[llvm] r257015 - [RS4GC] Add an option to suppress vector splitting

[Philip Reames via llvm-commits]

Author: reames
Date: Wed Jan  6 20:20:11 2016
New Revision: 257015

URL: http://llvm.org/viewvc/llvm-project?rev=257015&view=rev
Log:
[RS4GC] Add an option to suppress vector splitting

At the moment, this is essentially a diangostic option so that I can start collecting failing test cases, but we will eventually migrate to removing the vector splitting code entirely.


Modified:
    llvm/trunk/lib/Transforms/Scalar/RewriteStatepointsForGC.cpp

Modified: llvm/trunk/lib/Transforms/Scalar/RewriteStatepointsForGC.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Scalar/RewriteStatepointsForGC.cpp?rev=257015&r1=257014&r2=257015&view=diff
==============================================================================
--- llvm/trunk/lib/Transforms/Scalar/RewriteStatepointsForGC.cpp (original)
+++ llvm/trunk/lib/Transforms/Scalar/RewriteStatepointsForGC.cpp Wed Jan  6 20:20:11 2016
@@ -78,6 +78,13 @@ static cl::opt<bool>
     AllowStatepointWithNoDeoptInfo("rs4gc-allow-statepoint-with-no-deopt-info",
                                    cl::Hidden, cl::init(true));
 
+/// Should we split vectors of pointers into their individual elements?  This
+/// is known to be buggy, but the alternate implementation isn't yet ready.
+/// This is purely to provide a debugging and dianostic hook until the vector
+/// split is replaced with vector relocations.
+static cl::opt<bool> UseVectorSplit("rs4gc-split-vector-values", cl::Hidden,
+                                    cl::init(true));
+
 namespace {
 struct RewriteStatepointsForGC : public ModulePass {
   static char ID; // Pass identification, replacement for typeid
@@ -2380,15 +2387,19 @@ static bool insertParsePoints(Function &
 
   // Do a limited scalarization of any live at safepoint vector values which
   // contain pointers.  This enables this pass to run after vectorization at
-  // the cost of some possible performance loss.  TODO: it would be nice to
-  // natively support vectors all the way through the backend so we don't need
-  // to scalarize here.
-  for (size_t i = 0; i < Records.size(); i++) {
-    PartiallyConstructedSafepointRecord &Info = Records[i];
-    Instruction *Statepoint = ToUpdate[i].getInstruction();
-    splitVectorValues(cast<Instruction>(Statepoint), Info.LiveSet,
-                      Info.PointerToBase, DT);
-  }
+  // the cost of some possible performance loss.  Note: This is known to not
+  // handle updating of the side tables correctly which can lead to relocation
+  // bugs when the same vector is live at multiple statepoints.  We're in the
+  // process of implementing the alternate lowering - relocating the
+  // vector-of-pointers as first class item and updating the backend to
+  // understand that - but that's not yet complete.  
+  if (UseVectorSplit)
+    for (size_t i = 0; i < Records.size(); i++) {
+      PartiallyConstructedSafepointRecord &Info = Records[i];
+      Instruction *Statepoint = ToUpdate[i].getInstruction();
+      splitVectorValues(cast<Instruction>(Statepoint), Info.LiveSet,
+                        Info.PointerToBase, DT);
+    }
 
   // In order to reduce live set of statepoint we might choose to rematerialize
   // some values instead of relocating them. This is purely an optimization and