projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: d3a0422) | patch
[FastISel][AArch64] Don't fold the 'and' instruction into the 'tbz/tbnz' instruction...
authorJuergen Ributzka <juergen@apple.com>
Mon, 27 Oct 2014 19:16:48 +0000 (19:16 +0000)
committerJuergen Ributzka <juergen@apple.com>
Mon, 27 Oct 2014 19:16:48 +0000 (19:16 +0000)
commit5745cad86144b8a3d47f55568daff433033cf202
treea2153f88cb7d446ef3327faea73fb91428a88b6ftree | snapshot
parentd3a04223e84797a1432f10d7a153da6c258017a9commit | diff
[FastISel][AArch64] Don't fold the 'and' instruction into the 'tbz/tbnz' instruction if it is in a different basic block.

This fixes a bug where the input register was not defined for the 'tbz/tbnz'
instruction. This happened, because we folded the 'and' instruction from a
different basic block.

This fixes rdar://problem/18784013.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220704 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Target/AArch64/AArch64FastISel.cpp diff | blob | history
test/CodeGen/AArch64/fast-isel-tbz.ll diff | blob | history
C/C++ LLVM-based compilers that forbids OOTA behaviors