[MemCpyOpt] handle memcpy from memset in more cases #140954
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -1367,8 +1367,9 @@ bool MemCpyOptPass::processMemSetMemCpyDependence(MemCpyInst *MemCpy, | |
| return true; | ||
| } | ||
|
|
||
| /// Determine whether the pointer V had only undefined content (due to Def) up | ||
| /// to the given Size, either because it was freshly alloca'd or started its | ||
| /// lifetime. | ||
| static bool hasUndefContents(MemorySSA *MSSA, BatchAAResults &AA, Value *V, | ||
| MemoryDef *Def, Value *Size) { | ||
| if (MSSA->isLiveOnEntryDef(Def)) | ||
|
|
@@ -1403,6 +1404,24 @@ static bool hasUndefContents(MemorySSA *MSSA, BatchAAResults &AA, Value *V, | |
| return false; | ||
| } | ||
|
|
||
| // If the memcpy is larger than the previous, but the memory was undef prior to | ||
| // that, we can just ignore the tail. Technically we're only interested in the | ||
| // bytes from 0..MemSrcOffset and MemSrcLength+MemSrcOffset..CopySize here, but | ||
| // as we can't easily represent this location (hasUndefContents uses mustAlias | ||
| // which cannot deal with offsets), we use the full 0..CopySize range. | ||
| static bool overreadUndefContents(MemorySSA *MSSA, MemCpyInst *MemCpy, | ||
| MemIntrinsic *MemSrc, BatchAAResults &BAA) { | ||
| Value *CopySize = MemCpy->getLength(); | ||
| MemoryLocation MemCpyLoc = MemoryLocation::getForSource(MemCpy); | ||
| MemoryUseOrDef *MemSrcAccess = MSSA->getMemoryAccess(MemSrc); | ||
| MemoryAccess *Clobber = MSSA->getWalker()->getClobberingMemoryAccess( | ||
| MemSrcAccess->getDefiningAccess(), MemCpyLoc, BAA); | ||
| if (auto *MD = dyn_cast<MemoryDef>(Clobber)) | ||
| if (hasUndefContents(MSSA, BAA, MemCpy->getSource(), MD, CopySize)) | ||
| return true; | ||
| return false; | ||
| } | ||
|
|
||
| /// Transform memcpy to memset when its source was just memset. | ||
| /// In other words, turn: | ||
| /// \code | ||
|
|
@@ -1418,19 +1437,25 @@ static bool hasUndefContents(MemorySSA *MSSA, BatchAAResults &AA, Value *V, | |
| bool MemCpyOptPass::performMemCpyToMemSetOptzn(MemCpyInst *MemCpy, | ||
| MemSetInst *MemSet, | ||
| BatchAAResults &BAA) { | ||
| Value *MemSetSize = MemSet->getLength(); | ||
| Value *CopySize = MemCpy->getLength(); | ||
|
|
||
| int64_t MOffset = 0; | ||
dianqk marked this conversation as resolved.
Show resolved
Hide resolved
|
||
| 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. | ||
| if (MemCpy->getSource() != MemSet->getDest()) { | ||
| std::optional<int64_t> Offset = | ||
| MemCpy->getSource()->getPointerOffsetFrom(MemSet->getDest(), DL); | ||
| if (!Offset || *Offset < 0) | ||
| return false; | ||
| MOffset = *Offset; | ||
| } | ||
|
|
||
| if (MOffset != 0 || MemSetSize != CopySize) { | ||
dianqk marked this conversation as resolved.
Show resolved
Hide resolved
|
||
| // 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. | ||
| auto *CMemSetSize = dyn_cast<ConstantInt>(MemSetSize); | ||
| if (!CMemSetSize) | ||
| return false; | ||
|
|
@@ -1439,23 +1464,18 @@ bool MemCpyOptPass::performMemCpyToMemSetOptzn(MemCpyInst *MemCpy, | |
| auto *CCopySize = dyn_cast<ConstantInt>(CopySize); | ||
| if (!CCopySize) | ||
| return false; | ||
|
There was a problem hiding this comment. Should we also reduce the diff here? How about just modifying the comments? There was a problem hiding this comment. I think the helper function makes this function more readable, and the helper function should be more generally useful, since it implements a check that other functions in this file should also implement. In particular, I'm wanting to use this in There was a problem hiding this comment. In my opinion, making a new function is the more general approach when sharing code or when a piece of code grows significantly in size. --- a/llvm/lib/Transforms/Scalar/MemCpyOptimizer.cpp
+++ b/llvm/lib/Transforms/Scalar/MemCpyOptimizer.cpp
@@ -1790,6 +1790,9 @@ bool MemCpyOptPass::processMemCpy(MemCpyInst *M, BasicBlock::iterator &BBI) {
}
}
}
+ if (truncateUndefs(M, MI)) {
+ return true;
+ }
if (auto *MDep = dyn_cast<MemCpyInst>(MI))
if (processMemCpyMemCpyDependence(M, MDep, BAA))
return true;
There was a problem hiding this comment. I believe this helper function should also be used in |
||
| if (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); | ||
| } | ||
| } | ||
|
|
||
|
|
||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Uh oh!
There was an error while loading. Please reload this page.