Skip to content

[LAA] Fix type mismatch in getStartAndEndForAccess. #183116

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
kshitijvp merged 9 commits into from
Mar 11, 2026
Merged

[LAA] Fix type mismatch in getStartAndEndForAccess. #183116

Show file tree
Hide file tree
Changes from 6 commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
2b84803
Fix type mismatch
kshitijvp Feb 24, 2026
51892d4
Nitpick
kshitijvp Feb 28, 2026
9444d4f
Nitpick
kshitijvp Feb 28, 2026
372720b
fix tests
kshitijvp Feb 28, 2026
1f0a7a4
fix tests
kshitijvp Feb 28, 2026
525c392
fix tests
kshitijvp Feb 28, 2026
e2ae56e
nit
kshitijvp Mar 3, 2026
b4e4913
update tests
kshitijvp Mar 4, 2026
dd1818c
Minor changes
kshitijvp Mar 11, 2026
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
9 changes: 7 additions & 2 deletions llvm/lib/Analysis/LoopAccessAnalysis.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -250,8 +250,13 @@ static bool evaluatePtrAddRecAtMaxBTCWillNotWrap(
return true;
});
if (DerefRK) {
DerefBytesSCEV =
SE.getUMaxExpr(DerefBytesSCEV, SE.getSCEV(DerefRK.IRArgValue));
const SCEV *DerefRKSCEV = SE.getSCEV(DerefRK.IRArgValue);
// Ensure both operands have the same type
Comment thread
kshitijvp marked this conversation as resolved.
Outdated
Type *CommonTy =
SE.getWiderType(DerefBytesSCEV->getType(), DerefRKSCEV->getType());
Comment thread
kshitijvp marked this conversation as resolved.
DerefBytesSCEV = SE.getNoopOrAnyExtend(DerefBytesSCEV, CommonTy);
DerefRKSCEV = SE.getNoopOrAnyExtend(DerefRKSCEV, CommonTy);

@fhahn fhahn Mar 3, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Why any extend? Both quantities should be positive, so better always use Zext?

@kshitijvp kshitijvp Mar 3, 2026

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

done

DerefBytesSCEV = SE.getUMaxExpr(DerefBytesSCEV, DerefRKSCEV);
}

if (DerefBytesSCEV->isZero())
Expand Down
31 changes: 31 additions & 0 deletions llvm/test/Transforms/LoopVectorize/early_exit_store_legality.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -165,6 +165,7 @@ exit:
ret void
}

; Function Attrs: mustprogress norecurse nosync nounwind ssp memory(read, inaccessiblemem: write, target_mem0: none, target_mem1: none) uwtable(sync)
Comment thread
kshitijvp marked this conversation as resolved.
Outdated
define void @loop_contains_store_assumed_bounds(ptr noalias %array, ptr readonly %pred, i32 %n) {
; CHECK-LABEL: LV: Checking a loop in 'loop_contains_store_assumed_bounds'
; CHECK: LV: Not vectorizing: Loop may fault.
Expand Down Expand Up @@ -194,6 +195,36 @@ exit:
ret void
}

define void @test_assumed_bounds_type_mismatch(ptr noalias %array, ptr readonly %pred, i32 %n) nosync nofree {

@fhahn fhahn Mar 4, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

is there a reason we need a loop with a store? If not, would be good to remove the store and add to llvm/test/Transforms/LoopVectorize/dereferenceable-info-from-assumption-variable-size.ll. It should get vectorized with the crash fixed

@kshitijvp kshitijvp Mar 4, 2026

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

yes the store is redundant in this case, needs to be removed.

@kshitijvp kshitijvp Mar 4, 2026

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

removing the store instruction is causing faulting load, instead of being vectorized.

; CHECK-LABEL: LV: Checking a loop in 'test_assumed_bounds_type_mismatch'
; CHECK: LV: Not vectorizing: Loop may fault.
entry:
%n_bytes = mul nuw nsw i32 %n, 2
call void @llvm.assume(i1 true) [ "dereferenceable"(ptr %pred, i32 %n_bytes) ]
%tc = sext i32 %n to i64
br label %for.body

for.body:
%iv = phi i64 [ 0, %entry ], [ %iv.next, %for.inc ]
%st.addr = getelementptr inbounds i16, ptr %array, i64 %iv
%data = load i16, ptr %st.addr, align 2
%inc = add nsw i16 %data, 1
store i16 %inc, ptr %st.addr, align 2
%ee.addr = getelementptr inbounds i16, ptr %pred, i64 %iv
%ee.val = load i16, ptr %ee.addr, align 2
%ee.cond = icmp sgt i16 %ee.val, 500
br i1 %ee.cond, label %exit, label %for.inc

for.inc:
%iv.next = add nuw nsw i64 %iv, 1
%counted.cond = icmp eq i64 %iv.next, %tc
br i1 %counted.cond, label %exit, label %for.body

exit:
ret void
}


define void @loop_contains_store_to_pointer_with_no_deref_info(ptr align 2 dereferenceable(40) readonly %load.array, ptr align 2 noalias %array, ptr align 2 dereferenceable(40) readonly %pred) {
; CHECK-LABEL: LV: Checking a loop in 'loop_contains_store_to_pointer_with_no_deref_info'
; CHECK: LV: Not vectorizing: Writes to memory unsupported in early exit loops.
Expand Down