derek.holloway
7502bf64a4
CodeQL / Analyze (actions) (push) Canceled after 0s
CodeQL / Analyze (python) (push) Canceled after 0s
pre-commit / pre-commit (push) Canceled after 0s
Multi-Arch CI / setup (push) Canceled after 0s
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Setup (Multi-Arch) / setup (push) Canceled after 0s
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Multi-Arch CI / CI Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
105 lines
4.8 KiB
LLVM
105 lines
4.8 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
|
|
; RUN: llc < %s -mtriple arm64-apple-darwin -global-isel -stop-after=irtranslator -verify-machineinstrs | FileCheck %s
|
|
|
|
; Check that we don't try to tail-call with a non-forwarded sret parameter.
|
|
declare void @test_explicit_sret(ptr sret(i64))
|
|
|
|
; Forwarded explicit sret pointer => we can tail call.
|
|
define void @can_tail_call_forwarded_explicit_sret_ptr(ptr sret(i64) %arg) {
|
|
; CHECK-LABEL: name: can_tail_call_forwarded_explicit_sret_ptr
|
|
; CHECK: bb.1 (%ir-block.0):
|
|
; CHECK-NEXT: liveins: $x8
|
|
; CHECK-NEXT: {{ $}}
|
|
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(p0) = COPY $x8
|
|
; CHECK-NEXT: $x8 = COPY [[COPY]](p0)
|
|
; CHECK-NEXT: TCRETURNdi @test_explicit_sret, 0, csr_darwin_aarch64_aapcs, implicit $sp, implicit $x8
|
|
tail call void @test_explicit_sret(ptr %arg)
|
|
ret void
|
|
}
|
|
|
|
; Not marked as tail, so don't tail call.
|
|
define void @test_call_explicit_sret(ptr sret(i64) %arg) {
|
|
; CHECK-LABEL: name: test_call_explicit_sret
|
|
; CHECK: bb.1 (%ir-block.0):
|
|
; CHECK-NEXT: liveins: $x8
|
|
; CHECK-NEXT: {{ $}}
|
|
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(p0) = COPY $x8
|
|
; CHECK-NEXT: ADJCALLSTACKDOWN 0, 0, implicit-def $sp, implicit $sp
|
|
; CHECK-NEXT: $x8 = COPY [[COPY]](p0)
|
|
; CHECK-NEXT: BL @test_explicit_sret, csr_darwin_aarch64_aapcs, implicit-def $lr, implicit $sp, implicit $x8
|
|
; CHECK-NEXT: ADJCALLSTACKUP 0, 0, implicit-def $sp, implicit $sp
|
|
; CHECK-NEXT: RET_ReallyLR
|
|
call void @test_explicit_sret(ptr %arg)
|
|
ret void
|
|
}
|
|
|
|
define void @dont_tail_call_explicit_sret_alloca_unused() {
|
|
; CHECK-LABEL: name: dont_tail_call_explicit_sret_alloca_unused
|
|
; CHECK: bb.1 (%ir-block.0):
|
|
; CHECK-NEXT: [[FRAME_INDEX:%[0-9]+]]:_(p0) = G_FRAME_INDEX %stack.0.l
|
|
; CHECK-NEXT: ADJCALLSTACKDOWN 0, 0, implicit-def $sp, implicit $sp
|
|
; CHECK-NEXT: $x8 = COPY [[FRAME_INDEX]](p0)
|
|
; CHECK-NEXT: BL @test_explicit_sret, csr_darwin_aarch64_aapcs, implicit-def $lr, implicit $sp, implicit $x8
|
|
; CHECK-NEXT: ADJCALLSTACKUP 0, 0, implicit-def $sp, implicit $sp
|
|
; CHECK-NEXT: RET_ReallyLR
|
|
%l = alloca i64, align 8
|
|
tail call void @test_explicit_sret(ptr %l)
|
|
ret void
|
|
}
|
|
|
|
define void @dont_tail_call_explicit_sret_alloca_dummyusers(ptr %ptr) {
|
|
; CHECK-LABEL: name: dont_tail_call_explicit_sret_alloca_dummyusers
|
|
; CHECK: bb.1 (%ir-block.0):
|
|
; CHECK-NEXT: liveins: $x0
|
|
; CHECK-NEXT: {{ $}}
|
|
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(p0) = COPY $x0
|
|
; CHECK-NEXT: [[FRAME_INDEX:%[0-9]+]]:_(p0) = G_FRAME_INDEX %stack.0.l
|
|
; CHECK-NEXT: [[LOAD:%[0-9]+]]:_(i64) = G_LOAD [[COPY]](p0) :: (load (i64) from %ir.ptr)
|
|
; CHECK-NEXT: G_STORE [[LOAD]](i64), [[FRAME_INDEX]](p0) :: (store (i64) into %ir.l)
|
|
; CHECK-NEXT: ADJCALLSTACKDOWN 0, 0, implicit-def $sp, implicit $sp
|
|
; CHECK-NEXT: $x8 = COPY [[FRAME_INDEX]](p0)
|
|
; CHECK-NEXT: BL @test_explicit_sret, csr_darwin_aarch64_aapcs, implicit-def $lr, implicit $sp, implicit $x8
|
|
; CHECK-NEXT: ADJCALLSTACKUP 0, 0, implicit-def $sp, implicit $sp
|
|
; CHECK-NEXT: RET_ReallyLR
|
|
%l = alloca i64, align 8
|
|
%r = load i64, ptr %ptr, align 8
|
|
store i64 %r, ptr %l, align 8
|
|
tail call void @test_explicit_sret(ptr %l)
|
|
ret void
|
|
}
|
|
|
|
define void @dont_tail_call_tailcall_explicit_sret_gep(ptr %ptr) {
|
|
; CHECK-LABEL: name: dont_tail_call_tailcall_explicit_sret_gep
|
|
; CHECK: bb.1 (%ir-block.0):
|
|
; CHECK-NEXT: liveins: $x0
|
|
; CHECK-NEXT: {{ $}}
|
|
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(p0) = COPY $x0
|
|
; CHECK-NEXT: [[C:%[0-9]+]]:_(i64) = G_CONSTANT i64 8
|
|
; CHECK-NEXT: [[PTR_ADD:%[0-9]+]]:_(p0) = G_PTR_ADD [[COPY]], [[C]](i64)
|
|
; CHECK-NEXT: ADJCALLSTACKDOWN 0, 0, implicit-def $sp, implicit $sp
|
|
; CHECK-NEXT: $x8 = COPY [[PTR_ADD]](p0)
|
|
; CHECK-NEXT: BL @test_explicit_sret, csr_darwin_aarch64_aapcs, implicit-def $lr, implicit $sp, implicit $x8
|
|
; CHECK-NEXT: ADJCALLSTACKUP 0, 0, implicit-def $sp, implicit $sp
|
|
; CHECK-NEXT: RET_ReallyLR
|
|
%ptr2 = getelementptr i64, ptr %ptr, i32 1
|
|
tail call void @test_explicit_sret(ptr %ptr2)
|
|
ret void
|
|
}
|
|
|
|
define i64 @dont_tail_call_sret_alloca_returned() {
|
|
; CHECK-LABEL: name: dont_tail_call_sret_alloca_returned
|
|
; CHECK: bb.1 (%ir-block.0):
|
|
; CHECK-NEXT: [[FRAME_INDEX:%[0-9]+]]:_(p0) = G_FRAME_INDEX %stack.0.l
|
|
; CHECK-NEXT: ADJCALLSTACKDOWN 0, 0, implicit-def $sp, implicit $sp
|
|
; CHECK-NEXT: $x8 = COPY [[FRAME_INDEX]](p0)
|
|
; CHECK-NEXT: BL @test_explicit_sret, csr_darwin_aarch64_aapcs, implicit-def $lr, implicit $sp, implicit $x8
|
|
; CHECK-NEXT: ADJCALLSTACKUP 0, 0, implicit-def $sp, implicit $sp
|
|
; CHECK-NEXT: [[LOAD:%[0-9]+]]:_(i64) = G_LOAD [[FRAME_INDEX]](p0) :: (dereferenceable load (i64) from %ir.l)
|
|
; CHECK-NEXT: $x0 = COPY [[LOAD]](i64)
|
|
; CHECK-NEXT: RET_ReallyLR implicit $x0
|
|
%l = alloca i64, align 8
|
|
tail call void @test_explicit_sret(ptr %l)
|
|
%r = load i64, ptr %l, align 8
|
|
ret i64 %r
|
|
}
|