Skip to content

[MemCpyOpt] handle memcpy from memset for non-constant sizes #143727

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 2 commits into from
Jun 12, 2025
Merged

[MemCpyOpt] handle memcpy from memset for non-constant sizes #143727

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
9fa18cd
[memcpyopt] handle memcpy from memset for non-constant sizes
vtjnash Jun 2, 2025
09c87de
Update variable-sized-memset-memcpy.ll
vtjnash Jun 11, 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
38 changes: 19 additions & 19 deletions llvm/lib/Transforms/Scalar/MemCpyOptimizer.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1440,7 +1440,7 @@ bool MemCpyOptPass::performMemCpyToMemSetOptzn(MemCpyInst *MemCpy,
int64_t MOffset = 0;
const DataLayout &DL = MemCpy->getModule()->getDataLayout();
// We can only transforms memcpy's where the dest of one is the source of the
// other, or the memory transfer has a known offset from the memset.
// other, or they have a known offset.
if (MemCpy->getSource() != MemSet->getDest()) {
std::optional<int64_t> Offset =
MemCpy->getSource()->getPointerOffsetFrom(MemSet->getDest(), DL);
Expand All @@ -1451,28 +1451,28 @@ bool MemCpyOptPass::performMemCpyToMemSetOptzn(MemCpyInst *MemCpy,

if (MOffset != 0 || MemSetSize != CopySize) {
// Make sure the memcpy doesn't read any more than what the memset wrote,
// other than undef. Don't worry about sizes larger than i64. A known memset
// size is required.
// other than undef. Don't worry about sizes larger than i64.
auto *CMemSetSize = dyn_cast<ConstantInt>(MemSetSize);
if (!CMemSetSize)
return false;

// A known memcpy size is also required.
auto *CCopySize = dyn_cast<ConstantInt>(CopySize);
if (!CCopySize)
return false;
if (CCopySize->getZExtValue() + MOffset > CMemSetSize->getZExtValue()) {
if (!CMemSetSize || !CCopySize ||
CCopySize->getZExtValue() + MOffset > CMemSetSize->getZExtValue()) {
if (!overreadUndefContents(MSSA, MemCpy, MemSet, BAA))
return false;
// Clip the memcpy to the bounds of the memset
if (MOffset == 0)
CopySize = MemSetSize;
else
CopySize =
ConstantInt::get(CopySize->getType(),
CMemSetSize->getZExtValue() <= (uint64_t)MOffset
? 0
: CMemSetSize->getZExtValue() - MOffset);

if (CMemSetSize && CCopySize) {
// If both have constant sizes and offsets, clip the memcpy to the
// bounds of the memset if applicable.
assert(CCopySize->getZExtValue() + MOffset >
CMemSetSize->getZExtValue());
if (MOffset == 0)
CopySize = MemSetSize;
else
CopySize =
ConstantInt::get(CopySize->getType(),
CMemSetSize->getZExtValue() <= (uint64_t)MOffset
? 0
: CMemSetSize->getZExtValue() - MOffset);
}
}
}

Expand Down
6 changes: 3 additions & 3 deletions llvm/test/Transforms/MemCpyOpt/variable-sized-memset-memcpy.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,12 +19,12 @@ define void @test(ptr %src, i8 %c, i64 %size) {
}

; Differing sizes, but would be UB if size1 < size2 since the memcpy would reference outside of the first alloca
define void @negative_test(ptr %src, i8 %c, i64 %size1, i64 %size2) {
; CHECK-LABEL: @negative_test(
define void @dynsize_test(ptr %src, i8 %c, i64 %size1, i64 %size2) {
; CHECK-LABEL: @dynsize_test(
; CHECK-NEXT: [[DST1:%.*]] = alloca i8, i64 [[SIZE1:%.*]], align 1
; CHECK-NEXT: [[DST2:%.*]] = alloca i8, i64 [[SIZE2:%.*]], align 1
; CHECK-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[DST1]], i8 [[C:%.*]], i64 [[SIZE1]], i1 false)
; CHECK-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[DST2]], ptr align 8 [[DST1]], i64 [[SIZE2]], i1 false)
; CHECK-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[DST2]], i8 [[C]], i64 [[SIZE2]], i1 false)
Copy link
Contributor

Choose a reason for hiding this comment

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

nit: Change the test name, this is no longer a negative test.

; CHECK-NEXT: ret void
;
%dst1 = alloca i8, i64 %size1
Expand Down