[PATCH] D34056: Tail merge size

[Simon Pilgrim via Phabricator via llvm-commits]

RKSimon added reviewers: davide, filcab.
RKSimon added a comment.

A few minors, but I'd prefer someone more knowledgable in this area to review it.

You may need to bring in some ARM/AMDGPU/PowerPC guys too ?



================
Comment at: test/CodeGen/X86/loop-search.ll:2
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
 ; RUN: llc < %s -mtriple=x86_64-apple-darwin | FileCheck %s
----------------
Remove this line - we should only be generating with update_llc_test_checks.py


================
Comment at: test/CodeGen/X86/loop-search.ll:34
 ; CHECK-NEXT:    retq
-;
+;            ### FIXME: BB#3 and LBB0_1 should be merged
 entry:
----------------
I think this comment can go now?


================
Comment at: test/CodeGen/X86/tail-opts.ll:109
 ; with only a branch in common, regardless of the fallthrough situation.
-
 ; CHECK-LABEL: dont_merge_oddly:
----------------
Keep these whitespace if you can - minimize diff


================
Comment at: test/CodeGen/X86/tail-opts.ll:415
 ; Now two instructions are enough to be tail-duplicated.
-
 ; CHECK-LABEL: two_nosize:
----------------
Keep these whitespace if you can - minimize diff


https://reviews.llvm.org/D34056