diff options
author | Rob Clark <[email protected]> | 2015-04-22 13:07:33 -0400 |
---|---|---|
committer | Rob Clark <[email protected]> | 2015-04-22 13:20:28 -0400 |
commit | 632ea2a1139f4b228ca55331e411dbae9920c28d (patch) | |
tree | 90bf43853f598af9aa8a347a740d10394c6f778b | |
parent | 13527df143be1ca6e2f69bf4728f69efaebb3b13 (diff) |
freedreno/nir: sysval fixes
Signed-off-by: Rob Clark <[email protected]>
-rw-r--r-- | src/gallium/drivers/freedreno/a4xx/fd4_emit.c | 2 | ||||
-rw-r--r-- | src/gallium/drivers/freedreno/ir3/ir3_compiler_nir.c | 15 |
2 files changed, 12 insertions, 5 deletions
diff --git a/src/gallium/drivers/freedreno/a4xx/fd4_emit.c b/src/gallium/drivers/freedreno/a4xx/fd4_emit.c index 7a244d8a4c5..353873daa05 100644 --- a/src/gallium/drivers/freedreno/a4xx/fd4_emit.c +++ b/src/gallium/drivers/freedreno/a4xx/fd4_emit.c @@ -302,7 +302,7 @@ fd4_emit_vertex_bufs(struct fd_ringbuffer *ring, struct fd4_emit *emit) for (i = 0; i < vp->inputs_count; i++) { uint8_t semantic = sem2name(vp->inputs[i].semantic); - if (semantic == TGSI_SEMANTIC_VERTEXID) + if (semantic == TGSI_SEMANTIC_VERTEXID_NOBASE) vertex_regid = vp->inputs[i].regid; else if (semantic == TGSI_SEMANTIC_INSTANCEID) instance_regid = vp->inputs[i].regid; diff --git a/src/gallium/drivers/freedreno/ir3/ir3_compiler_nir.c b/src/gallium/drivers/freedreno/ir3/ir3_compiler_nir.c index e5f6c2a5e83..05e7049ad55 100644 --- a/src/gallium/drivers/freedreno/ir3/ir3_compiler_nir.c +++ b/src/gallium/drivers/freedreno/ir3/ir3_compiler_nir.c @@ -1140,6 +1140,7 @@ static void add_sysval_input(struct ir3_compile *ctx, unsigned name, so->inputs[n].interpolate = TGSI_INTERPOLATE_CONSTANT; so->total_in++; + ctx->block->ninputs = MAX2(ctx->block->ninputs, r + 1); ctx->block->inputs[r] = instr; } @@ -1785,16 +1786,22 @@ emit_instructions(struct ir3_compile *ctx) unsigned noutputs = exec_list_length(&ctx->s->outputs) * 4; /* we need to allocate big enough outputs array so that - * we can stuff the kill's at the end: + * we can stuff the kill's at the end. Likewise for vtx + * shaders, we need to leave room for sysvals: */ - if (ctx->so->type == SHADER_FRAGMENT) + if (ctx->so->type == SHADER_FRAGMENT) { noutputs += ARRAY_SIZE(ctx->kill); + } else if (ctx->so->type == SHADER_VERTEX) { + ninputs += 8; + } ctx->block = ir3_block_create(ctx->ir, 0, ninputs, noutputs); - if (ctx->so->type == SHADER_FRAGMENT) + if (ctx->so->type == SHADER_FRAGMENT) { ctx->block->noutputs -= ARRAY_SIZE(ctx->kill); - + } else if (ctx->so->type == SHADER_VERTEX) { + ctx->block->ninputs -= 8; + } /* for fragment shader, we have a single input register (usually * r0.xy) which is used as the base for bary.f varying fetch instrs: |