[LoongArch] Make ceil,floor,trunc,roundeven legal for lsx/lasx (llvm#165217)

ylzsx · Debadri Basak · commit 8653ec190bb0 · 2025-11-03T11:35:52.000Z
diff --git a/llvm/lib/Target/LoongArch/LoongArchISelLowering.cpp b/llvm/lib/Target/LoongArch/LoongArchISelLowering.cpp
@@ -371,6 +371,10 @@ LoongArchTargetLowering::LoongArchTargetLowering(const TargetMachine &TM,
                          ISD::SETUGE, ISD::SETUGT},
                         VT, Expand);
       setOperationAction(ISD::SCALAR_TO_VECTOR, VT, Legal);
+      setOperationAction(ISD::FCEIL, VT, Legal);
+      setOperationAction(ISD::FFLOOR, VT, Legal);
+      setOperationAction(ISD::FTRUNC, VT, Legal);
+      setOperationAction(ISD::FROUNDEVEN, VT, Legal);
     }
     setOperationAction(ISD::CTPOP, GRLenVT, Legal);
     setOperationAction(ISD::FCEIL, {MVT::f32, MVT::f64}, Legal);
@@ -453,6 +457,10 @@ LoongArchTargetLowering::LoongArchTargetLowering(const TargetMachine &TM,
                          ISD::SETUGE, ISD::SETUGT},
                         VT, Expand);
       setOperationAction(ISD::SCALAR_TO_VECTOR, VT, Legal);
+      setOperationAction(ISD::FCEIL, VT, Legal);
+      setOperationAction(ISD::FFLOOR, VT, Legal);
+      setOperationAction(ISD::FTRUNC, VT, Legal);
+      setOperationAction(ISD::FROUNDEVEN, VT, Legal);
     }
   }
 
diff --git a/llvm/lib/Target/LoongArch/LoongArchLASXInstrInfo.td b/llvm/lib/Target/LoongArch/LoongArchLASXInstrInfo.td
@@ -2424,6 +2424,12 @@ def : Pat<(int_loongarch_lasx_xvpickve_w_f v8f32:$xj, timm:$imm),
 def : Pat<(int_loongarch_lasx_xvpickve_d_f v4f64:$xj, timm:$imm),
           (XVPICKVE_D v4f64:$xj, (to_valid_timm timm:$imm))>;
 
+// Vector floating-point conversion
+defm : PatXrF<fceil, "XVFRINTRP">;
+defm : PatXrF<ffloor, "XVFRINTRM">;
+defm : PatXrF<ftrunc, "XVFRINTRZ">;
+defm : PatXrF<froundeven, "XVFRINTRNE">;
+
 // load
 def : Pat<(int_loongarch_lasx_xvld GPR:$rj, timm:$imm),
           (XVLD GPR:$rj, (to_valid_timm timm:$imm))>;
diff --git a/llvm/lib/Target/LoongArch/LoongArchLSXInstrInfo.td b/llvm/lib/Target/LoongArch/LoongArchLSXInstrInfo.td
@@ -2552,6 +2552,11 @@ def : Pat<(f64 (froundeven FPR64:$fj)),
           (f64 (EXTRACT_SUBREG (VFRINTRNE_D (VREPLVEI_D
                (SUBREG_TO_REG (i64 0), FPR64:$fj, sub_64), 0)), sub_64))>;
 
+defm : PatVrF<fceil, "VFRINTRP">;
+defm : PatVrF<ffloor, "VFRINTRM">;
+defm : PatVrF<ftrunc, "VFRINTRZ">;
+defm : PatVrF<froundeven, "VFRINTRNE">;
+
 // load
 def : Pat<(int_loongarch_lsx_vld GPR:$rj, timm:$imm),
           (VLD GPR:$rj, (to_valid_timm timm:$imm))>;
diff --git a/llvm/test/CodeGen/LoongArch/lasx/fp-rounding.ll b/llvm/test/CodeGen/LoongArch/lasx/fp-rounding.ll
@@ -7,38 +7,8 @@ define void @ceil_v8f32(ptr %res, ptr %a0) nounwind {
 ; CHECK-LABEL: ceil_v8f32:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    xvld $xr0, $a1, 0
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 5
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrp.s $vr1, $vr1
-; CHECK-NEXT:    xvpickve.w $xr2, $xr0, 4
-; CHECK-NEXT:    vreplvei.w $vr2, $vr2, 0
-; CHECK-NEXT:    vfrintrp.s $vr2, $vr2
-; CHECK-NEXT:    vextrins.w $vr2, $vr1, 16
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 6
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrp.s $vr1, $vr1
-; CHECK-NEXT:    vextrins.w $vr2, $vr1, 32
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 7
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrp.s $vr1, $vr1
-; CHECK-NEXT:    vextrins.w $vr2, $vr1, 48
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 1
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrp.s $vr1, $vr1
-; CHECK-NEXT:    xvpickve.w $xr3, $xr0, 0
-; CHECK-NEXT:    vreplvei.w $vr3, $vr3, 0
-; CHECK-NEXT:    vfrintrp.s $vr3, $vr3
-; CHECK-NEXT:    vextrins.w $vr3, $vr1, 16
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 2
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrp.s $vr1, $vr1
-; CHECK-NEXT:    vextrins.w $vr3, $vr1, 32
-; CHECK-NEXT:    xvpickve.w $xr0, $xr0, 3
-; CHECK-NEXT:    vreplvei.w $vr0, $vr0, 0
-; CHECK-NEXT:    vfrintrp.s $vr0, $vr0
-; CHECK-NEXT:    vextrins.w $vr3, $vr0, 48
-; CHECK-NEXT:    xvpermi.q $xr3, $xr2, 2
-; CHECK-NEXT:    xvst $xr3, $a0, 0
+; CHECK-NEXT:    xvfrintrp.s $xr0, $xr0
+; CHECK-NEXT:    xvst $xr0, $a0, 0
 ; CHECK-NEXT:    ret
 entry:
   %v0 = load <8 x float>, ptr %a0
@@ -52,21 +22,7 @@ define void @ceil_v4f64(ptr %res, ptr %a0) nounwind {
 ; CHECK-LABEL: ceil_v4f64:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    xvld $xr0, $a1, 0
-; CHECK-NEXT:    xvpickve.d $xr1, $xr0, 3
-; CHECK-NEXT:    vreplvei.d $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrp.d $vr1, $vr1
-; CHECK-NEXT:    xvpickve.d $xr2, $xr0, 2
-; CHECK-NEXT:    vreplvei.d $vr2, $vr2, 0
-; CHECK-NEXT:    vfrintrp.d $vr2, $vr2
-; CHECK-NEXT:    vextrins.d $vr2, $vr1, 16
-; CHECK-NEXT:    xvpickve.d $xr1, $xr0, 1
-; CHECK-NEXT:    vreplvei.d $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrp.d $vr1, $vr1
-; CHECK-NEXT:    xvpickve.d $xr0, $xr0, 0
-; CHECK-NEXT:    vreplvei.d $vr0, $vr0, 0
-; CHECK-NEXT:    vfrintrp.d $vr0, $vr0
-; CHECK-NEXT:    vextrins.d $vr0, $vr1, 16
-; CHECK-NEXT:    xvpermi.q $xr0, $xr2, 2
+; CHECK-NEXT:    xvfrintrp.d $xr0, $xr0
 ; CHECK-NEXT:    xvst $xr0, $a0, 0
 ; CHECK-NEXT:    ret
 entry:
@@ -81,38 +37,8 @@ define void @floor_v8f32(ptr %res, ptr %a0) nounwind {
 ; CHECK-LABEL: floor_v8f32:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    xvld $xr0, $a1, 0
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 5
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrm.s $vr1, $vr1
-; CHECK-NEXT:    xvpickve.w $xr2, $xr0, 4
-; CHECK-NEXT:    vreplvei.w $vr2, $vr2, 0
-; CHECK-NEXT:    vfrintrm.s $vr2, $vr2
-; CHECK-NEXT:    vextrins.w $vr2, $vr1, 16
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 6
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrm.s $vr1, $vr1
-; CHECK-NEXT:    vextrins.w $vr2, $vr1, 32
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 7
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrm.s $vr1, $vr1
-; CHECK-NEXT:    vextrins.w $vr2, $vr1, 48
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 1
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrm.s $vr1, $vr1
-; CHECK-NEXT:    xvpickve.w $xr3, $xr0, 0
-; CHECK-NEXT:    vreplvei.w $vr3, $vr3, 0
-; CHECK-NEXT:    vfrintrm.s $vr3, $vr3
-; CHECK-NEXT:    vextrins.w $vr3, $vr1, 16
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 2
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrm.s $vr1, $vr1
-; CHECK-NEXT:    vextrins.w $vr3, $vr1, 32
-; CHECK-NEXT:    xvpickve.w $xr0, $xr0, 3
-; CHECK-NEXT:    vreplvei.w $vr0, $vr0, 0
-; CHECK-NEXT:    vfrintrm.s $vr0, $vr0
-; CHECK-NEXT:    vextrins.w $vr3, $vr0, 48
-; CHECK-NEXT:    xvpermi.q $xr3, $xr2, 2
-; CHECK-NEXT:    xvst $xr3, $a0, 0
+; CHECK-NEXT:    xvfrintrm.s $xr0, $xr0
+; CHECK-NEXT:    xvst $xr0, $a0, 0
 ; CHECK-NEXT:    ret
 entry:
   %v0 = load <8 x float>, ptr %a0
@@ -126,21 +52,7 @@ define void @floor_v4f64(ptr %res, ptr %a0) nounwind {
 ; CHECK-LABEL: floor_v4f64:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    xvld $xr0, $a1, 0
-; CHECK-NEXT:    xvpickve.d $xr1, $xr0, 3
-; CHECK-NEXT:    vreplvei.d $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrm.d $vr1, $vr1
-; CHECK-NEXT:    xvpickve.d $xr2, $xr0, 2
-; CHECK-NEXT:    vreplvei.d $vr2, $vr2, 0
-; CHECK-NEXT:    vfrintrm.d $vr2, $vr2
-; CHECK-NEXT:    vextrins.d $vr2, $vr1, 16
-; CHECK-NEXT:    xvpickve.d $xr1, $xr0, 1
-; CHECK-NEXT:    vreplvei.d $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrm.d $vr1, $vr1
-; CHECK-NEXT:    xvpickve.d $xr0, $xr0, 0
-; CHECK-NEXT:    vreplvei.d $vr0, $vr0, 0
-; CHECK-NEXT:    vfrintrm.d $vr0, $vr0
-; CHECK-NEXT:    vextrins.d $vr0, $vr1, 16
-; CHECK-NEXT:    xvpermi.q $xr0, $xr2, 2
+; CHECK-NEXT:    xvfrintrm.d $xr0, $xr0
 ; CHECK-NEXT:    xvst $xr0, $a0, 0
 ; CHECK-NEXT:    ret
 entry:
@@ -155,38 +67,8 @@ define void @trunc_v8f32(ptr %res, ptr %a0) nounwind {
 ; CHECK-LABEL: trunc_v8f32:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    xvld $xr0, $a1, 0
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 5
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrz.s $vr1, $vr1
-; CHECK-NEXT:    xvpickve.w $xr2, $xr0, 4
-; CHECK-NEXT:    vreplvei.w $vr2, $vr2, 0
-; CHECK-NEXT:    vfrintrz.s $vr2, $vr2
-; CHECK-NEXT:    vextrins.w $vr2, $vr1, 16
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 6
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrz.s $vr1, $vr1
-; CHECK-NEXT:    vextrins.w $vr2, $vr1, 32
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 7
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrz.s $vr1, $vr1
-; CHECK-NEXT:    vextrins.w $vr2, $vr1, 48
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 1
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrz.s $vr1, $vr1
-; CHECK-NEXT:    xvpickve.w $xr3, $xr0, 0
-; CHECK-NEXT:    vreplvei.w $vr3, $vr3, 0
-; CHECK-NEXT:    vfrintrz.s $vr3, $vr3
-; CHECK-NEXT:    vextrins.w $vr3, $vr1, 16
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 2
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrz.s $vr1, $vr1
-; CHECK-NEXT:    vextrins.w $vr3, $vr1, 32
-; CHECK-NEXT:    xvpickve.w $xr0, $xr0, 3
-; CHECK-NEXT:    vreplvei.w $vr0, $vr0, 0
-; CHECK-NEXT:    vfrintrz.s $vr0, $vr0
-; CHECK-NEXT:    vextrins.w $vr3, $vr0, 48
-; CHECK-NEXT:    xvpermi.q $xr3, $xr2, 2
-; CHECK-NEXT:    xvst $xr3, $a0, 0
+; CHECK-NEXT:    xvfrintrz.s $xr0, $xr0
+; CHECK-NEXT:    xvst $xr0, $a0, 0
 ; CHECK-NEXT:    ret
 entry:
   %v0 = load <8 x float>, ptr %a0
@@ -200,21 +82,7 @@ define void @trunc_v4f64(ptr %res, ptr %a0) nounwind {
 ; CHECK-LABEL: trunc_v4f64:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    xvld $xr0, $a1, 0
-; CHECK-NEXT:    xvpickve.d $xr1, $xr0, 3
-; CHECK-NEXT:    vreplvei.d $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrz.d $vr1, $vr1
-; CHECK-NEXT:    xvpickve.d $xr2, $xr0, 2
-; CHECK-NEXT:    vreplvei.d $vr2, $vr2, 0
-; CHECK-NEXT:    vfrintrz.d $vr2, $vr2
-; CHECK-NEXT:    vextrins.d $vr2, $vr1, 16
-; CHECK-NEXT:    xvpickve.d $xr1, $xr0, 1
-; CHECK-NEXT:    vreplvei.d $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrz.d $vr1, $vr1
-; CHECK-NEXT:    xvpickve.d $xr0, $xr0, 0
-; CHECK-NEXT:    vreplvei.d $vr0, $vr0, 0
-; CHECK-NEXT:    vfrintrz.d $vr0, $vr0
-; CHECK-NEXT:    vextrins.d $vr0, $vr1, 16
-; CHECK-NEXT:    xvpermi.q $xr0, $xr2, 2
+; CHECK-NEXT:    xvfrintrz.d $xr0, $xr0
 ; CHECK-NEXT:    xvst $xr0, $a0, 0
 ; CHECK-NEXT:    ret
 entry:
@@ -229,38 +97,8 @@ define void @roundeven_v8f32(ptr %res, ptr %a0) nounwind {
 ; CHECK-LABEL: roundeven_v8f32:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    xvld $xr0, $a1, 0
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 5
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrne.s $vr1, $vr1
-; CHECK-NEXT:    xvpickve.w $xr2, $xr0, 4
-; CHECK-NEXT:    vreplvei.w $vr2, $vr2, 0
-; CHECK-NEXT:    vfrintrne.s $vr2, $vr2
-; CHECK-NEXT:    vextrins.w $vr2, $vr1, 16
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 6
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrne.s $vr1, $vr1
-; CHECK-NEXT:    vextrins.w $vr2, $vr1, 32
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 7
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrne.s $vr1, $vr1
-; CHECK-NEXT:    vextrins.w $vr2, $vr1, 48
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 1
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrne.s $vr1, $vr1
-; CHECK-NEXT:    xvpickve.w $xr3, $xr0, 0
-; CHECK-NEXT:    vreplvei.w $vr3, $vr3, 0
-; CHECK-NEXT:    vfrintrne.s $vr3, $vr3
-; CHECK-NEXT:    vextrins.w $vr3, $vr1, 16
-; CHECK-NEXT:    xvpickve.w $xr1, $xr0, 2
-; CHECK-NEXT:    vreplvei.w $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrne.s $vr1, $vr1
-; CHECK-NEXT:    vextrins.w $vr3, $vr1, 32
-; CHECK-NEXT:    xvpickve.w $xr0, $xr0, 3
-; CHECK-NEXT:    vreplvei.w $vr0, $vr0, 0
-; CHECK-NEXT:    vfrintrne.s $vr0, $vr0
-; CHECK-NEXT:    vextrins.w $vr3, $vr0, 48
-; CHECK-NEXT:    xvpermi.q $xr3, $xr2, 2
-; CHECK-NEXT:    xvst $xr3, $a0, 0
+; CHECK-NEXT:    xvfrintrne.s $xr0, $xr0
+; CHECK-NEXT:    xvst $xr0, $a0, 0
 ; CHECK-NEXT:    ret
 entry:
   %v0 = load <8 x float>, ptr %a0
@@ -274,21 +112,7 @@ define void @roundeven_v4f64(ptr %res, ptr %a0) nounwind {
 ; CHECK-LABEL: roundeven_v4f64:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    xvld $xr0, $a1, 0
-; CHECK-NEXT:    xvpickve.d $xr1, $xr0, 3
-; CHECK-NEXT:    vreplvei.d $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrne.d $vr1, $vr1
-; CHECK-NEXT:    xvpickve.d $xr2, $xr0, 2
-; CHECK-NEXT:    vreplvei.d $vr2, $vr2, 0
-; CHECK-NEXT:    vfrintrne.d $vr2, $vr2
-; CHECK-NEXT:    vextrins.d $vr2, $vr1, 16
-; CHECK-NEXT:    xvpickve.d $xr1, $xr0, 1
-; CHECK-NEXT:    vreplvei.d $vr1, $vr1, 0
-; CHECK-NEXT:    vfrintrne.d $vr1, $vr1
-; CHECK-NEXT:    xvpickve.d $xr0, $xr0, 0
-; CHECK-NEXT:    vreplvei.d $vr0, $vr0, 0
-; CHECK-NEXT:    vfrintrne.d $vr0, $vr0
-; CHECK-NEXT:    vextrins.d $vr0, $vr1, 16
-; CHECK-NEXT:    xvpermi.q $xr0, $xr2, 2
+; CHECK-NEXT:    xvfrintrne.d $xr0, $xr0
 ; CHECK-NEXT:    xvst $xr0, $a0, 0
 ; CHECK-NEXT:    ret
 entry:
diff --git a/llvm/test/CodeGen/LoongArch/lsx/fp-rounding.ll b/llvm/test/CodeGen/LoongArch/lsx/fp-rounding.ll
