[llvm] r267384 - ThinLTO: Move createNameAnonFunctionPass insertion in PassManagerBuilder (NFC)
Mehdi Amini via llvm-commits
llvm-commits at lists.llvm.org
Mon Apr 25 01:47:37 PDT 2016
Author: mehdi_amini
Date: Mon Apr 25 03:47:37 2016
New Revision: 267384
URL: http://llvm.org/viewvc/llvm-project?rev=267384&view=rev
Log:
ThinLTO: Move createNameAnonFunctionPass insertion in PassManagerBuilder (NFC)
It is just code motion, but makes more sense this way.
From: Mehdi Amini <mehdi.amini at apple.com>
Modified:
llvm/trunk/lib/Transforms/IPO/PassManagerBuilder.cpp
Modified: llvm/trunk/lib/Transforms/IPO/PassManagerBuilder.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/IPO/PassManagerBuilder.cpp?rev=267384&r1=267383&r2=267384&view=diff
==============================================================================
--- llvm/trunk/lib/Transforms/IPO/PassManagerBuilder.cpp (original)
+++ llvm/trunk/lib/Transforms/IPO/PassManagerBuilder.cpp Mon Apr 25 03:47:37 2016
@@ -249,8 +249,6 @@ void PassManagerBuilder::addFunctionSimp
if (PrepareForThinLTO) {
MPM.add(createAggressiveDCEPass()); // Delete dead instructions
addInstructionCombiningPass(MPM); // Combine silly seq's
- // Rename anon function to export them
- MPM.add(createNameAnonFunctionPass());
return;
}
// Rotate Loop - disable header duplication at -Oz
@@ -405,8 +403,11 @@ void PassManagerBuilder::populateModuleP
// If we are planning to perform ThinLTO later, let's not bloat the code with
// unrolling/vectorization/... now. We'll first run the inliner + CGSCC passes
// during ThinLTO and perform the rest of the optimizations afterward.
- if (PrepareForThinLTO)
+ if (PrepareForThinLTO) {
+ // Rename anon function to be able to export them in the summary.
+ MPM.add(createNameAnonFunctionPass());
return;
+ }
// FIXME: This is a HACK! The inliner pass above implicitly creates a CGSCC
// pass manager that we are specifically trying to avoid. To prevent this
More information about the llvm-commits
mailing list