cmd/compile: cache CFG-dependent computations

We compute a lot of stuff based off the CFG: postorder traversal,
dominators, dominator tree, loop nest.  Multiple phases use this
information and we end up recomputing some of it.  Add a cache
for this information so if the CFG hasn't changed, we can reuse
the previous computation.

Change-Id: I9b5b58af06830bd120afbee9cfab395a0a2f74b2
Reviewed-on: https://go-review.googlesource.com/29356
Reviewed-by: David Chase <drchase@google.com>

1 files changed, 2 insertions, 6 deletions

diff --git a/src/cmd/compile/internal/ssa/flagalloc.go b/src/cmd/compile/internal/ssa/flagalloc.go
index 5d1ced42b4..24b6a0ec89 100644
--- a/src/cmd/compile/internal/ssa/flagalloc.go
+++ b/src/cmd/compile/internal/ssa/flagalloc.go
@@ -11,14 +11,10 @@ func flagalloc(f *Func) {
 	// Compute the in-register flag value we want at the end of
 	// each block. This is basically a best-effort live variable
 	// analysis, so it can be much simpler than a full analysis.
-	// TODO: do we really need to keep flag values live across blocks?
-	// Could we force the flags register to be unused at basic block
-	// boundaries?  Then we wouldn't need this computation.
 	end := make([]*Value, f.NumBlocks())
+	po := f.postorder()
 	for n := 0; n < 2; n++ {
-		// Walk blocks backwards. Poor-man's postorder traversal.
-		for i := len(f.Blocks) - 1; i >= 0; i-- {
-			b := f.Blocks[i]
+		for _, b := range po {
 			// Walk values backwards to figure out what flag
 			// value we want in the flag register at the start
 			// of the block.