Index: lib/Analysis/ProfileSummaryInfo.cpp
===================================================================
--- lib/Analysis/ProfileSummaryInfo.cpp
+++ lib/Analysis/ProfileSummaryInfo.cpp
@@ -73,13 +73,18 @@
                                     BlockFrequencyInfo *BFI) {
   if (!Inst)
     return None;
+  if (!computeSummary())
+    return None;
   assert((isa<CallInst>(Inst) || isa<InvokeInst>(Inst)) &&
          "We can only get profile count for call/invoke instruction.");
-  // Check if there is a profile metadata on the instruction. If it is present,
-  // determine hotness solely based on that.
-  uint64_t TotalCount;
-  if (Inst->extractProfTotalWeight(TotalCount))
-    return TotalCount;
+  if (Summary->getKind() == ProfileSummary::PSK_Sample) {
+    // In sample PGO mode, check if there is a profile metadata on the
+    // instruction. If it is present, determine hotness solely based on that,
+    // since the sampled entry count may not be accurate.
+    uint64_t TotalCount;
+    if (Inst->extractProfTotalWeight(TotalCount))
+      return TotalCount;
+  }
   if (BFI)
     return BFI->getBlockProfileCount(Inst->getParent());
   return None;
Index: test/Bitcode/thinlto-function-summary-callgraph-pgo.ll
===================================================================
--- test/Bitcode/thinlto-function-summary-callgraph-pgo.ll
+++ test/Bitcode/thinlto-function-summary-callgraph-pgo.ll
@@ -17,7 +17,7 @@
 ; CHECK:       <GLOBALVAL_SUMMARY_BLOCK
 ; CHECK-NEXT:    <VERSION
 ; See if the call to func is registered, using the expected hotness type.
-; CHECK-NEXT:    <PERMODULE_PROFILE {{.*}} op4=1 op5=2/>
+; CHECK-NEXT:    <PERMODULE_PROFILE {{.*}} op4=1 op5=3/>
 ; CHECK-NEXT:  </GLOBALVAL_SUMMARY_BLOCK>
 ; CHECK: <STRTAB_BLOCK
 ; CHECK-NEXT: blob data = 'mainfunc'
@@ -28,8 +28,8 @@
 ; COMBINED-NEXT:    <VALUE_GUID
 ; COMBINED-NEXT:    <COMBINED
 ; See if the call to func is registered, using the expected hotness type.
-; op6=2 which is hotnessType::None.
-; COMBINED-NEXT:    <COMBINED_PROFILE {{.*}} op5=[[FUNCID]] op6=2/>
+; op6=2 which is hotnessType::Hot.
+; COMBINED-NEXT:    <COMBINED_PROFILE {{.*}} op5=[[FUNCID]] op6=3/>
 ; COMBINED-NEXT:  </GLOBALVAL_SUMMARY_BLOCK>
 
 ; ModuleID = 'thinlto-function-summary-callgraph.ll'
@@ -37,7 +37,7 @@
 target triple = "x86_64-unknown-linux-gnu"
 
 ; Function Attrs: nounwind uwtable
-define i32 @main() #0 !prof !2 {
+define i32 @main() #0 !prof !28 {
 entry:
     call void (...) @func()
     ret i32 0
@@ -45,7 +45,37 @@
 
 declare void @func(...) #1
 
-!2 = !{!"function_entry_count", i64 1}
+!llvm.module.flags = !{!1}
+
+!1 = !{i32 1, !"ProfileSummary", !2}
+!2 = !{!3, !4, !5, !6, !7, !8, !9, !10}
+!3 = !{!"ProfileFormat", !"InstrProf"}
+!4 = !{!"TotalCount", i64 2}
+!5 = !{!"MaxCount", i64 1}
+!6 = !{!"MaxInternalCount", i64 0}
+!7 = !{!"MaxFunctionCount", i64 1}
+!8 = !{!"NumCounts", i64 2}
+!9 = !{!"NumFunctions", i64 2}
+!10 = !{!"DetailedSummary", !11}
+!11 = !{!12, !13, !14, !15, !16, !17, !17, !18, !18, !19, !20, !21, !22, !23, !24, !25, !26, !27}
+!12 = !{i32 10000, i64 0, i32 0}
+!13 = !{i32 100000, i64 0, i32 0}
+!14 = !{i32 200000, i64 0, i32 0}
+!15 = !{i32 300000, i64 0, i32 0}
+!16 = !{i32 400000, i64 0, i32 0}
+!17 = !{i32 500000, i64 1, i32 2}
+!18 = !{i32 600000, i64 1, i32 2}
+!19 = !{i32 700000, i64 1, i32 2}
+!20 = !{i32 800000, i64 1, i32 2}
+!21 = !{i32 900000, i64 1, i32 2}
+!22 = !{i32 950000, i64 1, i32 2}
+!23 = !{i32 990000, i64 1, i32 2}
+!24 = !{i32 999000, i64 1, i32 2}
+!25 = !{i32 999900, i64 1, i32 2}
+!26 = !{i32 999990, i64 1, i32 2}
+!27 = !{i32 999999, i64 1, i32 2}
+!28 = !{!"function_entry_count", i64 1}
+
 
 ; OLD: Index {{.*}} contains 1 nodes (1 functions, 0 alias, 0 globals) and 1 edges (0 refs and 1 calls)
 ; OLD-COMBINED: Index {{.*}} contains 2 nodes (2 functions, 0 alias, 0 globals) and 1 edges (0 refs and 1 calls)
Index: test/Bitcode/thinlto-function-summary-callgraph-profile-summary.ll
===================================================================
--- test/Bitcode/thinlto-function-summary-callgraph-profile-summary.ll
+++ test/Bitcode/thinlto-function-summary-callgraph-profile-summary.ll
@@ -29,7 +29,7 @@
 ; CHECK-NEXT:    <VERSION
 ; CHECK-NEXT:    <VALUE_GUID op0=25 op1=123/>
 ; op4=hot1 op6=cold op8=hot2 op10=hot4 op12=none1 op14=hot3 op16=none2 op18=none3 op20=123
-; CHECK-NEXT:    <PERMODULE_PROFILE {{.*}} op4=1 op5=3 op6=5 op7=1 op8=2 op9=3 op10=4 op11=3 op12=6 op13=2 op14=3 op15=3 op16=7 op17=2 op18=8 op19=2 op20=25 op21=3/>
+; CHECK-NEXT:    <PERMODULE_PROFILE {{.*}} op4=1 op5=3 op6=5 op7=1 op8=2 op9=3 op10=4 op11=1 op12=6 op13=2 op14=3 op15=3 op16=7 op17=2 op18=8 op19=2 op20=25 op21=3/>
 ; CHECK-NEXT:  </GLOBALVAL_SUMMARY_BLOCK>
 
 ; CHECK: <STRTAB_BLOCK
Index: test/Bitcode/thinlto-function-summary-callgraph-sample-profile-summary.ll
===================================================================
--- test/Bitcode/thinlto-function-summary-callgraph-sample-profile-summary.ll
+++ test/Bitcode/thinlto-function-summary-callgraph-sample-profile-summary.ll
@@ -106,7 +106,7 @@
 
 !1 = !{i32 1, !"ProfileSummary", !2}
 !2 = !{!3, !4, !5, !6, !7, !8, !9, !10}
-!3 = !{!"ProfileFormat", !"InstrProf"}
+!3 = !{!"ProfileFormat", !"SampleProfile"}
 !4 = !{!"TotalCount", i64 10000}
 !5 = !{!"MaxCount", i64 10}
 !6 = !{!"MaxInternalCount", i64 1}
Index: test/Transforms/CodeGenPrepare/section-samplepgo.ll
===================================================================
--- test/Transforms/CodeGenPrepare/section-samplepgo.ll
+++ test/Transforms/CodeGenPrepare/section-samplepgo.ll
@@ -39,7 +39,7 @@
 !llvm.module.flags = !{!1}
 !1 = !{i32 1, !"ProfileSummary", !2}
 !2 = !{!3, !4, !5, !6, !7, !8, !9, !10}
-!3 = !{!"ProfileFormat", !"InstrProf"}
+!3 = !{!"ProfileFormat", !"SampleProfile"}
 !4 = !{!"TotalCount", i64 10000}
 !5 = !{!"MaxCount", i64 1000}
 !6 = !{!"MaxInternalCount", i64 1}
Index: test/Transforms/CodeGenPrepare/section.ll
===================================================================
--- test/Transforms/CodeGenPrepare/section.ll
+++ test/Transforms/CodeGenPrepare/section.ll
@@ -10,26 +10,26 @@
   ret void
 }
 
-; CHECK: hot_call_func{{.*}}!section_prefix ![[HOT_ID]]
-; The sum of 2 callsites are hot
-define void @hot_call_func() !prof !16 {
+; For instrumentation based PGO, we should only look at entry counts,
+; not call site VP metadata (which can exist on value profiled memcpy,
+; or possibly left behind after static analysis based devirtualization).
+; CHECK: cold_func1{{.*}}!section_prefix ![[COLD_ID:[0-9]+]]
+define void @cold_func1() !prof !16 {
   call void @hot_func(), !prof !17
   call void @hot_func(), !prof !17
   ret void
 }
 
-; CHECK-NOT: normal_func{{.*}}!section_prefix
-; The sum of all callsites are neither hot or cold
-define void @normal_func() !prof !16 {
+; CHECK: cold_func2{{.*}}!section_prefix
+define void @cold_func2() !prof !16 {
   call void @hot_func(), !prof !17
   call void @hot_func(), !prof !18
   call void @hot_func(), !prof !18
   ret void
 }
 
-; CHECK: cold_func{{.*}}!section_prefix ![[COLD_ID:[0-9]+]]
-; The entry and the callsite are both cold
-define void @cold_func() !prof !16 {
+; CHECK: cold_func3{{.*}}!section_prefix ![[COLD_ID]]
+define void @cold_func3() !prof !16 {
   call void @hot_func(), !prof !18
   ret void
 }
Index: test/Transforms/Inline/function-count-update-2.ll
===================================================================
--- test/Transforms/Inline/function-count-update-2.ll
+++ test/Transforms/Inline/function-count-update-2.ll
@@ -4,11 +4,11 @@
 ; has been inlined into a two callsites.
 
 ; CHECK: @callee() !prof [[COUNT:![0-9]+]]
-define i32 @callee() !prof !1 {
+define i32 @callee() !prof !15 {
   ret i32 0
 }
 
-define i32 @caller1() !prof !2 {
+define i32 @caller1() !prof !16 {
 ; CHECK-LABEL: @caller1
 ; CHECK-NOT: callee
 ; CHECK: ret
@@ -16,7 +16,7 @@
   ret i32 %i
 }
 
-define i32 @caller2() !prof !3 {
+define i32 @caller2() !prof !17 {
 ; CHECK-LABEL: @caller2
 ; CHECK-NOT: callee
 ; CHECK: ret
@@ -24,10 +24,24 @@
   ret i32 %i
 }
 
-!llvm.module.flags = !{!0}
+!llvm.module.flags = !{!1}
 ; CHECK: [[COUNT]] = !{!"function_entry_count", i64 0}
 !0 = !{i32 1, !"MaxFunctionCount", i32 1000}
-!1 = !{!"function_entry_count", i64 1000}
-!2 = !{!"function_entry_count", i64 600}
-!3 = !{!"function_entry_count", i64 400}
 
+!1 = !{i32 1, !"ProfileSummary", !2}
+!2 = !{!3, !4, !5, !6, !7, !8, !9, !10}
+!3 = !{!"ProfileFormat", !"InstrProf"}
+!4 = !{!"TotalCount", i64 2000}
+!5 = !{!"MaxCount", i64 1000}
+!6 = !{!"MaxInternalCount", i64 0}
+!7 = !{!"MaxFunctionCount", i64 1000}
+!8 = !{!"NumCounts", i64 3}
+!9 = !{!"NumFunctions", i64 3}
+!10 = !{!"DetailedSummary", !11}
+!11 = !{!12, !13, !14}
+!12 = !{i32 10000, i64 1000, i32 1}
+!13 = !{i32 999900, i64 600, i32 2}
+!14 = !{i32 999999, i64 600, i32 2}
+!15 = !{!"function_entry_count", i64 1000}
+!16 = !{!"function_entry_count", i64 600}
+!17 = !{!"function_entry_count", i64 400}
Index: test/Transforms/Inline/function-count-update-3.ll
===================================================================
--- test/Transforms/Inline/function-count-update-3.ll
+++ test/Transforms/Inline/function-count-update-3.ll
@@ -18,12 +18,12 @@
 
 @data = external global i32
 
-define i32 @a(i32 %a1) !prof !1 {
+define i32 @a(i32 %a1) !prof !15 {
   %a2 = call i32 @c(i32 %a1, i32 1)
   ret i32 %a2
 }
 
-define i32 @b(i32 %b1) !prof !2 {
+define i32 @b(i32 %b1) !prof !16 {
   %b2 = call i32 @c(i32 %b1, i32 %b1)
   ret i32 %b2
 }
@@ -31,7 +31,7 @@
 declare void @ext();
 
 ; CHECK: @c(i32 %c1, i32 %c100) !prof [[COUNT1:![0-9]+]]
-define i32 @c(i32 %c1, i32 %c100) !prof !3 {
+define i32 @c(i32 %c1, i32 %c100) !prof !17 {
   call void @ext()
   %cond = icmp sle i32 %c1, 1
   br i1 %cond, label %cond_true, label %cond_false
@@ -46,7 +46,7 @@
 
 
 ; CHECK: @e(i32 %c1) !prof [[COUNT2:![0-9]+]]
-define i32 @e(i32 %c1) !prof !4 {
+define i32 @e(i32 %c1) !prof !18 {
   %cond = icmp sle i32 %c1, 1
   br i1 %cond, label %cond_true, label %cond_false
 
@@ -67,12 +67,24 @@
   ret i32 0
 }
 
-!llvm.module.flags = !{!0}
+!llvm.module.flags = !{!1}
 ; CHECK: [[COUNT1]] = !{!"function_entry_count", i64 0}
 ; CHECK: [[COUNT2]] = !{!"function_entry_count", i64 400}
-!0 = !{i32 1, !"MaxFunctionCount", i32 5000}
-!1 = !{!"function_entry_count", i64 200}
-!2 = !{!"function_entry_count", i64 300}
-!3 = !{!"function_entry_count", i64 500}
-!4 = !{!"function_entry_count", i64 500}
-
+!1 = !{i32 1, !"ProfileSummary", !2}
+!2 = !{!3, !4, !5, !6, !7, !8, !9, !10}
+!3 = !{!"ProfileFormat", !"InstrProf"}
+!4 = !{!"TotalCount", i64 6500}
+!5 = !{!"MaxCount", i64 5000}
+!6 = !{!"MaxInternalCount", i64 0}
+!7 = !{!"MaxFunctionCount", i64 5000}
+!8 = !{!"NumCounts", i64 4}
+!9 = !{!"NumFunctions", i64 4}
+!10 = !{!"DetailedSummary", !11}
+!11 = !{!12, !13, !14}
+!12 = !{i32 10000, i64 5000, i32 1}
+!13 = !{i32 999900, i64 5000, i32 1}
+!14 = !{i32 999999, i64 5000, i32 1}
+!15 = !{!"function_entry_count", i64 200}
+!16 = !{!"function_entry_count", i64 300}
+!17 = !{!"function_entry_count", i64 500}
+!18 = !{!"function_entry_count", i64 500}
Index: test/Transforms/Inline/function-count-update.ll
===================================================================
--- test/Transforms/Inline/function-count-update.ll
+++ test/Transforms/Inline/function-count-update.ll
@@ -9,7 +9,7 @@
 ; containing the calls have a relative block frequency of 0.5).
 
 ; CHECK: @callee1(i32 %n) #0 !prof [[COUNT1:![0-9]+]]
-define i32 @callee1(i32 %n) #0 !prof !1 {
+define i32 @callee1(i32 %n) #0 !prof !15 {
   %cond = icmp sle i32 %n, 10
   br i1 %cond, label %cond_true, label %cond_false
 
@@ -22,12 +22,12 @@
 }
 
 ; CHECK: @callee2(i32 %n) #0 !prof [[COUNT2:![0-9]+]]
-define i32 @callee2(i32 %n) #0 !prof !2 {
+define i32 @callee2(i32 %n) #0 !prof !16 {
   %r1 = add i32 %n, 1
   ret i32 %r1
 }
 
-define i32 @caller(i32 %n) !prof !3 {
+define i32 @caller(i32 %n) !prof !17 {
   %cond = icmp sle i32 %n, 100
   br i1 %cond, label %cond_true, label %cond_false
 
@@ -39,12 +39,25 @@
   ret i32 0
 }
 
-!llvm.module.flags = !{!0}
+!llvm.module.flags = !{!1}
 ; CHECK: [[COUNT1]] = !{!"function_entry_count", i64 800}
 ; CHECK: [[COUNT2]] = !{!"function_entry_count", i64 1800}
-!0 = !{i32 1, !"MaxFunctionCount", i32 1000}
-!1 = !{!"function_entry_count", i64 1000}
-!2 = !{!"function_entry_count", i64 2000}
-!3 = !{!"function_entry_count", i64 400}
+!1 = !{i32 1, !"ProfileSummary", !2}
+!2 = !{!3, !4, !5, !6, !7, !8, !9, !10}
+!3 = !{!"ProfileFormat", !"InstrProf"}
+!4 = !{!"TotalCount", i64 3400}
+!5 = !{!"MaxCount", i64 2000}
+!6 = !{!"MaxInternalCount", i64 400}
+!7 = !{!"MaxFunctionCount", i64 2000}
+!8 = !{!"NumCounts", i64 3}
+!9 = !{!"NumFunctions", i64 3}
+!10 = !{!"DetailedSummary", !11}
+!11 = !{!12, !13, !14}
+!12 = !{i32 10000, i64 2000, i32 1}
+!13 = !{i32 999900, i64 1000, i32 2}
+!14 = !{i32 999999, i64 1000, i32 2}
+!15 = !{!"function_entry_count", i64 1000}
+!16 = !{!"function_entry_count", i64 2000}
+!17 = !{!"function_entry_count", i64 400}
 attributes #0 = { alwaysinline }
 
Index: test/Transforms/Inline/prof-update.ll
===================================================================
--- test/Transforms/Inline/prof-update.ll
+++ test/Transforms/Inline/prof-update.ll
@@ -6,21 +6,21 @@
 @func = global void ()* null
 
 ; CHECK: define void @callee(i32 %n) !prof ![[ENTRY_COUNT:[0-9]*]]
-define void  @callee(i32 %n) !prof !1 {
+define void  @callee(i32 %n) !prof !15 {
   %cond = icmp sle i32 %n, 10
   br i1 %cond, label %cond_true, label %cond_false
 cond_true:
 ; ext1 is optimized away, thus not updated.
 ; CHECK: call void @ext1(), !prof ![[COUNT_CALLEE1:[0-9]*]]
-  call void @ext1(), !prof !2
+  call void @ext1(), !prof !16
   ret void
 cond_false:
 ; ext is cloned and updated.
 ; CHECK: call void @ext(), !prof ![[COUNT_CALLEE:[0-9]*]]
-  call void @ext(), !prof !2
+  call void @ext(), !prof !16
   %f = load void ()*, void ()** @func
 ; CHECK: call void %f(), !prof ![[COUNT_IND_CALLEE:[0-9]*]] 
-  call void %f(), !prof !4
+  call void %f(), !prof !18
   ret void
 }
 
@@ -28,16 +28,29 @@
 define void @caller() {
 ; CHECK: call void @ext(), !prof ![[COUNT_CALLER:[0-9]*]]
 ; CHECK: call void %f.i(), !prof ![[COUNT_IND_CALLER:[0-9]*]]
-  call void @callee(i32 15), !prof !3
+  call void @callee(i32 15), !prof !17
   ret void
 }
 
-!llvm.module.flags = !{!0}
-!0 = !{i32 1, !"MaxFunctionCount", i32 2000}
-!1 = !{!"function_entry_count", i64 1000}
-!2 = !{!"branch_weights", i64 2000}
-!3 = !{!"branch_weights", i64 400}
-!4 = !{!"VP", i32 0, i64 140, i64 111, i64 80, i64 222, i64 40, i64 333, i64 20}
+!llvm.module.flags = !{!1}
+!1 = !{i32 1, !"ProfileSummary", !2}
+!2 = !{!3, !4, !5, !6, !7, !8, !9, !10}
+!3 = !{!"ProfileFormat", !"SampleProfile"}
+!4 = !{!"TotalCount", i64 10000}
+!5 = !{!"MaxCount", i64 10}
+!6 = !{!"MaxInternalCount", i64 1}
+!7 = !{!"MaxFunctionCount", i64 2000}
+!8 = !{!"NumCounts", i64 2}
+!9 = !{!"NumFunctions", i64 2}
+!10 = !{!"DetailedSummary", !11}
+!11 = !{!12, !13, !14}
+!12 = !{i32 10000, i64 100, i32 1}
+!13 = !{i32 999000, i64 100, i32 1}
+!14 = !{i32 999999, i64 1, i32 2}
+!15 = !{!"function_entry_count", i64 1000}
+!16 = !{!"branch_weights", i64 2000}
+!17 = !{!"branch_weights", i64 400}
+!18 = !{!"VP", i32 0, i64 140, i64 111, i64 80, i64 222, i64 40, i64 333, i64 20}
 attributes #0 = { alwaysinline }
 ; CHECK: ![[ENTRY_COUNT]] = !{!"function_entry_count", i64 600}
 ; CHECK: ![[COUNT_CALLEE1]] = !{!"branch_weights", i64 2000}
Index: unittests/Analysis/ProfileSummaryInfoTest.cpp
===================================================================
--- unittests/Analysis/ProfileSummaryInfoTest.cpp
+++ unittests/Analysis/ProfileSummaryInfoTest.cpp
@@ -162,6 +162,12 @@
 
   EXPECT_TRUE(PSI.isHotCallSite(CS1, &BFI));
   EXPECT_FALSE(PSI.isHotCallSite(CS2, &BFI));
+
+  // Test that adding an MD_prof metadata with a hot count on CS2 does not
+  // change its hotness as it has no effect in instrumented profiling.
+  MDBuilder MDB(M->getContext());
+  CI2->setMetadata(llvm::LLVMContext::MD_prof, MDB.createBranchWeights({400}));
+  EXPECT_FALSE(PSI.isHotCallSite(CS2, &BFI));
 }
 
 TEST_F(ProfileSummaryInfoTest, SampleProf) {