[SelectionDAG] Allow FREEZE to be hoisted before FP SETCC. #84358
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
@llvm/pr-subscribers-backend-risc-v @llvm/pr-subscribers-llvm-selectiondag Author: Craig Topper (topperc) ChangesNo nans/infs in SelectionDAG is complicated. Hopefully I've captured Recovers some regression from #84232. Stacked on #84241 and #84232 Patch is 232.06 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/84358.diff 20 Files Affected:
diff --git a/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp b/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
index f7ace79e8c51d4..022d097f48d700 100644
--- a/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
@@ -4998,6 +4998,7 @@ bool SelectionDAG::isGuaranteedNotToBeUndefOrPoison(SDValue Op,
return true;
switch (Opcode) {
+ case ISD::CONDCODE:
case ISD::VALUETYPE:
case ISD::FrameIndex:
case ISD::TargetFrameIndex:
@@ -5090,6 +5091,25 @@ bool SelectionDAG::canCreateUndefOrPoison(SDValue Op, const APInt &DemandedElts,
case ISD::BUILD_PAIR:
return false;
+ case ISD::SETCC: {
+ // Integer setcc cannot create undef or poison.
+ if (Op.getOperand(0).getValueType().isInteger())
+ return false;
+
+ // FP compares are more complicated. They can create poison for nan/infinity
+ // based on options and flags. The options and flags also cause special
+ // nonan condition codes to be used. Those condition codes may be preserved
+ // even if the nonan flag is dropped somewhere.
+ ISD::CondCode CCCode = cast<CondCodeSDNode>(Op.getOperand(2))->get();
+ if (((unsigned)CCCode & 0x10U))
+ return true;
+
+ const TargetOptions &Options = getTarget().Options;
+ return Options.NoNaNsFPMath || Options.NoInfsFPMath ||
+ (ConsiderFlags && (Op->getFlags().hasNoNaNs() ||
+ Op->getFlags().hasNoInfs()));
+ }
+
// Matches hasPoisonGeneratingFlags().
case ISD::ZERO_EXTEND:
return ConsiderFlags && Op->getFlags().hasNonNeg();
diff --git a/llvm/lib/Target/RISCV/RISCVISelLowering.cpp b/llvm/lib/Target/RISCV/RISCVISelLowering.cpp
index 4c3dc63afd878d..e3d1fd8d78f3a6 100644
--- a/llvm/lib/Target/RISCV/RISCVISelLowering.cpp
+++ b/llvm/lib/Target/RISCV/RISCVISelLowering.cpp
@@ -7246,25 +7246,25 @@ static SDValue combineSelectToBinOp(SDNode *N, SelectionDAG &DAG,
// (select c, -1, y) -> -c | y
if (isAllOnesConstant(TrueV)) {
SDValue Neg = DAG.getNegative(CondV, DL, VT);
- return DAG.getNode(ISD::OR, DL, VT, Neg, FalseV);
+ return DAG.getNode(ISD::OR, DL, VT, Neg, DAG.getFreeze(FalseV));
}
// (select c, y, -1) -> (c-1) | y
if (isAllOnesConstant(FalseV)) {
SDValue Neg = DAG.getNode(ISD::ADD, DL, VT, CondV,
DAG.getAllOnesConstant(DL, VT));
- return DAG.getNode(ISD::OR, DL, VT, Neg, TrueV);
+ return DAG.getNode(ISD::OR, DL, VT, Neg, DAG.getFreeze(TrueV));
}
// (select c, 0, y) -> (c-1) & y
if (isNullConstant(TrueV)) {
SDValue Neg = DAG.getNode(ISD::ADD, DL, VT, CondV,
DAG.getAllOnesConstant(DL, VT));
- return DAG.getNode(ISD::AND, DL, VT, Neg, FalseV);
+ return DAG.getNode(ISD::AND, DL, VT, Neg, DAG.getFreeze(FalseV));
}
// (select c, y, 0) -> -c & y
if (isNullConstant(FalseV)) {
SDValue Neg = DAG.getNegative(CondV, DL, VT);
- return DAG.getNode(ISD::AND, DL, VT, Neg, TrueV);
+ return DAG.getNode(ISD::AND, DL, VT, Neg, DAG.getFreeze(TrueV));
}
}
diff --git a/llvm/test/CodeGen/RISCV/alu64.ll b/llvm/test/CodeGen/RISCV/alu64.ll
index f032756e007b68..d2ee80e6aa9513 100644
--- a/llvm/test/CodeGen/RISCV/alu64.ll
+++ b/llvm/test/CodeGen/RISCV/alu64.ll
@@ -57,8 +57,8 @@ define i64 @sltiu(i64 %a) nounwind {
;
; RV32I-LABEL: sltiu:
; RV32I: # %bb.0:
-; RV32I-NEXT: sltiu a0, a0, 3
; RV32I-NEXT: seqz a1, a1
+; RV32I-NEXT: sltiu a0, a0, 3
; RV32I-NEXT: and a0, a1, a0
; RV32I-NEXT: li a1, 0
; RV32I-NEXT: ret
diff --git a/llvm/test/CodeGen/RISCV/atomicrmw-uinc-udec-wrap.ll b/llvm/test/CodeGen/RISCV/atomicrmw-uinc-udec-wrap.ll
index aa962d68fc5285..5914e45a153302 100644
--- a/llvm/test/CodeGen/RISCV/atomicrmw-uinc-udec-wrap.ll
+++ b/llvm/test/CodeGen/RISCV/atomicrmw-uinc-udec-wrap.ll
@@ -372,10 +372,10 @@ define i32 @atomicrmw_uinc_wrap_i32(ptr %ptr, i32 %val) {
; RV32IA-NEXT: # =>This Loop Header: Depth=1
; RV32IA-NEXT: # Child Loop BB2_3 Depth 2
; RV32IA-NEXT: mv a3, a2
-; RV32IA-NEXT: addi a2, a2, 1
-; RV32IA-NEXT: sltu a4, a3, a1
-; RV32IA-NEXT: neg a4, a4
-; RV32IA-NEXT: and a4, a4, a2
+; RV32IA-NEXT: addi a4, a2, 1
+; RV32IA-NEXT: sltu a2, a2, a1
+; RV32IA-NEXT: neg a2, a2
+; RV32IA-NEXT: and a4, a2, a4
; RV32IA-NEXT: .LBB2_3: # %atomicrmw.start
; RV32IA-NEXT: # Parent Loop BB2_1 Depth=1
; RV32IA-NEXT: # => This Inner Loop Header: Depth=2
@@ -607,10 +607,10 @@ define i64 @atomicrmw_uinc_wrap_i64(ptr %ptr, i64 %val) {
; RV64IA-NEXT: # =>This Loop Header: Depth=1
; RV64IA-NEXT: # Child Loop BB3_3 Depth 2
; RV64IA-NEXT: mv a3, a2
-; RV64IA-NEXT: addi a2, a2, 1
-; RV64IA-NEXT: sltu a4, a3, a1
-; RV64IA-NEXT: neg a4, a4
-; RV64IA-NEXT: and a4, a4, a2
+; RV64IA-NEXT: addi a4, a2, 1
+; RV64IA-NEXT: sltu a2, a2, a1
+; RV64IA-NEXT: neg a2, a2
+; RV64IA-NEXT: and a4, a2, a4
; RV64IA-NEXT: .LBB3_3: # %atomicrmw.start
; RV64IA-NEXT: # Parent Loop BB3_1 Depth=1
; RV64IA-NEXT: # => This Inner Loop Header: Depth=2
diff --git a/llvm/test/CodeGen/RISCV/bfloat-convert.ll b/llvm/test/CodeGen/RISCV/bfloat-convert.ll
index d533607ad54e38..d5041c2a7ca78b 100644
--- a/llvm/test/CodeGen/RISCV/bfloat-convert.ll
+++ b/llvm/test/CodeGen/RISCV/bfloat-convert.ll
@@ -456,80 +456,92 @@ define i64 @fcvt_l_bf16(bfloat %a) nounwind {
define i64 @fcvt_l_bf16_sat(bfloat %a) nounwind {
; RV32IZFBFMIN-LABEL: fcvt_l_bf16_sat:
; RV32IZFBFMIN: # %bb.0: # %start
-; RV32IZFBFMIN-NEXT: addi sp, sp, -16
-; RV32IZFBFMIN-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
-; RV32IZFBFMIN-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
-; RV32IZFBFMIN-NEXT: fsw fs0, 4(sp) # 4-byte Folded Spill
+; RV32IZFBFMIN-NEXT: addi sp, sp, -32
+; RV32IZFBFMIN-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
+; RV32IZFBFMIN-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
+; RV32IZFBFMIN-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
+; RV32IZFBFMIN-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
+; RV32IZFBFMIN-NEXT: sw s3, 12(sp) # 4-byte Folded Spill
+; RV32IZFBFMIN-NEXT: fsw fs0, 8(sp) # 4-byte Folded Spill
+; RV32IZFBFMIN-NEXT: lui a0, %hi(.LCPI10_0)
+; RV32IZFBFMIN-NEXT: flw fa5, %lo(.LCPI10_0)(a0)
; RV32IZFBFMIN-NEXT: fcvt.s.bf16 fs0, fa0
+; RV32IZFBFMIN-NEXT: flt.s s0, fa5, fs0
+; RV32IZFBFMIN-NEXT: neg s1, s0
; RV32IZFBFMIN-NEXT: lui a0, 913408
; RV32IZFBFMIN-NEXT: fmv.w.x fa5, a0
-; RV32IZFBFMIN-NEXT: fle.s s0, fa5, fs0
+; RV32IZFBFMIN-NEXT: fle.s s2, fa5, fs0
+; RV32IZFBFMIN-NEXT: neg s3, s2
; RV32IZFBFMIN-NEXT: fmv.s fa0, fs0
; RV32IZFBFMIN-NEXT: call __fixsfdi
+; RV32IZFBFMIN-NEXT: and a0, s3, a0
+; RV32IZFBFMIN-NEXT: or a0, s1, a0
+; RV32IZFBFMIN-NEXT: feq.s a2, fs0, fs0
+; RV32IZFBFMIN-NEXT: neg a2, a2
; RV32IZFBFMIN-NEXT: lui a4, 524288
-; RV32IZFBFMIN-NEXT: lui a2, 524288
-; RV32IZFBFMIN-NEXT: beqz s0, .LBB10_2
+; RV32IZFBFMIN-NEXT: lui a3, 524288
+; RV32IZFBFMIN-NEXT: beqz s2, .LBB10_2
; RV32IZFBFMIN-NEXT: # %bb.1: # %start
-; RV32IZFBFMIN-NEXT: mv a2, a1
+; RV32IZFBFMIN-NEXT: mv a3, a1
; RV32IZFBFMIN-NEXT: .LBB10_2: # %start
-; RV32IZFBFMIN-NEXT: lui a1, %hi(.LCPI10_0)
-; RV32IZFBFMIN-NEXT: flw fa5, %lo(.LCPI10_0)(a1)
-; RV32IZFBFMIN-NEXT: flt.s a3, fa5, fs0
-; RV32IZFBFMIN-NEXT: beqz a3, .LBB10_4
+; RV32IZFBFMIN-NEXT: and a0, a2, a0
+; RV32IZFBFMIN-NEXT: beqz s0, .LBB10_4
; RV32IZFBFMIN-NEXT: # %bb.3:
-; RV32IZFBFMIN-NEXT: addi a2, a4, -1
+; RV32IZFBFMIN-NEXT: addi a3, a4, -1
; RV32IZFBFMIN-NEXT: .LBB10_4: # %start
-; RV32IZFBFMIN-NEXT: feq.s a1, fs0, fs0
-; RV32IZFBFMIN-NEXT: neg a4, a1
-; RV32IZFBFMIN-NEXT: and a1, a4, a2
-; RV32IZFBFMIN-NEXT: neg a2, a3
-; RV32IZFBFMIN-NEXT: neg a3, s0
-; RV32IZFBFMIN-NEXT: and a0, a3, a0
-; RV32IZFBFMIN-NEXT: or a0, a2, a0
-; RV32IZFBFMIN-NEXT: and a0, a4, a0
-; RV32IZFBFMIN-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
-; RV32IZFBFMIN-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
-; RV32IZFBFMIN-NEXT: flw fs0, 4(sp) # 4-byte Folded Reload
-; RV32IZFBFMIN-NEXT: addi sp, sp, 16
+; RV32IZFBFMIN-NEXT: and a1, a2, a3
+; RV32IZFBFMIN-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
+; RV32IZFBFMIN-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
+; RV32IZFBFMIN-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
+; RV32IZFBFMIN-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
+; RV32IZFBFMIN-NEXT: lw s3, 12(sp) # 4-byte Folded Reload
+; RV32IZFBFMIN-NEXT: flw fs0, 8(sp) # 4-byte Folded Reload
+; RV32IZFBFMIN-NEXT: addi sp, sp, 32
; RV32IZFBFMIN-NEXT: ret
;
; R32IDZFBFMIN-LABEL: fcvt_l_bf16_sat:
; R32IDZFBFMIN: # %bb.0: # %start
-; R32IDZFBFMIN-NEXT: addi sp, sp, -16
-; R32IDZFBFMIN-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
-; R32IDZFBFMIN-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
+; R32IDZFBFMIN-NEXT: addi sp, sp, -32
+; R32IDZFBFMIN-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
+; R32IDZFBFMIN-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
+; R32IDZFBFMIN-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
+; R32IDZFBFMIN-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
+; R32IDZFBFMIN-NEXT: sw s3, 12(sp) # 4-byte Folded Spill
; R32IDZFBFMIN-NEXT: fsd fs0, 0(sp) # 8-byte Folded Spill
+; R32IDZFBFMIN-NEXT: lui a0, %hi(.LCPI10_0)
+; R32IDZFBFMIN-NEXT: flw fa5, %lo(.LCPI10_0)(a0)
; R32IDZFBFMIN-NEXT: fcvt.s.bf16 fs0, fa0
+; R32IDZFBFMIN-NEXT: flt.s s0, fa5, fs0
+; R32IDZFBFMIN-NEXT: neg s1, s0
; R32IDZFBFMIN-NEXT: lui a0, 913408
; R32IDZFBFMIN-NEXT: fmv.w.x fa5, a0
-; R32IDZFBFMIN-NEXT: fle.s s0, fa5, fs0
+; R32IDZFBFMIN-NEXT: fle.s s2, fa5, fs0
+; R32IDZFBFMIN-NEXT: neg s3, s2
; R32IDZFBFMIN-NEXT: fmv.s fa0, fs0
; R32IDZFBFMIN-NEXT: call __fixsfdi
+; R32IDZFBFMIN-NEXT: and a0, s3, a0
+; R32IDZFBFMIN-NEXT: or a0, s1, a0
+; R32IDZFBFMIN-NEXT: feq.s a2, fs0, fs0
+; R32IDZFBFMIN-NEXT: neg a2, a2
; R32IDZFBFMIN-NEXT: lui a4, 524288
-; R32IDZFBFMIN-NEXT: lui a2, 524288
-; R32IDZFBFMIN-NEXT: beqz s0, .LBB10_2
+; R32IDZFBFMIN-NEXT: lui a3, 524288
+; R32IDZFBFMIN-NEXT: beqz s2, .LBB10_2
; R32IDZFBFMIN-NEXT: # %bb.1: # %start
-; R32IDZFBFMIN-NEXT: mv a2, a1
+; R32IDZFBFMIN-NEXT: mv a3, a1
; R32IDZFBFMIN-NEXT: .LBB10_2: # %start
-; R32IDZFBFMIN-NEXT: lui a1, %hi(.LCPI10_0)
-; R32IDZFBFMIN-NEXT: flw fa5, %lo(.LCPI10_0)(a1)
-; R32IDZFBFMIN-NEXT: flt.s a3, fa5, fs0
-; R32IDZFBFMIN-NEXT: beqz a3, .LBB10_4
+; R32IDZFBFMIN-NEXT: and a0, a2, a0
+; R32IDZFBFMIN-NEXT: beqz s0, .LBB10_4
; R32IDZFBFMIN-NEXT: # %bb.3:
-; R32IDZFBFMIN-NEXT: addi a2, a4, -1
+; R32IDZFBFMIN-NEXT: addi a3, a4, -1
; R32IDZFBFMIN-NEXT: .LBB10_4: # %start
-; R32IDZFBFMIN-NEXT: feq.s a1, fs0, fs0
-; R32IDZFBFMIN-NEXT: neg a4, a1
-; R32IDZFBFMIN-NEXT: and a1, a4, a2
-; R32IDZFBFMIN-NEXT: neg a2, a3
-; R32IDZFBFMIN-NEXT: neg a3, s0
-; R32IDZFBFMIN-NEXT: and a0, a3, a0
-; R32IDZFBFMIN-NEXT: or a0, a2, a0
-; R32IDZFBFMIN-NEXT: and a0, a4, a0
-; R32IDZFBFMIN-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
-; R32IDZFBFMIN-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
+; R32IDZFBFMIN-NEXT: and a1, a2, a3
+; R32IDZFBFMIN-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
+; R32IDZFBFMIN-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
+; R32IDZFBFMIN-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
+; R32IDZFBFMIN-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
+; R32IDZFBFMIN-NEXT: lw s3, 12(sp) # 4-byte Folded Reload
; R32IDZFBFMIN-NEXT: fld fs0, 0(sp) # 8-byte Folded Reload
-; R32IDZFBFMIN-NEXT: addi sp, sp, 16
+; R32IDZFBFMIN-NEXT: addi sp, sp, 32
; R32IDZFBFMIN-NEXT: ret
;
; RV32ID-LABEL: fcvt_l_bf16_sat:
diff --git a/llvm/test/CodeGen/RISCV/double-convert.ll b/llvm/test/CodeGen/RISCV/double-convert.ll
index eb8ffe75ef7697..3700a18bafc612 100644
--- a/llvm/test/CodeGen/RISCV/double-convert.ll
+++ b/llvm/test/CodeGen/RISCV/double-convert.ll
@@ -749,40 +749,47 @@ define i64 @fcvt_l_d(double %a) nounwind {
define i64 @fcvt_l_d_sat(double %a) nounwind {
; RV32IFD-LABEL: fcvt_l_d_sat:
; RV32IFD: # %bb.0: # %start
-; RV32IFD-NEXT: addi sp, sp, -16
-; RV32IFD-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
-; RV32IFD-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
+; RV32IFD-NEXT: addi sp, sp, -32
+; RV32IFD-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
+; RV32IFD-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
+; RV32IFD-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
+; RV32IFD-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
+; RV32IFD-NEXT: sw s3, 12(sp) # 4-byte Folded Spill
; RV32IFD-NEXT: fsd fs0, 0(sp) # 8-byte Folded Spill
; RV32IFD-NEXT: lui a0, %hi(.LCPI12_0)
; RV32IFD-NEXT: fld fa5, %lo(.LCPI12_0)(a0)
+; RV32IFD-NEXT: lui a0, %hi(.LCPI12_1)
+; RV32IFD-NEXT: fld fa4, %lo(.LCPI12_1)(a0)
; RV32IFD-NEXT: fmv.d fs0, fa0
-; RV32IFD-NEXT: fle.d s0, fa5, fa0
+; RV32IFD-NEXT: flt.d s0, fa5, fa0
+; RV32IFD-NEXT: neg s1, s0
+; RV32IFD-NEXT: fle.d s2, fa4, fa0
+; RV32IFD-NEXT: neg s3, s2
; RV32IFD-NEXT: call __fixdfdi
+; RV32IFD-NEXT: and a0, s3, a0
+; RV32IFD-NEXT: or a0, s1, a0
+; RV32IFD-NEXT: feq.d a2, fs0, fs0
+; RV32IFD-NEXT: neg a2, a2
; RV32IFD-NEXT: lui a4, 524288
-; RV32IFD-NEXT: lui a2, 524288
-; RV32IFD-NEXT: beqz s0, .LBB12_2
+; RV32IFD-NEXT: li a5, 1
+; RV32IFD-NEXT: lui a3, 524288
+; RV32IFD-NEXT: bne s2, a5, .LBB12_2
; RV32IFD-NEXT: # %bb.1: # %start
-; RV32IFD-NEXT: mv a2, a1
+; RV32IFD-NEXT: mv a3, a1
; RV32IFD-NEXT: .LBB12_2: # %start
-; RV32IFD-NEXT: lui a1, %hi(.LCPI12_1)
-; RV32IFD-NEXT: fld fa5, %lo(.LCPI12_1)(a1)
-; RV32IFD-NEXT: flt.d a3, fa5, fs0
-; RV32IFD-NEXT: beqz a3, .LBB12_4
+; RV32IFD-NEXT: and a0, a2, a0
+; RV32IFD-NEXT: beqz s0, .LBB12_4
; RV32IFD-NEXT: # %bb.3:
-; RV32IFD-NEXT: addi a2, a4, -1
+; RV32IFD-NEXT: addi a3, a4, -1
; RV32IFD-NEXT: .LBB12_4: # %start
-; RV32IFD-NEXT: feq.d a1, fs0, fs0
-; RV32IFD-NEXT: neg a4, a1
-; RV32IFD-NEXT: and a1, a4, a2
-; RV32IFD-NEXT: neg a2, a3
-; RV32IFD-NEXT: neg a3, s0
-; RV32IFD-NEXT: and a0, a3, a0
-; RV32IFD-NEXT: or a0, a2, a0
-; RV32IFD-NEXT: and a0, a4, a0
-; RV32IFD-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
-; RV32IFD-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
+; RV32IFD-NEXT: and a1, a2, a3
+; RV32IFD-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
+; RV32IFD-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
+; RV32IFD-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
+; RV32IFD-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
+; RV32IFD-NEXT: lw s3, 12(sp) # 4-byte Folded Reload
; RV32IFD-NEXT: fld fs0, 0(sp) # 8-byte Folded Reload
-; RV32IFD-NEXT: addi sp, sp, 16
+; RV32IFD-NEXT: addi sp, sp, 32
; RV32IFD-NEXT: ret
;
; RV64IFD-LABEL: fcvt_l_d_sat:
@@ -800,40 +807,45 @@ define i64 @fcvt_l_d_sat(double %a) nounwind {
; RV32IZFINXZDINX-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
; RV32IZFINXZDINX-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
; RV32IZFINXZDINX-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
-; RV32IZFINXZDINX-NEXT: sw a0, 8(sp)
-; RV32IZFINXZDINX-NEXT: sw a1, 12(sp)
-; RV32IZFINXZDINX-NEXT: lw s0, 8(sp)
-; RV32IZFINXZDINX-NEXT: lw s1, 12(sp)
-; RV32IZFINXZDINX-NEXT: call __fixdfdi
+; RV32IZFINXZDINX-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
+; RV32IZFINXZDINX-NEXT: sw s3, 12(sp) # 4-byte Folded Spill
+; RV32IZFINXZDINX-NEXT: sw a0, 0(sp)
+; RV32IZFINXZDINX-NEXT: sw a1, 4(sp)
+; RV32IZFINXZDINX-NEXT: lw s0, 0(sp)
+; RV32IZFINXZDINX-NEXT: lw s1, 4(sp)
; RV32IZFINXZDINX-NEXT: lui a2, %hi(.LCPI12_0)
; RV32IZFINXZDINX-NEXT: lw a3, %lo(.LCPI12_0+4)(a2)
; RV32IZFINXZDINX-NEXT: lw a2, %lo(.LCPI12_0)(a2)
-; RV32IZFINXZDINX-NEXT: fle.d a2, a2, s0
+; RV32IZFINXZDINX-NEXT: fle.d s2, a2, s0
+; RV32IZFINXZDINX-NEXT: neg s3, s2
+; RV32IZFINXZDINX-NEXT: call __fixdfdi
+; RV32IZFINXZDINX-NEXT: lui a2, %hi(.LCPI12_1)
+; RV32IZFINXZDINX-NEXT: lw a3, %lo(.LCPI12_1+4)(a2)
+; RV32IZFINXZDINX-NEXT: lw a2, %lo(.LCPI12_1)(a2)
+; RV32IZFINXZDINX-NEXT: and a0, s3, a0
+; RV32IZFINXZDINX-NEXT: flt.d a3, a2, s0
+; RV32IZFINXZDINX-NEXT: neg a2, a3
+; RV32IZFINXZDINX-NEXT: or a0, a2, a0
+; RV32IZFINXZDINX-NEXT: feq.d a2, s0, s0
+; RV32IZFINXZDINX-NEXT: neg a2, a2
; RV32IZFINXZDINX-NEXT: lui a5, 524288
-; RV32IZFINXZDINX-NEXT: lui a3, 524288
-; RV32IZFINXZDINX-NEXT: beqz a2, .LBB12_2
+; RV32IZFINXZDINX-NEXT: li a6, 1
+; RV32IZFINXZDINX-NEXT: lui a4, 524288
+; RV32IZFINXZDINX-NEXT: bne s2, a6, .LBB12_2
; RV32IZFINXZDINX-NEXT: # %bb.1: # %start
-; RV32IZFINXZDINX-NEXT: mv a3, a1
+; RV32IZFINXZDINX-NEXT: mv a4, a1
; RV32IZFINXZDINX-NEXT: .LBB12_2: # %start
-; RV32IZFINXZDINX-NEXT: lui a1, %hi(.LCPI12_1)
-; RV32IZFINXZDINX-NEXT: lw a6, %lo(.LCPI12_1)(a1)
-; RV32IZFINXZDINX-NEXT: lw a7, %lo(.LCPI12_1+4)(a1)
-; RV32IZFINXZDINX-NEXT: flt.d a4, a6, s0
-; RV32IZFINXZDINX-NEXT: beqz a4, .LBB12_4
+; RV32IZFINXZDINX-NEXT: and a0, a2, a0
+; RV32IZFINXZDINX-NEXT: beqz a3, .LBB12_4
; RV32IZFINXZDINX-NEXT: # %bb.3:
-; RV32IZFINXZDINX-NEXT: addi a3, a5, -1
+; RV32IZFINXZDINX-NEXT: addi a4, a5, -1
; RV32IZFINXZDINX-NEXT: .LBB12_4: # %start
-; RV32IZFINXZDINX-NEXT: feq.d a1, s0, s0
-; RV32IZFINXZDINX-NEXT: neg a5, a1
-; RV32IZFINXZDINX-NEXT: and a1, a5, a3
-; RV32IZFINXZDINX-NEXT: neg a2, a2
-; RV32IZFINXZDINX-NEXT: and a0, a2, a0
-; RV32IZFINXZDINX-NEXT: neg a2, a4
-; RV32IZFINXZDINX-NEXT: or a0, a2, a0
-; RV32IZFINXZDINX-NEXT: and a0, a5, a0
+; RV32IZFINXZDINX-NEXT: and a1, a2, a4
; RV32IZFINXZDINX-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
; RV32IZFINXZDINX-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
; RV32IZFINXZDINX-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
+; RV32IZFINXZDINX-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
+; RV32IZFINXZDINX-NEXT: lw s3, 12(sp) # 4-byte Folded Reload
; RV32IZFINXZDINX-NEXT: addi sp, sp, 32
; RV32IZFINXZDINX-NEXT: ret
;
@@ -856,32 +868,33 @@ define i64 @fcvt_l_d_sat(double %a) nounwind {
; RV32I-NEXT: sw s3, 12(sp) # 4-byte Folded Spill
; RV32I-NEXT: sw s4, 8(sp) # 4-byte Folded Spill
; RV32I-NEXT: sw s5, 4(sp) # 4-byte Folded Spill
+; RV32I-NEXT: sw s6, 0(sp) # 4-byte Folded Spill
; RV32I-NEXT: mv s0, a1
; RV32I-NEXT: mv s1, a0
-; RV32I-NEXT: lui a3, 278016
-; RV32I-NEXT: addi a3, a3, -1
-; RV32I-NEXT: li a2, -1
-; RV32I-NEXT: call __gtdf2
-; RV32I-NEXT: mv s2, a0
; RV32I-NEXT: lui a3, 802304
-; RV32I-NEXT: mv a0, s1
-; RV32I-NEXT: mv a1, s0
; RV32I-NEXT: li a2, 0
; RV32I-NEXT: call __gedf2
-; RV32I-NEXT: mv s3, a0
+; RV32I-NEXT: mv s2, a0
; RV32I-NEXT: mv a0, s1
; RV32I-NEXT: mv a1, s0
; RV32I-NEXT: call __fixdfdi
-; RV32I-NEXT: mv s4, a0
-; RV32I-NEXT: mv s5, a1
-; RV32I-NEXT: lui a0, 524288
-; RV32I-NEXT: bgez s3, .LBB12_2
+; RV32I-NEXT: mv s3, a0
+; RV32I-NEXT: mv s4, a1
+; RV32I-NEXT: lui s6, 524288
+; RV32I-NEXT: bgez s2, .LBB12_2
; RV32I-NEXT: # %bb.1: # %start
-; RV32I-NEXT: lui s5, 524288
+; RV32I-NEXT: lui s4, 524288
; RV32I-NEXT: .LBB12_2: # %start
-; RV32I-NEXT: blez s2, .LBB12_4
+; RV32I-NEXT: lui a3, 278016
+; RV32I-NEXT: addi a3, a3, -1
+; RV32I-NEXT: li a2, -1
+; RV32I-NEXT: mv a0, s1
+; RV32I-NEXT: mv a1, s0
+; RV32I-NEXT: call __gtdf2
+; RV32I-NEXT: mv s5, a0
+; RV32I-NEXT: blez a0, .LBB12_4
; RV32I-NEXT: # %bb.3: # %start
-; RV32I-NEXT: addi s5, a0, -1
+; RV32I-NEXT: addi s4, s6, -1
; RV32I-NEXT: .LBB12_4: # %start
; RV32I-NEXT: mv a0, s1
; RV32I-NEXT: mv a1, s0
@@ -890,11 +903,11 @@ define i64 @fcvt_l_d_sat(double %a) nounwind {
; RV32I-NEXT: call __unorddf2
; RV32I-NEXT: snez a0, a0
; RV32I-NEXT: addi a0, a0, -1
-; RV32I-NEXT: and a1, a0, s5
-; RV32I-NEXT: slti a2, s3, 0
+; RV32I-NEXT: and a1, a0, s4
+; RV32I-NEXT: slti a2, s2, 0
; RV32I-NEXT: addi a2, a2, -1
-; RV32I-NEXT: and a2, a2, s4
-; RV32I-NEXT: sgtz a3, s2
+; RV32I-NEXT: and a2, a2, s3
+; RV32I-NEXT: sgtz a3, s5
; RV32I-NEXT: neg a3, a3
; RV32I-NEXT: or a2, a3, a2
; RV32I-NEXT: and a0, a0, a2
@@ -905,6 +918,7 @@ define i64 @fcvt_l_d_sat(double %a) nounwind {
; RV32I-NEXT: lw s3, 12(sp) # 4-byte Folded Reload
; RV32I-NEXT: lw ...
[truncated]
|
topperc
force-pushed
the
pr/fp-setcc-freeze
branch
from
542c9ea to
a2c021b
Compare
|
✅ With the latest revision this PR passed the C/C++ code formatter. |
arsenm
approved these changes
Mar 8, 2024
No nans/infs in SelectionDAG is complicated. Hopefully I've captured all of the cases. I've only applied to ConsiderFlags to the SDNodeFlags since those are the only ones that will be droped by hoisting. The condition code and TargetOptions would still be in effect. Recovers some regression from llvm#84232.
topperc
force-pushed
the
pr/fp-setcc-freeze
branch
from
a2c021b to
2f62181
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No nans/infs in SelectionDAG is complicated. Hopefully I've captured
all of the cases. I've only applied to ConsiderFlags to the SDNodeFlags
since those are the only ones that will be droped by hoisting. The
condition code and TargetOptions would still be in effect.
Recovers some regression from #84232.
Stacked on #84241 and #84232