From: Sanjay Patel <spatel@rotateright.com>
Date: Fri, 14 Aug 2015 17:53:40 +0000 (+0000)
Subject: [x86] fix allowsMisalignedMemoryAccess() implementation
X-Git-Url: http://demsky.eecs.uci.edu/git/?a=commitdiff_plain;h=09ec2d3ce34ef1b1e1683b324428fd156542601d;p=oota-llvm.git

[x86] fix allowsMisalignedMemoryAccess() implementation

This patch fixes the x86 implementation of allowsMisalignedMemoryAccess() to correctly
return the 'Fast' output parameter for 32-byte accesses. To test that, an existing load
merging optimization is changed to use the TLI hook. This exposes a shortcoming in the
current logic and results in the regression test update. Changing other direct users of
the isUnalignedMem32Slow() x86 CPU attribute would be a follow-on patch.

Without the fix in allowsMisalignedMemoryAccesses(), we will infinite loop when targeting
SandyBridge because LowerINSERT_SUBVECTOR() creates 32-byte loads from two 16-byte loads
while PerformLOADCombine() splits them back into 16-byte loads.

Differential Revision: http://reviews.llvm.org/D10662



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@245075 91177308-0d34-0410-b5e6-96231b3b80d8
---

diff --git a/lib/Target/X86/X86ISelLowering.cpp b/lib/Target/X86/X86ISelLowering.cpp
index 9a790010f8c..68a15e021cd 100644
--- a/lib/Target/X86/X86ISelLowering.cpp
+++ b/lib/Target/X86/X86ISelLowering.cpp
@@ -1915,8 +1915,14 @@ X86TargetLowering::allowsMisalignedMemoryAccesses(EVT VT,
                                                   unsigned,
                                                   unsigned,
                                                   bool *Fast) const {
-  if (Fast)
-    *Fast = Subtarget->isUnalignedMemAccessFast();
+  if (Fast) {
+    // FIXME: We should be checking 128-bit accesses separately from smaller
+    // accesses.
+    if (VT.getSizeInBits() == 256)
+      *Fast = !Subtarget->isUnalignedMem32Slow();
+    else
+      *Fast = Subtarget->isUnalignedMemAccessFast();
+  }
   return true;
 }
 
@@ -11259,14 +11265,25 @@ static SDValue LowerINSERT_SUBVECTOR(SDValue Op, const X86Subtarget *Subtarget,
   // --> load32 addr
   if ((IdxVal == OpVT.getVectorNumElements() / 2) &&
       Vec.getOpcode() == ISD::INSERT_SUBVECTOR &&
-      OpVT.is256BitVector() && SubVecVT.is128BitVector() &&
-      !Subtarget->isUnalignedMem32Slow()) {
-    SDValue SubVec2 = Vec.getOperand(1);
-    if (auto *Idx2 = dyn_cast<ConstantSDNode>(Vec.getOperand(2))) {
-      if (Idx2->getZExtValue() == 0) {
-        SDValue Ops[] = { SubVec2, SubVec };
-        if (SDValue Ld = EltsFromConsecutiveLoads(OpVT, Ops, dl, DAG, false))
-          return Ld;
+      OpVT.is256BitVector() && SubVecVT.is128BitVector()) {
+    auto *Idx2 = dyn_cast<ConstantSDNode>(Vec.getOperand(2));
+    if (Idx2 && Idx2->getZExtValue() == 0) {
+      SDValue SubVec2 = Vec.getOperand(1);
+      // If needed, look through a bitcast to get to the load.
+      if (SubVec2.getNode() && SubVec2.getOpcode() == ISD::BITCAST)
+        SubVec2 = SubVec2.getOperand(0);
+      
+      if (auto *FirstLd = dyn_cast<LoadSDNode>(SubVec2)) {
+        bool Fast;
+        unsigned Alignment = FirstLd->getAlignment();
+        unsigned AS = FirstLd->getAddressSpace();
+        const X86TargetLowering *TLI = Subtarget->getTargetLowering();
+        if (TLI->allowsMemoryAccess(*DAG.getContext(), DAG.getDataLayout(),
+                                    OpVT, AS, Alignment, &Fast) && Fast) {
+          SDValue Ops[] = { SubVec2, SubVec };
+          if (SDValue Ld = EltsFromConsecutiveLoads(OpVT, Ops, dl, DAG, false))
+            return Ld;
+        }
       }
     }
   }
diff --git a/test/CodeGen/X86/unaligned-32-byte-memops.ll b/test/CodeGen/X86/unaligned-32-byte-memops.ll
index d979c16f4ab..608c6e72fbe 100644
--- a/test/CodeGen/X86/unaligned-32-byte-memops.ll
+++ b/test/CodeGen/X86/unaligned-32-byte-memops.ll
@@ -75,12 +75,12 @@ define <8 x float> @combine_16_byte_loads_no_intrinsic(<4 x float>* %ptr) {
   ret <8 x float> %v3
 }
 
+; If the first load is 32-byte aligned, then the loads should be merged in all cases.
+
 define <8 x float> @combine_16_byte_loads_aligned(<4 x float>* %ptr) {
-;; FIXME: The first load is 32-byte aligned, so the second load should get merged.
 ; AVXSLOW-LABEL: combine_16_byte_loads_aligned:
 ; AVXSLOW:       # BB#0:
-; AVXSLOW-NEXT:    vmovaps 48(%rdi), %xmm0
-; AVXSLOW-NEXT:    vinsertf128 $1, 64(%rdi), %ymm0, %ymm0
+; AVXSLOW-NEXT:    vmovaps 48(%rdi), %ymm0
 ; AVXSLOW-NEXT:    retq
 ;
 ; AVXFAST-LABEL: combine_16_byte_loads_aligned: