Remove gflags

.github/workflows/ci.yml

@@ -22,7 +22,6 @@ jobs:
       fail-fast: false
       matrix:
         image:
-          - { name: 'ubuntu', tag: '18.04' }
           - { name: 'ubuntu', tag: '20.04' }
         llvm: ['12', '13']
 
@@ -40,7 +39,9 @@ jobs:
 
       - name: Build with build script
         shell: bash
-        run: ./scripts/build.sh --llvm-version ${{ matrix.llvm }}
+        run: |
+          sudo apt-get update && sudo apt-get install g++-8
+          ./scripts/build.sh --llvm-version ${{ matrix.llvm }}
       - name: Build with build-presets script
         shell: bash
         run: |
@@ -179,27 +180,12 @@ jobs:
 
       - name: Group the packages by platform
         run: |
-          zip -r9 remill_ubuntu-18.04_packages.zip \
-                  ubuntu-18.04*
-
           zip -r9 remill_ubuntu-20.04_packages.zip \
                   ubuntu-20.04*
 
           zip -r9 remill_macos-11_packages.zip \
                   macos-11*
 
-      - name: Upload the Ubuntu 18.04 packages
-        uses: actions/upload-release-asset@v1
-
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-
-        with:
-          upload_url: ${{ steps.create_release.outputs.upload_url }}
-          asset_path: remill_ubuntu-18.04_packages.zip
-          asset_name: remill_ubuntu-18.04_packages.zip
-          asset_content_type: application/gzip
-
       - name: Upload the Ubuntu 20.04 packages
         uses: actions/upload-release-asset@v1
 
@@ -230,7 +216,7 @@ jobs:
     strategy:
       matrix:
         llvm: ["12", "13"]
-        ubuntu: ["18.04"]
+        ubuntu: ["20.04"]
     steps:
     - uses: actions/checkout@v2
     - name: Build LLVM ${{ matrix.llvm }} on ${{ matrix.ubuntu }}

Dockerfile

@@ -1,7 +1,7 @@
 # Choose your LLVM version
 ARG LLVM_VERSION=13
 ARG ARCH=amd64
-ARG UBUNTU_VERSION=18.04
+ARG UBUNTU_VERSION=20.04
 ARG DISTRO_BASE=ubuntu${UBUNTU_VERSION}
 ARG BUILD_BASE=ubuntu:${UBUNTU_VERSION}
 ARG LIBRARIES=/opt/trailofbits

README.md

@@ -25,7 +25,7 @@ If you are experiencing undocumented problems with Remill then ask for help in t
 
 ## Supported Platforms
 
-Remill is supported on Linux platforms and has been tested on Ubuntu 14.04, 16.04, and 18.04. Remill also works on macOS, and has experimental support for Windows.
+Remill is supported on Linux platforms and has been tested on Ubuntu 20.04. Remill also works on macOS, and has experimental support for Windows.
 
 Remill's Linux version can also be built via Docker for quicker testing.
 
@@ -55,7 +55,7 @@ Remill now comes with a Dockerfile for easier testing. This Dockerfile reference
 
 The Dockerfile allows for quick builds of multiple supported LLVM, architecture, and Linux configurations.
 
-Quickstart (builds Remill against LLVM 12 on Ubuntu 18.04 for AMD64):
+Quickstart (builds Remill against LLVM 12 on Ubuntu 20.04 for AMD64):
 
 Clone Remill:
 ```shell
@@ -69,7 +69,7 @@ Build Remill Docker container:
 # do the build
 docker build . -t remill \
      -f Dockerfile \
-     --build-arg UBUNTU_VERSION=18.04 \
+     --build-arg UBUNTU_VERSION=20.04 \
      --build-arg ARCH=amd64 \
      --build-arg LLVM_VERSION=12
 ```

bin/lift/Lift.cpp

@@ -47,8 +47,13 @@
 #include <string>
 #include <system_error>
 
-DECLARE_string(arch);
-DECLARE_string(os);
+DEFINE_string(os, REMILL_OS,
+              "Operating system name of the code being "
+              "translated. Valid OSes: linux, macos, windows, solaris.");
+DEFINE_string(arch, REMILL_ARCH,
+              "Architecture of the code being translated. "
+              "Valid architectures: x86, amd64 (with or without "
+              "`_avx` or `_avx512` appended), aarch64, aarch32");
 
 DEFINE_uint64(address, 0,
               "Address at which we should assume the bytes are"
@@ -252,14 +257,14 @@ int main(int argc, char *argv[]) {
     return EXIT_FAILURE;
   }
 
-  std::unique_ptr<llvm::Module> module(remill::LoadArchSemantics(arch));
+  std::unique_ptr<llvm::Module> module(remill::LoadArchSemantics(arch.get()));
 
   const auto state_ptr_type = arch->StatePointerType();
   const auto mem_ptr_type = arch->MemoryPointerType();
 
   Memory memory = UnhexlifyInputBytes(addr_mask);
   SimpleTraceManager manager(memory);
-  remill::IntrinsicTable intrinsics(module);
+  remill::IntrinsicTable intrinsics(module.get());
   remill::InstructionLifter inst_lifter(arch, intrinsics);
   remill::TraceLifter trace_lifter(inst_lifter, manager);
 

include/remill/BC/IntrinsicTable.h

@@ -31,9 +31,6 @@ namespace remill {
 
 class IntrinsicTable {
  public:
-  inline explicit IntrinsicTable(const std::unique_ptr<llvm::Module> &module)
-      : IntrinsicTable(module.get()) {}
-
   explicit IntrinsicTable(llvm::Module *module);
 
   llvm::Function *const error;

include/remill/BC/Util.h

@@ -32,8 +32,10 @@
 // clang-format on
 
 #include <array>
+#include <filesystem>
 #include <functional>
 #include <memory>
+#include <optional>
 #include <string>
 #include <unordered_map>
 #include <vector>
@@ -148,20 +150,25 @@ llvm::GlobalVariable *FindGlobaVariable(llvm::Module *M, std::string_view name);
 
 // Try to verify a module.
 bool VerifyModule(llvm::Module *module);
+// Returns diagnostic message if verify failed.
+std::optional<std::string> VerifyModuleMsg(llvm::Module *module);
 
 // Parses and loads a bitcode file into memory.
+std::unique_ptr<llvm::Module>
+LoadModuleFromFile(llvm::LLVMContext *context,
+                   std::string_view file_name,
+                   bool allow_failure = false) __attribute__((__deprecated__));
+
 std::unique_ptr<llvm::Module> LoadModuleFromFile(llvm::LLVMContext *context,
-                                                 std::string_view file_name,
-                                                 bool allow_failure = false);
+                                                 std::filesystem::path file_name);
 
 // Loads the semantics for the `arch`-specific machine, i.e. the machine of the
 // code that we want to lift.
 std::unique_ptr<llvm::Module> LoadArchSemantics(const Arch *arch);
-
-inline std::unique_ptr<llvm::Module>
-LoadArchSemantics(const std::unique_ptr<const Arch> &arch) {
-  return LoadArchSemantics(arch.get());
-}
+// `sem_dirs` is forwarded to `FindSemanticsBitcodeFile`.
+std::unique_ptr<llvm::Module>
+LoadArchSemantics(const Arch *arch,
+                  const std::vector<std::filesystem::path> &sem_dirs);
 
 // Store an LLVM module into a file.
 bool StoreModuleToFile(llvm::Module *module, std::string_view file_name,
@@ -171,15 +178,16 @@ bool StoreModuleToFile(llvm::Module *module, std::string_view file_name,
 bool StoreModuleIRToFile(llvm::Module *module, std::string_view file_name,
                          bool allow_failure = false);
 
-// Find the path to the semantics bitcode file associated with `FLAGS_arch`.
-std::string FindTargetSemanticsBitcodeFile(void) __attribute__((deprecated));
-
-// Find the path to the semantics bitcode file associated with `REMILL_ARCH`,
-// the architecture on which remill is compiled.
-std::string FindHostSemanticsBitcodeFile(void) __attribute__((deprecated));
-
 // Find a semantics bitcode file for the architecture `arch`.
-std::string FindSemanticsBitcodeFile(std::string_view arch);
+// Default compile-time created list of directories is searched.
+std::optional<std::filesystem::path> FindSemanticsBitcodeFile(std::string_view arch);
+// List of directories to search is provided as second argument - default compile time
+// created list is used as fallback only if `fallback_to_defaults` is set.
+// A "shallow" search happens, searching for file `arch` + ".bc".
+std::optional<std::filesystem::path>
+FindSemanticsBitcodeFile(std::string_view arch,
+                         const std::vector<std::filesystem::path> &dirs,
+                         bool fallback_to_defaults=true);
 
 // Return a pointer to the Nth argument (N=0 is the first argument).
 llvm::Argument *NthArgument(llvm::Function *func, size_t index);

lib/Arch/Arch.cpp

@@ -43,11 +43,6 @@
 #include "remill/BC/Version.h"
 #include "remill/OS/OS.h"
 
-DEFINE_string(arch, REMILL_ARCH,
-              "Architecture of the code being translated. "
-              "Valid architectures: x86, amd64 (with or without "
-              "`_avx` or `_avx512` appended), aarch64, aarch32");
-
 namespace remill {
 namespace {