[InlineCost] Small changes to early exit condition. NFC.

[haicheng]

Summary

If Cost < Threshold, we will try to inline the callee. So, we can early exit when Cost >= Threshold which is earlier than Cost > Threshold.

[eastig]

lib/Analysis/InlineCost.cpp

1209 ↗

(On Diff #112481)

I don't think this is NFC. In visitSwitchInst I see the cost is calculating based on a result of TTI.getEstimatedNumberOfCaseClusters. If threshold is big (-O1 and above) this might not be a problem but if it is low (-Os, -Oz) this might be a problem.

[haicheng]

Keep the check Evgeny has question about unchanged.

[junbuml]

lib/Analysis/InlineCost.cpp

1209 ↗

(On Diff #112481)

Specifically for this part, we cannot say that this is NFC in case when "CostLowerBound == Threshold" and the switch ends up with a bit test. See above "FIXME". However, I cannot see any big problem using ">=" here.

[eastig]

LGTM
Thank you, Haicheng.

[eastig]

lib/Analysis/InlineCost.cpp

1209 ↗

(On Diff #112481)

I agree it does not look like a big problem but to be sure some benchmarking, especially code size, needs to be done.

llvm/trunk/lib/Analysis/InlineCost.cpp

if (IsCallerRecursive &&
             << NV("Callee", &F)
             << " is recursive and allocates too much stack space. Cost is "
                "not fully computed");
       return false;
     }
 
     // Check if we've past the maximum possible threshold so we don't spin in
     // huge basic blocks that will never inline.
-    if (Cost > Threshold && !ComputeFullInlineCost)
+    if (Cost >= Threshold && !ComputeFullInlineCost)
       return false;
   }
 
   return true;
 }
 
 /// \brief Compute the base pointer and cumulative constant offsets for V.
 ///
bool CallAnalyzer::analyzeCall(CallSite CS) {
   Cost -= getCallsiteCost(CS, DL);
 
   // If this function uses the coldcc calling convention, prefer not to inline
   // it.
   if (F.getCallingConv() == CallingConv::Cold)
     Cost += InlineConstants::ColdccPenalty;
 
   // Check if we're done. This can happen due to bonuses and penalties.
-  if (Cost > Threshold && !ComputeFullInlineCost)
+  if (Cost >= Threshold && !ComputeFullInlineCost)
     return false;
 
   if (F.empty())
     return true;
 
   Function *Caller = CS.getInstruction()->getParent()->getParent();
   // Check if the caller function is recursive itself.
   for (User *U : Caller->users()) {
typedef SetVector<BasicBlock *, SmallVector<BasicBlock *, 16>,
       BBSetVector;
   BBSetVector BBWorklist;
   BBWorklist.insert(&F.getEntryBlock());
   bool SingleBB = true;
   // Note that we *must not* cache the size, this loop grows the worklist.
   for (unsigned Idx = 0; Idx != BBWorklist.size(); ++Idx) {
     // Bail out the moment we cross the threshold. This means we'll under-count
     // the cost, but only when undercounting doesn't matter.
-    if (Cost > Threshold && !ComputeFullInlineCost)
+    if (Cost >= Threshold && !ComputeFullInlineCost)
       break;
 
     BasicBlock *BB = BBWorklist[Idx];
     if (BB->empty())
       continue;
 
     // Disallow inlining a blockaddress. A blockaddress only has defined
     // behavior for an indirect branch in the same function, and we do not