a5xx(screen))
>                         return 4;
>                 return 0;
>
> diff --git a/src/gallium/drivers/freedreno/ir3/ir3_compiler_nir.c b/src/gallium/drivers/freedreno/ir3/ir3_compiler_nir.c
> index da4aeaa7acb..c97df4f1d63 100644
> --- a/src/gallium/drivers/freedreno/ir3/ir3_compiler_nir.c
> +++ b/src/gallium/drivers/freedreno/ir3/ir3_compiler_nir.c
> @@ -2399,9 +2399,12 @@ emit_tex(struct ir3_context *ctx, nir_tex_instr *tex)
>          */
>         if (has_off | has_lod | has_bias) {
>                 if (has_off) {
> -                       for (i = 0; i < coords; i++)
> +                       unsigned off_coords = coords;
> +                       if (tex->sampler_dim == GLSL_SAMPLER_DIM_CUBE)
> +                               off_coords--;
> +                       for (i = 0; i < off_coords; i++)
>                                 src1[nsrc1++] = off[i];
> -                       if (coords < 2)
> +                       if (off_coords < 2)
>                                 src1[nsrc1++] = create_immed(b, fui(0.0));
>                         flags |= IR3_INSTR_O;
>                 }
> diff --git a/src/gallium/drivers/freedreno/ir3/ir3_nir.c b/src/gallium/drivers/freedreno/ir3/ir3_nir.c
> index 7dd24e5f4ee..39d3bf15fde 100644
> --- a/src/gallium/drivers/freedreno/ir3/ir3_nir.c
> +++ b/src/gallium/drivers/freedreno/ir3/ir3_nir.c
> @@ -188,6 +188,8 @@ ir3_optimize_nir(struct ir3_shader *shader, nir_shader *s,
>
>         OPT_V(s, nir_lower_tex, &tex_options);
>         OPT_V(s, nir_lower_load_const_to_scalar);
> +       if (shader->compiler->gpu_id < 500)
> +               OPT_V(s, ir3_nir_lower_tg4_to_tex);
>
>         ir3_optimize_loop(s);
>
> diff --git a/src/gallium/drivers/freedreno/ir3/ir3_nir.h b/src/gallium/drivers/freedreno/ir3/ir3_nir.h
> index e0e3108e328..ca873407fae 100644
> --- a/src/gallium/drivers/freedreno/ir3/ir3_nir.h
> +++ b/src/gallium/drivers/freedreno/ir3/ir3_nir.h
> @@ -38,6 +38,7 @@ void ir3_nir_scan_driver_consts(nir_shader *shader, struct ir3_driver_const_layo
>
>  bool ir3_nir_lower_if_else(nir_shader *shader);
>  bool ir3_nir_apply_trig_workarounds(nir_shader *shader);
> +bool ir3_nir_lower_tg4_to_tex(nir_shader *shader);
>
>  struct nir_shader * ir3_tgsi_to_nir(const struct tgsi_token *tokens);
>  const nir_shader_compiler_options * ir3_get_compiler_options(struct ir3_compiler *compiler);
> diff --git a/src/gallium/drivers/freedreno/ir3/ir3_nir_lower_tg4_to_tex.c b/src/gallium/drivers/freedreno/ir3/ir3_nir_lower_tg4_to_tex.c
> new file mode 100644
> index 00000000000..087688cd234
> --- /dev/null
> +++ b/src/gallium/drivers/freedreno/ir3/ir3_nir_lower_tg4_to_tex.c
> @@ -0,0 +1,139 @@
> +/*
> + * Copyright © 2017 Ilia Mirkin
> + *
> + * Permission is hereby granted, free of charge, to any person obtaining a
> + * copy of this software and associated documentation files (the "Software"),
> + * to deal in the Software without restriction, including without limitation
> + * the rights to use, copy, modify, merge, publish, distribute, sublicense,
> + * and/or sell copies of the Software, and to permit persons to whom the
> + * Software is furnished to do so, subject to the following conditions:
> + *
> + * The above copyright notice and this permission notice (including the next
> + * paragraph) shall be included in all copies or substantial portions of the
> + * Software.
> + *
> + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
> + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
> + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
> + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
> + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
> + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
> + * IN THE SOFTWARE.
> + *
> + */
> +
> +#include "ir3_nir.h"
> +#include "compiler/nir/nir_builder.h"
> +
> +/* A4XX has a broken GATHER4 operation. It performs the texture swizzle on the
> + * gather results, rather than before. As a result, it must be emulated with
> + * direct texture calls.
> + */
> +
> +static bool
> +lower_tg4(nir_block *block, nir_builder *b, void *mem_ctx)
> +{
> +       bool progress = false;
> +
> +       static const int offsets[3][2] = { {0, 1}, {1, 1}, {1, 0} };
> +
> +       nir_foreach_instr_safe(instr, block) {
> +               if (instr->type != nir_instr_type_tex)
> +                       continue;
> +
> +        nir_tex_instr *tg4 = (nir_tex_instr *)instr;
> +
> +               if (tg4->op != nir_texop_tg4)
> +                       continue;
> +
> +               b->cursor = nir_before_instr(&tg4->instr);
> +
> +               nir_ssa_def *results[4];
> +               int offset_index = nir_tex_instr_src_index(tg4, nir_tex_src_offset);
> +               for (int i = 0; i < 4; i++) {
> +                       int num_srcs = tg4->num_srcs + 1 /* lod */;
> +                       if (offset_index < 0 && i < 3)
> +                               num_srcs++;
> +
> +                       nir_tex_instr *tex = nir_tex_instr_create(b->shader, num_srcs);
> +                       tex->op = nir_texop_txl;
> +                       tex->sampler_dim = tg4->sampler_dim;
> +                       tex->coord_components = tg4->coord_components;
> +                       tex->is_array = tg4->is_array;
> +                       tex->is_shadow = tg4->is_shadow;
> +                       tex->is_new_style_shadow = tg4->is_new_style_shadow;
> +                       tex->texture_index = tg4->texture_index;
> +                       tex->texture = nir_deref_var_clone(tg4->texture, tex);
> +                       tex->sampler_index = tg4->sampler_index;
> +                       tex->sampler = nir_deref_var_clone(tg4->sampler, tex);
> +                       tex->dest_type = tg4->dest_type;
> +
> +                       for (int j = 0; j < tg4->num_srcs; j++) {
> +                               nir_src_copy(&tex->src[j].src, &tg4->src[j].src, tex);
> +                               tex->src[j].src_type = tg4->src[j].src_type;
> +                       }
> +                       if (i != 3) {
> +                               nir_ssa_def *offset =
> +                                       nir_vec2(b, nir_imm_int(b, offsets[i][0]),
> +                                                        nir_imm_int(b, offsets[i][1]));
> +                               if (offset_index < 0) {
> +                                       tex->src[tg4->num_srcs].src = nir_src_for_ssa(offset);
> +                                       tex->src[tg4->num_srcs].src_type = nir_tex_src_offset;
> +                               } else {
> +                                       assert(tex->src[offset_index].src.is_ssa);
> +                                       nir_ssa_def *orig = tex->src[offset_index].src.ssa;

nir_ssa_for_src()?

> +                                       tex->src[offset_index].src =
> +                                               nir_src_for_ssa(nir_iadd(b, orig, offset));
> +                               }
> +                       }
> +                       tex->src[num_srcs - 1].src = nir_src_for_ssa(nir_imm_float(b, 0));
> +                       tex->src[num_srcs - 1].src_type = nir_tex_src_lod;
> +
> +                       nir_ssa_dest_init(&tex->instr, &tex->dest,
> +                                                         nir_tex_instr_dest_size(tex), 32, NULL);
> +                       nir_builder_instr_insert(b, &tex->instr);
> +
> +                       results[i] = nir_channel(b, &tex->dest.ssa, tg4->component);
> +               }
> +
> +               nir_ssa_def *result = nir_vec4(b, results[0], results[1], results[2], results[3]);
> +               nir_ssa_def_rewrite_uses(&tg4->dest.ssa, nir_src_for_ssa(result));
> +
> +               nir_instr_remove(&tg4->instr);
> +
> +               progress = true;
> +       }
> +
> +       return progress;
> +}
> +
> +static bool
> +lower_tg4_func(nir_function_impl *impl)
> +{
> +       void *mem_ctx = ralloc_parent(impl);
> +       nir_builder b;
> +       nir_builder_init(&b, impl);
> +
> +       bool progress = false;
> +       nir_foreach_block_safe(block, impl) {
> +               progress |= lower_tg4(block, &b, mem_ctx);
> +       }
> +
> +       if (progress)
> +               nir_metadata_preserve(impl, nir_metadata_none);

we aren't changing flow control, so I think:

   nir_metadata_preserve(impl, nir_metadata_block_index |
                               nir_metadata_dominance);

Other than that, looks reasonable.

BR,
-R

> +
> +       return progress;
> +}
> +
> +bool
> +ir3_nir_lower_tg4_to_tex(nir_shader *shader)
> +{
> +       bool progress = false;
> +
> +       nir_foreach_function(function, shader) {
> +               if (function->impl)
> +                       progress |= lower_tg4_func(function->impl);
> +       }
> +
> +       return progress;
> +}
> diff --git a/src/gallium/drivers/freedreno/meson.build b/src/gallium/drivers/freedreno/meson.build
> index d2b901334d0..01b5836be29 100644
> --- a/src/gallium/drivers/freedreno/meson.build
> +++ b/src/gallium/drivers/freedreno/meson.build
> @@ -188,6 +188,7 @@ files_libfreedreno = files(
>    'ir3/ir3_nir.c',
>    'ir3/ir3_nir.h',
>    'ir3/ir3_nir_lower_if_else.c',
> +  'ir3/ir3_nir_lower_tg4_to_tex.c',
>    'ir3/ir3_print.c',
>    'ir3/ir3_ra.c',
>    'ir3/ir3_sched.c',
> --
> 2.13.6
>
> _______________________________________________
> Freedreno mailing list
> Freedreno at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/freedreno


More information about the Freedreno mailing list