diff options
author | Evan Cheng <evan.cheng@apple.com> | 2010-01-18 21:38:44 +0000 |
---|---|---|
committer | Evan Cheng <evan.cheng@apple.com> | 2010-01-18 21:38:44 +0000 |
commit | 1ad0e8b5769f25ca442f3d77bf374d03469010bc (patch) | |
tree | a4413ab41fe7bc56f036ec89728f80bde599b205 /test/CodeGen | |
parent | 2d53c21eca34b77bbe12a5208af9da780d6465c8 (diff) | |
download | llvm-1ad0e8b5769f25ca442f3d77bf374d03469010bc.tar.gz llvm-1ad0e8b5769f25ca442f3d77bf374d03469010bc.tar.bz2 llvm-1ad0e8b5769f25ca442f3d77bf374d03469010bc.tar.xz |
Canonicalize -1 - x to ~x.
Instcombine does this but apparently there are situations where this pattern will escape the optimizer and / or created by isel. Here is a case that's seen in JavaScriptCore:
%t1 = sub i32 0, %a
%t2 = add i32 %t1, -1
The dag combiner pattern: ((c1-A)+c2) -> (c1+c2)-A
will fold it to -1 - %a.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@93773 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'test/CodeGen')
-rw-r--r-- | test/CodeGen/X86/xor.ll | 11 |
1 files changed, 11 insertions, 0 deletions
diff --git a/test/CodeGen/X86/xor.ll b/test/CodeGen/X86/xor.ll index 7bd06bba4c..9bfff8a06a 100644 --- a/test/CodeGen/X86/xor.ll +++ b/test/CodeGen/X86/xor.ll @@ -131,3 +131,14 @@ bb12: ; X32: andl {{.*}}[[REG]] } +define i32 @test8(i32 %a) nounwind { +; rdar://7553032 +entry: + %t1 = sub i32 0, %a + %t2 = add i32 %t1, -1 + ret i32 %t2 +; X64: test8: +; X64: notl %eax +; X32: test8: +; X32: notl %eax +} |