caio/clang-p2996
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
890c4bece26e005cd9fa5511fe0efa7307794de5
clang-p2996/llvm/test/CodeGen/AArch64/aarch64-fold-lslfast.ll
Him188 e4a2d74e09 [AArch64][GISel] Always fold G_SHL into addressing mode where possible, unless the subtarget has addr-lsl-slow-14 (#96603) 
Before this patch, we fold G_SHL into addressing mode lsl only when
there is exactly one usage, or all the usages are memory ops, or we are
optimizing for size. However, lsl is free on all aarch64 targets except
those with FeatureAddrLSLSlow14.

This patch uses this fact and always folds G_SHL into lsl for memory
ops, with exceptions for FeatureAddrLSLSlow14.


This patch also fixes GISel 15% regression in TSVC kernel s482, and
brings regression in s291 from 20% to 10%.
2024-07-18 10:22:27 +01:00

451 lines
18 KiB
LLVM
Raw Blame History

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=aarch64-linux-gnu -mattr=+addr-lsl-slow-14 | FileCheck %s --check-prefixes=CHECK,CHECK0,CHECK0-SDAG
; RUN: llc < %s -mtriple=aarch64-linux-gnu -mattr=+addr-lsl-slow-14 -global-isel=1 -global-isel-abort=1 | FileCheck %s --check-prefixes=CHECK,CHECK0,CHECK0-GISEL
; RUN: llc < %s -mtriple=aarch64-linux-gnu | FileCheck %s --check-prefixes=CHECK,CHECK3,CHECK3-SDAG
; RUN: llc < %s -mtriple=aarch64-linux-gnu -global-isel=1 -global-isel-abort=1 | FileCheck %s --check-prefixes=CHECK,CHECK3,CHECK3-GISEL
%struct.a = type [256 x i16]
%struct.b = type [256 x i32]
%struct.c = type [256 x i64]
declare void @foo()
define i16 @halfword(ptr %ctx, i32 %xor72) nounwind {
; CHECK0-SDAG-LABEL: halfword:
; CHECK0-SDAG: // %bb.0:
; CHECK0-SDAG-NEXT: stp x30, x21, [sp, #-32]! // 16-byte Folded Spill
; CHECK0-SDAG-NEXT: // kill: def $w1 killed $w1 def $x1
; CHECK0-SDAG-NEXT: ubfx x8, x1, #9, #8
; CHECK0-SDAG-NEXT: stp x20, x19, [sp, #16] // 16-byte Folded Spill
; CHECK0-SDAG-NEXT: mov x19, x0
; CHECK0-SDAG-NEXT: lsl x21, x8, #1
; CHECK0-SDAG-NEXT: ldrh w20, [x0, x21]
; CHECK0-SDAG-NEXT: bl foo
; CHECK0-SDAG-NEXT: mov w0, w20
; CHECK0-SDAG-NEXT: strh w20, [x19, x21]
; CHECK0-SDAG-NEXT: ldp x20, x19, [sp, #16] // 16-byte Folded Reload
; CHECK0-SDAG-NEXT: ldp x30, x21, [sp], #32 // 16-byte Folded Reload
; CHECK0-SDAG-NEXT: ret
;
; CHECK0-GISEL-LABEL: halfword:
; CHECK0-GISEL: // %bb.0:
; CHECK0-GISEL-NEXT: str x30, [sp, #-32]! // 8-byte Folded Spill
; CHECK0-GISEL-NEXT: lsr w8, w1, #9
; CHECK0-GISEL-NEXT: stp x20, x19, [sp, #16] // 16-byte Folded Spill
; CHECK0-GISEL-NEXT: add x20, x0, w8, uxtb #1
; CHECK0-GISEL-NEXT: ldrh w19, [x20]
; CHECK0-GISEL-NEXT: bl foo
; CHECK0-GISEL-NEXT: mov w0, w19
; CHECK0-GISEL-NEXT: strh w19, [x20]
; CHECK0-GISEL-NEXT: ldp x20, x19, [sp, #16] // 16-byte Folded Reload
; CHECK0-GISEL-NEXT: ldr x30, [sp], #32 // 8-byte Folded Reload
; CHECK0-GISEL-NEXT: ret
;
; CHECK3-SDAG-LABEL: halfword:
; CHECK3-SDAG: // %bb.0:
; CHECK3-SDAG-NEXT: stp x30, x21, [sp, #-32]! // 16-byte Folded Spill
; CHECK3-SDAG-NEXT: // kill: def $w1 killed $w1 def $x1
; CHECK3-SDAG-NEXT: ubfx x21, x1, #9, #8
; CHECK3-SDAG-NEXT: stp x20, x19, [sp, #16] // 16-byte Folded Spill
; CHECK3-SDAG-NEXT: mov x19, x0
; CHECK3-SDAG-NEXT: ldrh w20, [x0, x21, lsl #1]
; CHECK3-SDAG-NEXT: bl foo
; CHECK3-SDAG-NEXT: mov w0, w20
; CHECK3-SDAG-NEXT: strh w20, [x19, x21, lsl #1]
; CHECK3-SDAG-NEXT: ldp x20, x19, [sp, #16] // 16-byte Folded Reload
; CHECK3-SDAG-NEXT: ldp x30, x21, [sp], #32 // 16-byte Folded Reload
; CHECK3-SDAG-NEXT: ret
;
; CHECK3-GISEL-LABEL: halfword:
; CHECK3-GISEL: // %bb.0:
; CHECK3-GISEL-NEXT: stp x30, x21, [sp, #-32]! // 16-byte Folded Spill
; CHECK3-GISEL-NEXT: lsr w8, w1, #9
; CHECK3-GISEL-NEXT: stp x20, x19, [sp, #16] // 16-byte Folded Spill
; CHECK3-GISEL-NEXT: mov x19, x0
; CHECK3-GISEL-NEXT: and x21, x8, #0xff
; CHECK3-GISEL-NEXT: ldrh w20, [x0, x21, lsl #1]
; CHECK3-GISEL-NEXT: bl foo
; CHECK3-GISEL-NEXT: mov w0, w20
; CHECK3-GISEL-NEXT: strh w20, [x19, x21, lsl #1]
; CHECK3-GISEL-NEXT: ldp x20, x19, [sp, #16] // 16-byte Folded Reload
; CHECK3-GISEL-NEXT: ldp x30, x21, [sp], #32 // 16-byte Folded Reload
; CHECK3-GISEL-NEXT: ret
%shr81 = lshr i32 %xor72, 9
%conv82 = zext i32 %shr81 to i64
%idxprom83 = and i64 %conv82, 255
%arrayidx86 = getelementptr inbounds %struct.a, ptr %ctx, i64 0, i64 %idxprom83
%result = load i16, ptr %arrayidx86, align 2
call void @foo()
store i16 %result, ptr %arrayidx86, align 2
ret i16 %result
}
define i32 @word(ptr %ctx, i32 %xor72) nounwind {
; CHECK0-SDAG-LABEL: word:
; CHECK0-SDAG: // %bb.0:
; CHECK0-SDAG-NEXT: stp x30, x21, [sp, #-32]! // 16-byte Folded Spill
; CHECK0-SDAG-NEXT: // kill: def $w1 killed $w1 def $x1
; CHECK0-SDAG-NEXT: ubfx x21, x1, #9, #8
; CHECK0-SDAG-NEXT: stp x20, x19, [sp, #16] // 16-byte Folded Spill
; CHECK0-SDAG-NEXT: mov x19, x0
; CHECK0-SDAG-NEXT: ldr w20, [x0, x21, lsl #2]
; CHECK0-SDAG-NEXT: bl foo
; CHECK0-SDAG-NEXT: mov w0, w20
; CHECK0-SDAG-NEXT: str w20, [x19, x21, lsl #2]
; CHECK0-SDAG-NEXT: ldp x20, x19, [sp, #16] // 16-byte Folded Reload
; CHECK0-SDAG-NEXT: ldp x30, x21, [sp], #32 // 16-byte Folded Reload
; CHECK0-SDAG-NEXT: ret
;
; CHECK0-GISEL-LABEL: word:
; CHECK0-GISEL: // %bb.0:
; CHECK0-GISEL-NEXT: stp x30, x21, [sp, #-32]! // 16-byte Folded Spill
; CHECK0-GISEL-NEXT: lsr w8, w1, #9
; CHECK0-GISEL-NEXT: stp x20, x19, [sp, #16] // 16-byte Folded Spill
; CHECK0-GISEL-NEXT: mov x19, x0
; CHECK0-GISEL-NEXT: and x21, x8, #0xff
; CHECK0-GISEL-NEXT: ldr w20, [x0, x21, lsl #2]
; CHECK0-GISEL-NEXT: bl foo
; CHECK0-GISEL-NEXT: mov w0, w20
; CHECK0-GISEL-NEXT: str w20, [x19, x21, lsl #2]
; CHECK0-GISEL-NEXT: ldp x20, x19, [sp, #16] // 16-byte Folded Reload
; CHECK0-GISEL-NEXT: ldp x30, x21, [sp], #32 // 16-byte Folded Reload
; CHECK0-GISEL-NEXT: ret
;
; CHECK3-SDAG-LABEL: word:
; CHECK3-SDAG: // %bb.0:
; CHECK3-SDAG-NEXT: stp x30, x21, [sp, #-32]! // 16-byte Folded Spill
; CHECK3-SDAG-NEXT: // kill: def $w1 killed $w1 def $x1
; CHECK3-SDAG-NEXT: ubfx x21, x1, #9, #8
; CHECK3-SDAG-NEXT: stp x20, x19, [sp, #16] // 16-byte Folded Spill
; CHECK3-SDAG-NEXT: mov x19, x0
; CHECK3-SDAG-NEXT: ldr w20, [x0, x21, lsl #2]
; CHECK3-SDAG-NEXT: bl foo
; CHECK3-SDAG-NEXT: mov w0, w20
; CHECK3-SDAG-NEXT: str w20, [x19, x21, lsl #2]
; CHECK3-SDAG-NEXT: ldp x20, x19, [sp, #16] // 16-byte Folded Reload
; CHECK3-SDAG-NEXT: ldp x30, x21, [sp], #32 // 16-byte Folded Reload
; CHECK3-SDAG-NEXT: ret
;
; CHECK3-GISEL-LABEL: word:
; CHECK3-GISEL: // %bb.0:
; CHECK3-GISEL-NEXT: stp x30, x21, [sp, #-32]! // 16-byte Folded Spill
; CHECK3-GISEL-NEXT: lsr w8, w1, #9
; CHECK3-GISEL-NEXT: stp x20, x19, [sp, #16] // 16-byte Folded Spill
; CHECK3-GISEL-NEXT: mov x19, x0
; CHECK3-GISEL-NEXT: and x21, x8, #0xff
; CHECK3-GISEL-NEXT: ldr w20, [x0, x21, lsl #2]
; CHECK3-GISEL-NEXT: bl foo
; CHECK3-GISEL-NEXT: mov w0, w20
; CHECK3-GISEL-NEXT: str w20, [x19, x21, lsl #2]
; CHECK3-GISEL-NEXT: ldp x20, x19, [sp, #16] // 16-byte Folded Reload
; CHECK3-GISEL-NEXT: ldp x30, x21, [sp], #32 // 16-byte Folded Reload
; CHECK3-GISEL-NEXT: ret
%shr81 = lshr i32 %xor72, 9
%conv82 = zext i32 %shr81 to i64
%idxprom83 = and i64 %conv82, 255
%arrayidx86 = getelementptr inbounds %struct.b, ptr %ctx, i64 0, i64 %idxprom83
%result = load i32, ptr %arrayidx86, align 4
call void @foo()
store i32 %result, ptr %arrayidx86, align 4
ret i32 %result
}
define i64 @doubleword(ptr %ctx, i32 %xor72) nounwind {
; CHECK0-SDAG-LABEL: doubleword:
; CHECK0-SDAG: // %bb.0:
; CHECK0-SDAG-NEXT: stp x30, x21, [sp, #-32]! // 16-byte Folded Spill
; CHECK0-SDAG-NEXT: // kill: def $w1 killed $w1 def $x1
; CHECK0-SDAG-NEXT: ubfx x21, x1, #9, #8
; CHECK0-SDAG-NEXT: stp x20, x19, [sp, #16] // 16-byte Folded Spill
; CHECK0-SDAG-NEXT: mov x19, x0
; CHECK0-SDAG-NEXT: ldr x20, [x0, x21, lsl #3]
; CHECK0-SDAG-NEXT: bl foo
; CHECK0-SDAG-NEXT: mov x0, x20
; CHECK0-SDAG-NEXT: str x20, [x19, x21, lsl #3]
; CHECK0-SDAG-NEXT: ldp x20, x19, [sp, #16] // 16-byte Folded Reload
; CHECK0-SDAG-NEXT: ldp x30, x21, [sp], #32 // 16-byte Folded Reload
; CHECK0-SDAG-NEXT: ret
;
; CHECK0-GISEL-LABEL: doubleword:
; CHECK0-GISEL: // %bb.0:
; CHECK0-GISEL-NEXT: stp x30, x21, [sp, #-32]! // 16-byte Folded Spill
; CHECK0-GISEL-NEXT: lsr w8, w1, #9
; CHECK0-GISEL-NEXT: stp x20, x19, [sp, #16] // 16-byte Folded Spill
; CHECK0-GISEL-NEXT: mov x19, x0
; CHECK0-GISEL-NEXT: and x21, x8, #0xff
; CHECK0-GISEL-NEXT: ldr x20, [x0, x21, lsl #3]
; CHECK0-GISEL-NEXT: bl foo
; CHECK0-GISEL-NEXT: mov x0, x20
; CHECK0-GISEL-NEXT: str x20, [x19, x21, lsl #3]
; CHECK0-GISEL-NEXT: ldp x20, x19, [sp, #16] // 16-byte Folded Reload
; CHECK0-GISEL-NEXT: ldp x30, x21, [sp], #32 // 16-byte Folded Reload
; CHECK0-GISEL-NEXT: ret
;
; CHECK3-SDAG-LABEL: doubleword:
; CHECK3-SDAG: // %bb.0:
; CHECK3-SDAG-NEXT: stp x30, x21, [sp, #-32]! // 16-byte Folded Spill
; CHECK3-SDAG-NEXT: // kill: def $w1 killed $w1 def $x1
; CHECK3-SDAG-NEXT: ubfx x21, x1, #9, #8
; CHECK3-SDAG-NEXT: stp x20, x19, [sp, #16] // 16-byte Folded Spill
; CHECK3-SDAG-NEXT: mov x19, x0
; CHECK3-SDAG-NEXT: ldr x20, [x0, x21, lsl #3]
; CHECK3-SDAG-NEXT: bl foo
; CHECK3-SDAG-NEXT: mov x0, x20
; CHECK3-SDAG-NEXT: str x20, [x19, x21, lsl #3]
; CHECK3-SDAG-NEXT: ldp x20, x19, [sp, #16] // 16-byte Folded Reload
; CHECK3-SDAG-NEXT: ldp x30, x21, [sp], #32 // 16-byte Folded Reload
; CHECK3-SDAG-NEXT: ret
;
; CHECK3-GISEL-LABEL: doubleword:
; CHECK3-GISEL: // %bb.0:
; CHECK3-GISEL-NEXT: stp x30, x21, [sp, #-32]! // 16-byte Folded Spill
; CHECK3-GISEL-NEXT: lsr w8, w1, #9
; CHECK3-GISEL-NEXT: stp x20, x19, [sp, #16] // 16-byte Folded Spill
; CHECK3-GISEL-NEXT: mov x19, x0
; CHECK3-GISEL-NEXT: and x21, x8, #0xff
; CHECK3-GISEL-NEXT: ldr x20, [x0, x21, lsl #3]
; CHECK3-GISEL-NEXT: bl foo
; CHECK3-GISEL-NEXT: mov x0, x20
; CHECK3-GISEL-NEXT: str x20, [x19, x21, lsl #3]
; CHECK3-GISEL-NEXT: ldp x20, x19, [sp, #16] // 16-byte Folded Reload
; CHECK3-GISEL-NEXT: ldp x30, x21, [sp], #32 // 16-byte Folded Reload
; CHECK3-GISEL-NEXT: ret
%shr81 = lshr i32 %xor72, 9
%conv82 = zext i32 %shr81 to i64
%idxprom83 = and i64 %conv82, 255
%arrayidx86 = getelementptr inbounds %struct.c, ptr %ctx, i64 0, i64 %idxprom83
%result = load i64, ptr %arrayidx86, align 8
call void @foo()
store i64 %result, ptr %arrayidx86, align 8
ret i64 %result
}
define i16 @multi_use_half_word(ptr %ctx, i32 %xor72) {
; CHECK0-SDAG-LABEL: multi_use_half_word:
; CHECK0-SDAG: // %bb.0: // %entry
; CHECK0-SDAG-NEXT: str x30, [sp, #-48]! // 8-byte Folded Spill
; CHECK0-SDAG-NEXT: stp x22, x21, [sp, #16] // 16-byte Folded Spill
; CHECK0-SDAG-NEXT: stp x20, x19, [sp, #32] // 16-byte Folded Spill
; CHECK0-SDAG-NEXT: .cfi_def_cfa_offset 48
; CHECK0-SDAG-NEXT: .cfi_offset w19, -8
; CHECK0-SDAG-NEXT: .cfi_offset w20, -16
; CHECK0-SDAG-NEXT: .cfi_offset w21, -24
; CHECK0-SDAG-NEXT: .cfi_offset w22, -32
; CHECK0-SDAG-NEXT: .cfi_offset w30, -48
; CHECK0-SDAG-NEXT: // kill: def $w1 killed $w1 def $x1
; CHECK0-SDAG-NEXT: ubfx x8, x1, #9, #8
; CHECK0-SDAG-NEXT: mov x19, x0
; CHECK0-SDAG-NEXT: lsl x21, x8, #1
; CHECK0-SDAG-NEXT: ldrh w20, [x0, x21]
; CHECK0-SDAG-NEXT: add w22, w20, #1
; CHECK0-SDAG-NEXT: bl foo
; CHECK0-SDAG-NEXT: mov w0, w20
; CHECK0-SDAG-NEXT: strh w22, [x19, x21]
; CHECK0-SDAG-NEXT: ldp x20, x19, [sp, #32] // 16-byte Folded Reload
; CHECK0-SDAG-NEXT: ldp x22, x21, [sp, #16] // 16-byte Folded Reload
; CHECK0-SDAG-NEXT: ldr x30, [sp], #48 // 8-byte Folded Reload
; CHECK0-SDAG-NEXT: ret
;
; CHECK0-GISEL-LABEL: multi_use_half_word:
; CHECK0-GISEL: // %bb.0: // %entry
; CHECK0-GISEL-NEXT: stp x30, x21, [sp, #-32]! // 16-byte Folded Spill
; CHECK0-GISEL-NEXT: stp x20, x19, [sp, #16] // 16-byte Folded Spill
; CHECK0-GISEL-NEXT: .cfi_def_cfa_offset 32
; CHECK0-GISEL-NEXT: .cfi_offset w19, -8
; CHECK0-GISEL-NEXT: .cfi_offset w20, -16
; CHECK0-GISEL-NEXT: .cfi_offset w21, -24
; CHECK0-GISEL-NEXT: .cfi_offset w30, -32
; CHECK0-GISEL-NEXT: lsr w8, w1, #9
; CHECK0-GISEL-NEXT: add x20, x0, w8, uxtb #1
; CHECK0-GISEL-NEXT: ldrh w19, [x20]
; CHECK0-GISEL-NEXT: add w21, w19, #1
; CHECK0-GISEL-NEXT: bl foo
; CHECK0-GISEL-NEXT: strh w19, [x20]
; CHECK0-GISEL-NEXT: mov w0, w19
; CHECK0-GISEL-NEXT: strh w21, [x20]
; CHECK0-GISEL-NEXT: ldp x20, x19, [sp, #16] // 16-byte Folded Reload
; CHECK0-GISEL-NEXT: ldp x30, x21, [sp], #32 // 16-byte Folded Reload
; CHECK0-GISEL-NEXT: ret
;
; CHECK3-SDAG-LABEL: multi_use_half_word:
; CHECK3-SDAG: // %bb.0: // %entry
; CHECK3-SDAG-NEXT: str x30, [sp, #-48]! // 8-byte Folded Spill
; CHECK3-SDAG-NEXT: stp x22, x21, [sp, #16] // 16-byte Folded Spill
; CHECK3-SDAG-NEXT: stp x20, x19, [sp, #32] // 16-byte Folded Spill
; CHECK3-SDAG-NEXT: .cfi_def_cfa_offset 48
; CHECK3-SDAG-NEXT: .cfi_offset w19, -8
; CHECK3-SDAG-NEXT: .cfi_offset w20, -16
; CHECK3-SDAG-NEXT: .cfi_offset w21, -24
; CHECK3-SDAG-NEXT: .cfi_offset w22, -32
; CHECK3-SDAG-NEXT: .cfi_offset w30, -48
; CHECK3-SDAG-NEXT: // kill: def $w1 killed $w1 def $x1
; CHECK3-SDAG-NEXT: ubfx x21, x1, #9, #8
; CHECK3-SDAG-NEXT: mov x19, x0
; CHECK3-SDAG-NEXT: ldrh w20, [x0, x21, lsl #1]
; CHECK3-SDAG-NEXT: add w22, w20, #1
; CHECK3-SDAG-NEXT: bl foo
; CHECK3-SDAG-NEXT: mov w0, w20
; CHECK3-SDAG-NEXT: strh w22, [x19, x21, lsl #1]
; CHECK3-SDAG-NEXT: ldp x20, x19, [sp, #32] // 16-byte Folded Reload
; CHECK3-SDAG-NEXT: ldp x22, x21, [sp, #16] // 16-byte Folded Reload
; CHECK3-SDAG-NEXT: ldr x30, [sp], #48 // 8-byte Folded Reload
; CHECK3-SDAG-NEXT: ret
;
; CHECK3-GISEL-LABEL: multi_use_half_word:
; CHECK3-GISEL: // %bb.0: // %entry
; CHECK3-GISEL-NEXT: str x30, [sp, #-48]! // 8-byte Folded Spill
; CHECK3-GISEL-NEXT: stp x22, x21, [sp, #16] // 16-byte Folded Spill
; CHECK3-GISEL-NEXT: stp x20, x19, [sp, #32] // 16-byte Folded Spill
; CHECK3-GISEL-NEXT: .cfi_def_cfa_offset 48
; CHECK3-GISEL-NEXT: .cfi_offset w19, -8
; CHECK3-GISEL-NEXT: .cfi_offset w20, -16
; CHECK3-GISEL-NEXT: .cfi_offset w21, -24
; CHECK3-GISEL-NEXT: .cfi_offset w22, -32
; CHECK3-GISEL-NEXT: .cfi_offset w30, -48
; CHECK3-GISEL-NEXT: lsr w8, w1, #9
; CHECK3-GISEL-NEXT: mov x19, x0
; CHECK3-GISEL-NEXT: and x21, x8, #0xff
; CHECK3-GISEL-NEXT: ldrh w20, [x0, x21, lsl #1]
; CHECK3-GISEL-NEXT: add w22, w20, #1
; CHECK3-GISEL-NEXT: bl foo
; CHECK3-GISEL-NEXT: strh w20, [x19, x21, lsl #1]
; CHECK3-GISEL-NEXT: mov w0, w20
; CHECK3-GISEL-NEXT: strh w22, [x19, x21, lsl #1]
; CHECK3-GISEL-NEXT: ldp x20, x19, [sp, #32] // 16-byte Folded Reload
; CHECK3-GISEL-NEXT: ldp x22, x21, [sp, #16] // 16-byte Folded Reload
; CHECK3-GISEL-NEXT: ldr x30, [sp], #48 // 8-byte Folded Reload
; CHECK3-GISEL-NEXT: ret
entry:
%shr81 = lshr i32 %xor72, 9
%conv82 = zext i32 %shr81 to i64
%idxprom83 = and i64 %conv82, 255
%arrayidx86 = getelementptr inbounds %struct.a, ptr %ctx, i64 0, i64 %idxprom83
%result = load i16, ptr %arrayidx86, align 2
%result2 = add i16 %result, 1
call void @foo()
store i16 %result, ptr %arrayidx86, align 2
store i16 %result2, ptr %arrayidx86, align 2
ret i16 %result
}
define i64 @multi_use_non_memory_call(i64 %a, i64 %b) {
; CHECK-LABEL: multi_use_non_memory_call:
; CHECK: // %bb.0: // %entry
; CHECK-NEXT: lsl x8, x0, #3
; CHECK-NEXT: lsl x9, x1, #3
; CHECK-NEXT: cmp x8, x9
; CHECK-NEXT: b.lt .LBB4_2
; CHECK-NEXT: // %bb.1: // %falsebb
; CHECK-NEXT: csel x0, x8, x9, gt
; CHECK-NEXT: ret
; CHECK-NEXT: .LBB4_2: // %truebb
; CHECK-NEXT: str x30, [sp, #-16]! // 8-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 16
; CHECK-NEXT: .cfi_offset w30, -16
; CHECK-NEXT: bl foo
entry:
%mul1 = shl i64 %a, 3
%mul2 = shl i64 %b, 3
%cmp = icmp slt i64 %mul1, %mul2
br i1 %cmp, label %truebb, label %falsebb
truebb:
tail call void @foo()
unreachable
falsebb:
%cmp2 = icmp sgt i64 %mul1, %mul2
br i1 %cmp2, label %exitbb, label %endbb
exitbb:
ret i64 %mul1
endbb:
ret i64 %mul2
}
define i64 @gep3(ptr %p, i64 %b) {
; CHECK-LABEL: gep3:
; CHECK: // %bb.0:
; CHECK-NEXT: mov x8, x0
; CHECK-NEXT: ldr x0, [x0, x1, lsl #3]
; CHECK-NEXT: str x1, [x8, x1, lsl #3]
; CHECK-NEXT: ret
%g = getelementptr inbounds i64, ptr %p, i64 %b
%l = load i64, ptr %g
store i64 %b, ptr %g
ret i64 %l
}
define i128 @gep4(ptr %p, i128 %a, i64 %b) {
; CHECK0-SDAG-LABEL: gep4:
; CHECK0-SDAG: // %bb.0:
; CHECK0-SDAG-NEXT: add x8, x0, x4, lsl #4
; CHECK0-SDAG-NEXT: ldp x0, x1, [x8]
; CHECK0-SDAG-NEXT: stp x2, x3, [x8]
; CHECK0-SDAG-NEXT: ret
;
; CHECK0-GISEL-LABEL: gep4:
; CHECK0-GISEL: // %bb.0:
; CHECK0-GISEL-NEXT: add x8, x0, x4, lsl #4
; CHECK0-GISEL-NEXT: mov v0.d[0], x2
; CHECK0-GISEL-NEXT: ldr q1, [x8]
; CHECK0-GISEL-NEXT: mov d2, v1.d[1]
; CHECK0-GISEL-NEXT: mov v0.d[1], x3
; CHECK0-GISEL-NEXT: fmov x0, d1
; CHECK0-GISEL-NEXT: fmov x1, d2
; CHECK0-GISEL-NEXT: str q0, [x8]
; CHECK0-GISEL-NEXT: ret
;
; CHECK3-SDAG-LABEL: gep4:
; CHECK3-SDAG: // %bb.0:
; CHECK3-SDAG-NEXT: add x8, x0, x4, lsl #4
; CHECK3-SDAG-NEXT: ldp x0, x1, [x8]
; CHECK3-SDAG-NEXT: stp x2, x3, [x8]
; CHECK3-SDAG-NEXT: ret
;
; CHECK3-GISEL-LABEL: gep4:
; CHECK3-GISEL: // %bb.0:
; CHECK3-GISEL-NEXT: ldr q1, [x0, x4, lsl #4]
; CHECK3-GISEL-NEXT: mov v0.d[0], x2
; CHECK3-GISEL-NEXT: mov x8, x0
; CHECK3-GISEL-NEXT: mov d2, v1.d[1]
; CHECK3-GISEL-NEXT: fmov x0, d1
; CHECK3-GISEL-NEXT: mov v0.d[1], x3
; CHECK3-GISEL-NEXT: fmov x1, d2
; CHECK3-GISEL-NEXT: str q0, [x8, x4, lsl #4]
; CHECK3-GISEL-NEXT: ret
%g = getelementptr inbounds i128, ptr %p, i64 %b
%l = load i128, ptr %g
store i128 %a, ptr %g
ret i128 %l
}
define i64 @addlsl3(i64 %a, i64 %b) {
; CHECK-LABEL: addlsl3:
; CHECK: // %bb.0:
; CHECK-NEXT: lsl x8, x0, #3
; CHECK-NEXT: add x9, x1, x8
; CHECK-NEXT: sub x8, x1, x8
; CHECK-NEXT: eor x0, x9, x8
; CHECK-NEXT: ret
%x = shl i64 %a, 3
%y = add i64 %b, %x
%z = sub i64 %b, %x
%r = xor i64 %y, %z
ret i64 %r
}
define i64 @addlsl4(i64 %a, i64 %b) {
; CHECK-LABEL: addlsl4:
; CHECK: // %bb.0:
; CHECK-NEXT: lsl x8, x0, #4
; CHECK-NEXT: add x9, x1, x8
; CHECK-NEXT: sub x8, x1, x8
; CHECK-NEXT: eor x0, x9, x8
; CHECK-NEXT: ret
%x = shl i64 %a, 4
%y = add i64 %b, %x
%z = sub i64 %b, %x
%r = xor i64 %y, %z
ret i64 %r
}
;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
; CHECK0: {{.*}}
; CHECK3: {{.*}}
Reference in New Issue View Git Blame Copy Permalink