From f52ce6a0ca73d1cd89091689efd8ea2e14748723 Mon Sep 17 00:00:00 2001 From: Chris Forbes Date: Thu, 24 Jan 2013 21:35:15 +1300 Subject: [PATCH] i965: add a new virtual opcode: SHADER_OPCODE_TXF_MS This is very similar to the TXF opcode, but lowers to `ld2dms` rather than `ld` on Gen7. V4: - add SHADER_OPCODE_TXF_MS to is_tex() functions, so regalloc thinks it actually writes the correct number of registers. Otherwise in nontrivial shaders some of the registers tend to get clobbered, producing bad results. Signed-off-by: Chris Forbes Reviewed-by: Paul Berry --- src/mesa/drivers/dri/i965/brw_defines.h | 1 + src/mesa/drivers/dri/i965/brw_fs.cpp | 2 ++ src/mesa/drivers/dri/i965/brw_fs_emit.cpp | 7 +++++++ src/mesa/drivers/dri/i965/brw_vec4.cpp | 1 + src/mesa/drivers/dri/i965/brw_vec4_emit.cpp | 7 +++++++ 5 files changed, 18 insertions(+) diff --git a/src/mesa/drivers/dri/i965/brw_defines.h b/src/mesa/drivers/dri/i965/brw_defines.h index d0794c841ab..d9b7f9aeb15 100644 --- a/src/mesa/drivers/dri/i965/brw_defines.h +++ b/src/mesa/drivers/dri/i965/brw_defines.h @@ -709,6 +709,7 @@ enum opcode { SHADER_OPCODE_TXL, SHADER_OPCODE_TXS, FS_OPCODE_TXB, + SHADER_OPCODE_TXF_MS, SHADER_OPCODE_SHADER_TIME_ADD, diff --git a/src/mesa/drivers/dri/i965/brw_fs.cpp b/src/mesa/drivers/dri/i965/brw_fs.cpp index fe347018472..7cc49347d83 100644 --- a/src/mesa/drivers/dri/i965/brw_fs.cpp +++ b/src/mesa/drivers/dri/i965/brw_fs.cpp @@ -336,6 +336,7 @@ fs_inst::is_tex() opcode == FS_OPCODE_TXB || opcode == SHADER_OPCODE_TXD || opcode == SHADER_OPCODE_TXF || + opcode == SHADER_OPCODE_TXF_MS || opcode == SHADER_OPCODE_TXL || opcode == SHADER_OPCODE_TXS); } @@ -740,6 +741,7 @@ fs_visitor::implied_mrf_writes(fs_inst *inst) case FS_OPCODE_TXB: case SHADER_OPCODE_TXD: case SHADER_OPCODE_TXF: + case SHADER_OPCODE_TXF_MS: case SHADER_OPCODE_TXL: case SHADER_OPCODE_TXS: return 1; diff --git a/src/mesa/drivers/dri/i965/brw_fs_emit.cpp b/src/mesa/drivers/dri/i965/brw_fs_emit.cpp index a25f594d9d8..2391ad12026 100644 --- a/src/mesa/drivers/dri/i965/brw_fs_emit.cpp +++ b/src/mesa/drivers/dri/i965/brw_fs_emit.cpp @@ -398,6 +398,12 @@ fs_generator::generate_tex(fs_inst *inst, struct brw_reg dst, struct brw_reg src case SHADER_OPCODE_TXF: msg_type = GEN5_SAMPLER_MESSAGE_SAMPLE_LD; break; + case SHADER_OPCODE_TXF_MS: + if (intel->gen >= 7) + msg_type = GEN7_SAMPLER_MESSAGE_SAMPLE_LD2DMS; + else + msg_type = GEN5_SAMPLER_MESSAGE_SAMPLE_LD; + break; default: assert(!"not reached"); break; @@ -1236,6 +1242,7 @@ fs_generator::generate_code(exec_list *instructions) case FS_OPCODE_TXB: case SHADER_OPCODE_TXD: case SHADER_OPCODE_TXF: + case SHADER_OPCODE_TXF_MS: case SHADER_OPCODE_TXL: case SHADER_OPCODE_TXS: generate_tex(inst, dst, src[0]); diff --git a/src/mesa/drivers/dri/i965/brw_vec4.cpp b/src/mesa/drivers/dri/i965/brw_vec4.cpp index 70f6b9d37f5..8182dccb02f 100644 --- a/src/mesa/drivers/dri/i965/brw_vec4.cpp +++ b/src/mesa/drivers/dri/i965/brw_vec4.cpp @@ -146,6 +146,7 @@ vec4_instruction::is_tex() return (opcode == SHADER_OPCODE_TEX || opcode == SHADER_OPCODE_TXD || opcode == SHADER_OPCODE_TXF || + opcode == SHADER_OPCODE_TXF_MS || opcode == SHADER_OPCODE_TXL || opcode == SHADER_OPCODE_TXS); } diff --git a/src/mesa/drivers/dri/i965/brw_vec4_emit.cpp b/src/mesa/drivers/dri/i965/brw_vec4_emit.cpp index cc2f5d4c8b0..b73711cf7ad 100644 --- a/src/mesa/drivers/dri/i965/brw_vec4_emit.cpp +++ b/src/mesa/drivers/dri/i965/brw_vec4_emit.cpp @@ -289,6 +289,12 @@ vec4_generator::generate_tex(vec4_instruction *inst, case SHADER_OPCODE_TXF: msg_type = GEN5_SAMPLER_MESSAGE_SAMPLE_LD; break; + case SHADER_OPCODE_TXF_MS: + if (intel->gen >= 7) + msg_type = GEN7_SAMPLER_MESSAGE_SAMPLE_LD2DMS; + else + msg_type = GEN5_SAMPLER_MESSAGE_SAMPLE_LD; + break; case SHADER_OPCODE_TXS: msg_type = GEN5_SAMPLER_MESSAGE_SAMPLE_RESINFO; break; @@ -645,6 +651,7 @@ vec4_generator::generate_vs_instruction(vec4_instruction *instruction, case SHADER_OPCODE_TEX: case SHADER_OPCODE_TXD: case SHADER_OPCODE_TXF: + case SHADER_OPCODE_TXF_MS: case SHADER_OPCODE_TXL: case SHADER_OPCODE_TXS: generate_tex(inst, dst, src[0]); -- 2.30.2