diff options
| author | John Brawn <john.brawn@arm.com> | 2017-05-26 13:59:12 +0000 |
|---|---|---|
| committer | John Brawn <john.brawn@arm.com> | 2017-05-26 13:59:12 +0000 |
| commit | 9009d2905debfc210859e5d20a1fb3ec319a9ce7 (patch) | |
| tree | 93ad3b86753bdbce6d9657acf75739ae0d92f12d /llvm/test/CodeGen/ARM/memcpy-inline.ll | |
| parent | ba9d8ba82aab927dd8e132c947cd64efff4e09d1 (diff) | |
| download | bcm5719-llvm-9009d2905debfc210859e5d20a1fb3ec319a9ce7.tar.gz bcm5719-llvm-9009d2905debfc210859e5d20a1fb3ec319a9ce7.zip | |
[ARM] Fix lowering of misaligned memcpy/memset
Currently getOptimalMemOpType returns i32 for large enough sizes without
checking for alignment, leading to poor code generation when misaligned accesses
aren't permitted as we generate a word store then later split it up into byte
stores. This means we inadvertantly go over the MaxStoresPerMemcpy limit and for
memset we splat the memset value into a word then immediately split it up
again.
Fix this by leaving it up to FindOptimalMemOpLowering to figure out which type
to use, but also fix a bug there where it wasn't correctly checking if
misaligned memory accesses are allowed.
Differential Revision: https://reviews.llvm.org/D33442
llvm-svn: 303990
Diffstat (limited to 'llvm/test/CodeGen/ARM/memcpy-inline.ll')
| -rw-r--r-- | llvm/test/CodeGen/ARM/memcpy-inline.ll | 5 |
1 files changed, 1 insertions, 4 deletions
diff --git a/llvm/test/CodeGen/ARM/memcpy-inline.ll b/llvm/test/CodeGen/ARM/memcpy-inline.ll index 436e49b9f39..b447497b270 100644 --- a/llvm/test/CodeGen/ARM/memcpy-inline.ll +++ b/llvm/test/CodeGen/ARM/memcpy-inline.ll @@ -95,10 +95,7 @@ entry: ; CHECK: movt [[REG7:r[0-9]+]], #22866 ; CHECK: str [[REG7]] ; CHECK-T1-LABEL: t5: -; CHECK-T1: movs [[TREG3:r[0-9]]], -; CHECK-T1: strb [[TREG3]], -; CHECK-T1: movs [[TREG4:r[0-9]]], -; CHECK-T1: strb [[TREG4]], +; CHECK-T1: bl _memcpy tail call void @llvm.memcpy.p0i8.p0i8.i64(i8* %C, i8* getelementptr inbounds ([7 x i8], [7 x i8]* @.str5, i64 0, i64 0), i64 7, i32 1, i1 false) ret void } |
