[Mesa-dev] [PATCH 4/4] i965/vec4: Perform CSE on CMP(N) instructions.

[Matt Turner]

Port of commit b16b3c87 to the vec4 code.

No shader-db improvements, but might as well. The fs backend saw an
improvement because it's scalar and multiple identical CMP instructions
were generated by the SEL peepholes.
---
 src/mesa/drivers/dri/i965/brw_vec4_cse.cpp | 17 ++++++++++++++++-
 1 file changed, 16 insertions(+), 1 deletion(-)

diff --git a/src/mesa/drivers/dri/i965/brw_vec4_cse.cpp b/src/mesa/drivers/dri/i965/brw_vec4_cse.cpp
index 8013517..1456e3a 100644
--- a/src/mesa/drivers/dri/i965/brw_vec4_cse.cpp
+++ b/src/mesa/drivers/dri/i965/brw_vec4_cse.cpp
@@ -56,6 +56,8 @@ is_expression(const vec4_instruction *const inst)
    case BRW_OPCODE_SHR:
    case BRW_OPCODE_SHL:
    case BRW_OPCODE_ASR:
+   case BRW_OPCODE_CMP:
+   case BRW_OPCODE_CMPN:
    case BRW_OPCODE_ADD:
    case BRW_OPCODE_MUL:
    case BRW_OPCODE_FRC:
@@ -135,7 +137,7 @@ vec4_visitor::opt_cse_local(bblock_t *block, exec_list *aeb)
 
       /* Skip some cases. */
       if (is_expression(inst) && !inst->predicate && inst->mlen == 0 &&
-          !inst->conditional_mod)
+          (inst->dst.file != HW_REG || inst->dst.is_null()))
       {
          bool found = false;
 
@@ -200,6 +202,19 @@ vec4_visitor::opt_cse_local(bblock_t *block, exec_list *aeb)
       foreach_list_safe(entry_node, aeb) {
          aeb_entry *entry = (aeb_entry *)entry_node;
 
+         /* Kill all AEB entries that write a different value to or read from
+          * the flag register if we just wrote it.
+          */
+         if (inst->writes_flag()) {
+            if (entry->generator->reads_flag() ||
+                (entry->generator->writes_flag() &&
+                 !instructions_match(inst, entry->generator))) {
+               entry->remove();
+               ralloc_free(entry);
+               continue;
+            }
+         }
+
          for (int i = 0; i < 3; i++) {
             src_reg *src = &entry->generator->src[i];
 
-- 
1.8.3.2