projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: a00ef11) | patch
Fixed a couple of optimization with EXTRACT_VECTOR_ELT that assumes the result
authorMon P Wang <wangmp@apple.com>
Mon, 1 Feb 2010 19:03:18 +0000 (19:03 +0000)
committerMon P Wang <wangmp@apple.com>
Mon, 1 Feb 2010 19:03:18 +0000 (19:03 +0000)
commitc6654ec498698919ae8388c53fe91d0366ac24c7
treec3608b4f5d1cf0405958b6e73288a8c83a750884tree | snapshot
parenta00ef1122c6364c7f54a1c428654fea5d9097e56commit | diff
Fixed a couple of optimization with EXTRACT_VECTOR_ELT that assumes the result
type is the same as the element type of the vector.  EXTRACT_VECTOR_ELT can
be used to extended the width of an integer type.  This fixes a bug for
Generic/vector-casts.ll on a ppc750.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@94990 91177308-0d34-0410-b5e6-96231b3b80d8
lib/CodeGen/SelectionDAG/DAGCombiner.cpp diff | blob | history
lib/CodeGen/SelectionDAG/SelectionDAG.cpp diff | blob | history
C/C++ LLVM-based compilers that forbids OOTA behaviors