Index: llvm/lib/CodeGen/SelectionDAG/LegalizeIntegerTypes.cpp =================================================================== --- llvm/lib/CodeGen/SelectionDAG/LegalizeIntegerTypes.cpp +++ llvm/lib/CodeGen/SelectionDAG/LegalizeIntegerTypes.cpp @@ -2296,16 +2296,20 @@ // An i1 vecreduce_or is equivalent to vecreduce_umax, use that instead if // vecreduce_or is not legal else if (Opcode == ISD::VECREDUCE_OR && OrigEltVT == MVT::i1 && - !TLI.isOperationLegalOrCustom(ISD::VECREDUCE_OR, InVT) && - TLI.isOperationLegalOrCustom(ISD::VECREDUCE_UMAX, InVT)) + !TLI.isOperationLegalOrCustom(ISD::VECREDUCE_OR, InVT) && + TLI.isOperationLegalOrCustom(ISD::VECREDUCE_UMAX, InVT)) { Opcode = ISD::VECREDUCE_UMAX; + Op = SExtPromotedInteger(N->getOperand(0)); + } // An i1 vecreduce_and is equivalent to vecreduce_umin, use that instead if // vecreduce_and is not legal else if (Opcode == ISD::VECREDUCE_AND && OrigEltVT == MVT::i1 && - !TLI.isOperationLegalOrCustom(ISD::VECREDUCE_AND, InVT) && - TLI.isOperationLegalOrCustom(ISD::VECREDUCE_UMIN, InVT)) + !TLI.isOperationLegalOrCustom(ISD::VECREDUCE_AND, InVT) && + TLI.isOperationLegalOrCustom(ISD::VECREDUCE_UMIN, InVT)) { Opcode = ISD::VECREDUCE_UMIN; + Op = SExtPromotedInteger(N->getOperand(0)); + } if (ResVT.bitsGE(EltVT)) return DAG.getNode(Opcode, SDLoc(N), ResVT, Op); Index: llvm/test/CodeGen/AArch64/illegal-floating-point-vector-compares.ll =================================================================== --- llvm/test/CodeGen/AArch64/illegal-floating-point-vector-compares.ll +++ llvm/test/CodeGen/AArch64/illegal-floating-point-vector-compares.ll @@ -8,7 +8,7 @@ ; CHECK-LABEL: unordered_floating_point_compare_on_v8f32: ; CHECK: // %bb.0: ; CHECK-NEXT: fcmgt v1.4s, v1.4s, #0.0 -; CHECK-NEXT: mov w8, #1 +; CHECK-NEXT: mov w8, #1 // =0x1 ; CHECK-NEXT: fcmgt v0.4s, v0.4s, #0.0 ; CHECK-NEXT: uzp1 v0.8h, v0.8h, v1.8h ; CHECK-NEXT: mvn v0.16b, v0.16b @@ -27,7 +27,7 @@ ; CHECK-LABEL: unordered_floating_point_compare_on_v16f32: ; CHECK: // %bb.0: ; CHECK-NEXT: fcmgt v3.4s, v3.4s, #0.0 -; CHECK-NEXT: mov w8, #1 +; CHECK-NEXT: mov w8, #1 // =0x1 ; CHECK-NEXT: fcmgt v2.4s, v2.4s, #0.0 ; CHECK-NEXT: fcmgt v1.4s, v1.4s, #0.0 ; CHECK-NEXT: fcmgt v0.4s, v0.4s, #0.0 @@ -49,7 +49,7 @@ ; CHECK-LABEL: unordered_floating_point_compare_on_v32f32: ; CHECK: // %bb.0: ; CHECK-NEXT: fcmgt v3.4s, v3.4s, #0.0 -; CHECK-NEXT: mov w9, #1 +; CHECK-NEXT: mov w9, #1 // =0x1 ; CHECK-NEXT: fcmgt v2.4s, v2.4s, #0.0 ; CHECK-NEXT: fcmgt v1.4s, v1.4s, #0.0 ; CHECK-NEXT: fcmgt v0.4s, v0.4s, #0.0 Index: llvm/test/CodeGen/AArch64/reduce-and.ll =================================================================== --- llvm/test/CodeGen/AArch64/reduce-and.ll +++ llvm/test/CodeGen/AArch64/reduce-and.ll @@ -20,6 +20,8 @@ define i1 @test_redand_v2i1(<2 x i1> %a) { ; CHECK-LABEL: test_redand_v2i1: ; CHECK: // %bb.0: +; CHECK-NEXT: shl v0.2s, v0.2s, #31 +; CHECK-NEXT: cmlt v0.2s, v0.2s, #0 ; CHECK-NEXT: uminp v0.2s, v0.2s, v0.2s ; CHECK-NEXT: fmov w8, s0 ; CHECK-NEXT: and w0, w8, #0x1 @@ -41,6 +43,8 @@ define i1 @test_redand_v4i1(<4 x i1> %a) { ; CHECK-LABEL: test_redand_v4i1: ; CHECK: // %bb.0: +; CHECK-NEXT: shl v0.4h, v0.4h, #15 +; CHECK-NEXT: cmlt v0.4h, v0.4h, #0 ; CHECK-NEXT: uminv h0, v0.4h ; CHECK-NEXT: fmov w8, s0 ; CHECK-NEXT: and w0, w8, #0x1 @@ -68,6 +72,8 @@ define i1 @test_redand_v8i1(<8 x i1> %a) { ; CHECK-LABEL: test_redand_v8i1: ; CHECK: // %bb.0: +; CHECK-NEXT: shl v0.8b, v0.8b, #7 +; CHECK-NEXT: cmlt v0.8b, v0.8b, #0 ; CHECK-NEXT: uminv b0, v0.8b ; CHECK-NEXT: fmov w8, s0 ; CHECK-NEXT: and w0, w8, #0x1 @@ -107,6 +113,8 @@ define i1 @test_redand_v16i1(<16 x i1> %a) { ; CHECK-LABEL: test_redand_v16i1: ; CHECK: // %bb.0: +; CHECK-NEXT: shl v0.16b, v0.16b, #7 +; CHECK-NEXT: cmlt v0.16b, v0.16b, #0 ; CHECK-NEXT: uminv b0, v0.16b ; CHECK-NEXT: fmov w8, s0 ; CHECK-NEXT: and w0, w8, #0x1 @@ -169,6 +177,8 @@ define <16 x i1> @test_redand_ins_v16i1(<16 x i1> %a) { ; CHECK-LABEL: test_redand_ins_v16i1: ; CHECK: // %bb.0: +; CHECK-NEXT: shl v0.16b, v0.16b, #7 +; CHECK-NEXT: cmlt v0.16b, v0.16b, #0 ; CHECK-NEXT: uminv b0, v0.16b ; CHECK-NEXT: ret ; Index: llvm/test/CodeGen/AArch64/reduce-or.ll =================================================================== --- llvm/test/CodeGen/AArch64/reduce-or.ll +++ llvm/test/CodeGen/AArch64/reduce-or.ll @@ -20,6 +20,8 @@ define i1 @test_redor_v2i1(<2 x i1> %a) { ; CHECK-LABEL: test_redor_v2i1: ; CHECK: // %bb.0: +; CHECK-NEXT: shl v0.2s, v0.2s, #31 +; CHECK-NEXT: cmlt v0.2s, v0.2s, #0 ; CHECK-NEXT: umaxp v0.2s, v0.2s, v0.2s ; CHECK-NEXT: fmov w8, s0 ; CHECK-NEXT: and w0, w8, #0x1 @@ -41,6 +43,8 @@ define i1 @test_redor_v4i1(<4 x i1> %a) { ; CHECK-LABEL: test_redor_v4i1: ; CHECK: // %bb.0: +; CHECK-NEXT: shl v0.4h, v0.4h, #15 +; CHECK-NEXT: cmlt v0.4h, v0.4h, #0 ; CHECK-NEXT: umaxv h0, v0.4h ; CHECK-NEXT: fmov w8, s0 ; CHECK-NEXT: and w0, w8, #0x1 @@ -68,6 +72,8 @@ define i1 @test_redor_v8i1(<8 x i1> %a) { ; CHECK-LABEL: test_redor_v8i1: ; CHECK: // %bb.0: +; CHECK-NEXT: shl v0.8b, v0.8b, #7 +; CHECK-NEXT: cmlt v0.8b, v0.8b, #0 ; CHECK-NEXT: umaxv b0, v0.8b ; CHECK-NEXT: fmov w8, s0 ; CHECK-NEXT: and w0, w8, #0x1 @@ -107,6 +113,8 @@ define i1 @test_redor_v16i1(<16 x i1> %a) { ; CHECK-LABEL: test_redor_v16i1: ; CHECK: // %bb.0: +; CHECK-NEXT: shl v0.16b, v0.16b, #7 +; CHECK-NEXT: cmlt v0.16b, v0.16b, #0 ; CHECK-NEXT: umaxv b0, v0.16b ; CHECK-NEXT: fmov w8, s0 ; CHECK-NEXT: and w0, w8, #0x1 @@ -169,6 +177,8 @@ define <16 x i1> @test_redor_ins_v16i1(<16 x i1> %a) { ; CHECK-LABEL: test_redor_ins_v16i1: ; CHECK: // %bb.0: +; CHECK-NEXT: shl v0.16b, v0.16b, #7 +; CHECK-NEXT: cmlt v0.16b, v0.16b, #0 ; CHECK-NEXT: umaxv b0, v0.16b ; CHECK-NEXT: ret ; Index: llvm/test/CodeGen/AArch64/reduce-xor.ll =================================================================== --- llvm/test/CodeGen/AArch64/reduce-xor.ll +++ llvm/test/CodeGen/AArch64/reduce-xor.ll @@ -19,6 +19,8 @@ define i1 @test_redxor_v2i1(<2 x i1> %a) { ; CHECK-LABEL: test_redxor_v2i1: ; CHECK: // %bb.0: +; CHECK-NEXT: shl v0.2s, v0.2s, #31 +; CHECK-NEXT: cmlt v0.2s, v0.2s, #0 ; CHECK-NEXT: addp v0.2s, v0.2s, v0.2s ; CHECK-NEXT: fmov w8, s0 ; CHECK-NEXT: and w0, w8, #0x1 @@ -40,6 +42,8 @@ define i1 @test_redxor_v4i1(<4 x i1> %a) { ; CHECK-LABEL: test_redxor_v4i1: ; CHECK: // %bb.0: +; CHECK-NEXT: shl v0.4h, v0.4h, #15 +; CHECK-NEXT: cmlt v0.4h, v0.4h, #0 ; CHECK-NEXT: addv h0, v0.4h ; CHECK-NEXT: fmov w8, s0 ; CHECK-NEXT: and w0, w8, #0x1 @@ -67,6 +71,8 @@ define i1 @test_redxor_v8i1(<8 x i1> %a) { ; CHECK-LABEL: test_redxor_v8i1: ; CHECK: // %bb.0: +; CHECK-NEXT: shl v0.8b, v0.8b, #7 +; CHECK-NEXT: cmlt v0.8b, v0.8b, #0 ; CHECK-NEXT: addv b0, v0.8b ; CHECK-NEXT: fmov w8, s0 ; CHECK-NEXT: and w0, w8, #0x1 @@ -106,6 +112,8 @@ define i1 @test_redxor_v16i1(<16 x i1> %a) { ; CHECK-LABEL: test_redxor_v16i1: ; CHECK: // %bb.0: +; CHECK-NEXT: shl v0.16b, v0.16b, #7 +; CHECK-NEXT: cmlt v0.16b, v0.16b, #0 ; CHECK-NEXT: addv b0, v0.16b ; CHECK-NEXT: fmov w8, s0 ; CHECK-NEXT: and w0, w8, #0x1 @@ -168,6 +176,8 @@ define <16 x i1> @test_redxor_ins_v16i1(<16 x i1> %a) { ; CHECK-LABEL: test_redxor_ins_v16i1: ; CHECK: // %bb.0: +; CHECK-NEXT: shl v0.16b, v0.16b, #7 +; CHECK-NEXT: cmlt v0.16b, v0.16b, #0 ; CHECK-NEXT: addv b0, v0.16b ; CHECK-NEXT: ret ; Index: llvm/test/CodeGen/AArch64/sve-fixed-length-ptest.ll =================================================================== --- llvm/test/CodeGen/AArch64/sve-fixed-length-ptest.ll +++ llvm/test/CodeGen/AArch64/sve-fixed-length-ptest.ll @@ -4,7 +4,7 @@ define i1 @ptest_v16i1_256bit_min_sve(ptr %a, ptr %b) vscale_range(2, 0) { ; CHECK-LABEL: ptest_v16i1_256bit_min_sve: ; CHECK: // %bb.0: -; CHECK-NEXT: mov x8, #8 +; CHECK-NEXT: mov x8, #8 // =0x8 ; CHECK-NEXT: ptrue p0.s, vl8 ; CHECK-NEXT: ld1w { z0.s }, p0/z, [x0, x8, lsl #2] ; CHECK-NEXT: ld1w { z1.s }, p0/z, [x0] Index: llvm/test/CodeGen/AArch64/vecreduce-add-legalization.ll =================================================================== --- llvm/test/CodeGen/AArch64/vecreduce-add-legalization.ll +++ llvm/test/CodeGen/AArch64/vecreduce-add-legalization.ll @@ -123,6 +123,8 @@ define i1 @test_v4i1(<4 x i1> %a) nounwind { ; CHECK-LABEL: test_v4i1: ; CHECK: // %bb.0: +; CHECK-NEXT: shl v0.4h, v0.4h, #15 +; CHECK-NEXT: cmlt v0.4h, v0.4h, #0 ; CHECK-NEXT: addv h0, v0.4h ; CHECK-NEXT: fmov w8, s0 ; CHECK-NEXT: and w0, w8, #0x1 Index: llvm/test/CodeGen/AArch64/vecreduce-and-legalization.ll =================================================================== --- llvm/test/CodeGen/AArch64/vecreduce-and-legalization.ll +++ llvm/test/CodeGen/AArch64/vecreduce-and-legalization.ll @@ -96,7 +96,7 @@ define i8 @test_v9i8(<9 x i8> %a) nounwind { ; CHECK-LABEL: test_v9i8: ; CHECK: // %bb.0: -; CHECK-NEXT: mov w8, #-1 +; CHECK-NEXT: mov w8, #-1 // =0xffffffff ; CHECK-NEXT: umov w14, v0.b[6] ; CHECK-NEXT: mov v1.16b, v0.16b ; CHECK-NEXT: mov v1.b[9], w8 @@ -140,6 +140,8 @@ define i1 @test_v4i1(<4 x i1> %a) nounwind { ; CHECK-LABEL: test_v4i1: ; CHECK: // %bb.0: +; CHECK-NEXT: shl v0.4h, v0.4h, #15 +; CHECK-NEXT: cmlt v0.4h, v0.4h, #0 ; CHECK-NEXT: uminv h0, v0.4h ; CHECK-NEXT: fmov w8, s0 ; CHECK-NEXT: and w0, w8, #0x1 Index: llvm/test/CodeGen/AArch64/vecreduce-umax-legalization.ll =================================================================== --- llvm/test/CodeGen/AArch64/vecreduce-umax-legalization.ll +++ llvm/test/CodeGen/AArch64/vecreduce-umax-legalization.ll @@ -138,6 +138,8 @@ define i1 @test_v4i1(<4 x i1> %a) nounwind { ; CHECK-LABEL: test_v4i1: ; CHECK: // %bb.0: +; CHECK-NEXT: shl v0.4h, v0.4h, #15 +; CHECK-NEXT: cmlt v0.4h, v0.4h, #0 ; CHECK-NEXT: umaxv h0, v0.4h ; CHECK-NEXT: fmov w8, s0 ; CHECK-NEXT: and w0, w8, #0x1