diff options
| author | Dan Gohman <gohman@apple.com> | 2009-06-04 22:49:04 +0000 |
|---|---|---|
| committer | Dan Gohman <gohman@apple.com> | 2009-06-04 22:49:04 +0000 |
| commit | a5b9645c4b7a1d8be5e41081b99f27b49b8aa8cf (patch) | |
| tree | 7f5a9f6633be0c4e77a1bb00e5bfcfcca14e219b /llvm/test/CodeGen/ARM/fixunsdfdi.ll | |
| parent | 72a4d2fec138ad6d2becbc69c6d034246a056d09 (diff) | |
| download | bcm5719-llvm-a5b9645c4b7a1d8be5e41081b99f27b49b8aa8cf.tar.gz bcm5719-llvm-a5b9645c4b7a1d8be5e41081b99f27b49b8aa8cf.zip | |
Split the Add, Sub, and Mul instruction opcodes into separate
integer and floating-point opcodes, introducing
FAdd, FSub, and FMul.
For now, the AsmParser, BitcodeReader, and IRBuilder all preserve
backwards compatability, and the Core LLVM APIs preserve backwards
compatibility for IR producers. Most front-ends won't need to change
immediately.
This implements the first step of the plan outlined here:
http://nondot.org/sabre/LLVMNotes/IntegerOverflow.txt
llvm-svn: 72897
Diffstat (limited to 'llvm/test/CodeGen/ARM/fixunsdfdi.ll')
| -rw-r--r-- | llvm/test/CodeGen/ARM/fixunsdfdi.ll | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/llvm/test/CodeGen/ARM/fixunsdfdi.ll b/llvm/test/CodeGen/ARM/fixunsdfdi.ll index d3038b9af76..777a3d69a19 100644 --- a/llvm/test/CodeGen/ARM/fixunsdfdi.ll +++ b/llvm/test/CodeGen/ARM/fixunsdfdi.ll @@ -13,7 +13,7 @@ bb5: ; preds = %bb3 %u.in.mask = and i64 %x14, -4294967296 ; <i64> [#uses=1] %.ins = or i64 0, %u.in.mask ; <i64> [#uses=1] %0 = bitcast i64 %.ins to double ; <double> [#uses=1] - %1 = sub double %x, %0 ; <double> [#uses=1] + %1 = fsub double %x, %0 ; <double> [#uses=1] %2 = fptosi double %1 to i32 ; <i32> [#uses=1] %3 = add i32 %2, 0 ; <i32> [#uses=1] %4 = zext i32 %3 to i64 ; <i64> [#uses=1] |
