Handle early inline for hot callsites that reside in the same basic block.

[danielcdh]

Summary

Callsites in the same basic block should share the same hotness. This patch checks for the hottest callsite in the same basic block, and use the hotness for all callsites in that basic block for early inline decisions. It also fixes the test to add "-S" so theat the "CHECK-NOT" is actually checking the content.

[dnovillo]

LGTM.

lib/Transforms/IPO/SampleProfile.cpp

bool SampleProfileLoader::inlineHotFunctions(Function &F) {
   bool Changed = false;
   LLVMContext &Ctx = F.getContext();
   std::function<AssumptionCache &(Function &)> GetAssumptionCache = [&](
       Function &F) -> AssumptionCache & { return ACT->getAssumptionCache(F); };
   while (true) {
     bool LocalChanged = false;
     SmallVector<Instruction *, 10> CIS;
     for (auto &BB : F) {
+      bool Hot = false;
+      SmallVector<Instruction *, 10> Candidates;
       for (auto &I : BB.getInstList()) {
         const FunctionSamples *FS = nullptr;
         if ((isa<CallInst>(I) || isa<InvokeInst>(I)) &&
             (FS = findCalleeFunctionSamples(I))) {
-
+          Candidates.push_back(&I);
           if (callsiteIsHot(Samples, FS))
-            CIS.push_back(&I);
-        }
+            Hot = true;
+        }
+      if (Hot) {
+        CIS.insert(CIS.begin(), Candidates.begin(), Candidates.end());
       }
     }
     for (auto I : CIS) {
       InlineFunctionInfo IFI(nullptr, ACT ? &GetAssumptionCache : nullptr);
       CallInst *CI = dyn_cast<CallInst>(I);
       InvokeInst *II = dyn_cast<InvokeInst>(I);
       Function *CalledFunction =
           (CI == nullptr ? II->getCalledFunction() : CI->getCalledFunction());

test/Transforms/SampleProfile/Inputs/einline.prof

 _Z3foov:200:100
- 1: _Z3barv:100
+ 1: _Z3barv:0
  3: _Z3barv:100

test/Transforms/SampleProfile/early-inline.ll

-; RUN: opt < %s -instcombine -sample-profile -sample-profile-file=%S/Inputs/einline.prof | FileCheck %s
+; RUN: opt < %s -instcombine -sample-profile -sample-profile-file=%S/Inputs/einline.prof -S | FileCheck %s
 
 ; Checks if both call and invoke can be inlined early if their inlined
 ; instances are hot in profile.
 
 target triple = "x86_64-unknown-linux-gnu"
 
 @_ZTIi = external constant i8*