summaryrefslogtreecommitdiffstats
path: root/llvm/lib/Transforms/Scalar/MemCpyOptimizer.cpp
diff options
context:
space:
mode:
authorAndrea Di Biagio <Andrea_DiBiagio@sn.scee.net>2015-10-09 10:53:41 +0000
committerAndrea Di Biagio <Andrea_DiBiagio@sn.scee.net>2015-10-09 10:53:41 +0000
commit99493df257a69faaf1eab1c9b1bbc204f1a2f557 (patch)
treeb584249b309380b7a49af9a9868a81b94147b022 /llvm/lib/Transforms/Scalar/MemCpyOptimizer.cpp
parent8112dd995175268d9dd39191a175943eba0f04af (diff)
downloadbcm5719-llvm-99493df257a69faaf1eab1c9b1bbc204f1a2f557.tar.gz
bcm5719-llvm-99493df257a69faaf1eab1c9b1bbc204f1a2f557.zip
[MemCpyOpt] Fix wrong merging adjacent nontemporal stores into memset calls.
Pass MemCpyOpt doesn't check if a store instruction is nontemporal. As a consequence, adjacent nontemporal stores are always merged into a memset call. Example: ;;; define void @foo(<4 x float>* nocapture %p) { entry: store <4 x float> zeroinitializer, <4 x float>* %p, align 16, !nontemporal !0 %p1 = getelementptr inbounds <4 x float>, <4 x float>* %dst, i64 1 store <4 x float> zeroinitializer, <4 x float>* %p1, align 16, !nontemporal !0 ret void } !0 = !{i32 1} ;;; In this example, the two nontemporal stores are combined to a memset of zero which does not preserve the nontemporal hint. Later on the backend (tested on a x86-64 corei7) expands that memset call into a sequence of two normal 16-byte aligned vector stores. opt -memcpyopt example.ll -S -o - | llc -mcpu=corei7 -o - Before: xorps %xmm0, %xmm0 movaps %xmm0, 16(%rdi) movaps %xmm0, (%rdi) With this patch, we no longer merge nontemporal stores into calls to memset. In this example, llc correctly expands the two stores into two movntps: xorps %xmm0, %xmm0 movntps %xmm0, 16(%rdi) movntps %xmm0, (%rdi) In theory, we could extend the usage of !nontemporal metadata to memcpy/memset calls. However a change like that would only have the effect of forcing the backend to expand !nontemporal memsets back to sequences of store instructions. A memset library call would not have exactly the same semantic of a builtin !nontemporal memset call. So, SelectionDAG will have to conservatively expand it back to a sequence of !nontemporal stores (effectively undoing the merging). Differential Revision: http://reviews.llvm.org/D13519 llvm-svn: 249820
Diffstat (limited to 'llvm/lib/Transforms/Scalar/MemCpyOptimizer.cpp')
-rw-r--r--llvm/lib/Transforms/Scalar/MemCpyOptimizer.cpp10
1 files changed, 10 insertions, 0 deletions
diff --git a/llvm/lib/Transforms/Scalar/MemCpyOptimizer.cpp b/llvm/lib/Transforms/Scalar/MemCpyOptimizer.cpp
index 4039c6f3534..64f8aafcda0 100644
--- a/llvm/lib/Transforms/Scalar/MemCpyOptimizer.cpp
+++ b/llvm/lib/Transforms/Scalar/MemCpyOptimizer.cpp
@@ -488,6 +488,16 @@ Instruction *MemCpyOpt::tryMergingIntoMemset(Instruction *StartInst,
bool MemCpyOpt::processStore(StoreInst *SI, BasicBlock::iterator &BBI) {
if (!SI->isSimple()) return false;
+
+ // Avoid merging nontemporal stores since the resulting
+ // memcpy/memset would not be able to preserve the nontemporal hint.
+ // In theory we could teach how to propagate the !nontemporal metadata to
+ // memset calls. However, that change would force the backend to
+ // conservatively expand !nontemporal memset calls back to sequences of
+ // store instructions (effectively undoing the merging).
+ if (SI->getMetadata(LLVMContext::MD_nontemporal))
+ return false;
+
const DataLayout &DL = SI->getModule()->getDataLayout();
// Detect cases where we're performing call slot forwarding, but
OpenPOWER on IntegriCloud