[llvm] [MVEGatherScatter] Fix GEP scale calculations (PR #185437)
via llvm-commits
llvm-commits at lists.llvm.org
Mon Mar 9 08:02:28 PDT 2026
llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT-->
@llvm/pr-subscribers-backend-arm
Author: Nikita Popov (nikic)
<details>
<summary>Changes</summary>
The GEP scale for a single index GEP is the type alloc size of the source element type. The pass was mostly computing it correctly, but two places were doing something different.
---
Patch is 31.89 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/185437.diff
2 Files Affected:
- (modified) llvm/lib/Target/ARM/MVEGatherScatterLowering.cpp (+2-2)
- (modified) llvm/test/CodeGen/Thumb2/mve-gather-increment.ll (+152-139)
``````````diff
diff --git a/llvm/lib/Target/ARM/MVEGatherScatterLowering.cpp b/llvm/lib/Target/ARM/MVEGatherScatterLowering.cpp
index 374a07882b485..869436a7e4964 100644
--- a/llvm/lib/Target/ARM/MVEGatherScatterLowering.cpp
+++ b/llvm/lib/Target/ARM/MVEGatherScatterLowering.cpp
@@ -230,7 +230,7 @@ Value *MVEGatherScatterLowering::decomposePtr(Value *Ptr, Value *&Offsets,
if (auto *GEP = dyn_cast<GetElementPtrInst>(Ptr)) {
if (Value *V = decomposeGEP(Offsets, Ty, GEP, Builder)) {
Scale =
- computeScale(GEP->getSourceElementType()->getPrimitiveSizeInBits(),
+ computeScale(DL->getTypeAllocSizeInBits(GEP->getSourceElementType()),
MemoryTy->getScalarSizeInBits());
return Scale == -1 ? nullptr : V;
}
@@ -752,7 +752,7 @@ Instruction *MVEGatherScatterLowering::tryCreateIncrementingGatScat(
// The gep was in charge of making sure the offsets are scaled correctly
// - calculate that factor so it can be applied by hand
int TypeScale =
- computeScale(DL->getTypeSizeInBits(GEP->getSourceElementType()),
+ computeScale(DL->getTypeAllocSizeInBits(GEP->getSourceElementType()),
DL->getTypeSizeInBits(GEP->getType()) /
cast<FixedVectorType>(GEP->getType())->getNumElements());
if (TypeScale == -1)
diff --git a/llvm/test/CodeGen/Thumb2/mve-gather-increment.ll b/llvm/test/CodeGen/Thumb2/mve-gather-increment.ll
index 0d86f22a321e0..5c673dda57aef 100644
--- a/llvm/test/CodeGen/Thumb2/mve-gather-increment.ll
+++ b/llvm/test/CodeGen/Thumb2/mve-gather-increment.ll
@@ -14,6 +14,19 @@ define arm_aapcs_vfpcc <4 x i32> @gather_inc_mini_4i32(ptr noalias nocapture rea
ret <4 x i32> %wide.masked.gather
}
+define arm_aapcs_vfpcc <4 x i32> @gather_inc_mini_4i32_gep_type(ptr noalias nocapture readonly %data, ptr noalias nocapture %dst, <4 x i32> %offs) {
+; CHECK-LABEL: gather_inc_mini_4i32_gep_type:
+; CHECK: @ %bb.0:
+; CHECK-NEXT: movs r1, #4
+; CHECK-NEXT: vadd.i32 q1, q0, r1
+; CHECK-NEXT: vldrw.u32 q0, [r0, q1, uxtw #2]
+; CHECK-NEXT: bx lr
+ %1 = add <4 x i32> %offs, <i32 4, i32 4, i32 4, i32 4>
+ %2 = getelementptr inbounds [4 x i8], ptr %data, <4 x i32> %1
+ %wide.masked.gather = call <4 x i32> @llvm.masked.gather.v4i32.v4p0(<4 x ptr> %2, i32 4, <4 x i1> <i1 true, i1 true, i1 true, i1 true>, <4 x i32> undef)
+ ret <4 x i32> %wide.masked.gather
+}
+
define arm_aapcs_vfpcc <4 x i32> @gather_inc_mini_4i32_i8(i32* noalias nocapture readonly %data, i32* noalias nocapture %dst, <4 x i32> %offs) {
; CHECK-LABEL: gather_inc_mini_4i32_i8:
; CHECK: @ %bb.0:
@@ -220,20 +233,20 @@ define arm_aapcs_vfpcc <16 x i8> @gather_inc_minipred_16i8(ptr noalias nocapture
define arm_aapcs_vfpcc void @gather_pre_inc(ptr noalias nocapture readonly %data, ptr noalias nocapture %dst, i32 %n.vec) {
; CHECK-LABEL: gather_pre_inc:
; CHECK: @ %bb.0: @ %vector.ph
-; CHECK-NEXT: adr r3, .LCPI7_0
+; CHECK-NEXT: adr r3, .LCPI8_0
; CHECK-NEXT: vldrw.u32 q0, [r3]
; CHECK-NEXT: vadd.i32 q0, q0, r0
-; CHECK-NEXT: .LBB7_1: @ %vector.body
+; CHECK-NEXT: .LBB8_1: @ %vector.body
; CHECK-NEXT: @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT: vldrw.u32 q1, [q0, #96]!
; CHECK-NEXT: subs r2, #4
; CHECK-NEXT: vstrb.8 q1, [r1], #16
-; CHECK-NEXT: bne .LBB7_1
+; CHECK-NEXT: bne .LBB8_1
; CHECK-NEXT: @ %bb.2: @ %end
; CHECK-NEXT: bx lr
; CHECK-NEXT: .p2align 4
; CHECK-NEXT: @ %bb.3:
-; CHECK-NEXT: .LCPI7_0:
+; CHECK-NEXT: .LCPI8_0:
; CHECK-NEXT: .long 4294967224 @ 0xffffffb8
; CHECK-NEXT: .long 4294967248 @ 0xffffffd0
; CHECK-NEXT: .long 4294967272 @ 0xffffffe8
@@ -262,20 +275,20 @@ end:
define arm_aapcs_vfpcc void @gather_pre_inc_i8(ptr noalias nocapture readonly %data, ptr noalias nocapture %dst, i32 %n.vec) {
; CHECK-LABEL: gather_pre_inc_i8:
; CHECK: @ %bb.0: @ %vector.ph
-; CHECK-NEXT: adr r3, .LCPI8_0
+; CHECK-NEXT: adr r3, .LCPI9_0
; CHECK-NEXT: vldrw.u32 q0, [r3]
; CHECK-NEXT: vadd.i32 q0, q0, r0
-; CHECK-NEXT: .LBB8_1: @ %vector.body
+; CHECK-NEXT: .LBB9_1: @ %vector.body
; CHECK-NEXT: @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT: vldrw.u32 q1, [q0, #24]!
; CHECK-NEXT: subs r2, #4
; CHECK-NEXT: vstrb.8 q1, [r1], #16
-; CHECK-NEXT: bne .LBB8_1
+; CHECK-NEXT: bne .LBB9_1
; CHECK-NEXT: @ %bb.2: @ %end
; CHECK-NEXT: bx lr
; CHECK-NEXT: .p2align 4
; CHECK-NEXT: @ %bb.3:
-; CHECK-NEXT: .LCPI8_0:
+; CHECK-NEXT: .LCPI9_0:
; CHECK-NEXT: .long 4294967278 @ 0xffffffee
; CHECK-NEXT: .long 4294967284 @ 0xfffffff4
; CHECK-NEXT: .long 4294967290 @ 0xfffffffa
@@ -304,20 +317,20 @@ end:
define arm_aapcs_vfpcc void @gather_post_inc(ptr noalias nocapture readonly %data, ptr noalias nocapture %dst, i32 %n.vec43) {
; CHECK-LABEL: gather_post_inc:
; CHECK: @ %bb.0: @ %vector.ph41
-; CHECK-NEXT: adr r3, .LCPI9_0
+; CHECK-NEXT: adr r3, .LCPI10_0
; CHECK-NEXT: vldrw.u32 q0, [r3]
; CHECK-NEXT: vadd.i32 q0, q0, r0
-; CHECK-NEXT: .LBB9_1: @ %vector.body39
+; CHECK-NEXT: .LBB10_1: @ %vector.body39
; CHECK-NEXT: @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT: vldrw.u32 q1, [q0, #96]!
; CHECK-NEXT: subs r2, #4
; CHECK-NEXT: vstrb.8 q1, [r1], #16
-; CHECK-NEXT: bne .LBB9_1
+; CHECK-NEXT: bne .LBB10_1
; CHECK-NEXT: @ %bb.2: @ %end
; CHECK-NEXT: bx lr
; CHECK-NEXT: .p2align 4
; CHECK-NEXT: @ %bb.3:
-; CHECK-NEXT: .LCPI9_0:
+; CHECK-NEXT: .LCPI10_0:
; CHECK-NEXT: .long 4294967200 @ 0xffffffa0
; CHECK-NEXT: .long 4294967224 @ 0xffffffb8
; CHECK-NEXT: .long 4294967248 @ 0xffffffd0
@@ -348,38 +361,38 @@ define arm_aapcs_vfpcc void @gather_inc_v4i32_simple(ptr noalias nocapture reado
; CHECK-NEXT: cmp r2, #1
; CHECK-NEXT: it lt
; CHECK-NEXT: bxlt lr
-; CHECK-NEXT: .LBB10_1: @ %vector.ph.preheader
+; CHECK-NEXT: .LBB11_1: @ %vector.ph.preheader
; CHECK-NEXT: .save {r4, lr}
; CHECK-NEXT: push {r4, lr}
; CHECK-NEXT: bic r12, r2, #3
; CHECK-NEXT: movs r3, #1
; CHECK-NEXT: sub.w lr, r12, #4
; CHECK-NEXT: add.w r4, r3, lr, lsr #2
-; CHECK-NEXT: adr r3, .LCPI10_0
+; CHECK-NEXT: adr r3, .LCPI11_0
; CHECK-NEXT: vldrw.u32 q0, [r3]
; CHECK-NEXT: vadd.i32 q0, q0, r0
-; CHECK-NEXT: .LBB10_2: @ %vector.ph
+; CHECK-NEXT: .LBB11_2: @ %vector.ph
; CHECK-NEXT: @ =>This Loop Header: Depth=1
-; CHECK-NEXT: @ Child Loop BB10_3 Depth 2
+; CHECK-NEXT: @ Child Loop BB11_3 Depth 2
; CHECK-NEXT: dls lr, r4
; CHECK-NEXT: mov r0, r1
; CHECK-NEXT: vmov q1, q0
-; CHECK-NEXT: .LBB10_3: @ %vector.body
-; CHECK-NEXT: @ Parent Loop BB10_2 Depth=1
+; CHECK-NEXT: .LBB11_3: @ %vector.body
+; CHECK-NEXT: @ Parent Loop BB11_2 Depth=1
; CHECK-NEXT: @ => This Inner Loop Header: Depth=2
; CHECK-NEXT: vldrw.u32 q2, [q1, #16]!
; CHECK-NEXT: vstrb.8 q2, [r0], #16
-; CHECK-NEXT: le lr, .LBB10_3
+; CHECK-NEXT: le lr, .LBB11_3
; CHECK-NEXT: @ %bb.4: @ %middle.block
-; CHECK-NEXT: @ in Loop: Header=BB10_2 Depth=1
+; CHECK-NEXT: @ in Loop: Header=BB11_2 Depth=1
; CHECK-NEXT: cmp r12, r2
-; CHECK-NEXT: bne .LBB10_2
+; CHECK-NEXT: bne .LBB11_2
; CHECK-NEXT: @ %bb.5:
; CHECK-NEXT: pop.w {r4, lr}
; CHECK-NEXT: bx lr
; CHECK-NEXT: .p2align 4
; CHECK-NEXT: @ %bb.6:
-; CHECK-NEXT: .LCPI10_0:
+; CHECK-NEXT: .LCPI11_0:
; CHECK-NEXT: .long 4294967280 @ 0xfffffff0
; CHECK-NEXT: .long 4294967284 @ 0xfffffff4
; CHECK-NEXT: .long 4294967288 @ 0xfffffff8
@@ -418,7 +431,7 @@ define arm_aapcs_vfpcc void @gather_inc_v4i32_complex(ptr noalias nocapture read
; CHECK-NEXT: cmp r2, #1
; CHECK-NEXT: it lt
; CHECK-NEXT: bxlt lr
-; CHECK-NEXT: .LBB11_1: @ %vector.ph.preheader
+; CHECK-NEXT: .LBB12_1: @ %vector.ph.preheader
; CHECK-NEXT: .save {r4, r5, r7, lr}
; CHECK-NEXT: push {r4, r5, r7, lr}
; CHECK-NEXT: .vsave {d8, d9, d10, d11, d12, d13, d14, d15}
@@ -426,26 +439,26 @@ define arm_aapcs_vfpcc void @gather_inc_v4i32_complex(ptr noalias nocapture read
; CHECK-NEXT: bic r12, r2, #3
; CHECK-NEXT: movs r3, #1
; CHECK-NEXT: sub.w lr, r12, #4
-; CHECK-NEXT: adr r4, .LCPI11_1
-; CHECK-NEXT: adr r5, .LCPI11_2
+; CHECK-NEXT: adr r4, .LCPI12_1
+; CHECK-NEXT: adr r5, .LCPI12_2
; CHECK-NEXT: vldrw.u32 q1, [r4]
; CHECK-NEXT: add.w r3, r3, lr, lsr #2
-; CHECK-NEXT: adr.w lr, .LCPI11_0
+; CHECK-NEXT: adr.w lr, .LCPI12_0
; CHECK-NEXT: vldrw.u32 q0, [r5]
; CHECK-NEXT: vldrw.u32 q2, [lr]
; CHECK-NEXT: vadd.i32 q1, q1, r0
; CHECK-NEXT: vadd.i32 q0, q0, r0
; CHECK-NEXT: vadd.i32 q2, q2, r0
-; CHECK-NEXT: .LBB11_2: @ %vector.ph
+; CHECK-NEXT: .LBB12_2: @ %vector.ph
; CHECK-NEXT: @ =>This Loop Header: Depth=1
-; CHECK-NEXT: @ Child Loop BB11_3 Depth 2
+; CHECK-NEXT: @ Child Loop BB12_3 Depth 2
; CHECK-NEXT: dls lr, r3
; CHECK-NEXT: mov r0, r1
; CHECK-NEXT: vmov q3, q1
; CHECK-NEXT: vmov q4, q0
; CHECK-NEXT: vmov q5, q2
-; CHECK-NEXT: .LBB11_3: @ %vector.body
-; CHECK-NEXT: @ Parent Loop BB11_2 Depth=1
+; CHECK-NEXT: .LBB12_3: @ %vector.body
+; CHECK-NEXT: @ Parent Loop BB12_2 Depth=1
; CHECK-NEXT: @ => This Inner Loop Header: Depth=2
; CHECK-NEXT: vldrw.u32 q6, [q5, #48]!
; CHECK-NEXT: vldrw.u32 q7, [q3, #48]!
@@ -453,28 +466,28 @@ define arm_aapcs_vfpcc void @gather_inc_v4i32_complex(ptr noalias nocapture read
; CHECK-NEXT: vldrw.u32 q7, [q4, #48]!
; CHECK-NEXT: vadd.i32 q6, q6, q7
; CHECK-NEXT: vstrb.8 q6, [r0], #16
-; CHECK-NEXT: le lr, .LBB11_3
+; CHECK-NEXT: le lr, .LBB12_3
; CHECK-NEXT: @ %bb.4: @ %middle.block
-; CHECK-NEXT: @ in Loop: Header=BB11_2 Depth=1
+; CHECK-NEXT: @ in Loop: Header=BB12_2 Depth=1
; CHECK-NEXT: cmp r12, r2
-; CHECK-NEXT: bne .LBB11_2
+; CHECK-NEXT: bne .LBB12_2
; CHECK-NEXT: @ %bb.5:
; CHECK-NEXT: vpop {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT: pop.w {r4, r5, r7, lr}
; CHECK-NEXT: bx lr
; CHECK-NEXT: .p2align 4
; CHECK-NEXT: @ %bb.6:
-; CHECK-NEXT: .LCPI11_0:
+; CHECK-NEXT: .LCPI12_0:
; CHECK-NEXT: .long 4294967248 @ 0xffffffd0
; CHECK-NEXT: .long 4294967260 @ 0xffffffdc
; CHECK-NEXT: .long 4294967272 @ 0xffffffe8
; CHECK-NEXT: .long 4294967284 @ 0xfffffff4
-; CHECK-NEXT: .LCPI11_1:
+; CHECK-NEXT: .LCPI12_1:
; CHECK-NEXT: .long 4294967252 @ 0xffffffd4
; CHECK-NEXT: .long 4294967264 @ 0xffffffe0
; CHECK-NEXT: .long 4294967276 @ 0xffffffec
; CHECK-NEXT: .long 4294967288 @ 0xfffffff8
-; CHECK-NEXT: .LCPI11_2:
+; CHECK-NEXT: .LCPI12_2:
; CHECK-NEXT: .long 4294967256 @ 0xffffffd8
; CHECK-NEXT: .long 4294967268 @ 0xffffffe4
; CHECK-NEXT: .long 4294967280 @ 0xfffffff0
@@ -522,38 +535,38 @@ define arm_aapcs_vfpcc void @gather_inc_v4i32_large(ptr noalias nocapture readon
; CHECK-NEXT: cmp r2, #1
; CHECK-NEXT: it lt
; CHECK-NEXT: bxlt lr
-; CHECK-NEXT: .LBB12_1: @ %vector.ph.preheader
+; CHECK-NEXT: .LBB13_1: @ %vector.ph.preheader
; CHECK-NEXT: .save {r4, lr}
; CHECK-NEXT: push {r4, lr}
; CHECK-NEXT: bic r12, r2, #3
; CHECK-NEXT: movs r3, #1
; CHECK-NEXT: sub.w lr, r12, #4
; CHECK-NEXT: add.w r4, r3, lr, lsr #2
-; CHECK-NEXT: adr r3, .LCPI12_0
+; CHECK-NEXT: adr r3, .LCPI13_0
; CHECK-NEXT: vldrw.u32 q0, [r3]
; CHECK-NEXT: vadd.i32 q0, q0, r0
-; CHECK-NEXT: .LBB12_2: @ %vector.ph
+; CHECK-NEXT: .LBB13_2: @ %vector.ph
; CHECK-NEXT: @ =>This Loop Header: Depth=1
-; CHECK-NEXT: @ Child Loop BB12_3 Depth 2
+; CHECK-NEXT: @ Child Loop BB13_3 Depth 2
; CHECK-NEXT: dls lr, r4
; CHECK-NEXT: mov r0, r1
; CHECK-NEXT: vmov q1, q0
-; CHECK-NEXT: .LBB12_3: @ %vector.body
-; CHECK-NEXT: @ Parent Loop BB12_2 Depth=1
+; CHECK-NEXT: .LBB13_3: @ %vector.body
+; CHECK-NEXT: @ Parent Loop BB13_2 Depth=1
; CHECK-NEXT: @ => This Inner Loop Header: Depth=2
; CHECK-NEXT: vldrw.u32 q2, [q1, #508]!
; CHECK-NEXT: vstrb.8 q2, [r0], #16
-; CHECK-NEXT: le lr, .LBB12_3
+; CHECK-NEXT: le lr, .LBB13_3
; CHECK-NEXT: @ %bb.4: @ %middle.block
-; CHECK-NEXT: @ in Loop: Header=BB12_2 Depth=1
+; CHECK-NEXT: @ in Loop: Header=BB13_2 Depth=1
; CHECK-NEXT: cmp r12, r2
-; CHECK-NEXT: bne .LBB12_2
+; CHECK-NEXT: bne .LBB13_2
; CHECK-NEXT: @ %bb.5:
; CHECK-NEXT: pop.w {r4, lr}
; CHECK-NEXT: bx lr
; CHECK-NEXT: .p2align 4
; CHECK-NEXT: @ %bb.6:
-; CHECK-NEXT: .LCPI12_0:
+; CHECK-NEXT: .LCPI13_0:
; CHECK-NEXT: .long 4294966788 @ 0xfffffe04
; CHECK-NEXT: .long 4294966792 @ 0xfffffe08
; CHECK-NEXT: .long 4294966796 @ 0xfffffe0c
@@ -598,7 +611,7 @@ define arm_aapcs_vfpcc void @gather_inc_v8i16_simple(ptr noalias nocapture reado
; CHECK-NEXT: sub sp, #28
; CHECK-NEXT: cmp r2, #1
; CHECK-NEXT: strd r1, r2, [sp, #4] @ 8-byte Folded Spill
-; CHECK-NEXT: blt .LBB13_5
+; CHECK-NEXT: blt .LBB14_5
; CHECK-NEXT: @ %bb.1: @ %vector.ph.preheader
; CHECK-NEXT: ldr r1, [sp, #8] @ 4-byte Reload
; CHECK-NEXT: movs r6, #1
@@ -608,16 +621,16 @@ define arm_aapcs_vfpcc void @gather_inc_v8i16_simple(ptr noalias nocapture reado
; CHECK-NEXT: str r1, [sp] @ 4-byte Spill
; CHECK-NEXT: sub.w r3, r1, #8
; CHECK-NEXT: add.w r8, r6, r3, lsr #3
-; CHECK-NEXT: adr r3, .LCPI13_0
+; CHECK-NEXT: adr r3, .LCPI14_0
; CHECK-NEXT: vldrw.u32 q0, [r3]
-; CHECK-NEXT: .LBB13_2: @ %vector.ph
+; CHECK-NEXT: .LBB14_2: @ %vector.ph
; CHECK-NEXT: @ =>This Loop Header: Depth=1
-; CHECK-NEXT: @ Child Loop BB13_3 Depth 2
+; CHECK-NEXT: @ Child Loop BB14_3 Depth 2
; CHECK-NEXT: dls lr, r8
; CHECK-NEXT: vmov q1, q0
; CHECK-NEXT: ldr r6, [sp, #4] @ 4-byte Reload
-; CHECK-NEXT: .LBB13_3: @ %vector.body
-; CHECK-NEXT: @ Parent Loop BB13_2 Depth=1
+; CHECK-NEXT: .LBB14_3: @ %vector.body
+; CHECK-NEXT: @ Parent Loop BB14_2 Depth=1
; CHECK-NEXT: @ => This Inner Loop Header: Depth=2
; CHECK-NEXT: vstrw.32 q1, [r2]
; CHECK-NEXT: mov r12, r2
@@ -650,19 +663,19 @@ define arm_aapcs_vfpcc void @gather_inc_v8i16_simple(ptr noalias nocapture reado
; CHECK-NEXT: vmov.16 q2[6], r7
; CHECK-NEXT: vmov.16 q2[7], r5
; CHECK-NEXT: vstrb.8 q2, [r6], #16
-; CHECK-NEXT: le lr, .LBB13_3
+; CHECK-NEXT: le lr, .LBB14_3
; CHECK-NEXT: @ %bb.4: @ %middle.block
-; CHECK-NEXT: @ in Loop: Header=BB13_2 Depth=1
+; CHECK-NEXT: @ in Loop: Header=BB14_2 Depth=1
; CHECK-NEXT: ldr r1, [sp, #8] @ 4-byte Reload
; CHECK-NEXT: ldr r3, [sp] @ 4-byte Reload
; CHECK-NEXT: cmp r3, r1
-; CHECK-NEXT: bne .LBB13_2
-; CHECK-NEXT: .LBB13_5: @ %for.cond.cleanup
+; CHECK-NEXT: bne .LBB14_2
+; CHECK-NEXT: .LBB14_5: @ %for.cond.cleanup
; CHECK-NEXT: add sp, #28
; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc}
; CHECK-NEXT: .p2align 4
; CHECK-NEXT: @ %bb.6:
-; CHECK-NEXT: .LCPI13_0:
+; CHECK-NEXT: .LCPI14_0:
; CHECK-NEXT: .short 0 @ 0x0
; CHECK-NEXT: .short 1 @ 0x1
; CHECK-NEXT: .short 2 @ 0x2
@@ -715,10 +728,10 @@ define arm_aapcs_vfpcc void @gather_inc_v8i16_complex(ptr noalias nocapture read
; CHECK-NEXT: sub sp, #120
; CHECK-NEXT: cmp r2, #1
; CHECK-NEXT: strd r1, r2, [sp, #64] @ 8-byte Folded Spill
-; CHECK-NEXT: blt.w .LBB14_5
+; CHECK-NEXT: blt.w .LBB15_5
; CHECK-NEXT: @ %bb.1: @ %vector.ph.preheader
; CHECK-NEXT: ldr r1, [sp, #68] @ 4-byte Reload
-; CHECK-NEXT: adr r3, .LCPI14_2
+; CHECK-NEXT: adr r3, .LCPI15_2
; CHECK-NEXT: vldrw.u32 q0, [r3]
; CHECK-NEXT: movs r2, #1
; CHECK-NEXT: bic r1, r1, #7
@@ -727,16 +740,16 @@ define arm_aapcs_vfpcc void @gather_inc_v8i16_complex(ptr noalias nocapture read
; CHECK-NEXT: vstrw.32 q0, [sp, #40] @ 16-byte Spill
; CHECK-NEXT: add.w r1, r2, r1, lsr #3
; CHECK-NEXT: str r1, [sp, #60] @ 4-byte Spill
-; CHECK-NEXT: adr r1, .LCPI14_0
-; CHECK-NEXT: adr r2, .LCPI14_1
+; CHECK-NEXT: adr r1, .LCPI15_0
+; CHECK-NEXT: adr r2, .LCPI15_1
; CHECK-NEXT: vldrw.u32 q0, [r1]
; CHECK-NEXT: vstrw.32 q0, [sp, #24] @ 16-byte Spill
; CHECK-NEXT: vldrw.u32 q0, [r2]
; CHECK-NEXT: add r2, sp, #104
; CHECK-NEXT: vstrw.32 q0, [sp, #8] @ 16-byte Spill
-; CHECK-NEXT: .LBB14_2: @ %vector.ph
+; CHECK-NEXT: .LBB15_2: @ %vector.ph
; CHECK-NEXT: @ =>This Loop Header: Depth=1
-; CHECK-NEXT: @ Child Loop BB14_3 Depth 2
+; CHECK-NEXT: @ Child Loop BB15_3 Depth 2
; CHECK-NEXT: ldr r1, [sp, #60] @ 4-byte Reload
; CHECK-NEXT: add.w r10, sp, #88
; CHECK-NEXT: dls lr, r1
@@ -744,8 +757,8 @@ define arm_aapcs_vfpcc void @gather_inc_v8i16_complex(ptr noalias nocapture read
; CHECK-NEXT: vldrw.u32 q4, [sp, #24] @ 16-byte Reload
; CHECK-NEXT: vldrw.u32 q5, [sp, #40] @ 16-byte Reload
; CHECK-NEXT: vldrw.u32 q6, [sp, #8] @ 16-byte Reload
-; CHECK-NEXT: .LBB14_3: @ %vector.body
-; CHECK-NEXT: @ Parent Loop BB14_2 Depth=1
+; CHECK-NEXT: .LBB15_3: @ %vector.body
+; CHECK-NEXT: @ Parent Loop BB15_2 Depth=1
; CHECK-NEXT: @ => This Inner Loop Header: Depth=2
; CHECK-NEXT: vstrw.32 q5, [r2]
; CHECK-NEXT: mov r8, r2
@@ -839,21 +852,21 @@ define arm_aapcs_vfpcc void @gather_inc_v8i16_complex(ptr noalias nocapture read
; CHECK-NEXT: vadd.i16 q0, q3, q1
; CHECK-NEXT: vadd.i16 q0, q0, q7
; CHECK-NEXT: vstrb.8 q0, [r7], #16
-; CHECK-NEXT: le lr, .LBB14_3
+; CHECK-NEXT: le lr, .LBB15_3
; CHECK-NEXT: @ %bb.4: @ %middle.block
-; CHECK-NEXT: @ in Loop: Header=BB14_2 Depth=1
+; CHECK-NEXT: @ in Loop: Header=BB15_2 Depth=1
; CHECK-NEXT: ldr r1, [sp, #4] @ 4-byte Reload
; CHECK-NEXT: ldr r3, [sp, #68] @ 4-byte Reload
; CHECK-NEXT: cmp r1, r3
-; CHECK-NEXT: bne.w .LBB14_2
-; CHECK-NEXT: .LBB14_5: @ %for.cond.cleanup
+; CHECK-NEXT: bne.w .LBB15_2
+; CHECK-NEXT: .LBB15_5: @ %for.cond.cleanup
; CHECK-NEXT: add sp, #120
; CHECK-NEXT: vpop {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT: add sp, #4
; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc}
; CHECK-NEXT: .p2align 4
; CHECK-NEXT: @ %bb.6:
-; CHECK-NEXT: .LCPI14_0:
+; CHECK-NEXT: .LCPI15_0:
; CHECK-NEXT: .short 1 @ 0x1
; CHECK-NEXT: .short 4 @ 0x4
; CHECK-NEXT: .short 7 @ 0x7
@@ -862,7 +875,7 @@ define arm_aapcs_vfpcc void @gather_inc_v8i16_complex(ptr noalias nocapture read
; CHECK-NEXT: .short 16 @ 0x10
; CHECK-NEXT: .short 19 @ 0x13
; CHECK-NEXT: .short 22 @ 0x16
-; CHECK-NEXT: .LCPI14_1:
+; CHECK-NEXT: .LCPI15_1:
; CHECK-NEXT: .short 0 @ 0x0
; CHECK-NEXT: .short 3 @ 0x3
; CHECK-NEXT: .short 6 @ 0x6
@@ -871,7 +884,7 @@ define arm_aapcs_vfpcc void @gather_inc_v8i16_complex(ptr noalias nocapture read
; CHECK-NEXT: .short 15 @ 0xf
; CHECK-NEXT: .short 18 @ 0x12
; CHECK-NEXT: .short 21 @ 0x15
-; CHECK-NEXT: .LCPI14_2:
+; CHECK-NEXT: .LCPI15_2:
; CHECK-NEXT: .short 2 @ 0x2
; CHECK-NEXT: .short 5 @ 0x5
; CHECK-NEXT: .short 8 @ 0x8
@@ -933,21 +946,21 @@ define arm_aapcs_vfpcc void @gather_inc_v16i8_complex(ptr noalias nocapture read
; CHECK-NEXT: sub sp, #312
; CHECK-NEXT: cmp r2, #1
; CHECK-NEXT: str r1, [sp, #116] @ 4-byte Spill
-; CHECK-NEXT: blt.w .LBB15_5
+; CHECK-NEXT: blt.w .LBB16_5
; CHECK-NEXT: @ %bb.1: @ %vector.ph.preheader
-; CHECK-NEXT: adr r1, .LCPI15_0
-; CHECK-NEXT: adr r6, .LCPI15_8
+; CHECK-NEXT: adr r1, .LCPI16_0
+; CHECK-NEXT: adr r6, .LCPI16_8
; CHECK-NEXT: vldrw.u32 q0, [r1]
-; CHECK-NEXT: adr r1, .LCPI15_1
-; CHECK-NEXT: adr r7, .LCPI15_7
-; CHECK-NEXT: adr r3, .LCPI15_6
+; CHECK-NEXT: adr r1, .LCPI16_1
+; CHECK-NEXT: adr r7, .LCPI16_7
+; CHECK-NEXT: adr r3, .LCPI16_6
; CHECK-NEXT: vstrw.32 q0, [sp, #96] @ 16-byte Spill
; CHECK-NEXT: vldrw.u32 q0, [r1]
-; CHECK-NEXT: adr r1, .LCPI15_5
+; CHECK-NEXT: adr r1, .LCPI16_5
; CHECK-NEXT: bic r10, r2, #7
; CHECK-NEXT: vstrw.32 q0, [sp, #80] @ 16-byte Spill
; CHECK-NEXT: vldrw.u32 q0, [r6]
-; CHECK-NEXT: adr r6, .LCPI15_9
+; CHECK-NEXT: adr r6, .LCPI16_9
; CHECK-NEXT: vstrw.32 q0, [sp, #64] @ 16-byte Spill
; CHECK-NEXT: vldrw.u32 q0, [r7]
; CHECK-NEXT: vstrw.32 q0, [sp, #48] @ 16-byte Spill
@@ -957,25 +970,25 @@ define arm_aapcs_vfpcc void @gather_inc_v16i8_complex(ptr noalias nocapture read
; CHECK-NEXT: vstrw.32 q0, [sp, #16] @ 16-byte Spill
; CHECK-NEXT: vldrw.u32 q0, [r3]
; CHECK-NEXT: vstrw.32 q0, [sp] @ 16-byte Spill
-; CHECK-NEXT: .LBB15_...
[truncated]
``````````
</details>
https://github.com/llvm/llvm-project/pull/185437
More information about the llvm-commits
mailing list