[PATCH] D69591: Devirtualize a call on alloca without waiting for post inline cleanup and next DevirtSCCRepeatedPass iteration.

[David Li via Phabricator via llvm-commits]

davidxl added inline comments.


================
Comment at: llvm/lib/Analysis/Loads.cpp:458
+      // alias analysis.
+      APInt LoadOffset(DL.getTypeSizeInBits(Ptr->getType()), 0);
+      APInt StoreOffset(DL.getTypeSizeInBits(
----------------
what is the change intended for? If it fixes some bugs, it should be split out.


================
Comment at: llvm/lib/Transforms/IPO/Inliner.cpp:897
+static void tryPromoteCall(CallSite &CS) {
+  assert(!CS.getCalledFunction());
+  Module *M = CS.getCaller()->getParent();
----------------
The pattern recognition part can be extracted as a utility/helper function and put int perhaps IndirectCallPromotionAnalysis.h|cpp file.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D69591/new/

https://reviews.llvm.org/D69591