forked from OSchip/llvm-project
[AArch64][GlobalISel] Don't allow s128 for G_ISNAN
getAPFloatFromSize doesn't support s128, so we can't lower this without asserting right now. To fix the buildbots, don't allow any scalars other than s16, s32, and s64.
This commit is contained in:
parent
b2e77cd095
commit
c22b64ef66
|
@ -727,7 +727,8 @@ AArch64LegalizerInfo::AArch64LegalizerInfo(const AArch64Subtarget &ST)
|
|||
// TODO: Vector types.
|
||||
getActionDefinitionsBuilder({G_SADDSAT, G_SSUBSAT}).lowerIf(isScalar(0));
|
||||
|
||||
getActionDefinitionsBuilder(G_ISNAN).lower();
|
||||
getActionDefinitionsBuilder(G_ISNAN).lowerIf(
|
||||
any(isVector(1), typeInSet(1, {s16, s32, s64})));
|
||||
|
||||
// TODO: Vector types.
|
||||
getActionDefinitionsBuilder({G_FMAXNUM, G_FMINNUM})
|
||||
|
|
Loading…
Reference in New Issue