[llvm] r264587 - [PGO] Fix name encoding for ObjC-like functions

[Vedant Kumar via llvm-commits]

Author: vedantk
Date: Mon Mar 28 10:52:08 2016
New Revision: 264587

URL: http://llvm.org/viewvc/llvm-project?rev=264587&view=rev
Log:
[PGO] Fix name encoding for ObjC-like functions

Function names in ObjC can have spaces in them. This interacts poorly
with name compression, which uses spaces to separate PGO names. Fix the
issue by using a different separator and update a test.

I chose "\01" as the separator because 1) it's non-printable, 2) we
strip it from PGO names, and 3) it's the next natural choice once "\00"
is discarded (that one's overloaded).

Differential Revision: http://reviews.llvm.org/D18516

Modified:
    llvm/trunk/include/llvm/ProfileData/InstrProf.h
    llvm/trunk/lib/ProfileData/InstrProf.cpp
    llvm/trunk/unittests/ProfileData/InstrProfTest.cpp

Modified: llvm/trunk/include/llvm/ProfileData/InstrProf.h
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/include/llvm/ProfileData/InstrProf.h?rev=264587&r1=264586&r2=264587&view=diff
==============================================================================
--- llvm/trunk/include/llvm/ProfileData/InstrProf.h (original)
+++ llvm/trunk/include/llvm/ProfileData/InstrProf.h Mon Mar 28 10:52:08 2016
@@ -151,6 +151,9 @@ inline StringRef getInstrProfFileOverrid
   return "__llvm_profile_override_default_filename";
 }
 
+/// Return the marker used to separate PGO names during serialization.
+inline StringRef getInstrProfNameSeparator() { return "\01"; }
+
 /// Return the modified name for function \c F suitable to be
 /// used the key for profile lookup.
 std::string getPGOFuncName(const Function &F,

Modified: llvm/trunk/lib/ProfileData/InstrProf.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/ProfileData/InstrProf.cpp?rev=264587&r1=264586&r2=264587&view=diff
==============================================================================
--- llvm/trunk/lib/ProfileData/InstrProf.cpp (original)
+++ llvm/trunk/lib/ProfileData/InstrProf.cpp Mon Mar 28 10:52:08 2016
@@ -160,7 +160,7 @@ int collectPGOFuncNameStrings(const std:
                               bool doCompression, std::string &Result) {
   uint8_t Header[16], *P = Header;
   std::string UncompressedNameStrings =
-      join(NameStrs.begin(), NameStrs.end(), StringRef(" "));
+      join(NameStrs.begin(), NameStrs.end(), getInstrProfNameSeparator());
 
   unsigned EncLen = encodeULEB128(UncompressedNameStrings.length(), P);
   P += EncLen;
@@ -238,7 +238,7 @@ int readPGOFuncNameStrings(StringRef Nam
     }
     // Now parse the name strings.
     SmallVector<StringRef, 0> Names;
-    NameStrings.split(Names, ' ');
+    NameStrings.split(Names, getInstrProfNameSeparator());
     for (StringRef &Name : Names)
       Symtab.addFuncName(Name);
 

Modified: llvm/trunk/unittests/ProfileData/InstrProfTest.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/unittests/ProfileData/InstrProfTest.cpp?rev=264587&r1=264586&r2=264587&view=diff
==============================================================================
--- llvm/trunk/unittests/ProfileData/InstrProfTest.cpp (original)
+++ llvm/trunk/unittests/ProfileData/InstrProfTest.cpp Mon Mar 28 10:52:08 2016
@@ -893,7 +893,7 @@ TEST_P(MaybeSparseInstrProfTest, instr_p
     OS << "func_" << I;
     FuncNames1.push_back(OS.str());
     str.clear();
-    OS << "fooooooooooooooo_" << I;
+    OS << "f oooooooooooooo_" << I;
     FuncNames1.push_back(OS.str());
     str.clear();
     OS << "BAR_" << I;
@@ -931,7 +931,7 @@ TEST_P(MaybeSparseInstrProfTest, instr_p
       StringRef R = Symtab.getFuncName(IndexedInstrProf::ComputeHash(FuncNames1[0]));
       ASSERT_EQ(StringRef("func_0"), R);
       R = Symtab.getFuncName(IndexedInstrProf::ComputeHash(FuncNames1[1]));
-      ASSERT_EQ(StringRef("fooooooooooooooo_0"), R);
+      ASSERT_EQ(StringRef("f oooooooooooooo_0"), R);
       for (int I = 0; I < 3; I++) {
         std::string N[4];
         N[0] = FuncNames1[2 * I];