Introduce the llvm-cfi-verify tool.

[hctim]

Summary

Introduces the llvm-cfi-verify tool to llvm. Includes the design document (docs/CFIVerify.rst). Current implementation of the tool is simply a disassembler that identifies and prints the indirect control flow instructions.

[vlad.tsyrklevich]

docs/CFIVerify.rst

38

ensures

56

nit: bytecode is traditionally used to refer to instructions run by a VM/interpreter, replace with 'machine code' instead

tools/llvm-cfi-verify/llvm-cfi-verify.cpp

2

Run clang-format

213

errs()

236

This can assert() on a bad opcode, should we check if (BadInstruction) before this?

244

I believe you can use a range-based for loop here: for (auto &Op : Instruction)

[hctim]

Updates from Vlad's comments.

tools/llvm-cfi-verify/llvm-cfi-verify.cpp

244

Thanks for the catch, had a look at the source for a range-iterator from ::Operands.begin() to ::Operands.end() under the name "operands()" but didn't realise that they were simply under the object itself.

[vlad.tsyrklevich]

tools/llvm-cfi-verify/llvm-cfi-verify.cpp

154

Could we change these messages to 'Failed to initialize MCRegisterInfo' instead?

[kcc]

docs/CFIVerify.rst

66

I thought the call should always be fallthrough. Did you see the opposite case?

[hctim]

Changed error reporting text slightly. Single line if-else statements now no longer have braces. Utilised commit in D37719 to remove helper function for Triple deduction from an ObjectFile.

[hctim]

docs/CFIVerify.rst

66

Fallthrough-to-ud2 is currently present. Will write some code to evaluate this case and provide report.

docs/CFIVerify.rst

+==============================================
+Control Flow Verification Tool Design Document
+==============================================
+
+Objective
+=========
+
+This document provides an overview of an external tool to verify the protection
+mechanisms implemented by Clang's *Control Flow Integrity* (CFI) schemes
+(``-fsanitize=cfi``). This tool, provided a binary or DSO, should infer whether
+indirect control flow operations are protected by CFI, and should output these
+results in a human-readable form.
+
+This tool should also be added as part of Clang's continuous integration testing
+framework, where modifications to the compiler ensure that CFI protection
+schemes are still present in the final binary.
+
+Location
+========
+
+This tool will be present as a part of the LLVM toolchain, and will reside in
+the "/llvm/tools/llvm-cfi-verify" directory, relative to the LLVM trunk. It will
+be tested in two methods:
+
+- Unit tests to validate code sections, present in "/llvm/unittests/llvm-cfi-
+  verify".
+- Integration tests, present in "/llvm/tools/clang/test/LLVMCFIVerify". These
+  integration tests are part of clang as part of a continuous integration
+  framework, ensuring updates to the compiler that reduce CFI coverage on
+  indirect control flow instructions are identified.
+
+Background
+==========
+
+This tool will continuously validate that CFI directives are properly
+implemented around all indirect control flows by analysing the output machine
+code. The analysis of machine code is important as it ensures that any bugs
+present in linker or compiler do not subvert CFI protections in the final
+shipped binary.
+
+Unprotected indirect control flow instructions will be flagged for manual
+review. These unexpected control flows may simply have not been accounted for in
+the compiler implementation of CFI (e.g. indirect jumps to facilitate switch
+statements may not be fully protected).
+
+It may be possible in the future to extend this tool to flag unnecessary CFI
+directives (e.g. CFI directives around a static call to a non-polymorphic base
+type). This type of directive has no security implications, but may present
+performance impacts.
+
+Design Ideas
+============
+
+This tool will disassemble binaries and DSO's from their machine code format and
+analyse the disassembled machine code. The tool will inspect virtual calls and
+indirect function calls. This tool will also inspect indirect jumps, as inlined
+functions and jump tables should also be subject to CFI protections. Non-virtual
+calls (``-fsanitize=cfi-nvcall``) and cast checks (``-fsanitize=cfi-*cast*``)
+are not implemented due to a lack of information provided by the bytecode.
+
+The tool would operate by searching for indirect control flow instructions in
+the disassembly. A control flow graph would be generated from a small buffer of
+the instructions surrounding the 'target' control flow instruction. If the
+target instruction is branched-to, the fallthrough of the branch should be the
+CFI trap (on x86, this is a ``ud2`` instruction). If the target instruction is
+the fallthrough (i.e. immediately succeeds) of a conditional jump, the
+conditional jump target should be the CFI trap. If an indirect control flow
+instruction does not conform to one of these formats, the target will be noted
+as being CFI-unprotected.
+
+Note that in the second case outlined above (where the target instruction is the
+fallthrough of a conditional jump), if the target represents a vcall that takes
+arguments, these arguments may be pushed to the stack after the branch but
+before the target instruction. In these cases, a secondary 'spill graph' in
+constructed, to ensure the register argument used by the indirect jump/call is
+not spilled from the stack at any point in the interim period. If there are no
+spills that affect the target register, the target is marked as CFI-protected.
+
+Other Design Notes
+~~~~~~~~~~~~~~~~~~
+
+Only machine code sections that are marked as executable will be subject to this
+analysis. Non-executable sections do not require analysis as any execution
+present in these sections has already violated the control flow integrity.
+
+Suitable extensions may be made at a later date to include anaylsis for indirect
+control flow operations across DSO boundaries. Currently, these CFI features are
+only experimental with an unstable ABI, making them unsuitable for analysis.

tools/LLVMBuild.txt

  bugpoint
  dsymutil
  llc
  lli
  llvm-ar
  llvm-as
  llvm-bcanalyzer
  llvm-cat
+ llvm-cfi-verify
  llvm-cov
  llvm-cvtres
  llvm-diff
  llvm-dis
  llvm-dwarfdump
  llvm-dwp
  llvm-extract
  llvm-jitlistener

tools/llvm-cfi-verify/CMakeLists.txt

+set(LLVM_LINK_COMPONENTS
+  AllTargetsAsmPrinters
+  AllTargetsAsmParsers
+  AllTargetsDescs
+  AllTargetsDisassemblers
+  AllTargetsInfos
+  MC
+  MCParser
+  Support
+  )
+
+add_llvm_tool(llvm-cfi-verify
+  llvm-cfi-verify.cpp
+  )

tools/llvm-cfi-verify/LLVMBuild.txt

+;===- ./tools/llvm-cfi-verify/LLVMBuild.txt --------------------*- Conf -*--===;
+;
+;                     The LLVM Compiler Infrastructure
+;
+; This file is distributed under the University of Illinois Open Source
+; License. See LICENSE.TXT for details.
+;
+;===------------------------------------------------------------------------===;
+;
+; This is an LLVMBuild description file for the components in this subdirectory.
+;
+; For more information on the LLVMBuild system, please see:
+;
+;   http://llvm.org/docs/LLVMBuild.html
+;
+;===------------------------------------------------------------------------===;
+
+[component_0]
+type = Tool
+name = llvm-cfi-verify
+parent = Tools
+required_libraries = MC MCDisassembler MCParser Support all-targets

tools/llvm-cfi-verify/llvm-cfi-verify.cpp

+//===-- llvm-cfi-verify.cpp - CFI Verification tool for LLVM --------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This tool verifies Control Flow Integrity (CFI) instrumentation by static
+// binary anaylsis. See the design document in /docs/CFIVerify.rst for more
+// information.
+//
+// This tool is currently incomplete. It currently only does disassembly for
+// object files, and searches through the code for indirect control flow
+// instructions, printing them once found.
+//
+//===----------------------------------------------------------------------===//
+
+#include "llvm/MC/MCAsmInfo.h"
+#include "llvm/MC/MCContext.h"
+#include "llvm/MC/MCDisassembler/MCDisassembler.h"
+#include "llvm/MC/MCInst.h"
+#include "llvm/MC/MCInstPrinter.h"
+#include "llvm/MC/MCInstrAnalysis.h"
+#include "llvm/MC/MCInstrDesc.h"
+#include "llvm/MC/MCInstrInfo.h"
+#include "llvm/MC/MCObjectFileInfo.h"
+#include "llvm/MC/MCRegisterInfo.h"
+#include "llvm/MC/MCSubtargetInfo.h"
+#include "llvm/Object/Binary.h"
+#include "llvm/Object/COFF.h"
+#include "llvm/Object/ObjectFile.h"
+#include "llvm/Support/Casting.h"
+#include "llvm/Support/CommandLine.h"
+#include "llvm/Support/MemoryBuffer.h"
+#include "llvm/Support/TargetRegistry.h"
+#include "llvm/Support/TargetSelect.h"
+#include "llvm/Support/raw_ostream.h"
+
+#include <cassert>
+#include <cstdlib>
+
+using namespace llvm;
+using namespace llvm::object;
+
+cl::opt<bool> ArgDumpSymbols("sym", cl::desc("Dump the symbol table."));
+cl::opt<std::string> InputFilename(cl::Positional, cl::desc("<input file>"),
+                                   cl::Required);
+
+static void printSymbols(const ObjectFile *Object) {
+  for (const SymbolRef &Symbol : Object->symbols()) {
+    outs() << "Symbol [" << format_hex_no_prefix(Symbol.getValue(), 2)
+           << "] = ";
+
+    auto SymbolName = Symbol.getName();
+    if (SymbolName)
+      outs() << *SymbolName;
+    else
+      outs() << "UNKNOWN";
+
+    if (Symbol.getFlags() & SymbolRef::SF_Hidden)
+      outs() << " .hidden";
+
+    outs() << " (Section = ";
+
+    auto SymbolSection = Symbol.getSection();
+    if (SymbolSection) {
+      StringRef SymbolSectionName;
+      if ((*SymbolSection)->getName(SymbolSectionName))
+        outs() << "UNKNOWN)";
+      else
+        outs() << SymbolSectionName << ")";
+    } else {
+      outs() << "N/A)";
+    }
+
+    outs() << "\n";
+  }
+}
+
+int main(int argc, char **argv) {
+  cl::ParseCommandLineOptions(argc, argv);
+
+  InitializeAllTargetInfos();
+  InitializeAllTargetMCs();
+  InitializeAllAsmParsers();
+  InitializeAllDisassemblers();
+
+  Expected<OwningBinary<Binary>> BinaryOrErr = createBinary(InputFilename);
+  if (!BinaryOrErr) {
+    errs() << "Failed to open file.\n";
+    return EXIT_FAILURE;
+  }
+
+  Binary &Binary = *BinaryOrErr.get().getBinary();
+  ObjectFile *Object = dyn_cast<ObjectFile>(&Binary);
+  if (!Object) {
+    errs() << "Disassembling of non-objects not currently supported.\n";
+    return EXIT_FAILURE;
+  }
+
+  Triple TheTriple = Object->makeTriple();
+  std::string TripleName = TheTriple.getTriple();
+  std::string ArchName = "";
+  std::string ErrorString;
+
+  const Target *TheTarget =
+      TargetRegistry::lookupTarget(ArchName, TheTriple, ErrorString);
+
+  if (!TheTarget) {
+    errs() << "Couldn't find target \"" << TheTriple.getTriple()
+           << "\", failed with error: " << ErrorString << ".\n";
+    return EXIT_FAILURE;
+  }
+
+  SubtargetFeatures Features = Object->getFeatures();
+
+  std::unique_ptr<const MCRegisterInfo> RegisterInfo(
+      TheTarget->createMCRegInfo(TripleName));
+  if (!RegisterInfo) {
+    errs() << "Failed to initialise RegisterInfo.\n";
+    return EXIT_FAILURE;
+  }
+
+  std::unique_ptr<const MCAsmInfo> AsmInfo(
+      TheTarget->createMCAsmInfo(*RegisterInfo, TripleName));
+  if (!AsmInfo) {
+    errs() << "Failed to initialise AsmInfo.\n";
+    return EXIT_FAILURE;
+  }
+
+  std::string MCPU = "";
+  std::unique_ptr<MCSubtargetInfo> SubtargetInfo(
+      TheTarget->createMCSubtargetInfo(TripleName, MCPU, Features.getString()));
+  if (!SubtargetInfo) {
+    errs() << "Failed to initialise SubtargetInfo.\n";
+    return EXIT_FAILURE;
+  }
+
+  std::unique_ptr<const MCInstrInfo> MII(TheTarget->createMCInstrInfo());
+  if (!MII) {
+    errs() << "Failed to initialise MII.\n";
+    return EXIT_FAILURE;
+  }
+
+  MCObjectFileInfo MOFI;
+  MCContext Context(AsmInfo.get(), RegisterInfo.get(), &MOFI);
+
+  std::unique_ptr<const MCDisassembler> Disassembler(
+      TheTarget->createMCDisassembler(*SubtargetInfo, Context));
+
+  if (!Disassembler) {
+    errs() << "No disassembler available for target.";
+    return EXIT_FAILURE;
+  }
+
+  std::unique_ptr<const MCInstrAnalysis> MIA(
+      TheTarget->createMCInstrAnalysis(MII.get()));
+
+  std::unique_ptr<MCInstPrinter> Printer(
+      TheTarget->createMCInstPrinter(TheTriple, AsmInfo->getAssemblerDialect(),
+                                     *AsmInfo, *MII, *RegisterInfo));
+
+  if (ArgDumpSymbols)
+    printSymbols(Object);
+
+  for (const SectionRef &Section : Object->sections()) {
+    outs() << "Section [" << format_hex_no_prefix(Section.getAddress(), 2)
+           << "] = ";
+    StringRef SectionName;
+
+    if (Section.getName(SectionName))
+      outs() << "UNKNOWN.\n";
+    else
+      outs() << SectionName << "\n";
+
+    StringRef SectionContents;
+    if (Section.getContents(SectionContents)) {
+      errs() << "Failed to retrieve section contents.\n";
+      return EXIT_FAILURE;
+    }
+
+    MCInst Instruction;
+    size_t InstructionSize;
+
+    ArrayRef<uint8_t> SectionBytes((const uint8_t *)SectionContents.data(),
+                                   Section.getSize());
+
+    for (size_t Byte = 0; Byte < Section.getSize();) {
+      bool BadInstruction = false;
+
+      // Disassemble the instruction.
+      if (Disassembler->getInstruction(
+              Instruction, InstructionSize, SectionBytes.drop_front(Byte), 0,
+              nulls(), outs()) != MCDisassembler::Success) {
+        BadInstruction = true;
+      }
+
+      Byte += InstructionSize;
+
+      if (BadInstruction)
+        continue;
+
+      // Skip instructions that do not affect the control flow.
+      const auto &InstrDesc = MII->get(Instruction.getOpcode());
+      if (!InstrDesc.mayAffectControlFlow(Instruction, *RegisterInfo))
+        continue;
+
+      // Skip instructions that do not operate on register operands.
+      bool UsesRegisterOperand = false;
+      for (const auto &Operand : Instruction) {
+        if (Operand.isReg())
+          UsesRegisterOperand = true;
+      }
+
+      if (!UsesRegisterOperand)
+        continue;
+
+      // Print the instruction address.
+      outs() << "    "
+             << format_hex(Section.getAddress() + Byte - InstructionSize, 2)
+             << ": ";
+
+      // Print the instruction bytes.
+      for (size_t i = 0; i < InstructionSize; ++i) {
+        outs() << format_hex_no_prefix(SectionBytes[Byte - InstructionSize + i],
+                                       2)
+               << " ";
+      }
+
+      // Print the instruction.
+      outs() << " | " << MII->getName(Instruction.getOpcode()) << " ";
+      Instruction.dump_pretty(outs(), Printer.get());
+
+      outs() << "\n";
+    }
+  }
+
+  return EXIT_SUCCESS;
+}