//===- llvm/Support/Unix/Path.inc - Unix Path Implementation ----*- C++ -*-===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception // //===----------------------------------------------------------------------===// // // This file implements the Unix specific implementation of the Path API. // //===----------------------------------------------------------------------===// //===----------------------------------------------------------------------===// //=== WARNING: Implementation here must contain only generic UNIX code that //=== is guaranteed to work on *all* UNIX variants. //===----------------------------------------------------------------------===// #include "Unix.h" #include #include #if HAVE_SYS_STAT_H #include #endif #if HAVE_FCNTL_H #include #endif #ifdef HAVE_UNISTD_H #include #endif #ifdef HAVE_SYS_MMAN_H #include #endif #include #include #include #ifdef __APPLE__ #include #include #include #if __has_include() #include #endif #elif defined(__FreeBSD__) #include #if __FreeBSD_version >= 1300057 #include #else #include extern char **environ; #endif #elif defined(__DragonFly__) #include #elif defined(__MVS__) #include "llvm/Support/AutoConvert.h" #include #endif // Both stdio.h and cstdio are included via different paths and // stdcxx's cstdio doesn't include stdio.h, so it doesn't #undef the macros // either. #undef ferror #undef feof #if !defined(PATH_MAX) // For GNU Hurd #if defined(__GNU__) #define PATH_MAX 4096 #elif defined(__MVS__) #define PATH_MAX _XOPEN_PATH_MAX #endif #endif #include #if !defined(__APPLE__) && !defined(__OpenBSD__) && !defined(__FreeBSD__) && \ !defined(__linux__) && !defined(__FreeBSD_kernel__) && !defined(_AIX) #include #define STATVFS statvfs #define FSTATVFS fstatvfs #define STATVFS_F_FRSIZE(vfs) vfs.f_frsize #else #if defined(__OpenBSD__) || defined(__FreeBSD__) #include #include #elif defined(__linux__) #if defined(HAVE_LINUX_MAGIC_H) #include #else #if defined(HAVE_LINUX_NFS_FS_H) #include #endif #if defined(HAVE_LINUX_SMB_H) #include #endif #endif #include #elif defined(_AIX) #include // depends on `uint` to be a typedef from to // `uint_t`; however, does not always declare `uint`. We provide // the typedef prior to including to work around this issue. typedef uint_t uint; #include #else #include #endif #define STATVFS statfs #define FSTATVFS fstatfs #define STATVFS_F_FRSIZE(vfs) static_cast(vfs.f_bsize) #endif #if defined(__NetBSD__) || defined(__DragonFly__) || defined(__GNU__) || \ defined(__MVS__) #define STATVFS_F_FLAG(vfs) (vfs).f_flag #else #define STATVFS_F_FLAG(vfs) (vfs).f_flags #endif using namespace llvm; namespace llvm { namespace sys { namespace fs { const file_t kInvalidFile = -1; #if defined(__FreeBSD__) || defined(__NetBSD__) || defined(__OpenBSD__) || \ defined(__FreeBSD_kernel__) || defined(__linux__) || defined(__CYGWIN__) || \ defined(__DragonFly__) || defined(_AIX) || defined(__GNU__) || \ (defined(__sun__) && defined(__svr4__) || defined(__HAIKU__)) static int test_dir(char ret[PATH_MAX], const char *dir, const char *bin) { struct stat sb; char fullpath[PATH_MAX]; int chars = snprintf(fullpath, PATH_MAX, "%s/%s", dir, bin); // We cannot write PATH_MAX characters because the string will be terminated // with a null character. Fail if truncation happened. if (chars >= PATH_MAX) return 1; if (!realpath(fullpath, ret)) return 1; if (stat(fullpath, &sb) != 0) return 1; return 0; } static char *getprogpath(char ret[PATH_MAX], const char *bin) { if (bin == nullptr) return nullptr; /* First approach: absolute path. */ if (bin[0] == '/') { if (test_dir(ret, "/", bin) == 0) return ret; return nullptr; } /* Second approach: relative path. */ if (strchr(bin, '/')) { char cwd[PATH_MAX]; if (!getcwd(cwd, PATH_MAX)) return nullptr; if (test_dir(ret, cwd, bin) == 0) return ret; return nullptr; } /* Third approach: $PATH */ char *pv; if ((pv = getenv("PATH")) == nullptr) return nullptr; char *s = strdup(pv); if (!s) return nullptr; char *state; for (char *t = strtok_r(s, ":", &state); t != nullptr; t = strtok_r(nullptr, ":", &state)) { if (test_dir(ret, t, bin) == 0) { free(s); return ret; } } free(s); return nullptr; } #endif // __FreeBSD__ || __NetBSD__ || __FreeBSD_kernel__ /// GetMainExecutable - Return the path to the main executable, given the /// value of argv[0] from program startup. std::string getMainExecutable(const char *argv0, void *MainAddr) { #if defined(__APPLE__) // On OS X the executable path is saved to the stack by dyld. Reading it // from there is much faster than calling dladdr, especially for large // binaries with symbols. char exe_path[PATH_MAX]; uint32_t size = sizeof(exe_path); if (_NSGetExecutablePath(exe_path, &size) == 0) { char link_path[PATH_MAX]; if (realpath(exe_path, link_path)) return link_path; } #elif defined(__FreeBSD__) // On FreeBSD if the exec path specified in ELF auxiliary vectors is // preferred, if available. /proc/curproc/file and the KERN_PROC_PATHNAME // sysctl may not return the desired path if there are multiple hardlinks // to the file. char exe_path[PATH_MAX]; #if __FreeBSD_version >= 1300057 if (elf_aux_info(AT_EXECPATH, exe_path, sizeof(exe_path)) == 0) { char link_path[PATH_MAX]; if (realpath(exe_path, link_path)) return link_path; } #else // elf_aux_info(AT_EXECPATH, ... is not available in all supported versions, // fall back to finding the ELF auxiliary vectors after the process's // environment. char **p = ::environ; while (*p++ != 0) ; // Iterate through auxiliary vectors for AT_EXECPATH. for (Elf_Auxinfo *aux = (Elf_Auxinfo *)p; aux->a_type != AT_NULL; aux++) { if (aux->a_type == AT_EXECPATH) { char link_path[PATH_MAX]; if (realpath((char *)aux->a_un.a_ptr, link_path)) return link_path; } } #endif // Fall back to argv[0] if auxiliary vectors are not available. if (getprogpath(exe_path, argv0) != NULL) return exe_path; #elif defined(_AIX) || defined(__DragonFly__) || defined(__FreeBSD_kernel__) || \ defined(__NetBSD__) const char *curproc = "/proc/curproc/file"; char exe_path[PATH_MAX]; if (sys::fs::exists(curproc)) { ssize_t len = readlink(curproc, exe_path, sizeof(exe_path)); if (len > 0) { // Null terminate the string for realpath. readlink never null // terminates its output. len = std::min(len, ssize_t(sizeof(exe_path) - 1)); exe_path[len] = '\0'; return exe_path; } } // If we don't have procfs mounted, fall back to argv[0] if (getprogpath(exe_path, argv0) != NULL) return exe_path; #elif defined(__linux__) || defined(__CYGWIN__) || defined(__gnu_hurd__) char exe_path[PATH_MAX]; const char *aPath = "/proc/self/exe"; if (sys::fs::exists(aPath)) { // /proc is not always mounted under Linux (chroot for example). ssize_t len = readlink(aPath, exe_path, sizeof(exe_path)); if (len < 0) return ""; // Null terminate the string for realpath. readlink never null // terminates its output. len = std::min(len, ssize_t(sizeof(exe_path) - 1)); exe_path[len] = '\0'; // On Linux, /proc/self/exe always looks through symlinks. However, on // GNU/Hurd, /proc/self/exe is a symlink to the path that was used to start // the program, and not the eventual binary file. Therefore, call realpath // so this behaves the same on all platforms. #if _POSIX_VERSION >= 200112 || defined(__GLIBC__) if (char *real_path = realpath(exe_path, nullptr)) { std::string ret = std::string(real_path); free(real_path); return ret; } #else char real_path[PATH_MAX]; if (realpath(exe_path, real_path)) return std::string(real_path); #endif } // Fall back to the classical detection. if (getprogpath(exe_path, argv0)) return exe_path; #elif defined(__OpenBSD__) || defined(__HAIKU__) char exe_path[PATH_MAX]; // argv[0] only if (getprogpath(exe_path, argv0) != NULL) return exe_path; #elif defined(__sun__) && defined(__svr4__) char exe_path[PATH_MAX]; const char *aPath = "/proc/self/execname"; if (sys::fs::exists(aPath)) { int fd = open(aPath, O_RDONLY); if (fd == -1) return ""; if (read(fd, exe_path, sizeof(exe_path)) < 0) return ""; return exe_path; } // Fall back to the classical detection. if (getprogpath(exe_path, argv0) != NULL) return exe_path; #elif defined(__MVS__) int token = 0; W_PSPROC buf; char exe_path[PS_PATHBLEN]; pid_t pid = getpid(); memset(&buf, 0, sizeof(buf)); buf.ps_pathptr = exe_path; buf.ps_pathlen = sizeof(exe_path); while (true) { if ((token = w_getpsent(token, &buf, sizeof(buf))) <= 0) break; if (buf.ps_pid != pid) continue; char real_path[PATH_MAX]; if (realpath(exe_path, real_path)) return std::string(real_path); break; // Found entry, but realpath failed. } #elif defined(HAVE_DLFCN_H) && defined(HAVE_DLADDR) // Use dladdr to get executable path if available. Dl_info DLInfo; int err = dladdr(MainAddr, &DLInfo); if (err == 0) return ""; // If the filename is a symlink, we need to resolve and return the location of // the actual executable. char link_path[PATH_MAX]; if (realpath(DLInfo.dli_fname, link_path)) return link_path; #else #error GetMainExecutable is not implemented on this host yet. #endif return ""; } TimePoint<> basic_file_status::getLastAccessedTime() const { return toTimePoint(fs_st_atime, fs_st_atime_nsec); } TimePoint<> basic_file_status::getLastModificationTime() const { return toTimePoint(fs_st_mtime, fs_st_mtime_nsec); } UniqueID file_status::getUniqueID() const { return UniqueID(fs_st_dev, fs_st_ino); } uint32_t file_status::getLinkCount() const { return fs_st_nlinks; } ErrorOr disk_space(const Twine &Path) { struct STATVFS Vfs; if (::STATVFS(const_cast(Path.str().c_str()), &Vfs)) return errnoAsErrorCode(); auto FrSize = STATVFS_F_FRSIZE(Vfs); space_info SpaceInfo; SpaceInfo.capacity = static_cast(Vfs.f_blocks) * FrSize; SpaceInfo.free = static_cast(Vfs.f_bfree) * FrSize; SpaceInfo.available = static_cast(Vfs.f_bavail) * FrSize; return SpaceInfo; } std::error_code current_path(SmallVectorImpl &result) { result.clear(); const char *pwd = ::getenv("PWD"); llvm::sys::fs::file_status PWDStatus, DotStatus; if (pwd && llvm::sys::path::is_absolute(pwd) && !llvm::sys::fs::status(pwd, PWDStatus) && !llvm::sys::fs::status(".", DotStatus) && PWDStatus.getUniqueID() == DotStatus.getUniqueID()) { result.append(pwd, pwd + strlen(pwd)); return std::error_code(); } result.resize_for_overwrite(PATH_MAX); while (true) { if (::getcwd(result.data(), result.size()) == nullptr) { // See if there was a real error. if (errno != ENOMEM) { result.clear(); return errnoAsErrorCode(); } // Otherwise there just wasn't enough space. result.resize_for_overwrite(result.capacity() * 2); } else break; } result.truncate(strlen(result.data())); return std::error_code(); } std::error_code set_current_path(const Twine &path) { SmallString<128> path_storage; StringRef p = path.toNullTerminatedStringRef(path_storage); if (::chdir(p.begin()) == -1) return errnoAsErrorCode(); return std::error_code(); } std::error_code create_directory(const Twine &path, bool IgnoreExisting, perms Perms) { SmallString<128> path_storage; StringRef p = path.toNullTerminatedStringRef(path_storage); if (::mkdir(p.begin(), Perms) == -1) { if (errno != EEXIST || !IgnoreExisting) return errnoAsErrorCode(); } return std::error_code(); } // Note that we are using symbolic link because hard links are not supported by // all filesystems (SMB doesn't). std::error_code create_link(const Twine &to, const Twine &from) { // Get arguments. SmallString<128> from_storage; SmallString<128> to_storage; StringRef f = from.toNullTerminatedStringRef(from_storage); StringRef t = to.toNullTerminatedStringRef(to_storage); if (::symlink(t.begin(), f.begin()) == -1) return errnoAsErrorCode(); return std::error_code(); } std::error_code create_hard_link(const Twine &to, const Twine &from) { // Get arguments. SmallString<128> from_storage; SmallString<128> to_storage; StringRef f = from.toNullTerminatedStringRef(from_storage); StringRef t = to.toNullTerminatedStringRef(to_storage); if (::link(t.begin(), f.begin()) == -1) return errnoAsErrorCode(); return std::error_code(); } std::error_code remove(const Twine &path, bool IgnoreNonExisting) { SmallString<128> path_storage; StringRef p = path.toNullTerminatedStringRef(path_storage); struct stat buf; if (lstat(p.begin(), &buf) != 0) { if (errno != ENOENT || !IgnoreNonExisting) return errnoAsErrorCode(); return std::error_code(); } // Note: this check catches strange situations. In all cases, LLVM should // only be involved in the creation and deletion of regular files. This // check ensures that what we're trying to erase is a regular file. It // effectively prevents LLVM from erasing things like /dev/null, any block // special file, or other things that aren't "regular" files. if (!S_ISREG(buf.st_mode) && !S_ISDIR(buf.st_mode) && !S_ISLNK(buf.st_mode)) return make_error_code(errc::operation_not_permitted); if (::remove(p.begin()) == -1) { if (errno != ENOENT || !IgnoreNonExisting) return errnoAsErrorCode(); } return std::error_code(); } static bool is_local_impl(struct STATVFS &Vfs) { #if defined(__linux__) || defined(__GNU__) #ifndef NFS_SUPER_MAGIC #define NFS_SUPER_MAGIC 0x6969 #endif #ifndef SMB_SUPER_MAGIC #define SMB_SUPER_MAGIC 0x517B #endif #ifndef CIFS_MAGIC_NUMBER #define CIFS_MAGIC_NUMBER 0xFF534D42 #endif #ifdef __GNU__ switch ((uint32_t)Vfs.__f_type) { #else switch ((uint32_t)Vfs.f_type) { #endif case NFS_SUPER_MAGIC: case SMB_SUPER_MAGIC: case CIFS_MAGIC_NUMBER: return false; default: return true; } #elif defined(__CYGWIN__) // Cygwin doesn't expose this information; would need to use Win32 API. return false; #elif defined(__Fuchsia__) // Fuchsia doesn't yet support remote filesystem mounts. return true; #elif defined(__EMSCRIPTEN__) // Emscripten doesn't currently support remote filesystem mounts. return true; #elif defined(__HAIKU__) // Haiku doesn't expose this information. return false; #elif defined(__sun) // statvfs::f_basetype contains a null-terminated FSType name of the mounted // target StringRef fstype(Vfs.f_basetype); // NFS is the only non-local fstype?? return fstype != "nfs"; #elif defined(_AIX) // Call mntctl; try more than twice in case of timing issues with a concurrent // mount. int Ret; size_t BufSize = 2048u; std::unique_ptr Buf; int Tries = 3; while (Tries--) { Buf = std::make_unique(BufSize); Ret = mntctl(MCTL_QUERY, BufSize, Buf.get()); if (Ret != 0) break; BufSize = *reinterpret_cast(Buf.get()); Buf.reset(); } if (Ret == -1) // There was an error; "remote" is the conservative answer. return false; // Look for the correct vmount entry. char *CurObjPtr = Buf.get(); while (Ret--) { struct vmount *Vp = reinterpret_cast(CurObjPtr); static_assert(sizeof(Vfs.f_fsid) == sizeof(Vp->vmt_fsid), "fsid length mismatch"); if (memcmp(&Vfs.f_fsid, &Vp->vmt_fsid, sizeof Vfs.f_fsid) == 0) return (Vp->vmt_flags & MNT_REMOTE) == 0; CurObjPtr += Vp->vmt_length; } // vmount entry not found; "remote" is the conservative answer. return false; #elif defined(__MVS__) // The file system can have an arbitrary structure on z/OS; must go with the // conservative answer. return false; #else return !!(STATVFS_F_FLAG(Vfs) & MNT_LOCAL); #endif } std::error_code is_local(const Twine &Path, bool &Result) { struct STATVFS Vfs; if (::STATVFS(const_cast(Path.str().c_str()), &Vfs)) return errnoAsErrorCode(); Result = is_local_impl(Vfs); return std::error_code(); } std::error_code is_local(int FD, bool &Result) { struct STATVFS Vfs; if (::FSTATVFS(FD, &Vfs)) return errnoAsErrorCode(); Result = is_local_impl(Vfs); return std::error_code(); } std::error_code rename(const Twine &from, const Twine &to) { // Get arguments. SmallString<128> from_storage; SmallString<128> to_storage; StringRef f = from.toNullTerminatedStringRef(from_storage); StringRef t = to.toNullTerminatedStringRef(to_storage); if (::rename(f.begin(), t.begin()) == -1) return errnoAsErrorCode(); return std::error_code(); } std::error_code resize_file(int FD, uint64_t Size) { // Use ftruncate as a fallback. It may or may not allocate space. At least on // OS X with HFS+ it does. if (::ftruncate(FD, Size) == -1) return errnoAsErrorCode(); return std::error_code(); } static int convertAccessMode(AccessMode Mode) { switch (Mode) { case AccessMode::Exist: return F_OK; case AccessMode::Write: return W_OK; case AccessMode::Execute: return R_OK | X_OK; // scripts also need R_OK. } llvm_unreachable("invalid enum"); } std::error_code access(const Twine &Path, AccessMode Mode) { SmallString<128> PathStorage; StringRef P = Path.toNullTerminatedStringRef(PathStorage); if (::access(P.begin(), convertAccessMode(Mode)) == -1) return errnoAsErrorCode(); if (Mode == AccessMode::Execute) { // Don't say that directories are executable. struct stat buf; if (0 != stat(P.begin(), &buf)) return errc::permission_denied; if (!S_ISREG(buf.st_mode)) return errc::permission_denied; } return std::error_code(); } bool can_execute(const Twine &Path) { return !access(Path, AccessMode::Execute); } bool equivalent(file_status A, file_status B) { assert(status_known(A) && status_known(B)); return A.fs_st_dev == B.fs_st_dev && A.fs_st_ino == B.fs_st_ino; } std::error_code equivalent(const Twine &A, const Twine &B, bool &result) { file_status fsA, fsB; if (std::error_code ec = status(A, fsA)) return ec; if (std::error_code ec = status(B, fsB)) return ec; result = equivalent(fsA, fsB); return std::error_code(); } static void expandTildeExpr(SmallVectorImpl &Path) { StringRef PathStr(Path.begin(), Path.size()); if (PathStr.empty() || !PathStr.starts_with("~")) return; PathStr = PathStr.drop_front(); StringRef Expr = PathStr.take_until([](char c) { return path::is_separator(c); }); StringRef Remainder = PathStr.substr(Expr.size() + 1); SmallString<128> Storage; if (Expr.empty()) { // This is just ~/..., resolve it to the current user's home dir. if (!path::home_directory(Storage)) { // For some reason we couldn't get the home directory. Just exit. return; } // Overwrite the first character and insert the rest. Path[0] = Storage[0]; Path.insert(Path.begin() + 1, Storage.begin() + 1, Storage.end()); return; } // This is a string of the form ~username/, look up this user's entry in the // password database. std::unique_ptr Buf; long BufSize = sysconf(_SC_GETPW_R_SIZE_MAX); if (BufSize <= 0) BufSize = 16384; Buf = std::make_unique(BufSize); struct passwd Pwd; std::string User = Expr.str(); struct passwd *Entry = nullptr; getpwnam_r(User.c_str(), &Pwd, Buf.get(), BufSize, &Entry); if (!Entry || !Entry->pw_dir) { // Unable to look up the entry, just return back the original path. return; } Storage = Remainder; Path.clear(); Path.append(Entry->pw_dir, Entry->pw_dir + strlen(Entry->pw_dir)); llvm::sys::path::append(Path, Storage); } void expand_tilde(const Twine &path, SmallVectorImpl &dest) { dest.clear(); if (path.isTriviallyEmpty()) return; path.toVector(dest); expandTildeExpr(dest); } static file_type typeForMode(mode_t Mode) { if (S_ISDIR(Mode)) return file_type::directory_file; else if (S_ISREG(Mode)) return file_type::regular_file; else if (S_ISBLK(Mode)) return file_type::block_file; else if (S_ISCHR(Mode)) return file_type::character_file; else if (S_ISFIFO(Mode)) return file_type::fifo_file; else if (S_ISSOCK(Mode)) return file_type::socket_file; else if (S_ISLNK(Mode)) return file_type::symlink_file; return file_type::type_unknown; } static std::error_code fillStatus(int StatRet, const struct stat &Status, file_status &Result) { if (StatRet != 0) { std::error_code EC = errnoAsErrorCode(); if (EC == errc::no_such_file_or_directory) Result = file_status(file_type::file_not_found); else Result = file_status(file_type::status_error); return EC; } uint32_t atime_nsec, mtime_nsec; #if defined(HAVE_STRUCT_STAT_ST_MTIMESPEC_TV_NSEC) atime_nsec = Status.st_atimespec.tv_nsec; mtime_nsec = Status.st_mtimespec.tv_nsec; #elif defined(HAVE_STRUCT_STAT_ST_MTIM_TV_NSEC) atime_nsec = Status.st_atim.tv_nsec; mtime_nsec = Status.st_mtim.tv_nsec; #else atime_nsec = mtime_nsec = 0; #endif perms Perms = static_cast(Status.st_mode) & all_perms; Result = file_status(typeForMode(Status.st_mode), Perms, Status.st_dev, Status.st_nlink, Status.st_ino, Status.st_atime, atime_nsec, Status.st_mtime, mtime_nsec, Status.st_uid, Status.st_gid, Status.st_size); return std::error_code(); } std::error_code status(const Twine &Path, file_status &Result, bool Follow) { SmallString<128> PathStorage; StringRef P = Path.toNullTerminatedStringRef(PathStorage); struct stat Status; int StatRet = (Follow ? ::stat : ::lstat)(P.begin(), &Status); return fillStatus(StatRet, Status, Result); } std::error_code status(int FD, file_status &Result) { struct stat Status; int StatRet = ::fstat(FD, &Status); return fillStatus(StatRet, Status, Result); } unsigned getUmask() { // Chose arbitary new mask and reset the umask to the old mask. // umask(2) never fails so ignore the return of the second call. unsigned Mask = ::umask(0); (void)::umask(Mask); return Mask; } std::error_code setPermissions(const Twine &Path, perms Permissions) { SmallString<128> PathStorage; StringRef P = Path.toNullTerminatedStringRef(PathStorage); if (::chmod(P.begin(), Permissions)) return errnoAsErrorCode(); return std::error_code(); } std::error_code setPermissions(int FD, perms Permissions) { if (::fchmod(FD, Permissions)) return errnoAsErrorCode(); return std::error_code(); } std::error_code setLastAccessAndModificationTime(int FD, TimePoint<> AccessTime, TimePoint<> ModificationTime) { #if defined(HAVE_FUTIMENS) timespec Times[2]; Times[0] = sys::toTimeSpec(AccessTime); Times[1] = sys::toTimeSpec(ModificationTime); if (::futimens(FD, Times)) return errnoAsErrorCode(); return std::error_code(); #elif defined(HAVE_FUTIMES) timeval Times[2]; Times[0] = sys::toTimeVal( std::chrono::time_point_cast(AccessTime)); Times[1] = sys::toTimeVal(std::chrono::time_point_cast( ModificationTime)); if (::futimes(FD, Times)) return errnoAsErrorCode(); return std::error_code(); #elif defined(__MVS__) attrib_t Attr; memset(&Attr, 0, sizeof(Attr)); Attr.att_atimechg = 1; Attr.att_atime = sys::toTimeT(AccessTime); Attr.att_mtimechg = 1; Attr.att_mtime = sys::toTimeT(ModificationTime); if (::__fchattr(FD, &Attr, sizeof(Attr)) != 0) return errnoAsErrorCode(); return std::error_code(); #else #warning Missing futimes() and futimens() return make_error_code(errc::function_not_supported); #endif } std::error_code mapped_file_region::init(int FD, uint64_t Offset, mapmode Mode) { assert(Size != 0); int flags = (Mode == readwrite) ? MAP_SHARED : MAP_PRIVATE; int prot = (Mode == readonly) ? PROT_READ : (PROT_READ | PROT_WRITE); #if defined(MAP_NORESERVE) flags |= MAP_NORESERVE; #endif #if defined(__APPLE__) //---------------------------------------------------------------------- // Newer versions of MacOSX have a flag that will allow us to read from // binaries whose code signature is invalid without crashing by using // the MAP_RESILIENT_CODESIGN flag. Also if a file from removable media // is mapped we can avoid crashing and return zeroes to any pages we try // to read if the media becomes unavailable by using the // MAP_RESILIENT_MEDIA flag. These flags are only usable when mapping // with PROT_READ, so take care not to specify them otherwise. //---------------------------------------------------------------------- if (Mode == readonly) { #if defined(MAP_RESILIENT_CODESIGN) flags |= MAP_RESILIENT_CODESIGN; #endif #if defined(MAP_RESILIENT_MEDIA) flags |= MAP_RESILIENT_MEDIA; #endif } #endif // #if defined (__APPLE__) Mapping = ::mmap(nullptr, Size, prot, flags, FD, Offset); if (Mapping == MAP_FAILED) return errnoAsErrorCode(); return std::error_code(); } mapped_file_region::mapped_file_region(int fd, mapmode mode, size_t length, uint64_t offset, std::error_code &ec) : Size(length), Mode(mode) { (void)Mode; ec = init(fd, offset, mode); if (ec) copyFrom(mapped_file_region()); } void mapped_file_region::unmapImpl() { if (Mapping) ::munmap(Mapping, Size); } void mapped_file_region::dontNeedImpl() { assert(Mode == mapped_file_region::readonly); if (!Mapping) return; #if defined(__MVS__) || defined(_AIX) // If we don't have madvise, or it isn't beneficial, treat this as a no-op. #elif defined(POSIX_MADV_DONTNEED) ::posix_madvise(Mapping, Size, POSIX_MADV_DONTNEED); #else ::madvise(Mapping, Size, MADV_DONTNEED); #endif } int mapped_file_region::alignment() { return Process::getPageSizeEstimate(); } std::error_code detail::directory_iterator_construct(detail::DirIterState &it, StringRef path, bool follow_symlinks) { SmallString<128> path_null(path); DIR *directory = ::opendir(path_null.c_str()); if (!directory) return errnoAsErrorCode(); it.IterationHandle = reinterpret_cast(directory); // Add something for replace_filename to replace. path::append(path_null, "."); it.CurrentEntry = directory_entry(path_null.str(), follow_symlinks); return directory_iterator_increment(it); } std::error_code detail::directory_iterator_destruct(detail::DirIterState &it) { if (it.IterationHandle) ::closedir(reinterpret_cast(it.IterationHandle)); it.IterationHandle = 0; it.CurrentEntry = directory_entry(); return std::error_code(); } static file_type direntType(dirent *Entry) { // Most platforms provide the file type in the dirent: Linux/BSD/Mac. // The DTTOIF macro lets us reuse our status -> type conversion. // Note that while glibc provides a macro to see if this is supported, // _DIRENT_HAVE_D_TYPE, it's not defined on BSD/Mac, so we test for the // d_type-to-mode_t conversion macro instead. #if defined(DTTOIF) return typeForMode(DTTOIF(Entry->d_type)); #else // Other platforms such as Solaris require a stat() to get the type. return file_type::type_unknown; #endif } std::error_code detail::directory_iterator_increment(detail::DirIterState &It) { errno = 0; dirent *CurDir = ::readdir(reinterpret_cast(It.IterationHandle)); if (CurDir == nullptr && errno != 0) { return errnoAsErrorCode(); } else if (CurDir != nullptr) { StringRef Name(CurDir->d_name); if ((Name.size() == 1 && Name[0] == '.') || (Name.size() == 2 && Name[0] == '.' && Name[1] == '.')) return directory_iterator_increment(It); It.CurrentEntry.replace_filename(Name, direntType(CurDir)); } else return directory_iterator_destruct(It); return std::error_code(); } ErrorOr directory_entry::status() const { file_status s; if (auto EC = fs::status(Path, s, FollowSymlinks)) return EC; return s; } // // FreeBSD optionally provides /proc/self/fd, but it is incompatible with // Linux. The thing to use is realpath. // #if !defined(__FreeBSD__) && !defined(__OpenBSD__) #define TRY_PROC_SELF_FD #endif #if !defined(F_GETPATH) && defined(TRY_PROC_SELF_FD) static bool hasProcSelfFD() { // If we have a /proc filesystem mounted, we can quickly establish the // real name of the file with readlink static const bool Result = (::access("/proc/self/fd", R_OK) == 0); return Result; } #endif static int nativeOpenFlags(CreationDisposition Disp, OpenFlags Flags, FileAccess Access) { int Result = 0; if (Access == FA_Read) Result |= O_RDONLY; else if (Access == FA_Write) Result |= O_WRONLY; else if (Access == (FA_Read | FA_Write)) Result |= O_RDWR; // This is for compatibility with old code that assumed OF_Append implied // would open an existing file. See Windows/Path.inc for a longer comment. if (Flags & OF_Append) Disp = CD_OpenAlways; if (Disp == CD_CreateNew) { Result |= O_CREAT; // Create if it doesn't exist. Result |= O_EXCL; // Fail if it does. } else if (Disp == CD_CreateAlways) { Result |= O_CREAT; // Create if it doesn't exist. Result |= O_TRUNC; // Truncate if it does. } else if (Disp == CD_OpenAlways) { Result |= O_CREAT; // Create if it doesn't exist. } else if (Disp == CD_OpenExisting) { // Nothing special, just don't add O_CREAT and we get these semantics. } // Using append mode with z/OS UTF-8 auto-conversion results in EINVAL when // calling write(). Instead we need to use lseek() to set offset to EOF after // open(). #ifndef __MVS__ if (Flags & OF_Append) Result |= O_APPEND; #endif #ifdef O_CLOEXEC if (!(Flags & OF_ChildInherit)) Result |= O_CLOEXEC; #endif return Result; } std::error_code openFile(const Twine &Name, int &ResultFD, CreationDisposition Disp, FileAccess Access, OpenFlags Flags, unsigned Mode) { int OpenFlags = nativeOpenFlags(Disp, Flags, Access); SmallString<128> Storage; StringRef P = Name.toNullTerminatedStringRef(Storage); // Call ::open in a lambda to avoid overload resolution in RetryAfterSignal // when open is overloaded, such as in Bionic. auto Open = [&]() { return ::open(P.begin(), OpenFlags, Mode); }; if ((ResultFD = sys::RetryAfterSignal(-1, Open)) < 0) return errnoAsErrorCode(); #ifndef O_CLOEXEC if (!(Flags & OF_ChildInherit)) { int r = fcntl(ResultFD, F_SETFD, FD_CLOEXEC); (void)r; assert(r == 0 && "fcntl(F_SETFD, FD_CLOEXEC) failed"); } #endif #ifdef __MVS__ /* Reason about auto-conversion and file tags. Setting the file tag only * applies if file is opened in write mode: * * Text file: * File exists File created * CD_CreateNew n/a conv: on * tag: set 1047 * CD_CreateAlways conv: auto conv: on * tag: auto 1047 tag: set 1047 * CD_OpenAlways conv: auto conv: on * tag: auto 1047 tag: set 1047 * CD_OpenExisting conv: auto n/a * tag: unchanged * * Binary file: * File exists File created * CD_CreateNew n/a conv: off * tag: set binary * CD_CreateAlways conv: off conv: off * tag: auto binary tag: set binary * CD_OpenAlways conv: off conv: off * tag: auto binary tag: set binary * CD_OpenExisting conv: off n/a * tag: unchanged * * Actions: * conv: off -> auto-conversion is turned off * conv: on -> auto-conversion is turned on * conv: auto -> auto-conversion is turned on if the file is untagged * tag: set 1047 -> set the file tag to text encoded in 1047 * tag: set binary -> set the file tag to binary * tag: auto 1047 -> set file tag to 1047 if not set * tag: auto binary -> set file tag to binary if not set * tag: unchanged -> do not care about the file tag * * It is not possible to distinguish between the cases "file exists" and * "file created". In the latter case, the file tag is not set and the file * size is zero. The decision table boils down to: * * the file tag is set if * - the file is opened for writing * - the create disposition is not equal to CD_OpenExisting * - the file tag is not set * - the file size is zero * * This only applies if the file is a regular file. E.g. enabling * auto-conversion for reading from /dev/null results in error EINVAL when * calling read(). * * Using append mode with z/OS UTF-8 auto-conversion results in EINVAL when * calling write(). Instead we need to use lseek() to set offset to EOF after * open(). */ if ((Flags & OF_Append) && lseek(ResultFD, 0, SEEK_END) == -1) return errnoAsErrorCode(); struct stat Stat; if (fstat(ResultFD, &Stat) == -1) return errnoAsErrorCode(); if (S_ISREG(Stat.st_mode)) { bool DoSetTag = (Access & FA_Write) && (Disp != CD_OpenExisting) && !Stat.st_tag.ft_txtflag && !Stat.st_tag.ft_ccsid && Stat.st_size == 0; if (Flags & OF_Text) { if (auto EC = llvm::enableAutoConversion(ResultFD)) return EC; if (DoSetTag) { if (auto EC = llvm::setFileTag(ResultFD, CCSID_IBM_1047, true)) return EC; } } else { if (auto EC = llvm::disableAutoConversion(ResultFD)) return EC; if (DoSetTag) { if (auto EC = llvm::setFileTag(ResultFD, FT_BINARY, false)) return EC; } } } #endif return std::error_code(); } Expected openNativeFile(const Twine &Name, CreationDisposition Disp, FileAccess Access, OpenFlags Flags, unsigned Mode) { int FD; std::error_code EC = openFile(Name, FD, Disp, Access, Flags, Mode); if (EC) return errorCodeToError(EC); return FD; } std::error_code openFileForRead(const Twine &Name, int &ResultFD, OpenFlags Flags, SmallVectorImpl *RealPath) { std::error_code EC = openFile(Name, ResultFD, CD_OpenExisting, FA_Read, Flags, 0666); if (EC) return EC; // Attempt to get the real name of the file, if the user asked if (!RealPath) return std::error_code(); RealPath->clear(); #if defined(F_GETPATH) // When F_GETPATH is availble, it is the quickest way to get // the real path name. char Buffer[PATH_MAX]; if (::fcntl(ResultFD, F_GETPATH, Buffer) != -1) RealPath->append(Buffer, Buffer + strlen(Buffer)); #else char Buffer[PATH_MAX]; #if defined(TRY_PROC_SELF_FD) if (hasProcSelfFD()) { char ProcPath[64]; snprintf(ProcPath, sizeof(ProcPath), "/proc/self/fd/%d", ResultFD); ssize_t CharCount = ::readlink(ProcPath, Buffer, sizeof(Buffer)); if (CharCount > 0) RealPath->append(Buffer, Buffer + CharCount); } else { #endif SmallString<128> Storage; StringRef P = Name.toNullTerminatedStringRef(Storage); // Use ::realpath to get the real path name if (::realpath(P.begin(), Buffer) != nullptr) RealPath->append(Buffer, Buffer + strlen(Buffer)); #if defined(TRY_PROC_SELF_FD) } #endif #endif return std::error_code(); } Expected openNativeFileForRead(const Twine &Name, OpenFlags Flags, SmallVectorImpl *RealPath) { file_t ResultFD; std::error_code EC = openFileForRead(Name, ResultFD, Flags, RealPath); if (EC) return errorCodeToError(EC); return ResultFD; } file_t getStdinHandle() { return 0; } file_t getStdoutHandle() { return 1; } file_t getStderrHandle() { return 2; } Expected readNativeFile(file_t FD, MutableArrayRef Buf) { #if defined(__APPLE__) size_t Size = std::min(Buf.size(), INT32_MAX); #else size_t Size = Buf.size(); #endif ssize_t NumRead = sys::RetryAfterSignal(-1, ::read, FD, Buf.data(), Size); if (ssize_t(NumRead) == -1) return errorCodeToError(errnoAsErrorCode()); // The underlying operation on these platforms allow opening directories // for reading in more cases than other platforms. #if defined(__MVS__) || defined(_AIX) struct stat Status; if (fstat(FD, &Status) == -1) return errorCodeToError(errnoAsErrorCode()); if (S_ISDIR(Status.st_mode)) return errorCodeToError(make_error_code(errc::is_a_directory)); #endif return NumRead; } Expected readNativeFileSlice(file_t FD, MutableArrayRef Buf, uint64_t Offset) { #if defined(__APPLE__) size_t Size = std::min(Buf.size(), INT32_MAX); #else size_t Size = Buf.size(); #endif #ifdef HAVE_PREAD ssize_t NumRead = sys::RetryAfterSignal(-1, ::pread, FD, Buf.data(), Size, Offset); #else if (lseek(FD, Offset, SEEK_SET) == -1) return errorCodeToError(errnoAsErrorCode()); ssize_t NumRead = sys::RetryAfterSignal(-1, ::read, FD, Buf.data(), Size); #endif if (NumRead == -1) return errorCodeToError(errnoAsErrorCode()); return NumRead; } std::error_code tryLockFile(int FD, std::chrono::milliseconds Timeout) { auto Start = std::chrono::steady_clock::now(); auto End = Start + Timeout; do { struct flock Lock; memset(&Lock, 0, sizeof(Lock)); Lock.l_type = F_WRLCK; Lock.l_whence = SEEK_SET; Lock.l_start = 0; Lock.l_len = 0; if (::fcntl(FD, F_SETLK, &Lock) != -1) return std::error_code(); int Error = errno; if (Error != EACCES && Error != EAGAIN) return std::error_code(Error, std::generic_category()); usleep(1000); } while (std::chrono::steady_clock::now() < End); return make_error_code(errc::no_lock_available); } std::error_code lockFile(int FD) { struct flock Lock; memset(&Lock, 0, sizeof(Lock)); Lock.l_type = F_WRLCK; Lock.l_whence = SEEK_SET; Lock.l_start = 0; Lock.l_len = 0; if (::fcntl(FD, F_SETLKW, &Lock) != -1) return std::error_code(); return errnoAsErrorCode(); } std::error_code unlockFile(int FD) { struct flock Lock; Lock.l_type = F_UNLCK; Lock.l_whence = SEEK_SET; Lock.l_start = 0; Lock.l_len = 0; if (::fcntl(FD, F_SETLK, &Lock) != -1) return std::error_code(); return errnoAsErrorCode(); } std::error_code closeFile(file_t &F) { file_t TmpF = F; F = kInvalidFile; return Process::SafelyCloseFileDescriptor(TmpF); } template static std::error_code remove_directories_impl(const T &Entry, bool IgnoreErrors) { std::error_code EC; directory_iterator Begin(Entry, EC, false); directory_iterator End; while (Begin != End) { auto &Item = *Begin; ErrorOr st = Item.status(); if (st) { if (is_directory(*st)) { EC = remove_directories_impl(Item, IgnoreErrors); if (EC && !IgnoreErrors) return EC; } EC = fs::remove(Item.path(), true); if (EC && !IgnoreErrors) return EC; } else if (!IgnoreErrors) { return st.getError(); } Begin.increment(EC); if (EC && !IgnoreErrors) return EC; } return std::error_code(); } std::error_code remove_directories(const Twine &path, bool IgnoreErrors) { auto EC = remove_directories_impl(path, IgnoreErrors); if (EC && !IgnoreErrors) return EC; EC = fs::remove(path, true); if (EC && !IgnoreErrors) return EC; return std::error_code(); } std::error_code real_path(const Twine &path, SmallVectorImpl &dest, bool expand_tilde) { dest.clear(); if (path.isTriviallyEmpty()) return std::error_code(); if (expand_tilde) { SmallString<128> Storage; path.toVector(Storage); expandTildeExpr(Storage); return real_path(Storage, dest, false); } SmallString<128> Storage; StringRef P = path.toNullTerminatedStringRef(Storage); char Buffer[PATH_MAX]; if (::realpath(P.begin(), Buffer) == nullptr) return errnoAsErrorCode(); dest.append(Buffer, Buffer + strlen(Buffer)); return std::error_code(); } std::error_code changeFileOwnership(int FD, uint32_t Owner, uint32_t Group) { auto FChown = [&]() { return ::fchown(FD, Owner, Group); }; // Retry if fchown call fails due to interruption. if ((sys::RetryAfterSignal(-1, FChown)) < 0) return errnoAsErrorCode(); return std::error_code(); } } // end namespace fs namespace path { bool home_directory(SmallVectorImpl &result) { std::unique_ptr Buf; char *RequestedDir = getenv("HOME"); if (!RequestedDir) { long BufSize = sysconf(_SC_GETPW_R_SIZE_MAX); if (BufSize <= 0) BufSize = 16384; Buf = std::make_unique(BufSize); struct passwd Pwd; struct passwd *pw = nullptr; getpwuid_r(getuid(), &Pwd, Buf.get(), BufSize, &pw); if (pw && pw->pw_dir) RequestedDir = pw->pw_dir; } if (!RequestedDir) return false; result.clear(); result.append(RequestedDir, RequestedDir + strlen(RequestedDir)); return true; } static bool getDarwinConfDir(bool TempDir, SmallVectorImpl &Result) { #if defined(_CS_DARWIN_USER_TEMP_DIR) && defined(_CS_DARWIN_USER_CACHE_DIR) // On Darwin, use DARWIN_USER_TEMP_DIR or DARWIN_USER_CACHE_DIR. // macros defined in on darwin >= 9 int ConfName = TempDir ? _CS_DARWIN_USER_TEMP_DIR : _CS_DARWIN_USER_CACHE_DIR; size_t ConfLen = confstr(ConfName, nullptr, 0); if (ConfLen > 0) { do { Result.resize(ConfLen); ConfLen = confstr(ConfName, Result.data(), Result.size()); } while (ConfLen > 0 && ConfLen != Result.size()); if (ConfLen > 0) { assert(Result.back() == 0); Result.pop_back(); return true; } Result.clear(); } #endif return false; } bool user_config_directory(SmallVectorImpl &result) { #ifdef __APPLE__ // Mac: ~/Library/Preferences/ if (home_directory(result)) { append(result, "Library", "Preferences"); return true; } #else // XDG_CONFIG_HOME as defined in the XDG Base Directory Specification: // http://standards.freedesktop.org/basedir-spec/basedir-spec-latest.html if (const char *RequestedDir = getenv("XDG_CONFIG_HOME")) { result.clear(); result.append(RequestedDir, RequestedDir + strlen(RequestedDir)); return true; } #endif // Fallback: ~/.config if (!home_directory(result)) { return false; } append(result, ".config"); return true; } bool cache_directory(SmallVectorImpl &result) { #ifdef __APPLE__ if (getDarwinConfDir(false /*tempDir*/, result)) { return true; } #else // XDG_CACHE_HOME as defined in the XDG Base Directory Specification: // http://standards.freedesktop.org/basedir-spec/basedir-spec-latest.html if (const char *RequestedDir = getenv("XDG_CACHE_HOME")) { result.clear(); result.append(RequestedDir, RequestedDir + strlen(RequestedDir)); return true; } #endif if (!home_directory(result)) { return false; } append(result, ".cache"); return true; } static const char *getEnvTempDir() { // Check whether the temporary directory is specified by an environment // variable. const char *EnvironmentVariables[] = {"TMPDIR", "TMP", "TEMP", "TEMPDIR"}; for (const char *Env : EnvironmentVariables) { if (const char *Dir = std::getenv(Env)) return Dir; } return nullptr; } static const char *getDefaultTempDir(bool ErasedOnReboot) { #ifdef P_tmpdir if ((bool)P_tmpdir) return P_tmpdir; #endif if (ErasedOnReboot) return "/tmp"; return "/var/tmp"; } void system_temp_directory(bool ErasedOnReboot, SmallVectorImpl &Result) { Result.clear(); if (ErasedOnReboot) { // There is no env variable for the cache directory. if (const char *RequestedDir = getEnvTempDir()) { Result.append(RequestedDir, RequestedDir + strlen(RequestedDir)); return; } } if (getDarwinConfDir(ErasedOnReboot, Result)) return; const char *RequestedDir = getDefaultTempDir(ErasedOnReboot); Result.append(RequestedDir, RequestedDir + strlen(RequestedDir)); } } // end namespace path namespace fs { #ifdef __APPLE__ /// This implementation tries to perform an APFS CoW clone of the file, /// which can be much faster and uses less space. /// Unfortunately fcopyfile(3) does not support COPYFILE_CLONE, so the /// file descriptor variant of this function still uses the default /// implementation. std::error_code copy_file(const Twine &From, const Twine &To) { std::string FromS = From.str(); std::string ToS = To.str(); #if __has_builtin(__builtin_available) if (__builtin_available(macos 10.12, *)) { // Optimistically try to use clonefile() and handle errors, rather than // calling stat() to see if it'll work. // // Note: It's okay if From is a symlink. In contrast to the behaviour of // copyfile() with COPYFILE_CLONE, clonefile() clones targets (not the // symlink itself) unless the flag CLONE_NOFOLLOW is passed. if (!clonefile(FromS.c_str(), ToS.c_str(), 0)) return std::error_code(); auto Errno = errno; switch (Errno) { case EEXIST: // To already exists. case ENOTSUP: // Device does not support cloning. case EXDEV: // From and To are on different devices. break; default: // Anything else will also break copyfile(). return std::error_code(Errno, std::generic_category()); } // TODO: For EEXIST, profile calling fs::generateUniqueName() and // clonefile() in a retry loop (then rename() on success) before falling // back to copyfile(). Depending on the size of the file this could be // cheaper. } #endif if (!copyfile(FromS.c_str(), ToS.c_str(), /*State=*/NULL, COPYFILE_DATA)) return std::error_code(); return errnoAsErrorCode(); } #endif // __APPLE__ } // end namespace fs } // end namespace sys } // end namespace llvm