[PowerPC] Forbid f128 SELECT_CC optimized into fsel #71497
Conversation
fsel only supports f32 and f64.
| @@ -8080,7 +8080,8 @@ SDValue PPCTargetLowering::LowerSELECT_CC(SDValue Op, SelectionDAG &DAG) const { | |||
| // For more information, see section F.3 of the 2.06 ISA specification. | |||
| // With ISA 3.0 | |||
| if ((!DAG.getTarget().Options.NoInfsFPMath && !Flags.hasNoInfs()) || | |||
| (!DAG.getTarget().Options.NoNaNsFPMath && !Flags.hasNoNaNs())) | |||
| (!DAG.getTarget().Options.NoNaNsFPMath && !Flags.hasNoNaNs()) || | |||
| ResVT == MVT::f128) | |||
There was a problem hiding this comment.
does that mean fsel also does not support ppc_fp128?
There was a problem hiding this comment.
yes, we don't have instruction for ppc_fp128 fsel. The motivating case exists from 097a95f , which sets select_cc f128 as custom, while we don't have that lowering for ppc_fp128.
There was a problem hiding this comment.
But 097a95f will not try to custom SELECT_CC to fsel for ppc_fp128, right? That commit is only for ieee long double.
Do we have other pattern that will generate fsel operating on ppc_fp128 type?
There was a problem hiding this comment.
No. PPCISD::FSEL is only generated in LowerSELECT_CC, while SELECT_CC ppc_fp128 is expand.
There was a problem hiding this comment.
This seems reasonable to me.
Please wait for one day in case @lei137 has other concerns. Thanks for fixing this.
fsel only supports f32 and f64.