1 # NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
2 # RUN: llc -mtriple=thumbv8.1m.main -mattr=+mve -run-pass=arm-low-overhead-loops --verify-machineinstrs %s -o - | FileCheck %s
4 define dso_local void @incorrect_sub_16(i16* noalias nocapture %A, i16* noalias nocapture readonly %B, i16* noalias nocapture readonly %C, i32 %N) local_unnamed_addr #0 {
6 %cmp8 = icmp sgt i32 %N, 0
12 %5 = add nuw nsw i32 %4, 1
13 br i1 %cmp8, label %vector.ph, label %for.cond.cleanup
15 vector.ph: ; preds = %entry
16 %start = call i32 @llvm.start.loop.iterations.i32(i32 %5)
19 vector.body: ; preds = %vector.body, %vector.ph
20 %lsr.iv17 = phi i16* [ %scevgep18, %vector.body ], [ %A, %vector.ph ]
21 %lsr.iv14 = phi i16* [ %scevgep15, %vector.body ], [ %C, %vector.ph ]
22 %lsr.iv = phi i16* [ %scevgep, %vector.body ], [ %B, %vector.ph ]
23 %6 = phi i32 [ %start, %vector.ph ], [ %11, %vector.body ]
24 %7 = phi i32 [ %N, %vector.ph ], [ %9, %vector.body ]
25 %lsr.iv13 = bitcast i16* %lsr.iv to <8 x i16>*
26 %lsr.iv1416 = bitcast i16* %lsr.iv14 to <8 x i16>*
27 %lsr.iv1719 = bitcast i16* %lsr.iv17 to <8 x i16>*
28 %8 = call <8 x i1> @llvm.arm.mve.vctp16(i32 %7)
30 %wide.masked.load = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %lsr.iv13, i32 4, <8 x i1> %8, <8 x i16> undef)
31 %wide.masked.load12 = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %lsr.iv1416, i32 4, <8 x i1> %8, <8 x i16> undef)
32 %10 = add nsw <8 x i16> %wide.masked.load12, %wide.masked.load
33 call void @llvm.masked.store.v8i16.p0v8i16(<8 x i16> %10, <8 x i16>* %lsr.iv1719, i32 4, <8 x i1> %8)
34 %scevgep = getelementptr i16, i16* %lsr.iv, i32 8
35 %scevgep15 = getelementptr i16, i16* %lsr.iv14, i32 8
36 %scevgep18 = getelementptr i16, i16* %lsr.iv17, i32 8
37 %11 = call i32 @llvm.loop.decrement.reg.i32.i32.i32(i32 %6, i32 1)
38 %12 = icmp ne i32 %11, 0
39 br i1 %12, label %vector.body, label %for.cond.cleanup
41 for.cond.cleanup: ; preds = %vector.body, %entry
44 declare i32 @llvm.start.loop.iterations.i32(i32)
45 declare <8 x i1> @llvm.arm.mve.vctp16(i32)
46 declare i32 @llvm.loop.decrement.reg.i32.i32.i32(i32, i32)
47 declare <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>*, i32 immarg, <8 x i1>, <8 x i16>)
48 declare void @llvm.masked.store.v8i16.p0v8i16(<8 x i16>, <8 x i16>*, i32 immarg, <8 x i1>)
51 name: incorrect_sub_16
53 exposesReturnsTwice: false
55 regBankSelected: false
58 tracksRegLiveness: true
62 - { reg: '$r0', virtual-reg: '' }
63 - { reg: '$r1', virtual-reg: '' }
64 - { reg: '$r2', virtual-reg: '' }
65 - { reg: '$r3', virtual-reg: '' }
67 isFrameAddressTaken: false
68 isReturnAddressTaken: false
78 cvBytesOfCalleeSavedRegisters: 0
79 hasOpaqueSPAdjustment: false
81 hasMustTailInVarArgFunc: false
87 - { id: 0, name: '', type: spill-slot, offset: -4, size: 4, alignment: 4,
88 stack-id: default, callee-saved-register: '$lr', callee-saved-restored: false,
89 debug-info-variable: '', debug-info-expression: '', debug-info-location: '' }
90 - { id: 1, name: '', type: spill-slot, offset: -8, size: 4, alignment: 4,
91 stack-id: default, callee-saved-register: '$r7', callee-saved-restored: true,
92 debug-info-variable: '', debug-info-expression: '', debug-info-location: '' }
95 machineFunctionInfo: {}
97 ; CHECK-LABEL: name: incorrect_sub_16
99 ; CHECK: successors: %bb.1(0x80000000)
100 ; CHECK: liveins: $lr, $r0, $r1, $r2, $r3, $r7
101 ; CHECK: frame-setup tPUSH 14 /* CC::al */, $noreg, killed $r7, killed $lr, implicit-def $sp, implicit $sp
102 ; CHECK: frame-setup CFI_INSTRUCTION def_cfa_offset 8
103 ; CHECK: frame-setup CFI_INSTRUCTION offset $lr, -4
104 ; CHECK: frame-setup CFI_INSTRUCTION offset $r7, -8
105 ; CHECK: tCMPi8 renamable $r3, 1, 14 /* CC::al */, $noreg, implicit-def $cpsr
106 ; CHECK: t2IT 11, 8, implicit-def $itstate
107 ; CHECK: tPOP_RET 11 /* CC::lt */, killed $cpsr, def $r7, def $pc, implicit killed $itstate
108 ; CHECK: bb.1.vector.ph:
109 ; CHECK: successors: %bb.2(0x80000000)
110 ; CHECK: liveins: $r0, $r1, $r2, $r3
111 ; CHECK: renamable $r12 = t2ADDri renamable $r3, 3, 14 /* CC::al */, $noreg, $noreg
112 ; CHECK: renamable $lr = t2MOVi 1, 14 /* CC::al */, $noreg, $noreg
113 ; CHECK: renamable $r12 = t2BICri killed renamable $r12, 3, 14 /* CC::al */, $noreg, $noreg
114 ; CHECK: renamable $r12 = t2SUBri killed renamable $r12, 4, 14 /* CC::al */, $noreg, $noreg
115 ; CHECK: renamable $lr = nuw nsw t2ADDrs killed renamable $lr, killed renamable $r12, 19, 14 /* CC::al */, $noreg, $noreg
116 ; CHECK: bb.2.vector.body:
117 ; CHECK: successors: %bb.2(0x7c000000), %bb.3(0x04000000)
118 ; CHECK: liveins: $lr, $r0, $r1, $r2, $r3
119 ; CHECK: renamable $vpr = MVE_VCTP16 renamable $r3, 0, $noreg
120 ; CHECK: MVE_VPST 4, implicit $vpr
121 ; CHECK: renamable $r1, renamable $q0 = MVE_VLDRHU16_post killed renamable $r1, 16, 1, renamable $vpr :: (load (s128) from %ir.lsr.iv13, align 4)
122 ; CHECK: renamable $r2, renamable $q1 = MVE_VLDRHU16_post killed renamable $r2, 16, 1, renamable $vpr :: (load (s128) from %ir.lsr.iv1416, align 4)
123 ; CHECK: renamable $r3, dead $cpsr = tSUBi8 killed renamable $r3, 7, 14 /* CC::al */, $noreg
124 ; CHECK: renamable $q0 = nsw MVE_VADDi16 killed renamable $q1, killed renamable $q0, 0, $noreg, undef renamable $q0
125 ; CHECK: MVE_VPST 8, implicit $vpr
126 ; CHECK: renamable $r0 = MVE_VSTRHU16_post killed renamable $q0, killed renamable $r0, 16, 1, killed renamable $vpr :: (store (s128) into %ir.lsr.iv1719, align 4)
127 ; CHECK: $lr = t2LEUpdate killed renamable $lr, %bb.2
128 ; CHECK: bb.3.for.cond.cleanup:
129 ; CHECK: tPOP_RET 14 /* CC::al */, $noreg, def $r7, def $pc
131 successors: %bb.1(0x80000000)
132 liveins: $r0, $r1, $r2, $r3, $r7, $lr
134 frame-setup tPUSH 14, $noreg, killed $r7, killed $lr, implicit-def $sp, implicit $sp
135 frame-setup CFI_INSTRUCTION def_cfa_offset 8
136 frame-setup CFI_INSTRUCTION offset $lr, -4
137 frame-setup CFI_INSTRUCTION offset $r7, -8
138 tCMPi8 renamable $r3, 1, 14, $noreg, implicit-def $cpsr
139 t2IT 11, 8, implicit-def $itstate
140 tPOP_RET 11, killed $cpsr, def $r7, def $pc, implicit killed $itstate
143 successors: %bb.2(0x80000000)
144 liveins: $r0, $r1, $r2, $r3, $r7, $lr
146 renamable $r12 = t2ADDri renamable $r3, 3, 14, $noreg, $noreg
147 renamable $lr = t2MOVi 1, 14, $noreg, $noreg
148 renamable $r12 = t2BICri killed renamable $r12, 3, 14, $noreg, $noreg
149 renamable $r12 = t2SUBri killed renamable $r12, 4, 14, $noreg, $noreg
150 renamable $lr = nuw nsw t2ADDrs killed renamable $lr, killed renamable $r12, 19, 14, $noreg, $noreg
151 $lr = t2DoLoopStart renamable $lr
154 successors: %bb.2(0x7c000000), %bb.3(0x04000000)
155 liveins: $lr, $r0, $r1, $r2, $r3
157 renamable $vpr = MVE_VCTP16 renamable $r3, 0, $noreg
158 MVE_VPST 4, implicit $vpr
159 renamable $r1, renamable $q0 = MVE_VLDRHU16_post killed renamable $r1, 16, 1, renamable $vpr :: (load (s128) from %ir.lsr.iv13, align 4)
160 renamable $r2, renamable $q1 = MVE_VLDRHU16_post killed renamable $r2, 16, 1, renamable $vpr :: (load (s128) from %ir.lsr.iv1416, align 4)
161 renamable $r3, dead $cpsr = tSUBi8 killed renamable $r3, 7, 14, $noreg
162 renamable $q0 = nsw MVE_VADDi16 killed renamable $q1, killed renamable $q0, 0, $noreg, undef renamable $q0
163 MVE_VPST 8, implicit $vpr
164 renamable $r0 = MVE_VSTRHU16_post killed renamable $q0, killed renamable $r0, 16, 1, killed renamable $vpr :: (store (s128) into %ir.lsr.iv1719, align 4)
165 renamable $lr = t2LoopDec killed renamable $lr, 1
166 t2LoopEnd renamable $lr, %bb.2, implicit-def dead $cpsr
169 bb.3.for.cond.cleanup:
170 tPOP_RET 14, $noreg, def $r7, def $pc