diff options
| author | Cameron Zwarich <zwarich@apple.com> | 2011-06-29 22:24:25 +0000 |
|---|---|---|
| committer | Cameron Zwarich <zwarich@apple.com> | 2011-06-29 22:24:25 +0000 |
| commit | faff12731968c2d3f1ff56a43749d27e1696aafa (patch) | |
| tree | 3f8acbcb39deb74a8e73625364f6d8bbfdd8fdf6 /test/CodeGen/ARM/2011-06-29-MergeGlobalsAlign.ll | |
| parent | ae218dee5e3a079becd6b9b8d47e67cf814b9b70 (diff) | |
| download | external_llvm-faff12731968c2d3f1ff56a43749d27e1696aafa.zip external_llvm-faff12731968c2d3f1ff56a43749d27e1696aafa.tar.gz external_llvm-faff12731968c2d3f1ff56a43749d27e1696aafa.tar.bz2 | |
In the ARM global merging pass, allow extraneous alignment specifiers. This pass
already makes the assumption, which is correct on ARM, that a type's alignment is
less than its alloc size. This improves codegen with Clang (which inserts a lot of
extraneous alignment specifiers) and fixes <rdar://problem/9695089>.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@134106 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'test/CodeGen/ARM/2011-06-29-MergeGlobalsAlign.ll')
| -rw-r--r-- | test/CodeGen/ARM/2011-06-29-MergeGlobalsAlign.ll | 12 |
1 files changed, 12 insertions, 0 deletions
diff --git a/test/CodeGen/ARM/2011-06-29-MergeGlobalsAlign.ll b/test/CodeGen/ARM/2011-06-29-MergeGlobalsAlign.ll new file mode 100644 index 0000000..1b5b8a9 --- /dev/null +++ b/test/CodeGen/ARM/2011-06-29-MergeGlobalsAlign.ll @@ -0,0 +1,12 @@ +; RUN: llc < %s -mtriple=thumbv7-apple-darwin10 | FileCheck %s +; CHECK: .zerofill __DATA,__bss,__MergedGlobals,16,2 + +%struct.config = type { i16, i16, i16, i16 } + +@prev = external global [0 x i16] +@max_lazy_match = internal unnamed_addr global i32 0, align 4 +@read_buf = external global i32 (i8*, i32)* +@window = external global [0 x i8] +@lookahead = internal unnamed_addr global i32 0, align 4 +@eofile.b = internal unnamed_addr global i1 false +@ins_h = internal unnamed_addr global i32 0, align 4 |