r600g: cayman fix integer multiplies

Looks insane, but it does appear we need a full slot per input/output.

This fixes another 180 or so piglit tests.

Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/src/gallium/drivers/r600/r600_shader.c b/src/gallium/drivers/r600/r600_shader.c
index 4389fe8..aac1cd3 100644 (file)
--- a/src/gallium/drivers/r600/r600_shader.c
+++ b/src/gallium/drivers/r600/r600_shader.c
@@ -1450,6 +1450,34 @@ static int cayman_emit_float_instr(struct r600_shader_ctx *ctx)
        return 0;
 }
 
+static int cayman_mul_int_instr(struct r600_shader_ctx *ctx)
+{
+       struct tgsi_full_instruction *inst = &ctx->parse.FullToken.FullInstruction;
+       int i, j, k, r;
+       struct r600_bytecode_alu alu;
+       int last_slot = (inst->Dst[0].Register.WriteMask & 0x8) ? 4 : 3;
+       for (k = 0; k < last_slot; k++) {
+               if (!(inst->Dst[0].Register.WriteMask & (1 << k)))
+                       continue;
+
+               for (i = 0 ; i < 4; i++) {
+                       memset(&alu, 0, sizeof(struct r600_bytecode_alu));
+                       alu.inst = ctx->inst_info->r600_opcode;
+                       for (j = 0; j < inst->Instruction.NumSrcRegs; j++) {
+                               r600_bytecode_src(&alu.src[j], &ctx->src[j], k);
+                       }
+                       tgsi_dst(ctx, &inst->Dst[0], i, &alu.dst);
+                       alu.dst.write = (i == k);
+                       if (i == 3)
+                               alu.last = 1;
+                       r = r600_bytecode_add_alu(ctx->bc, &alu);
+                       if (r)
+                               return r;
+               }
+       }
+       return 0;
+}
+
 /*
  * r600 - trunc to -PI..PI range
  * r700 - normalize by dividing by 2PI
@@ -4996,7 +5024,7 @@ static struct r600_shader_tgsi_instruction cm_shader_tgsi_instruction[] = {
        {TGSI_OPCODE_UMAX,      0, EG_V_SQ_ALU_WORD1_OP2_SQ_OP2_INST_MAX_UINT, tgsi_op2},
        {TGSI_OPCODE_UMIN,      0, EG_V_SQ_ALU_WORD1_OP2_SQ_OP2_INST_MIN_UINT, tgsi_op2},
        {TGSI_OPCODE_UMOD,      0, EG_V_SQ_ALU_WORD1_OP2_SQ_OP2_INST_NOP, tgsi_unsupported},
-       {TGSI_OPCODE_UMUL,      0, EG_V_SQ_ALU_WORD1_OP2_SQ_OP2_INST_MULLO_INT, tgsi_op2},
+       {TGSI_OPCODE_UMUL,      0, EG_V_SQ_ALU_WORD1_OP2_SQ_OP2_INST_MULLO_INT, cayman_mul_int_instr},
        {TGSI_OPCODE_USEQ,      0, EG_V_SQ_ALU_WORD1_OP2_SQ_OP2_INST_SETE_INT, tgsi_op2},
        {TGSI_OPCODE_USGE,      0, EG_V_SQ_ALU_WORD1_OP2_SQ_OP2_INST_SETGE_UINT, tgsi_op2},
        {TGSI_OPCODE_USHR,      0, EG_V_SQ_ALU_WORD1_OP2_SQ_OP2_INST_LSHR_INT, tgsi_op2},