i965/fs: Calculate interference better in register_coalesce.

Previously we simply considered two registers whose live ranges
overlapped to interfere. Cases such as

   set A     ------
   ...             |
   mov B, A  --    |
   ...         | B | A
   use B     --    |
   ...             |
   use A     ------

would be considered to interfere, even though B is an unmodified copy of
A whose live range fit wholly inside that of A.

If no writes to A or B occur between the mov B, A and the use of B then
we can safely coalesce them.

Instead of removing MOV instructions, we make them NOPs and remove them
at once after the main pass is finished in order to avoid recomputing
live intervals (which are needed to perform the previous step).

total instructions in shared programs: 1543768 -> 1513077 (-1.99%)
instructions in affected programs:     951563 -> 920872 (-3.23%)
GAINED:                                46
LOST:                                  22

Reviewed-by: Jordan Justen <[email protected]>

1 files changed, 72 insertions, 7 deletions

diff --git a/src/mesa/drivers/dri/i965/brw_fs.cpp b/src/mesa/drivers/dri/i965/brw_fs.cpp
index 914f0619481..b00e2b5726e 100644
--- a/src/mesa/drivers/dri/i965/brw_fs.cpp
+++ b/src/mesa/drivers/dri/i965/brw_fs.cpp
@@ -2282,7 +2282,7 @@ fs_visitor::register_coalesce()
    int reg_to_offset[MAX_SAMPLER_MESSAGE_SIZE];
    fs_inst *mov[MAX_SAMPLER_MESSAGE_SIZE];
 
-   foreach_list_safe(node, &this->instructions) {
+   foreach_list(node, &this->instructions) {
       fs_inst *inst = (fs_inst *)node;
 
       if (inst->opcode != BRW_OPCODE_MOV ||
@@ -2305,8 +2305,49 @@ fs_visitor::register_coalesce()
       int var_to = live_intervals->var_from_reg(&inst->dst);
 
       if (live_intervals->vars_interfere(var_from, var_to) &&
-          !inst->dst.equals(inst->src[0]))
-         continue;
+          !inst->dst.equals(inst->src[0])) {
+
+         /* We know that the live ranges of A (var_from) and B (var_to)
+          * interfere because of the ->vars_interfere() call above. If the end
+          * of B's live range is after the end of A's range, then we know two
+          * things:
+          *  - the start of B's live range must be in A's live range (since we
+          *    already know the two ranges interfere, this is the only remaining
+          *    possibility)
+          *  - the interference isn't of the form we're looking for (where B is
+          *    entirely inside A)
+          */
+         if (live_intervals->end[var_to] > live_intervals->end[var_from])
+            continue;
+
+         bool overwritten = false;
+         int scan_ip = -1;
+
+         foreach_list(n, &this->instructions) {
+            fs_inst *scan_inst = (fs_inst *)n;
+            scan_ip++;
+
+            if (scan_inst->is_control_flow()) {
+               overwritten = true;
+               break;
+            }
+
+            if (scan_ip <= live_intervals->start[var_to])
+               continue;
+
+            if (scan_ip > live_intervals->end[var_to])
+               break;
+
+            if (scan_inst->dst.equals(inst->dst) ||
+                scan_inst->dst.equals(inst->src[0])) {
+               overwritten = true;
+               break;
+            }
+         }
+
+         if (overwritten)
+            continue;
+      }
 
       if (reg_from != inst->src[0].reg) {
          reg_from = inst->src[0].reg;
@@ -2331,9 +2372,18 @@ fs_visitor::register_coalesce()
       if (channels_remaining)
          continue;
 
+      bool removed = false;
       for (int i = 0; i < src_size; i++) {
-         if (mov[i])
-            mov[i]->remove();
+         if (mov[i]) {
+            removed = true;
+
+            mov[i]->opcode = BRW_OPCODE_NOP;
+            mov[i]->conditional_mod = BRW_CONDITIONAL_NONE;
+            mov[i]->dst = reg_undef;
+            mov[i]->src[0] = reg_undef;
+            mov[i]->src[1] = reg_undef;
+            mov[i]->src[2] = reg_undef;
+         }
       }
 
       foreach_list(node, &this->instructions) {
@@ -2355,11 +2405,26 @@ fs_visitor::register_coalesce()
                      scan_inst->src[j].reg_offset = reg_to_offset[i];
                   }
                }
-
-               progress = true;
             }
          }
       }
+
+      if (removed) {
+         live_intervals->start[var_to] = MIN2(live_intervals->start[var_to],
+                                              live_intervals->start[var_from]);
+         live_intervals->end[var_to] = MAX2(live_intervals->end[var_to],
+                                            live_intervals->end[var_from]);
+         reg_from = -1;
+      }
+   }
+
+   foreach_list_safe(node, &this->instructions) {
+      fs_inst *inst = (fs_inst *)node;
+
+      if (inst->opcode == BRW_OPCODE_NOP) {
+         inst->remove();
+         progress = true;
+      }
    }
 
    if (progress)