projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 65aac0f) | patch
[AlignmentFromAssumptions] Don't crash just because the target is 32-bit
authorHal Finkel <hfinkel@anl.gov>
Thu, 11 Sep 2014 08:40:17 +0000 (08:40 +0000)
committerHal Finkel <hfinkel@anl.gov>
Thu, 11 Sep 2014 08:40:17 +0000 (08:40 +0000)
commit2c1a4875e8087ecee9744a609eff3b63e73eab42
tree76a4492522f62eaf52b15b7767cd3b280a6afdf3tree | snapshot
parent65aac0f8e350f36808f685459195d2ae99e8949fcommit | diff
[AlignmentFromAssumptions] Don't crash just because the target is 32-bit

We used to crash processing any relevant @llvm.assume on a 32-bit target
(because we'd ask SE to subtract expressions of differing types). I've copied
our 'simple.ll' test, but with the data layout from arm-linux-gnueabihf to get
some meaningful test coverage here.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@217574 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Transforms/Scalar/AlignmentFromAssumptions.cpp diff | blob | history
test/Transforms/AlignmentFromAssumptions/simple32.ll [new file with mode: 0644] blob
C/C++ LLVM-based compilers that forbids OOTA behaviors