radeonsi: Don't export unused clip distance vectors from vertex shader

E.g. the Source engine seems to always write to gl_ClipVertex, but normally
doesn't enable any GL_CLIP_DISTANCEn states. This change removes some
irrelevant parts from the generated vertex shader code in such cases.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>

diff --git a/src/gallium/drivers/radeonsi/radeonsi_shader.c b/src/gallium/drivers/radeonsi/radeonsi_shader.c
index dd9581dd95af16e5d643e182c7a7d22ffc3f4245..6bf4b05ec878eb33f6eb529defb1d592f651132e 100644 (file)
--- a/src/gallium/drivers/radeonsi/radeonsi_shader.c
+++ b/src/gallium/drivers/radeonsi/radeonsi_shader.c
@@ -565,6 +565,7 @@ static void si_llvm_emit_clipvertex(struct lp_build_tgsi_context * bld_base,
                                    LLVMValueRef (*pos)[9], unsigned index)
 {
        struct si_shader_context *si_shader_ctx = si_shader_context(bld_base);
+       struct si_pipe_shader *shader = si_shader_ctx->shader;
        struct lp_build_context *base = &bld_base->base;
        struct lp_build_context *uint = &si_shader_ctx->radeon_bld.soa.bld_base.uint_bld;
        unsigned reg_index;
@@ -583,6 +584,11 @@ static void si_llvm_emit_clipvertex(struct lp_build_tgsi_context * bld_base,
        for (reg_index = 0; reg_index < 2; reg_index ++) {
                LLVMValueRef *args = pos[2 + reg_index];
 
+               if (!(shader->key.vs.ucps_enabled & (1 << reg_index)))
+                       continue;
+
+               shader->shader.clip_dist_write |= 0xf << (4 * reg_index);
+
                args[5] =
                args[6] =
                args[7] =
@@ -709,13 +715,15 @@ handle_semantic:
                                }
                                break;
                        case TGSI_SEMANTIC_CLIPDIST:
+                               if (!(si_shader_ctx->shader->key.vs.ucps_enabled &
+                                     (1 << d->Semantic.Index)))
+                                       continue;
                                shader->clip_dist_write |=
                                        d->Declaration.UsageMask << (d->Semantic.Index << 2);
                                target = V_008DFC_SQ_EXP_POS + 2 + d->Semantic.Index;
                                break;
                        case TGSI_SEMANTIC_CLIPVERTEX:
                                si_llvm_emit_clipvertex(bld_base, pos_args, index);
-                               shader->clip_dist_write = 0xFF;
                                continue;
                        case TGSI_SEMANTIC_FOG:
                        case TGSI_SEMANTIC_GENERIC:

diff --git a/src/gallium/drivers/radeonsi/radeonsi_shader.h b/src/gallium/drivers/radeonsi/radeonsi_shader.h
index f28a0ea544399dfe574319c651fe20c0637339b4..2d4468a2f37a3e5721fc356c6d36f0ac3d46200f 100644 (file)
--- a/src/gallium/drivers/radeonsi/radeonsi_shader.h
+++ b/src/gallium/drivers/radeonsi/radeonsi_shader.h
@@ -128,6 +128,7 @@ union si_shader_key {
        } ps;
        struct {
                unsigned        instance_divisors[PIPE_MAX_ATTRIBS];
+               unsigned        ucps_enabled:2;
        } vs;
 };
 

diff --git a/src/gallium/drivers/radeonsi/si_state.c b/src/gallium/drivers/radeonsi/si_state.c
index 1dd51a8f93e49c859add0f58ecaf63e7b693aa8b..d1e3c9d5279962e8153c3740f9936c5fba122d5e 100644 (file)
--- a/src/gallium/drivers/radeonsi/si_state.c
+++ b/src/gallium/drivers/radeonsi/si_state.c
@@ -2040,6 +2040,10 @@ static INLINE void si_shader_selector_key(struct pipe_context *ctx,
                for (i = 0; i < rctx->vertex_elements->count; ++i)
                        key->vs.instance_divisors[i] = rctx->vertex_elements->elements[i].instance_divisor;
 
+               if (rctx->queued.named.rasterizer->clip_plane_enable & 0xf0)
+                       key->vs.ucps_enabled |= 0x2;
+               if (rctx->queued.named.rasterizer->clip_plane_enable & 0xf)
+                       key->vs.ucps_enabled |= 0x1;
        } else if (sel->type == PIPE_SHADER_FRAGMENT) {
                if (sel->fs_write_all)
                        key->ps.nr_cbufs = rctx->framebuffer.nr_cbufs;