Skip to content

GlobalISel: Fix defined register of invariant.start #125664

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 4 commits into from
Feb 4, 2025
Merged

GlobalISel: Fix defined register of invariant.start #125664

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
839080f
GlobalISel: Fix defined register of invariant.start
ro-i Feb 4, 2025
d8bdb59
simplify promote-dependency-on-invariant-result.ll
ro-i Feb 4, 2025
eab32fb
simplify promote-dependency-on-invariant-result.ll further
ro-i Feb 4, 2025
885344e
add store again to promote-dependency-on-invariant-result.ll
ro-i Feb 4, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 1 addition & 3 deletions llvm/lib/CodeGen/GlobalISel/IRTranslator.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2441,9 +2441,7 @@ bool IRTranslator::translateKnownIntrinsic(const CallInst &CI, Intrinsic::ID ID,
return true;
}
case Intrinsic::invariant_start: {
LLT PtrTy = getLLTForType(*CI.getArgOperand(0)->getType(), *DL);
Register Undef = MRI->createGenericVirtualRegister(PtrTy);
MIRBuilder.buildUndef(Undef);
MIRBuilder.buildUndef(getOrCreateVReg(CI));
return true;
}
case Intrinsic::invariant_end:
Expand Down
2 changes: 1 addition & 1 deletion llvm/test/CodeGen/AArch64/GlobalISel/arm64-irtranslator.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2262,7 +2262,7 @@ declare ptr @llvm.invariant.start.p0(i64, ptr nocapture) readonly nounwind
declare void @llvm.invariant.end.p0(ptr, i64, ptr nocapture) nounwind
define void @test_invariant_intrin() {
; CHECK-LABEL: name: test_invariant_intrin
; CHECK: %{{[0-9]+}}:_(s64) = G_IMPLICIT_DEF
; CHECK: %{{[0-9]+}}:_(p0) = G_IMPLICIT_DEF
; CHECK-NEXT: RET_ReallyLR
%x = alloca %t
%inv = call ptr @llvm.invariant.start.p0(i64 8, ptr %x)
Expand Down
24 changes: 24 additions & 0 deletions llvm/test/CodeGen/AMDGPU/GlobalISel/promote-dependency-on-invariant-result.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,24 @@
; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py UTC_ARGS: --version 5
; RUN: llc -mtriple=amdgcn -mcpu=gfx90a -O0 -global-isel=true --stop-after=irtranslator -o - %s | FileCheck %s
target datalayout = "A5"

declare ptr @llvm.invariant.start.p5(i64 immarg, ptr addrspace(5) nocapture)
declare void @llvm.invariant.end.p5(ptr, i64 immarg, ptr addrspace(5) nocapture)

define amdgpu_kernel void @use_invariant_promotable_lds(ptr addrspace(5) %arg, i32 %i) {
; CHECK-LABEL: name: use_invariant_promotable_lds
; CHECK: bb.1.bb:
; CHECK-NEXT: liveins: $sgpr4_sgpr5
; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(p4) = COPY $sgpr4_sgpr5
; CHECK-NEXT: [[INT:%[0-9]+]]:_(p4) = G_INTRINSIC intrinsic(@llvm.amdgcn.kernarg.segment.ptr)
; CHECK-NEXT: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 36
; CHECK-NEXT: %13:_(p4) = nuw nusw G_PTR_ADD [[INT]], [[C]](s64)
; CHECK-NEXT: [[LOAD:%[0-9]+]]:_(p5) = G_LOAD %13(p4) :: (dereferenceable invariant load (p5) from %ir.arg.kernarg.offset, addrspace 4)
; CHECK-NEXT: [[DEF:%[0-9]+]]:_(p0) = G_IMPLICIT_DEF
; CHECK-NEXT: S_ENDPGM 0
bb:
%tmp = call ptr @llvm.invariant.start.p5(i64 4, ptr addrspace(5) %arg)
call void @llvm.invariant.end.p5(ptr %tmp, i64 4, ptr addrspace(5) %arg)
ret void
}