[llvm] 6ccc9e5 - [AArch64] Add taildup test with computed gotos.
Florian Hahn via llvm-commits
llvm-commits at lists.llvm.org
Mon Jul 28 05:27:20 PDT 2025
Author: Florian Hahn
Date: 2025-07-28T13:26:51+01:00
New Revision: 6ccc9e559da8d0f07e496d375dbc02bc441e60d9
URL: https://github.com/llvm/llvm-project/commit/6ccc9e559da8d0f07e496d375dbc02bc441e60d9
DIFF: https://github.com/llvm/llvm-project/commit/6ccc9e559da8d0f07e496d375dbc02bc441e60d9.diff
LOG: [AArch64] Add taildup test with computed gotos.
Add a test case showing missed optimizations from early taildup with
computed gotos for https://github.com/llvm/llvm-project/pull/150911.
Added:
llvm/test/CodeGen/AArch64/late-taildup-computed-goto.ll
Modified:
Removed:
################################################################################
diff --git a/llvm/test/CodeGen/AArch64/late-taildup-computed-goto.ll b/llvm/test/CodeGen/AArch64/late-taildup-computed-goto.ll
new file mode 100644
index 0000000000000..c4a027c6d8070
--- /dev/null
+++ b/llvm/test/CodeGen/AArch64/late-taildup-computed-goto.ll
@@ -0,0 +1,162 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
+; RUN: llc -tail-dup-pred-size=2 -tail-dup-succ-size=2 -o - %s | FileCheck %s
+
+target triple = "arm64-apple-macosx13.0.0"
+
+ at opcode.targets = local_unnamed_addr constant [6 x ptr] [ptr blockaddress(@test_interp, %op1.bb), ptr blockaddress(@test_interp, %op6.bb), ptr blockaddress(@test_interp, %loop.header), ptr blockaddress(@test_interp, %op2.bb), ptr blockaddress(@test_interp, %op4.bb), ptr blockaddress(@test_interp, %op5.bb)]
+
+define void @test_interp(ptr %frame, ptr %dst) {
+; CHECK-LABEL: test_interp:
+; CHECK: ; %bb.0: ; %entry
+; CHECK-NEXT: stp x24, x23, [sp, #-64]! ; 16-byte Folded Spill
+; CHECK-NEXT: stp x22, x21, [sp, #16] ; 16-byte Folded Spill
+; CHECK-NEXT: stp x20, x19, [sp, #32] ; 16-byte Folded Spill
+; CHECK-NEXT: stp x29, x30, [sp, #48] ; 16-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 64
+; CHECK-NEXT: .cfi_offset w30, -8
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: .cfi_offset w19, -24
+; CHECK-NEXT: .cfi_offset w20, -32
+; CHECK-NEXT: .cfi_offset w21, -40
+; CHECK-NEXT: .cfi_offset w22, -48
+; CHECK-NEXT: .cfi_offset w23, -56
+; CHECK-NEXT: .cfi_offset w24, -64
+; CHECK-NEXT: Lloh0:
+; CHECK-NEXT: adrp x21, _opcode.targets at PAGE
+; CHECK-NEXT: Lloh1:
+; CHECK-NEXT: add x21, x21, _opcode.targets at PAGEOFF
+; CHECK-NEXT: mov x22, xzr
+; CHECK-NEXT: add x8, x21, xzr, lsl #3
+; CHECK-NEXT: mov x19, x1
+; CHECK-NEXT: mov x20, x0
+; CHECK-NEXT: add x23, x22, #1
+; CHECK-NEXT: br x8
+; CHECK-NEXT: Ltmp0: ; Block address taken
+; CHECK-NEXT: LBB0_1: ; %loop.header
+; CHECK-NEXT: ; =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: add x8, x21, x23, lsl #3
+; CHECK-NEXT: mov x20, xzr
+; CHECK-NEXT: mov x22, xzr
+; CHECK-NEXT: add x23, x23, #1
+; CHECK-NEXT: br x8
+; CHECK-NEXT: Ltmp1: ; Block address taken
+; CHECK-NEXT: LBB0_2: ; %op1.bb
+; CHECK-NEXT: ; =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: str xzr, [x19]
+; CHECK-NEXT: mov w8, #1 ; =0x1
+; CHECK-NEXT: ldr x0, [x20, #-8]!
+; CHECK-NEXT: ldr x9, [x0, #8]
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: ldr x8, [x9, #48]
+; CHECK-NEXT: blr x8
+; CHECK-NEXT: add x8, x21, x23, lsl #3
+; CHECK-NEXT: add x23, x23, #1
+; CHECK-NEXT: br x8
+; CHECK-NEXT: Ltmp2: ; Block address taken
+; CHECK-NEXT: LBB0_3: ; %op2.bb
+; CHECK-NEXT: ; =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: add x8, x21, x23, lsl #3
+; CHECK-NEXT: mov x20, xzr
+; CHECK-NEXT: add x23, x23, #1
+; CHECK-NEXT: str x22, [x19]
+; CHECK-NEXT: mov x22, xzr
+; CHECK-NEXT: br x8
+; CHECK-NEXT: Ltmp3: ; Block address taken
+; CHECK-NEXT: LBB0_4: ; %op4.bb
+; CHECK-NEXT: ; =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: str x22, [x19]
+; CHECK-NEXT: add x10, x21, x23, lsl #3
+; CHECK-NEXT: add x23, x23, #1
+; CHECK-NEXT: ldur x8, [x22, #12]
+; CHECK-NEXT: ldur x9, [x20, #-8]
+; CHECK-NEXT: add x22, x22, #20
+; CHECK-NEXT: stp x8, x9, [x20, #-8]
+; CHECK-NEXT: add x20, x20, #8
+; CHECK-NEXT: br x10
+; CHECK-NEXT: Ltmp4: ; Block address taken
+; CHECK-NEXT: LBB0_5: ; %op5.bb
+; CHECK-NEXT: ; =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: str x22, [x19]
+; CHECK-NEXT: add x10, x21, x23, lsl #3
+; CHECK-NEXT: add x23, x23, #1
+; CHECK-NEXT: ldur x8, [x22, #12]
+; CHECK-NEXT: ldur x9, [x20, #-8]
+; CHECK-NEXT: add x22, x22, #20
+; CHECK-NEXT: stp x8, x9, [x20, #-8]
+; CHECK-NEXT: add x20, x20, #8
+; CHECK-NEXT: br x10
+; CHECK-NEXT: Ltmp5: ; Block address taken
+; CHECK-NEXT: LBB0_6: ; %op6.bb
+; CHECK-NEXT: ; =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: ldr x0, [x20, #-8]!
+; CHECK-NEXT: mov w8, #1 ; =0x1
+; CHECK-NEXT: ldr x9, [x0, #8]
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: ldr x8, [x9, #48]
+; CHECK-NEXT: blr x8
+; CHECK-NEXT: add x8, x21, x23, lsl #3
+; CHECK-NEXT: add x23, x23, #1
+; CHECK-NEXT: br x8
+; CHECK-NEXT: .loh AdrpAdd Lloh0, Lloh1
+entry:
+ br label %loop.header
+
+loop.header:
+ %iv = phi i64 [ 0, %entry ], [ %iv.next, %op1.bb ], [ %iv.next, %op2.bb ], [ %iv.next, %op4.bb ], [ %iv.next, %op5.bb ], [ %iv.next, %op6.bb ], [ %iv.next, %loop.header ]
+ %stack.pointer = phi ptr [ %frame, %entry ], [ %stack.8, %op1.bb ], [ null, %op2.bb ], [ %stack.next, %op4.bb ], [ %stack.next.2, %op5.bb ], [ %stack.4, %op6.bb ], [ null, %loop.header ]
+ %next.instr = phi ptr [ null, %entry ], [ %next.instr, %op1.bb ], [ null, %op2.bb ], [ %next.instr.20, %op4.bb ], [ %next.instr.21, %op5.bb ], [ %next.instr, %op6.bb ], [ null, %loop.header ]
+ %iv.next = add i64 %iv, 1
+ %next_op = getelementptr [6 x ptr], ptr @opcode.targets, i64 0, i64 %iv
+ indirectbr ptr %next_op, [label %op1.bb, label %op6.bb, label %loop.header, label %op2.bb, label %op4.bb, label %op5.bb]
+
+op1.bb:
+ store ptr null, ptr %dst, align 8
+ %stack.8 = getelementptr i8, ptr %stack.pointer, i64 -8
+ %l.0 = load ptr, ptr %stack.8, align 8
+ store i64 1, ptr %l.0, align 8
+ %gep.0 = getelementptr i8, ptr %l.0, i64 8
+ %l.1 = load ptr, ptr %gep.0, align 8
+ %gep.1 = getelementptr i8, ptr %l.1, i64 48
+ %l.2 = load ptr, ptr %gep.1, align 8
+ tail call void %l.2(ptr nonnull %l.0)
+ br label %loop.header
+
+op2.bb:
+ store ptr %next.instr, ptr %dst, align 8
+ br label %loop.header
+
+op4.bb:
+ store ptr %next.instr, ptr %dst, align 8
+ %next.instr.20 = getelementptr i8, ptr %next.instr, i64 20
+ %stack.2 = getelementptr i8, ptr %stack.pointer, i64 -8
+ %l.3 = load ptr, ptr %stack.2, align 8
+ %next.instr.12 = getelementptr i8, ptr %next.instr, i64 12
+ %next.instr.12.val = load ptr, ptr %next.instr.12, align 2
+ store ptr %next.instr.12.val, ptr %stack.2, align 8
+ store ptr %l.3, ptr %stack.pointer, align 8
+ %stack.next = getelementptr i8, ptr %stack.pointer, i64 8
+ br label %loop.header
+
+op5.bb:
+ store ptr %next.instr, ptr %dst, align 8
+ %next.instr.21 = getelementptr i8, ptr %next.instr, i64 20
+ %stack.3 = getelementptr i8, ptr %stack.pointer, i64 -8
+ %l.4 = load ptr, ptr %stack.3, align 8
+ %next.instr.2 = getelementptr i8, ptr %next.instr, i64 12
+ %next.instr.2.val = load ptr, ptr %next.instr.2, align 2
+ store ptr %next.instr.2.val, ptr %stack.3, align 8
+ store ptr %l.4, ptr %stack.pointer, align 8
+ %stack.next.2 = getelementptr i8, ptr %stack.pointer, i64 8
+ br label %loop.header
+
+op6.bb:
+ %stack.4 = getelementptr i8, ptr %stack.pointer, i64 -8
+ %l.5 = load ptr, ptr %stack.4, align 8
+ store i64 1, ptr %l.5, align 8
+ %gep.5 = getelementptr i8, ptr %l.5, i64 8
+ %l.6 = load ptr, ptr %gep.5, align 8
+ %gep.6 = getelementptr i8, ptr %l.6, i64 48
+ %l.7 = load ptr, ptr %gep.6, align 8
+ tail call void %l.7(ptr nonnull %l.5)
+ br label %loop.header
+}
More information about the llvm-commits
mailing list