diff options
author | James Y Knight <jyknight@google.com> | 2015-11-02 18:48:08 +0000 |
---|---|---|
committer | James Y Knight <jyknight@google.com> | 2015-11-02 18:48:08 +0000 |
commit | 646c4032e7e9634ad102a00987fa92c1c5bacd5a (patch) | |
tree | 74859a861baf7e3560b726411b74071e9be710cf /llvm/lib/CodeGen | |
parent | 06fb04f1fa22b2d2398e3c426a204484fd524919 (diff) | |
download | bcm5719-llvm-646c4032e7e9634ad102a00987fa92c1c5bacd5a.tar.gz bcm5719-llvm-646c4032e7e9634ad102a00987fa92c1c5bacd5a.zip |
Fix two issues in MergeConsecutiveStores:
1) PR25154. This is basically a repeat of PR18102, which was fixed in
r200201, and broken again by r234430. The latter changed which of the
store nodes was merged into from the first to the last. Thus, we now
also need to prefer merging a later store at a given address into the
target node, instead of an earlier one.
2) While investigating that, I also realized I'd introduced a bug in
r236850. There, I removed a check for alignment -- not realizing that
nothing except the alignment check was ensuring that none of the stores
were overlapping! This is a really bogus way to ensure there's no
aliased stores.
A better solution to both of these issues is likely to always use the
code added in the 'if (UseAA)' branches which rearrange the chain based
on a more principled analysis. I'll look into whether that can be used
always, but in the interest of getting things back to working, I think a
minimal change makes sense.
llvm-svn: 251816
Diffstat (limited to 'llvm/lib/CodeGen')
-rw-r--r-- | llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp | 17 |
1 files changed, 15 insertions, 2 deletions
diff --git a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp index 31f0df99d47..a790282fada 100644 --- a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp +++ b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp @@ -11165,6 +11165,13 @@ void DAGCombiner::getStoreMergeAndAliasCandidates( if (Index->getMemoryVT() != MemVT) break; + // We do not allow under-aligned stores in order to prevent + // overriding stores. NOTE: this is a bad hack. Alignment SHOULD + // be irrelevant here; what MATTERS is that we not move memory + // operations that potentially overlap past each-other. + if (Index->getAlignment() < MemVT.getStoreSize()) + break; + // We found a potential memory operand to merge. StoreNodes.push_back(MemOpLink(Index, Ptr.Offset, Seq++)); @@ -11249,12 +11256,18 @@ bool DAGCombiner::MergeConsecutiveStores(StoreSDNode* St) { if (StoreNodes.size() < 2) return false; - // Sort the memory operands according to their distance from the base pointer. + // Sort the memory operands according to their distance from the + // base pointer. As a secondary criteria: make sure stores coming + // later in the code come first in the list. This is important for + // the non-UseAA case, because we're merging stores into the FINAL + // store along a chain which potentially contains aliasing stores. + // Thus, if there are multiple stores to the same address, the last + // one can be considered for merging but not the others. std::sort(StoreNodes.begin(), StoreNodes.end(), [](MemOpLink LHS, MemOpLink RHS) { return LHS.OffsetFromBase < RHS.OffsetFromBase || (LHS.OffsetFromBase == RHS.OffsetFromBase && - LHS.SequenceNum > RHS.SequenceNum); + LHS.SequenceNum < RHS.SequenceNum); }); // Scan the memory operations on the chain and find the first non-consecutive |