[llvm-profgen] Ignore the whole trace with the leading external branch

[wlei]

Summary

The first LBR entry can be an external branch, we should ignore the whole trace.

7f7448e889e4 0x7f7448e889e4/0x7f7448e88826/P/-/-/1  0x7f7448e8899f/0x7f7448e889d8/P/-/-/4  ...

[hoy]

Good catch, thx.

[wenlei]

llvm/tools/llvm-profgen/PerfReader.cpp

501

This means we always ignore any external branches, and we will no longer generate any artificial branches for internal->external->internal transition?

[wlei]

Updating D111749: [llvm-profgen] Also ignore the first external branch

[wenlei]

llvm/tools/llvm-profgen/PerfReader.cpp

493

As we discussed, we shouldn't relax this for CS profile - it can't handle leading external address during unwinding.

[wlei]

1. revert back to preivous version, ignore the whole trace
2. add test case for artificial branch

[wenlei]

llvm/test/tools/llvm-profgen/inline-noprobe2.test

1–4

can we avoid having test depending on debug prints as much as we can? can check the final result only in this case.

llvm/tools/llvm-profgen/PerfReader.cpp

495

nit: with the break at the end, it's probably more readable if we separate the continue by replacing the else if with if.

[wlei]

llvm/test/tools/llvm-profgen/inline-noprobe2.test

1–4

I see, the final result is enough for this case.

llvm/tools/llvm-profgen/PerfReader.cpp

495

Good point!

[wlei]

Updating D111749: [llvm-profgen] Also ignore the first external branch

[wenlei]

lgtm, thanks.

llvm/test/tools/llvm-profgen/Inputs/artificial-branch.perfscript

+           400870 0x400870/0x400540/P/-/-/1  0x4008bf/0x400870/P/-/-/7  0x7f7448e88cc1/0x400875/P/-/-/1  0x7f7448e88875/0x7f7448e88cbd/P/-/-/2  0x7f7448e88840/0x7f7448e8885c/P/-/-/1  0x7f7448e8882d/0x7f7448e8883a/P/-/-/1  0x7f7448e889e4/0x7f7448e88826/P/-/-/1  0x7f7448e8899f/0x7f7448e889d8/P/-/-/4  0x7f7448e88821/0x7f7448e88960/P/-/-/1  0x7f7448e887f9/0x7f7448e88815/P/-/-/2  0x7f7448e887e4/0x7f7448e887f2/P/-/-/2  0x7f7448e88cb8/0x7f7448e887c0/P/-/-/1  0x400540/0x7f7448e88cb0/P/-/-/1  0x400870/0x400540/P/-/-/1  0x4008bf/0x400870/P/-/-/5

llvm/test/tools/llvm-profgen/inline-noprobe2.test

-
+; RUN: llvm-profgen --format=text --perfscript=%S/Inputs/artificial-branch.perfscript --binary=%S/Inputs/inline-noprobe2.perfbin --output=%t  --skip-symbolization --use-offset=0
+; RUN: FileCheck %s --input-file %t --check-prefix=CHECK-ARTIFICIAL-BRANCH
+; RUN: llvm-profgen --format=text --perfscript=%S/Inputs/inline-noprobe2.perfscript --binary=%S/Inputs/inline-noprobe2.perfbin --output=%t  --skip-symbolization --use-offset=0
+; RUN: FileCheck %s --input-file %t --check-prefix=CHECK-RAW-PROFILE
 ; RUN: llvm-profgen --format=text --perfscript=%S/Inputs/inline-noprobe2.perfscript --binary=%S/Inputs/inline-noprobe2.perfbin --output=%t
 ; RUN: FileCheck %s --input-file %t --check-prefix=CHECK
 
 ; RUN: llvm-profgen --format=extbinary --perfscript=%S/Inputs/inline-noprobe2.perfscript --binary=%S/Inputs/inline-noprobe2.perfbin --output=%t
 ; RUN: llvm-profdata show -show-prof-sym-list -sample %t | FileCheck %s --check-prefix=CHECK-SYM-LIST
 
+; CHECK-ARTIFICIAL-BRANCH: 3
+; CHECK-ARTIFICIAL-BRANCH: 400540-400540:1
+; CHECK-ARTIFICIAL-BRANCH: 400870-400870:2
+; CHECK-ARTIFICIAL-BRANCH: 400875-4008bf:1
+; CHECK-ARTIFICIAL-BRANCH: 2
+; CHECK-ARTIFICIAL-BRANCH: 400870->400540:2
+; CHECK-ARTIFICIAL-BRANCH: 4008bf->400870:2
+
 ; CHECK-SYM-LIST: Dump profile symbol list
 ; CHECK-SYM-LIST: main
 ; CHECK-SYM-LIST: partition_pivot_first
 ; CHECK-SYM-LIST: partition_pivot_last
 ; CHECK-SYM-LIST: quick_sort
 
+
+;CHECK-RAW-PROFILE-NOT: 7f7448e889e4
+;CHECK-RAW-PROFILE-NOT: 7f7448e88826
+
 ;CHECK:       partition_pivot_first:1045:5
 ;CHECK-NEXT:  0: 5
 ;CHECK-NEXT:  1: 5
 ;CHECK-NEXT:  2: 5
 ;CHECK-NEXT:  3: 5
 ;CHECK-NEXT:  3.1: 83
 ;CHECK-NEXT:  4: 82
 ;CHECK-NEXT:  4.1: 26

llvm/tools/llvm-profgen/PerfReader.cpp

while (Index < Records.size()) {
     bool IsIncoming = !SrcIsInternal && DstIsInternal;
     bool IsOutgoing = SrcIsInternal && !DstIsInternal;
     bool IsArtificial = false;
 
     // Ignore branches outside the current binary. Ignore all remaining branches
     // if there's no incoming branch before the external branch in reverse
     // order.
     if (IsExternal) {
       if (PrevTrDst)
         continue;
-      else if (!LBRStack.empty()) {
+      if (!LBRStack.empty()) {
         WithColor::warning()
             << "Invalid transfer to external code in LBR record at line "
             << TraceIt.getLineNumber() << ": " << TraceIt.getCurrentLine()
             << "\n";
-        break;
       }
+      break;
     }
 
     if (IsOutgoing) {
       if (!PrevTrDst) {
         // This is unpaired outgoing jump which is likely due to interrupt or
         // incomplete LBR trace. Ignore current and subsequent entries since
         // they are likely in different contexts.
         break;