ef84452571
Previously we recursively looked through extends and truncates on both SourceValue and WideVal. SourceValue is the largest source found for each of the stores we are combining. WideVal is the source for the current store. Previously we could incorrectly look through a (zext (trunc X)) pair and incorrectly believe X to be a good source. I think we could also look through a zext on one store and a sext on another store and arbitrarily pick one of the extends as the final source. With this patch we only look through one level of extend or truncate. And we don't look through extends/truncs on both SourceValue and WideVal at the same time. This may lose some optimization cases, but keeps everything we had tests for. Fixes #90936.
50 lines
1.4 KiB
LLVM
50 lines
1.4 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 4
|
|
; RUN: llc < %s -mtriple=aarch64 | FileCheck %s
|
|
|
|
define void @f(i16 %arg, ptr %arg1) {
|
|
; CHECK-LABEL: f:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: ubfx w8, w0, #8, #6
|
|
; CHECK-NEXT: strb w0, [x1]
|
|
; CHECK-NEXT: strb w8, [x1, #1]
|
|
; CHECK-NEXT: ret
|
|
bb:
|
|
%i = trunc i16 %arg to i8
|
|
%i2 = trunc i16 %arg to i14
|
|
%i3 = lshr i14 %i2, 8
|
|
store i8 %i, ptr %arg1, align 1
|
|
%i4 = getelementptr i8, ptr %arg1, i64 1
|
|
%i5 = trunc i14 %i3 to i8
|
|
store i8 %i5, ptr %i4, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @g(i32 %arg, ptr %arg1) {
|
|
; CHECK-LABEL: g:
|
|
; CHECK: // %bb.0: // %bb
|
|
; CHECK-NEXT: lsr w8, w0, #8
|
|
; CHECK-NEXT: lsr w9, w0, #16
|
|
; CHECK-NEXT: strb w0, [x1]
|
|
; CHECK-NEXT: strb wzr, [x1, #3]
|
|
; CHECK-NEXT: strb w8, [x1, #1]
|
|
; CHECK-NEXT: strb w9, [x1, #2]
|
|
; CHECK-NEXT: ret
|
|
bb:
|
|
%i = trunc i32 %arg to i8
|
|
store i8 %i, ptr %arg1, align 1
|
|
%i2 = lshr i32 %arg, 8
|
|
%i3 = trunc i32 %i2 to i8
|
|
%i4 = getelementptr i8, ptr %arg1, i64 1
|
|
store i8 %i3, ptr %i4, align 1
|
|
%i5 = lshr i32 %arg, 16
|
|
%i6 = trunc i32 %i5 to i8
|
|
%i7 = getelementptr i8, ptr %arg1, i64 2
|
|
store i8 %i6, ptr %i7, align 1
|
|
%i8 = zext i8 %i to i32
|
|
%i9 = lshr i32 %i8, 24
|
|
%i10 = getelementptr i8, ptr %arg1, i64 3
|
|
%i11 = trunc i32 %i9 to i8
|
|
store i8 %i11, ptr %i10, align 1
|
|
ret void
|
|
}
|