diff options
| author | Sanjay Patel <spatel@rotateright.com> | 2017-02-03 21:43:34 +0000 |
|---|---|---|
| committer | Sanjay Patel <spatel@rotateright.com> | 2017-02-03 21:43:34 +0000 |
| commit | 73fc8ddb065112fbb7cad1453774ed96bbb7d6e3 (patch) | |
| tree | ba2c3ab04aba252f1c73bad9b5a480c056530f0f /llvm/test/Transforms/InstCombine/and.ll | |
| parent | 25324313325e906368fafd142bc6138c231e0b44 (diff) | |
| download | bcm5719-llvm-73fc8ddb065112fbb7cad1453774ed96bbb7d6e3.tar.gz bcm5719-llvm-73fc8ddb065112fbb7cad1453774ed96bbb7d6e3.zip | |
[InstCombine] fix operand-complexity-based canonicalization (PR28296)
The code comments didn't match the code logic, and we didn't actually distinguish the fake unary (not/neg/fneg)
operators from arguments. Adding another level to the weighting scheme provides more structure and can help
simplify the pattern matching in InstCombine and other places.
I fixed regressions that would have shown up from this change in:
rL290067
rL290127
But that doesn't mean there are no pattern-matching logic holes left; some combines may just be missing regression tests.
Should fix:
https://llvm.org/bugs/show_bug.cgi?id=28296
Differential Revision: https://reviews.llvm.org/D27933
llvm-svn: 294049
Diffstat (limited to 'llvm/test/Transforms/InstCombine/and.ll')
| -rw-r--r-- | llvm/test/Transforms/InstCombine/and.ll | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/llvm/test/Transforms/InstCombine/and.ll b/llvm/test/Transforms/InstCombine/and.ll index 8a30de1c07d..da71880b9bd 100644 --- a/llvm/test/Transforms/InstCombine/and.ll +++ b/llvm/test/Transforms/InstCombine/and.ll @@ -176,7 +176,7 @@ define i8 @test16(i8 %A) { define i8 @test17(i8 %X, i8 %Y) { ; CHECK-LABEL: @test17( ; CHECK-NEXT: [[Y_NOT:%.*]] = xor i8 %Y, -1 -; CHECK-NEXT: [[D:%.*]] = or i8 %X, [[Y_NOT]] +; CHECK-NEXT: [[D:%.*]] = or i8 [[Y_NOT]], %X ; CHECK-NEXT: ret i8 [[D]] ; %B = xor i8 %X, -1 |
