PEI: Delay checking requiresFrameIndexReplacementScavenging

[arsenm]

Summary

Currently this is called before the frame size is set on the
function. For AMDGPU, the scavenger is used for large frames where
part of the offset needs to be materialized in a register, so
estimating the frame size is useful for knowing whether the scavenger
is useful.

[thegameg]

It would be good to have a test case for this, but I assume you're planning on changing SIRegisterInfo::requiresFrameIndexReplacementScavenging? Maybe you can merge both patches since AMDGPU seems to be the only user of the hook.

[arsenm]

In D59836#1443502, @thegameg wrote:

It would be good to have a test case for this, but I assume you're planning on changing SIRegisterInfo::requiresFrameIndexReplacementScavenging? Maybe you can merge both patches since AMDGPU seems to be the only user of the hook.

I usually try to separate the target specific part when possible: D59840

[thegameg]

In D59836#1443555, @arsenm wrote:

I usually try to separate the target specific part when possible: D59840

Fair enough, LGTM.

[arsenm]

r357087

lib/CodeGen/PrologEpilogInserter.cpp

 bool PEI::runOnMachineFunction(MachineFunction &MF) {
   NumFuncSeen++;
   const Function &F = MF.getFunction();
   const TargetRegisterInfo *TRI = MF.getSubtarget().getRegisterInfo();
   const TargetFrameLowering *TFI = MF.getSubtarget().getFrameLowering();
 
   RS = TRI->requiresRegisterScavenging(MF) ? new RegScavenger() : nullptr;
   FrameIndexVirtualScavenging = TRI->requiresFrameIndexScavenging(MF);
-  FrameIndexEliminationScavenging = (RS && !FrameIndexVirtualScavenging) ||
-    TRI->requiresFrameIndexReplacementScavenging(MF);
   ORE = &getAnalysis<MachineOptimizationRemarkEmitterPass>().getORE();
 
   // Calculate the MaxCallFrameSize and AdjustsStack variables for the
   // function's frame information. Also eliminates call frame pseudo
   // instructions.
   calculateCallFrameInfo(MF);
 
   // Determine placement of CSR spill/restore code and prolog/epilog code:
void PEI::insertPrologEpilogCode(MachineFunction &MF) {
   if (MF.getFunction().getCallingConv() == CallingConv::HiPE)
     for (MachineBasicBlock *SaveBlock : SaveBlocks)
       TFI.adjustForHiPEPrologue(MF, *SaveBlock);
 }
 
 /// replaceFrameIndices - Replace all MO_FrameIndex operands with physical
 /// register references and actual offsets.
 void PEI::replaceFrameIndices(MachineFunction &MF) {
-  const TargetFrameLowering &TFI = *MF.getSubtarget().getFrameLowering();
-  if (!TFI.needsFrameIndexResolution(MF)) return;
+  const auto &ST = MF.getSubtarget();
+  const TargetFrameLowering &TFI = *ST.getFrameLowering();
+  if (!TFI.needsFrameIndexResolution(MF))
+    return;
+
+  const TargetRegisterInfo *TRI = ST.getRegisterInfo();
+
+  // Allow the target to determine this after knowing the frame size.
+  FrameIndexEliminationScavenging = (RS && !FrameIndexVirtualScavenging) ||
+    TRI->requiresFrameIndexReplacementScavenging(MF);
 
   // Store SPAdj at exit of a basic block.
   SmallVector<int, 8> SPState;
   SPState.resize(MF.getNumBlockIDs());
   df_iterator_default_set<MachineBasicBlock*> Reachable;
 
   // Iterate over the reachable blocks in DFS order.
   for (auto DFI = df_ext_begin(&MF, Reachable), DFE = df_ext_end(&MF, Reachable);