[PATCH] D11904: [X86] Disable mul -> shl + lea combine when compiling for minsize

[Sanjay Patel via llvm-commits]

spatel added inline comments.

================
Comment at: lib/Target/X86/X86ISelLowering.cpp:23521
@@ -23520,1 +23520,3 @@
                                  TargetLowering::DAGCombinerInfo &DCI) {
+  // An imul is usually smaller than the alternative sequence
+  if (DAG.getMachineFunction().getFunction()->optForMinSize())
----------------
Missing period.

================
Comment at: test/CodeGen/X86/imul.ll:112-119
@@ +111,10 @@
+
+define i32 @mul4_32_minsize(i32 %A) minsize {
+; X64-LABEL: mul4_32_minsize:
+; X64: leal
+; X86-LABEL: mul4_32_minsize:
+; X86: shll
+    %mul = mul i32 %A, 4
+    ret i32 %mul
+}
+
----------------
Does this test change with the patch? Is there a reason i386 should be different than x86-64?


http://reviews.llvm.org/D11904