Teach the llvm-readobj COFF dumper to dump debug line tables from object files

[timurrrr]

Summary

[Extracted from D2232]

[timurrrr]

Fix unused variable warning on Linux by adding an extra check

[timurrrr]

Add some comments, checks and polish to the code

[timurrrr]

Refined some boundary checks

[echristo]

Not so fond of the formatting, but seems at least reasonable.

tools/llvm-readobj/COFFDumper.cpp

669

4? An enum maybe or something else?

694

Same here with F2, F3, F4...

787

Preceded? Awesome.

[timurrrr]

Do I read it correctly as "LGTM but please use enum here and here?"

It's probably reasonable to put the enum in some coff related header shared
between lib/AsmPrinter and lib/Tools
19 дек. 2013 г. 1:35 пользователь "Eric Christopher" <[e-mail]>
написал:

Not so fond of the formatting, but seems at least reasonable.

Comment at: tools/llvm-readobj/COFFDumper.cpp:669
@@ +668,3 @@
+ W.printHex("Magic", Magic);
+ if (Magic != 4) {

+ error(object_error::parse_failed);

4? An enum maybe or something else?

Comment at: tools/llvm-readobj/COFFDumper.cpp:694
@@ +693,3 @@
+ switch (SubSectionType) {
+ case 0xF2: {
+ // F2 is a PC to file:line table. Some data to parse this

subsection is

Same here with F2, F3, F4...

Comment at: tools/llvm-readobj/COFFDumper.cpp:787
@@ +786,3 @@
+ StringTable.data()[FilenameOffsetInF3 - 1] != '\0') {
+ // Each string in an F3 subsection should be preceded by a null

+ // character.

Preceded? Awesome.

http://llvm-reviews.chandlerc.com/D2425

[timurrrr]

Addressed the review comments

[timurrrr]

Going to land this now

tools/llvm-readobj/COFFDumper.cpp

669

Done

694

Good point - added an enum to include/llvm/Support/COFF.h

787

Yes, each string is both preceded (checked here) and succeeded (that's why we check for the EOL null) by a null character.
Two adjacent strings share the delimiting null character like so:
|0xF3|Size|0x0|String1|0x0|String2|0x0|...|0x0|StringLast|0x0|

[timurrrr]

r197674.

[timurrrr]

And thanks for the review!

test/tools/llvm-readobj/codeview-linetables.test

+RUN: llvm-readobj -s -codeview-linetables %p/Inputs/multifunction-linetables.obj.coff-2012-i368 \
+RUN:   | FileCheck %s -check-prefix MFUN32
+RUN: llvm-readobj -s -codeview-linetables %p/Inputs/multifunction-linetables.obj.coff-2012-x86_64 \
+RUN:   | FileCheck %s -check-prefix MFUN64
+RUN: llvm-readobj -s -codeview-linetables %p/Inputs/multifile-linetables.obj.coff-2012-i368 \
+RUN:   | FileCheck %s -check-prefix MFILE32
+RUN: llvm-readobj -s -codeview-linetables %p/Inputs/multifile-linetables.obj.coff-2012-x86_64 \
+RUN:   | FileCheck %s -check-prefix MFILE64
+
+MFUN32:      CodeViewLineTables [
+MFUN32-NEXT:   Magic: 0x4
+MFUN32-NEXT:   Subsection [
+MFUN32-NEXT:     Type: 0xF1
+MFUN32-NEXT:     PayloadSize: 0x52
+MFUN32:        ]
+MFUN32-NEXT:   Subsection [
+MFUN32-NEXT:     Type: 0xF5
+MFUN32-NEXT:     PayloadSize: 0x24
+MFUN32:        ]
+MFUN32-NEXT:   Subsection [
+MFUN32-NEXT:     Type: 0xF1
+MFUN32-NEXT:     PayloadSize: 0x4B
+MFUN32:        ]
+MFUN32-NEXT:   Subsection [
+MFUN32-NEXT:     Type: 0xF2
+MFUN32-NEXT:     PayloadSize: 0x30
+MFUN32:          FunctionName: _x
+MFUN32-NEXT:   ]
+MFUN32-NEXT:   Subsection [
+MFUN32-NEXT:     Type: 0xF5
+MFUN32-NEXT:     PayloadSize: 0x24
+MFUN32:        ]
+MFUN32-NEXT:   Subsection [
+MFUN32-NEXT:     Type: 0xF1
+MFUN32-NEXT:     PayloadSize: 0x4B
+MFUN32:        ]
+MFUN32-NEXT:   Subsection [
+MFUN32-NEXT:     Type: 0xF2
+MFUN32-NEXT:     PayloadSize: 0x30
+MFUN32:          FunctionName: _y
+MFUN32-NEXT:   ]
+MFUN32-NEXT:   Subsection [
+MFUN32-NEXT:     Type: 0xF5
+MFUN32-NEXT:     PayloadSize: 0x24
+MFUN32:        ]
+MFUN32-NEXT:   Subsection [
+MFUN32-NEXT:     Type: 0xF1
+MFUN32-NEXT:     PayloadSize: 0x4B
+MFUN32:        ]
+MFUN32-NEXT:   Subsection [
+MFUN32-NEXT:     Type: 0xF2
+MFUN32-NEXT:     PayloadSize: 0x40
+MFUN32:          FunctionName: _f
+MFUN32-NEXT:   ]
+MFUN32-NEXT:   Subsection [
+MFUN32-NEXT:     Type: 0xF4
+MFUN32-NEXT:     PayloadSize: 0x18
+MFUN32:        ]
+MFUN32-NEXT:   Subsection [
+MFUN32-NEXT:     Type: 0xF3
+MFUN32-NEXT:     PayloadSize: 0x46
+MFUN32:        ]
+MFUN32-NEXT:   Subsection [
+MFUN32-NEXT:     Type: 0xF1
+MFUN32-NEXT:     PayloadSize: 0x8
+MFUN32:        ]
+MFUN32-NEXT:   FunctionLineTable [
+MFUN32-NEXT:     FunctionName: _x
+MFUN32-NEXT:     CodeSize: 0xA
+MFUN32-NEXT:     FilenameSegment [
+MFUN32-NEXT:       Filename: d:\source.c
+MFUN32-NEXT:       +0x0: 3
+MFUN32-NEXT:       +0x3: 4
+MFUN32-NEXT:       +0x8: 5
+MFUN32-NEXT:     ]
+MFUN32-NEXT:   ]
+MFUN32-NEXT:   FunctionLineTable [
+MFUN32-NEXT:     FunctionName: _y
+MFUN32-NEXT:     CodeSize: 0xA
+MFUN32-NEXT:     FilenameSegment [
+MFUN32-NEXT:       Filename: d:\source.c
+MFUN32-NEXT:       +0x0: 7
+MFUN32-NEXT:       +0x3: 8
+MFUN32-NEXT:       +0x8: 9
+MFUN32-NEXT:     ]
+MFUN32-NEXT:   ]
+MFUN32-NEXT:   FunctionLineTable [
+MFUN32-NEXT:     FunctionName: _f
+MFUN32-NEXT:     CodeSize: 0x14
+MFUN32-NEXT:     FilenameSegment [
+MFUN32-NEXT:       Filename: d:\source.c
+MFUN32-NEXT:       +0x0: 11
+MFUN32-NEXT:       +0x3: 12
+MFUN32-NEXT:       +0x8: 13
+MFUN32-NEXT:       +0xD: 14
+MFUN32-NEXT:       +0x12: 15
+MFUN32-NEXT:     ]
+MFUN32-NEXT:   ]
+MFUN32-NEXT: ]
+
+MFUN64:      CodeViewLineTables [
+MFUN64-NEXT:   Magic: 0x4
+MFUN64-NEXT:   Subsection [
+MFUN64-NEXT:     Type: 0xF1
+MFUN64-NEXT:     PayloadSize: 0x52
+MFUN64:        ]
+MFUN64-NEXT:   Subsection [
+MFUN64-NEXT:     Type: 0xF1
+MFUN64-NEXT:     PayloadSize: 0x4B
+MFUN64:        ]
+MFUN64-NEXT:   Subsection [
+MFUN64-NEXT:     Type: 0xF2
+MFUN64-NEXT:     PayloadSize: 0x30
+MFUN64:          FunctionName: x
+MFUN64-NEXT:   ]
+MFUN64-NEXT:   Subsection [
+MFUN64-NEXT:     Type: 0xF1
+MFUN64-NEXT:     PayloadSize: 0x4B
+MFUN64:        ]
+MFUN64-NEXT:   Subsection [
+MFUN64-NEXT:     Type: 0xF2
+MFUN64-NEXT:     PayloadSize: 0x30
+MFUN64:          FunctionName: y
+MFUN64-NEXT:   ]
+MFUN64-NEXT:   Subsection [
+MFUN64-NEXT:     Type: 0xF1
+MFUN64-NEXT:     PayloadSize: 0x4B
+MFUN64:        ]
+MFUN64-NEXT:   Subsection [
+MFUN64-NEXT:     Type: 0xF2
+MFUN64-NEXT:     PayloadSize: 0x40
+MFUN64:          FunctionName: f
+MFUN64-NEXT:   ]
+MFUN64-NEXT:   Subsection [
+MFUN64-NEXT:     Type: 0xF4
+MFUN64-NEXT:     PayloadSize: 0x18
+MFUN64:        ]
+MFUN64-NEXT:   Subsection [
+MFUN64-NEXT:     Type: 0xF3
+MFUN64-NEXT:     PayloadSize: 0xD
+MFUN64:        ]
+MFUN64-NEXT:   Subsection [
+MFUN64-NEXT:     Type: 0xF1
+MFUN64-NEXT:     PayloadSize: 0x8
+MFUN64:        ]
+MFUN64-NEXT:   FunctionLineTable [
+MFUN64-NEXT:     FunctionName: x
+MFUN64-NEXT:     CodeSize: 0xE
+MFUN64-NEXT:     FilenameSegment [
+MFUN64-NEXT:       Filename: d:\source.c
+MFUN64-NEXT:       +0x0: 3
+MFUN64-NEXT:       +0x4: 4
+MFUN64-NEXT:       +0x9: 5
+MFUN64-NEXT:     ]
+MFUN64-NEXT:   ]
+MFUN64-NEXT:   FunctionLineTable [
+MFUN64-NEXT:     FunctionName: y
+MFUN64-NEXT:     CodeSize: 0xE
+MFUN64-NEXT:     FilenameSegment [
+MFUN64-NEXT:       Filename: d:\source.c
+MFUN64-NEXT:       +0x0: 7
+MFUN64-NEXT:       +0x4: 8
+MFUN64-NEXT:       +0x9: 9
+MFUN64-NEXT:     ]
+MFUN64-NEXT:   ]
+MFUN64-NEXT:   FunctionLineTable [
+MFUN64-NEXT:     FunctionName: f
+MFUN64-NEXT:     CodeSize: 0x18
+MFUN64-NEXT:     FilenameSegment [
+MFUN64-NEXT:       Filename: d:\source.c
+MFUN64-NEXT:       +0x0: 11
+MFUN64-NEXT:       +0x4: 12
+MFUN64-NEXT:       +0x9: 13
+MFUN64-NEXT:       +0xE: 14
+MFUN64-NEXT:       +0x13: 15
+MFUN64-NEXT:     ]
+MFUN64-NEXT:   ]
+MFUN64-NEXT: ]
+
+MFILE32:      CodeViewLineTables [
+MFILE32-NEXT:   Magic: 0x4
+MFILE32-NEXT:   Subsection [
+MFILE32-NEXT:     Type: 0xF1
+MFILE32-NEXT:     PayloadSize: 0x51
+MFILE32:        ]
+MFILE32-NEXT:   Subsection [
+MFILE32-NEXT:     Type: 0xF5
+MFILE32-NEXT:     PayloadSize: 0x24
+MFILE32:        ]
+MFILE32-NEXT:   Subsection [
+MFILE32-NEXT:     Type: 0xF1
+MFILE32-NEXT:     PayloadSize: 0x4B
+MFILE32:        ]
+MFILE32-NEXT:   Subsection [
+MFILE32-NEXT:     Type: 0xF2
+MFILE32-NEXT:     PayloadSize: 0x64
+MFILE32:          FunctionName: _f
+MFILE32-NEXT:   ]
+MFILE32-NEXT:   Subsection [
+MFILE32-NEXT:     Type: 0xF4
+MFILE32-NEXT:     PayloadSize: 0x28
+MFILE32:        ]
+MFILE32-NEXT:   Subsection [
+MFILE32-NEXT:     Type: 0xF3
+MFILE32-NEXT:     PayloadSize: 0x57
+MFILE32:        ]
+MFILE32-NEXT:   Subsection [
+MFILE32-NEXT:     Type: 0xF1
+MFILE32-NEXT:     PayloadSize: 0x8
+MFILE32:        ]
+MFILE32-NEXT:   FunctionLineTable [
+MFILE32-NEXT:     FunctionName: _f
+MFILE32-NEXT:     CodeSize: 0x14
+MFILE32-NEXT:     FilenameSegment [
+MFILE32-NEXT:       Filename: d:\input.c
+MFILE32-NEXT:       +0x0: 3
+MFILE32-NEXT:     ]
+MFILE32-NEXT:     FilenameSegment [
+MFILE32-NEXT:       Filename: d:\one.c
+MFILE32-NEXT:       +0x3: 1
+MFILE32-NEXT:     ]
+MFILE32-NEXT:     FilenameSegment [
+MFILE32-NEXT:       Filename: d:\two.c
+MFILE32-NEXT:       +0x8: 2
+MFILE32-NEXT:     ]
+MFILE32-NEXT:     FilenameSegment [
+MFILE32-NEXT:       Filename: d:\one.c
+MFILE32-NEXT:       +0xD: 7
+MFILE32-NEXT:       +0x12: 8
+MFILE32-NEXT:     ]
+MFILE32-NEXT:   ]
+MFILE32-NEXT: ]
+
+MFILE64:      CodeViewLineTables [
+MFILE64-NEXT:   Magic: 0x4
+MFILE64-NEXT:   Subsection [
+MFILE64-NEXT:     Type: 0xF1
+MFILE64-NEXT:     PayloadSize: 0x51
+MFILE64:        ]
+MFILE64-NEXT:   Subsection [
+MFILE64-NEXT:     Type: 0xF1
+MFILE64-NEXT:     PayloadSize: 0x4B
+MFILE64:        ]
+MFILE64-NEXT:   Subsection [
+MFILE64-NEXT:     Type: 0xF2
+MFILE64-NEXT:     PayloadSize: 0x64
+MFILE64:          FunctionName: f
+MFILE64-NEXT:   ]
+MFILE64-NEXT:   Subsection [
+MFILE64-NEXT:     Type: 0xF4
+MFILE64-NEXT:     PayloadSize: 0x28
+MFILE64:        ]
+MFILE64-NEXT:   Subsection [
+MFILE64-NEXT:     Type: 0xF3
+MFILE64-NEXT:     PayloadSize: 0x1E
+MFILE64:        ]
+MFILE64-NEXT:   Subsection [
+MFILE64-NEXT:     Type: 0xF1
+MFILE64-NEXT:     PayloadSize: 0x8
+MFILE64:        ]
+MFILE64-NEXT:   FunctionLineTable [
+MFILE64-NEXT:     FunctionName: f
+MFILE64-NEXT:     CodeSize: 0x18
+MFILE64-NEXT:     FilenameSegment [
+MFILE64-NEXT:       Filename: d:\input.c
+MFILE64-NEXT:       +0x0: 3
+MFILE64-NEXT:     ]
+MFILE64-NEXT:     FilenameSegment [
+MFILE64-NEXT:       Filename: d:\one.c
+MFILE64-NEXT:       +0x4: 1
+MFILE64-NEXT:     ]
+MFILE64-NEXT:     FilenameSegment [
+MFILE64-NEXT:       Filename: d:\two.c
+MFILE64-NEXT:       +0x9: 2
+MFILE64-NEXT:     ]
+MFILE64-NEXT:     FilenameSegment [
+MFILE64-NEXT:       Filename: d:\one.c
+MFILE64-NEXT:       +0xE: 7
+MFILE64-NEXT:       +0x13: 8
+MFILE64-NEXT:     ]
+MFILE64-NEXT:   ]
+MFILE64-NEXT: ]

tools/llvm-readobj/COFFDumper.cpp

 #include "StreamWriter.h"
 
 #include "llvm/ADT/DenseMap.h"
 #include "llvm/ADT/SmallString.h"
 #include "llvm/Object/COFF.h"
 #include "llvm/Object/ObjectFile.h"
 #include "llvm/Support/Casting.h"
 #include "llvm/Support/Compiler.h"
+#include "llvm/Support/DataExtractor.h"
 #include "llvm/Support/Format.h"
 #include "llvm/Support/SourceMgr.h"
 #include "llvm/Support/Win64EH.h"
 #include "llvm/Support/raw_ostream.h"
 #include "llvm/Support/system_error.h"
 
 #include <algorithm>
 #include <cstring>
private:
 
   void printUnwindInfo(
     const Win64EH::UnwindInfo& UI,
     uint64_t OffsetInSection,
     const std::vector<RelocationRef> &Rels);
 
   void printUnwindCode(const Win64EH::UnwindInfo& UI, ArrayRef<UnwindCode> UCs);
 
+  void printCodeViewLineTables(section_iterator SecI);
+
   void cacheRelocations();
 
   error_code getSectionContents(
     const std::vector<RelocationRef> &Rels,
     uint64_t Offset,
     ArrayRef<uint8_t> &Contents,
     uint64_t &Addr);
 
if (PEHeader->NumberOfRvaAndSize > 0) {
 
       for (uint32_t i = 0; i < PEHeader->NumberOfRvaAndSize; ++i) {
         printDataDirectory(i, directory[i]);
       }
     }
   }
 }
 
+void COFFDumper::printCodeViewLineTables(section_iterator SecI) {
+  StringRef Data;
+  if (error(SecI->getContents(Data))) return;
+
+  SmallVector<StringRef, 10> FunctionNames;
+  StringMap<StringRef> FunctionLineTables;
+  StringRef FileIndexToStringOffsetTable;
+  StringRef StringTable;
+
+  ListScope D(W, "CodeViewLineTables");
+  {
+    DataExtractor DE(Data, true, 4);
+    uint32_t Offset = 0;
+
+    W.printHex("Magic", DE.getU32(&Offset));
+
+    bool Finished = false;
+    while (DE.isValidOffset(Offset) && !Finished) {
+      uint32_t SubSectionType = DE.getU32(&Offset),
+               PayloadSize = DE.getU32(&Offset);
+      ListScope S(W, "Subsection");
+      W.printHex("Type", SubSectionType);
+      W.printHex("PayloadSize", PayloadSize);
+      if (PayloadSize > Data.size() - Offset) {
+        error(object_error::parse_failed);
+        return;
+      }
+      StringRef Contents = Data.substr(Offset, PayloadSize);
+      W.printBinaryBlock("Contents", Contents);
+
+      switch (SubSectionType) {
+      case 0xF2: {
+        // There should be at least three words to store two function
+        // relocations and size of the code.
+        if (PayloadSize < 12) {
+          error(object_error::parse_failed);
+          return;
+        }
+
+        StringRef FunctionName;
+        if (error(resolveSymbolName(RelocMap[Obj->getCOFFSection(SecI)], Offset,
+                                    FunctionName)))
+          return;
+        W.printString("FunctionName", FunctionName);
+        if (FunctionLineTables.count(FunctionName) != 0) {
+          // Saw debug info for this function already?
+          error(object_error::parse_failed);
+          return;
+        }
+
+        FunctionLineTables[FunctionName] = Contents;
+        FunctionNames.push_back(FunctionName);
+        break;
+      }
+      case 0xF4:
+        if (PayloadSize == 0 || FileIndexToStringOffsetTable.data() != 0) {
+          // Empty or duplicate 0xF4 subsection.
+          error(object_error::parse_failed);
+          return;
+        }
+        FileIndexToStringOffsetTable = Contents;
+        break;
+      case 0xF3:
+        if (PayloadSize == 0 || StringTable.data() != 0 ||
+            Contents.back() != '\0') {
+          // Empty or duplicate or non-null-terminated 0xF3 subsection.
+          error(object_error::parse_failed);
+          return;
+        }
+        StringTable = Contents;
+        break;
+      }
+      Offset += PayloadSize;
+
+      // Align the reading pointer by 4.
+      Offset += (-Offset) % 4;
+    }
+  }
+
+  for (unsigned I = 0, E = FunctionNames.size(); I != E; ++I) {
+    StringRef Name = FunctionNames[I];
+    ListScope S(W, "FunctionLineTable");
+    W.printString("FunctionName", Name);
+
+    DataExtractor DE(FunctionLineTables[Name], true, 4);
+    uint32_t Offset = 8;  // Skip relocations
+    uint32_t FunctionSize = DE.getU32(&Offset);
+    W.printHex("CodeSize", FunctionSize);
+    while (DE.isValidOffset(Offset)) {
+      uint32_t OffsetInF4      = DE.getU32(&Offset),
+               SegmentLength   = DE.getU32(&Offset),
+               FullSegmentSize = DE.getU32(&Offset);
+      if (FullSegmentSize != 12 + 8 * SegmentLength) {
+        error(object_error::parse_failed);
+        return;
+      }
+
+      uint32_t OffsetInF3;
+      {
+        DataExtractor SDE(FileIndexToStringOffsetTable, true, 4);
+        uint32_t OffsetInSDE = OffsetInF4;
+        if (!SDE.isValidOffset(OffsetInSDE)) {
+          error(object_error::parse_failed);
+          return;
+        }
+        OffsetInF3 = SDE.getU32(&OffsetInSDE);
+      }
+
+      if (OffsetInF3 == 0 || StringTable.data()[OffsetInF3 - 1] != '\0') {
+        // Each string in an F3 subsection should be preceded by a null
+        // character.
+        error(object_error::parse_failed);
+        return;
+      }
+
+      StringRef Filename(StringTable.data() + OffsetInF3);
+      ListScope S(W, "FilenameSegment");
+      W.printString("Filename", Filename);
+      for (unsigned J = 0; J != SegmentLength; ++J) {
+        uint32_t PC = DE.getU32(&Offset),
+                 LineNumber = DE.getU32(&Offset) & 0x7fffffff;
+        char Buffer[32];
+        format("+0x%X", PC).snprint(Buffer, 32);
+        W.printNumber(Buffer, LineNumber);
+      }
+    }
+  }
+}
+
 void COFFDumper::printSections() {
   error_code EC;
 
   ListScope SectionsD(W, "Sections");
   int SectionNumber = 0;
   for (section_iterator SecI = Obj->begin_sections(),
                         SecE = Obj->end_sections();
                         SecI != SecE; SecI.increment(EC)) {
     if (error(EC))
       break;
 
     ++SectionNumber;
     const coff_section *Section = Obj->getCOFFSection(SecI);
if (opts::SectionSymbols) {
         bool Contained = false;
         if (SecI->containsSymbol(*SymI, Contained) || !Contained)
           continue;
 
         printSymbol(SymI);
       }
     }
 
+    if (Name == ".debug$S" && opts::CodeViewLineTables)
+      printCodeViewLineTables(SecI);
+
     if (opts::SectionData) {
       StringRef Data;
       if (error(SecI->getContents(Data))) break;
 
       W.printBinaryBlock("SectionData", Data);
     }
   }
 }

tools/llvm-readobj/llvm-readobj.h

namespace opts {
   extern llvm::cl::opt<bool> SectionRelocations;
   extern llvm::cl::opt<bool> SectionSymbols;
   extern llvm::cl::opt<bool> SectionData;
   extern llvm::cl::opt<bool> Relocations;
   extern llvm::cl::opt<bool> Symbols;
   extern llvm::cl::opt<bool> DynamicSymbols;
   extern llvm::cl::opt<bool> UnwindInfo;
   extern llvm::cl::opt<bool> ExpandRelocs;
+  extern llvm::cl::opt<bool> CodeViewLineTables;
 } // namespace opts
 
 #define LLVM_READOBJ_ENUM_ENT(ns, enum) \
   { #enum, ns::enum }
 
 #endif

tools/llvm-readobj/llvm-readobj.cpp

namespace opts {
 
   // -program-headers
   cl::opt<bool> ProgramHeaders("program-headers",
     cl::desc("Display ELF program headers"));
 
   // -expand-relocs
   cl::opt<bool> ExpandRelocs("expand-relocs",
     cl::desc("Expand each shown relocation to multiple lines"));
+
+  // -codeview-linetables
+  cl::opt<bool> CodeViewLineTables("codeview-linetables",
+    cl::desc("Display CodeView line table information"));
 } // namespace opts
 
 static int ReturnValue = EXIT_SUCCESS;
 
 namespace llvm {
 
 bool error(error_code EC) {
   if (!EC)