mirror of
https://github.com/RPCS3/llvm.git
synced 2024-12-15 16:09:57 +00:00
56a5661c60
Function 'ConstantFoldScalarCall' (in ConstantFolding.cpp) works under the
wrong assumption that a call to 'convert.from.fp16' returns a value of
type 'float'.
However, intrinsic 'convert.from.fp16' can be overloaded; for example, we
can call 'convert.from.fp16.f64' to convert from half to double; etc.
Before this patch, the following example would have triggered an assertion
failure in opt (with -constprop):
```
define double @foo() {
entry:
%0 = call double @llvm.convert.from.fp16.f64(i16 0)
ret double %0
}
```
This patch fixes the problem in ConstantFolding.cpp. When folding a call to
convert.from.fp16, we perform a different kind of conversion based on the call
return type.
Added test 'Transform/ConstProp/convert-from-fp16.ll'.
Differential Revision: http://reviews.llvm.org/D9771
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@237377 91177308-0d34-0410-b5e6-96231b3b80d8
|
||
|---|---|---|
| .. | ||
| 2002-05-03-DivideByZeroException.ll | ||
| 2002-05-03-NotOperator.ll | ||
| 2002-09-03-SetCC-Bools.ll | ||
| 2003-05-12-DivideError.ll | ||
| 2005-01-28-SetCCGEP.ll | ||
| 2006-11-30-vector-cast.ll | ||
| 2006-12-01-bool-casts.ll | ||
| 2006-12-01-TruncBoolBug.ll | ||
| 2007-02-05-BitCast.ll | ||
| 2007-02-23-sdiv.ll | ||
| 2007-11-23-cttz.ll | ||
| 2008-07-07-VectorCompare.ll | ||
| 2009-06-20-constexpr-zero-lhs.ll | ||
| 2009-09-01-GEP-Crash.ll | ||
| basictest.ll | ||
| bitcast.ll | ||
| bswap.ll | ||
| calls.ll | ||
| constant-expr.ll | ||
| convert-from-fp16.ll | ||
| div-zero.ll | ||
| extractvalue.ll | ||
| float-to-ptr-cast.ll | ||
| InsertElement.ll | ||
| insertvalue.ll | ||
| loads.ll | ||
| logicaltest.ll | ||
| overflow-ops.ll | ||
| phi.ll | ||
| remtest.ll | ||
| shift.ll | ||
| trunc_vec.ll | ||