I've noticed this while trying to move addRuntimeCheck to LoopAccessAnalysis.
I think that the intention was to early exit from the overflow checking before
the code for the memchecks. This is the entire reason why we compute
FirstCheckInst but then we don't use that as the splitting instruction but the
final check. Looks like an oversight.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@228056
91177308-0d34-0410-b5e6-
96231b3b80d8
if (MemRuntimeCheck) {
// Create a new block containing the memory check.
BasicBlock *CheckBlock =
- LastBypassBlock->splitBasicBlock(MemRuntimeCheck, "vector.memcheck");
+ LastBypassBlock->splitBasicBlock(FirstCheckInst, "vector.memcheck");
if (ParentLoop)
ParentLoop->addBasicBlockToLoop(CheckBlock, *LI);
LoopBypassBlocks.push_back(CheckBlock);