Skip to content

[SROA] Prevent load atomic vector from being generated #112432

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

Open
wants to merge 10 commits into
base: main
Choose a base branch
from
Open

[SROA] Prevent load atomic vector from being generated #112432

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
cb05b56
[SROA] Prevent load atomic vector from being generated
jofrn Oct 15, 2024
0c61505
Autogenerate assertions
jofrn Oct 15, 2024
083346d
Add i32,ptr tests and rename variables
jofrn Oct 16, 2024
cf47505
memcpy size must be >=8 to generate load atomic <1 x ptr>
jofrn Oct 17, 2024
ef4fe92
Use isValidAtomicTy
jofrn Oct 18, 2024
bdebaa8
Add AtomicOrdering to isValidAtomicTy interface
jofrn Oct 18, 2024
d68d923
Updated basictest.ll for checking getTypeSizeInBits
jofrn Oct 18, 2024
1c1aae7
Revert basictest.ll changes and DL check
jofrn Oct 18, 2024
67e0560
Add <2 x i32> test. Add DL as optional to isValidAtomicTy
jofrn Oct 20, 2024
e2e58ee
Added nonbyte test. Removed interface to isValidAtomicTy.
jofrn Oct 21, 2024
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
5 changes: 5 additions & 0 deletions llvm/lib/Transforms/Scalar/SROA.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2853,6 +2853,11 @@ class AllocaSliceRewriter : public InstVisitor<AllocaSliceRewriter, bool> {

bool visitLoadInst(LoadInst &LI) {
LLVM_DEBUG(dbgs() << " original: " << LI << "\n");

// load atomic vector would be generated, which is illegal
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
// load atomic vector would be generated, which is illegal
// Load atomic vector would be generated, which is illegal.

if (LI.isAtomic() && NewAI.getAllocatedType()->isVectorTy())
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is it illegal from LangRef's perspective?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It is at least illegal from the perspective of IR/Verifier.

Copy link
Contributor Author

@jofrn jofrn Oct 15, 2024 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It looks like this is written here.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The restriction is dumb and we should relax it. Instead of just hardcoding isVectorTy here, should have some kind of LoadInst::isValidAtomicType or something

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The verifier check should be moved into a LoadInst helper

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No alloca reference, just the type. The implementation should match Verifier::visitLoadInst, the isIntOrIntPtrTy || isFloatingPointTy and checkAtomicMemAccessSize (also add some non-atomic size tests?)

Copy link
Contributor Author

@jofrn jofrn Oct 16, 2024 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We need to ensure an atomic does not have a vector type. If the load is atomic and the alloca that will lend its type over has a vector type, then we will generate an atomic vector, which are illegal. We need to ensure that this doesn't occur.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No, you need to ensure the atomic is a valid type for an atomic load.

Alternatively you can do the load with the equivalent sized type and then bitcast (which is why this restriction is dumb in the first place, the lowering can always do the same)

Copy link
Contributor Author

@jofrn jofrn Oct 16, 2024 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We are ensuring it is a valid type in the case under question, by checking the alloca's type. The invalid load will not be generated if the alloca has a vector type as that type will overwrite the load's type.

Copy link
Contributor Author

@jofrn jofrn Oct 17, 2024 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We don't want any loads that are atomic to have a vector type, regardless of whether the atomic itself is already valid; so checking if the atomic is more valid before translating it in AllocaSliceRewriter will miss cases where we form an invalid atomic during SROA (and then later form a vector type with it in visitLoadInst). Even though it is not likely as SROA probably won't form these, it illustrates why we don't need the extra checks here.

return false;

Value *OldOp = LI.getOperand(0);
assert(OldOp == OldPtr);

Expand Down
19 changes: 19 additions & 0 deletions llvm/test/Transforms/SROA/atomic-vector.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,19 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
; RUN: opt < %s -passes='sroa' -S | FileCheck %s

define float @atomic_vector() {
; CHECK-LABEL: define float @atomic_vector() {
; CHECK-NEXT: [[TMP1:%.*]] = alloca <1 x float>, align 4
; CHECK-NEXT: store <1 x float> undef, ptr [[TMP1]], align 4
; CHECK-NEXT: [[TMP2:%.*]] = load atomic volatile float, ptr [[TMP1]] acquire, align 4
; CHECK-NEXT: ret float [[TMP2]]
;
%1 = alloca <1 x float>
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Use named values in tests

%2 = alloca <1 x float>
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Test an integer and pointer vector too

%3 = alloca ptr
call void @llvm.memcpy.p0.p0.i64(ptr %2, ptr %1, i64 4, i1 false)
store ptr %2, ptr %3
%4 = load ptr, ptr %3
%5 = load atomic volatile float, ptr %4 acquire, align 4
ret float %5
}
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Test a <2 x i16> or some other real vector. 1 x is a degenerate case

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Add test for the non-byte illegal case?

Loading