/* * Copyright (c) 2019-2022 Albert Schwarzkopf * * Permission to use, copy, modify, and distribute this software for any * purpose with or without fee is hereby granted, provided that the above * copyright notice and this permission notice appear in all copies. * * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ #include "exile.h" static struct syscall_vow_map exile_vow_map[] = { {EXILE_SYS(read), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(write), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(open), EXILE_SYSCALL_VOW_RPATH|EXILE_SYSCALL_VOW_WPATH}, {EXILE_SYS(close), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(stat), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(fstat), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(lstat), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(poll), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(lseek), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(mmap), EXILE_SYSCALL_VOW_STDIO|EXILE_SYSCALL_VOW_PROT_EXEC}, {EXILE_SYS(mprotect), EXILE_SYSCALL_VOW_STDIO|EXILE_SYSCALL_VOW_PROT_EXEC}, {EXILE_SYS(munmap), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(brk), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(rt_sigaction), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(rt_sigprocmask), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(rt_sigreturn), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(ioctl), EXILE_SYSCALL_VOW_STDIO|EXILE_SYSCALL_VOW_IOCTL}, {EXILE_SYS(pread64), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(pwrite64), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(readv), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(writev), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(access), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(pipe), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(select), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(sched_yield), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(mremap), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(msync), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(mincore), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(madvise), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(shmget), EXILE_SYSCALL_VOW_SHM}, {EXILE_SYS(shmat), EXILE_SYSCALL_VOW_SHM}, {EXILE_SYS(shmctl), EXILE_SYSCALL_VOW_SHM}, {EXILE_SYS(dup), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(dup2), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(pause), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(nanosleep), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(getitimer), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(alarm), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(setitimer), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(getpid), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(sendfile), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(socket), EXILE_SYSCALL_VOW_INET|EXILE_SYSCALL_VOW_UNIX}, {EXILE_SYS(connect), EXILE_SYSCALL_VOW_INET|EXILE_SYSCALL_VOW_UNIX}, {EXILE_SYS(accept), EXILE_SYSCALL_VOW_INET|EXILE_SYSCALL_VOW_UNIX}, {EXILE_SYS(sendto), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(recvfrom), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(sendmsg), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(recvmsg), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(shutdown), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(bind), EXILE_SYSCALL_VOW_INET|EXILE_SYSCALL_VOW_UNIX}, {EXILE_SYS(listen), EXILE_SYSCALL_VOW_INET|EXILE_SYSCALL_VOW_UNIX}, {EXILE_SYS(getsockname), EXILE_SYSCALL_VOW_INET|EXILE_SYSCALL_VOW_UNIX}, {EXILE_SYS(getpeername), EXILE_SYSCALL_VOW_INET|EXILE_SYSCALL_VOW_UNIX}, {EXILE_SYS(socketpair), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(setsockopt), EXILE_SYSCALL_VOW_INET|EXILE_SYSCALL_VOW_UNIX}, {EXILE_SYS(getsockopt), EXILE_SYSCALL_VOW_INET|EXILE_SYSCALL_VOW_UNIX}, {EXILE_SYS(clone), EXILE_SYSCALL_VOW_CLONE|EXILE_SYSCALL_VOW_THREAD}, {EXILE_SYS(fork), EXILE_SYSCALL_VOW_CLONE}, {EXILE_SYS(vfork), EXILE_SYSCALL_VOW_CLONE}, {EXILE_SYS(execve), EXILE_SYSCALL_VOW_EXEC}, {EXILE_SYS(exit), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(wait4), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(kill), EXILE_SYSCALL_VOW_PROC}, {EXILE_SYS(uname), EXILE_SYSCALL_VOW_PROC}, {EXILE_SYS(semget), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(semop), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(semctl), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(shmdt), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(msgget), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(msgsnd), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(msgrcv), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(msgctl), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(fcntl), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(flock), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(fsync), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(fdatasync), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(truncate), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(ftruncate), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(getdents), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(getcwd), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(chdir), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(fchdir), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(rename), EXILE_SYSCALL_VOW_CPATH}, {EXILE_SYS(mkdir), EXILE_SYSCALL_VOW_CPATH}, {EXILE_SYS(rmdir), EXILE_SYSCALL_VOW_CPATH}, {EXILE_SYS(creat), EXILE_SYSCALL_VOW_CPATH}, {EXILE_SYS(link), EXILE_SYSCALL_VOW_CPATH}, {EXILE_SYS(unlink), EXILE_SYSCALL_VOW_CPATH}, {EXILE_SYS(symlink), EXILE_SYSCALL_VOW_CPATH}, {EXILE_SYS(readlink), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(chmod), EXILE_SYSCALL_VOW_FATTR}, {EXILE_SYS(fchmod), EXILE_SYSCALL_VOW_FATTR}, {EXILE_SYS(chown), EXILE_SYSCALL_VOW_CHOWN}, {EXILE_SYS(fchown), EXILE_SYSCALL_VOW_CHOWN}, {EXILE_SYS(lchown), EXILE_SYSCALL_VOW_CHOWN}, {EXILE_SYS(umask), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(gettimeofday), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(getrlimit), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(getrusage), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(sysinfo), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(times), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(getuid), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(getgid), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(setuid), EXILE_SYSCALL_VOW_ID}, {EXILE_SYS(setgid), EXILE_SYSCALL_VOW_ID}, {EXILE_SYS(geteuid), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(getegid), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(setpgid), EXILE_SYSCALL_VOW_PROC}, {EXILE_SYS(getppid), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(getpgrp), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(setsid), EXILE_SYSCALL_VOW_PROC}, {EXILE_SYS(setreuid), EXILE_SYSCALL_VOW_ID}, {EXILE_SYS(setregid), EXILE_SYSCALL_VOW_ID}, {EXILE_SYS(getgroups), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(setgroups), EXILE_SYSCALL_VOW_ID}, {EXILE_SYS(setresuid), EXILE_SYSCALL_VOW_ID}, {EXILE_SYS(getresuid), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(setresgid), EXILE_SYSCALL_VOW_ID}, {EXILE_SYS(getresgid), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(getpgid), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(setfsuid), EXILE_SYSCALL_VOW_ID}, {EXILE_SYS(setfsgid), EXILE_SYSCALL_VOW_ID}, {EXILE_SYS(getsid), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(capget), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(rt_sigpending), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(rt_sigtimedwait), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(rt_sigqueueinfo), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(rt_sigsuspend), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(utime), EXILE_SYSCALL_VOW_FATTR}, {EXILE_SYS(mknod), EXILE_SYSCALL_VOW_DPATH}, {EXILE_SYS(uselib), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(ustat), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(statfs), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(fstatfs), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(getpriority), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(setpriority), EXILE_SYSCALL_VOW_SCHED|EXILE_SYSCALL_VOW_PROC}, {EXILE_SYS(sched_setparam), EXILE_SYSCALL_VOW_SCHED}, {EXILE_SYS(sched_getparam), EXILE_SYSCALL_VOW_SCHED}, {EXILE_SYS(sched_setscheduler), EXILE_SYSCALL_VOW_SCHED}, {EXILE_SYS(sched_getscheduler), EXILE_SYSCALL_VOW_SCHED}, {EXILE_SYS(sched_get_priority_max), EXILE_SYSCALL_VOW_SCHED}, {EXILE_SYS(sched_get_priority_min), EXILE_SYSCALL_VOW_SCHED}, {EXILE_SYS(sched_rr_get_interval), EXILE_SYSCALL_VOW_SCHED}, {EXILE_SYS(mlock), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(munlock), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(mlockall), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(munlockall), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(vhangup), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(prctl), EXILE_SYSCALL_VOW_STDIO|EXILE_SYSCALL_VOW_PRCTL|EXILE_SYSCALL_VOW_SECCOMP_INSTALL}, {EXILE_SYS(arch_prctl), EXILE_SYSCALL_VOW_STDIO|EXILE_SYSCALL_VOW_PRCTL}, {EXILE_SYS(setrlimit), EXILE_SYSCALL_VOW_PROC}, {EXILE_SYS(sync), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(gettid), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(readahead), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(setxattr), EXILE_SYSCALL_VOW_FATTR}, {EXILE_SYS(lsetxattr), EXILE_SYSCALL_VOW_FATTR}, {EXILE_SYS(fsetxattr), EXILE_SYSCALL_VOW_FATTR}, {EXILE_SYS(getxattr), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(lgetxattr), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(fgetxattr), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(listxattr), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(llistxattr), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(flistxattr), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(removexattr), EXILE_SYSCALL_VOW_FATTR}, {EXILE_SYS(lremovexattr), EXILE_SYSCALL_VOW_FATTR}, {EXILE_SYS(fremovexattr), EXILE_SYSCALL_VOW_FATTR}, {EXILE_SYS(tkill), EXILE_SYSCALL_VOW_PROC}, {EXILE_SYS(time), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(futex), EXILE_SYSCALL_VOW_THREAD}, {EXILE_SYS(sched_getaffinity), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(set_thread_area), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(get_thread_area), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(lookup_dcookie), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(epoll_create), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(epoll_ctl_old), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(epoll_wait_old), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(remap_file_pages), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(getdents64), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(set_tid_address), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(semtimedop), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(fadvise64), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(timer_create), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(timer_settime), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(timer_gettime), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(timer_getoverrun), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(timer_delete), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(clock_gettime), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(clock_getres), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(clock_nanosleep), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(exit_group), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(epoll_wait), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(epoll_ctl), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(tgkill), EXILE_SYSCALL_VOW_PROC}, {EXILE_SYS(utimes), EXILE_SYSCALL_VOW_FATTR}, {EXILE_SYS(mbind), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(get_mempolicy), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(mq_open), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(mq_unlink), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(mq_timedsend), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(mq_timedreceive), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(mq_notify), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(mq_getsetattr), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(waitid), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(inotify_init), EXILE_SYSCALL_VOW_FSNOTIFY}, {EXILE_SYS(inotify_add_watch), EXILE_SYSCALL_VOW_FSNOTIFY}, {EXILE_SYS(inotify_rm_watch), EXILE_SYSCALL_VOW_FSNOTIFY}, {EXILE_SYS(openat), EXILE_SYSCALL_VOW_RPATH|EXILE_SYSCALL_VOW_WPATH}, {EXILE_SYS(mkdirat), EXILE_SYSCALL_VOW_CPATH}, {EXILE_SYS(mknodat), EXILE_SYSCALL_VOW_DPATH}, {EXILE_SYS(fchownat), EXILE_SYSCALL_VOW_CHOWN}, {EXILE_SYS(futimesat), EXILE_SYSCALL_VOW_FATTR}, {EXILE_SYS(newfstatat), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(unlinkat), EXILE_SYSCALL_VOW_CPATH}, {EXILE_SYS(renameat), EXILE_SYSCALL_VOW_CPATH}, {EXILE_SYS(linkat), EXILE_SYSCALL_VOW_CPATH}, {EXILE_SYS(symlinkat), EXILE_SYSCALL_VOW_CPATH}, {EXILE_SYS(readlinkat), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(fchmodat), EXILE_SYSCALL_VOW_FATTR}, {EXILE_SYS(faccessat), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(pselect6), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(ppoll), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(set_robust_list), EXILE_SYSCALL_VOW_THREAD}, {EXILE_SYS(get_robust_list), EXILE_SYSCALL_VOW_THREAD}, {EXILE_SYS(splice), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(tee), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(sync_file_range), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(vmsplice), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(move_pages), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(utimensat), EXILE_SYSCALL_VOW_FATTR}, {EXILE_SYS(epoll_pwait), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(signalfd), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(timerfd_create), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(eventfd), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(fallocate), EXILE_SYSCALL_VOW_WPATH|EXILE_SYSCALL_VOW_CPATH}, {EXILE_SYS(timerfd_settime), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(timerfd_gettime), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(accept4), EXILE_SYSCALL_VOW_UNIX|EXILE_SYSCALL_VOW_INET}, {EXILE_SYS(signalfd4), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(eventfd2), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(epoll_create1), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(dup3), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(pipe2), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(inotify_init1), EXILE_SYSCALL_VOW_FSNOTIFY}, {EXILE_SYS(preadv), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(pwritev), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(recvmmsg), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(fanotify_init), EXILE_SYSCALL_VOW_FSNOTIFY}, {EXILE_SYS(fanotify_mark), EXILE_SYSCALL_VOW_FSNOTIFY}, {EXILE_SYS(prlimit64), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(open_by_handle_at), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(sendmmsg), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(getcpu), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(sched_setattr), EXILE_SYSCALL_VOW_SCHED}, {EXILE_SYS(sched_getattr), EXILE_SYSCALL_VOW_SCHED}, {EXILE_SYS(renameat2), EXILE_SYSCALL_VOW_CPATH}, {EXILE_SYS(getrandom), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(memfd_create), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(execveat), EXILE_SYSCALL_VOW_EXEC}, {EXILE_SYS(mlock2), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(copy_file_range), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(statx), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(rseq), EXILE_SYSCALL_VOW_THREAD}, {EXILE_SYS(clone3), EXILE_SYSCALL_VOW_CLONE|EXILE_SYSCALL_VOW_THREAD}, {EXILE_SYS(close_range), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(openat2), EXILE_SYSCALL_VOW_RPATH|EXILE_SYSCALL_VOW_WPATH}, {EXILE_SYS(faccessat2), EXILE_SYSCALL_VOW_RPATH}, {EXILE_SYS(process_madvise), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(epoll_pwait2), EXILE_SYSCALL_VOW_STDIO}, {EXILE_SYS(futex_waitv), EXILE_SYSCALL_VOW_THREAD} }; struct str_to_vow_map str_to_vow_map[] = { { "chown", EXILE_SYSCALL_VOW_CHOWN}, { "clone", EXILE_SYSCALL_VOW_CLONE}, { "cpath", EXILE_SYSCALL_VOW_CPATH}, { "dpath", EXILE_SYSCALL_VOW_DPATH}, { "exec", EXILE_SYSCALL_VOW_EXEC}, { "fattr", EXILE_SYSCALL_VOW_FATTR}, { "fsnotify", EXILE_SYSCALL_VOW_FSNOTIFY}, { "id", EXILE_SYSCALL_VOW_ID}, { "inet", EXILE_SYSCALL_VOW_INET}, { "ioctl", EXILE_SYSCALL_VOW_IOCTL}, { "prctl", EXILE_SYSCALL_VOW_PRCTL}, { "proc", EXILE_SYSCALL_VOW_PROC}, { "prot_exec", EXILE_SYSCALL_VOW_PROT_EXEC}, { "rpath", EXILE_SYSCALL_VOW_RPATH}, { "sched", EXILE_SYSCALL_VOW_SCHED}, { "seccomp_install", EXILE_SYSCALL_VOW_SECCOMP_INSTALL}, { "shm", EXILE_SYSCALL_VOW_SHM}, { "stdio", EXILE_SYSCALL_VOW_STDIO}, { "thread", EXILE_SYSCALL_VOW_THREAD}, { "unix", EXILE_SYSCALL_VOW_UNIX}, { "wpath", EXILE_SYSCALL_VOW_WPATH}, { "error", EXILE_SYSCALL_VOW_DENY_ERROR} }; /* Converts the whitespace separated vows strings to vows flags * * This mainly helps readability, as lots of flags ORed together is not * very readable. * * If an unkown string is found, abort() is called. */ uint64_t exile_vows_from_str(const char *str) { uint64_t result = 0; char current[64] = { 0 }; char *ptr = current; const char *end = ptr + sizeof(current)-1; do { while(ptr <= end && *str != '\0' && *str != ' ') { *ptr = *str; ++ptr; ++str; } int found = 0; for(size_t i = 0; i < sizeof(str_to_vow_map)/sizeof(str_to_vow_map[0]); i++) { if(strcmp(str_to_vow_map[i].str, current) == 0) { result |= str_to_vow_map[i].value; found = 1; break; } } if(!found) { EXILE_LOG_ERROR("No such vow: %s\n", current); abort(); } memset(current, 0, sizeof(current)); ptr = current; } while(*str++ != '\0'); return result; } inline int exile_landlock_is_available() { #if HAVE_LANDLOCK == 1 int ruleset = landlock_create_ruleset(NULL, 0, LANDLOCK_CREATE_RULESET_VERSION); return ruleset > 0; #endif return 0; } int exile_append_syscall_policy(struct exile_policy *exile_policy, long syscall, unsigned int syscall_policy, struct sock_filter *argfilters, size_t n) { struct exile_syscall_policy *newpolicy = (struct exile_syscall_policy *) calloc(1, sizeof(struct exile_syscall_policy)); if(newpolicy == NULL) { EXILE_LOG_ERROR("Failed to allocate memory for syscall policy\n"); exile_policy->exile_flags |= EXILE_FLAG_ADD_SYSCALL_POLICY_FAIL; return -1; } newpolicy->policy = syscall_policy; newpolicy->syscall = syscall; newpolicy->argfilterscount = n; if(n > EXILE_ARGFILTERS_COUNT) { EXILE_LOG_ERROR("Too many argfilters supplied\n"); exile_policy->exile_flags |= EXILE_FLAG_ADD_SYSCALL_POLICY_FAIL; free(newpolicy); return -1; } for(size_t i = 0; i < n; i++) { newpolicy->argfilters[i] = argfilters[i]; } newpolicy->next = NULL; *(exile_policy->syscall_policies_tail) = newpolicy; exile_policy->syscall_policies_tail = &(newpolicy->next); exile_policy->disable_syscall_filter = 0; return 0; } int exile_append_syscall_default_policy(struct exile_policy *exile_policy, unsigned int default_policy) { return exile_append_syscall_policy(exile_policy, EXILE_SYSCALL_MATCH_ALL, default_policy, NULL, 0); } #define COUNT_EXILE_SYSCALL_FILTER(f) \ sizeof(f)/sizeof(f[0]) #define EXILE_SYSCALL_FILTER_LOAD_ARG(val) \ { 0, EXILE_BPF_LOAD_SECCOMP_ARG(val), 0} /* Returns, for the specific syscall, the correct sock_filter struct for the provided vow_promises * * Returns: 0 if none copied, otherwise the number of entries in "filter". */ int get_vow_argfilter(long syscall, uint64_t vow_promises, struct sock_filter *filter , int *policy) { /* How to read this: * Keep in mind our default action is to deny, unless it's a syscall from a vow promise. Then it will be * accepted if the argument values are good (if we care about them at all). * EXILE_BPF_MATCH() means the argument value is good, and the syscall can be accepted without further checks * EXILE_BPF_NO_MATCH() means the syscall won't be allowed because the value is illegal * * First field (vowmask): The mask to check * Last field (whenset): If mask is set in vow_promises, then add this filter, otherwise don't. */ struct exile_syscall_filter mmap_filter[] = { EXILE_SYSCALL_FILTER_LOAD_ARG(2), { EXILE_SYSCALL_VOW_PROT_EXEC, EXILE_BPF_NO_MATCH_SET(PROT_EXEC), 0}, }; struct exile_syscall_filter ioctl_filter[] = { EXILE_SYSCALL_FILTER_LOAD_ARG(1), { EXILE_SYSCALL_VOW_IOCTL, EXILE_BPF_NO_MATCH_SET(TIOCSTI), 1 }, { EXILE_SYSCALL_VOW_IOCTL, EXILE_BPF_RETURN_MATCHING, 1 }, { EXILE_SYSCALL_VOW_STDIO, EXILE_BPF_MATCH(FIONREAD), 1}, { EXILE_SYSCALL_VOW_STDIO, EXILE_BPF_MATCH(FIONBIO), 1}, { EXILE_SYSCALL_VOW_STDIO, EXILE_BPF_MATCH(FIONREAD), 1}, { EXILE_SYSCALL_VOW_STDIO, EXILE_BPF_MATCH(FIOCLEX), 1}, { EXILE_SYSCALL_VOW_STDIO, EXILE_BPF_MATCH(FIONCLEX), 1}, { EXILE_SYSCALL_VOW_STDIO, EXILE_BPF_RETURN_NOT_MATCHING, 1} }; struct exile_syscall_filter open_filter[] = { EXILE_SYSCALL_FILTER_LOAD_ARG(1), { EXILE_SYSCALL_VOW_CPATH, EXILE_BPF_NO_MATCH_SET(O_CREAT), 0 }, { EXILE_SYSCALL_VOW_WPATH, EXILE_BPF_NO_MATCH_SET(O_TMPFILE),0 }, { EXILE_SYSCALL_VOW_WPATH, EXILE_BPF_NO_MATCH_SET(O_WRONLY),0 }, { EXILE_SYSCALL_VOW_WPATH, EXILE_BPF_NO_MATCH_SET(O_RDWR),0 }, { EXILE_SYSCALL_VOW_WPATH, EXILE_BPF_NO_MATCH_SET(O_APPEND),0 }, }; struct exile_syscall_filter socket_filter[] = { EXILE_SYSCALL_FILTER_LOAD_ARG(0), { EXILE_SYSCALL_VOW_UNIX, EXILE_BPF_MATCH(AF_UNIX), 1 }, { EXILE_SYSCALL_VOW_INET, EXILE_BPF_MATCH(AF_INET), 1 }, { EXILE_SYSCALL_VOW_INET, EXILE_BPF_MATCH(AF_INET6), 1 }, { 0, EXILE_BPF_RETURN_NOT_MATCHING, 0} }; struct exile_syscall_filter setsockopt_filter[] = { EXILE_SYSCALL_FILTER_LOAD_ARG(2), { 0, EXILE_BPF_NO_MATCH(SO_DEBUG), 0 }, { 0, EXILE_BPF_NO_MATCH(SO_SNDBUFFORCE), 0 } }; struct exile_syscall_filter clone_filter[] = { /* It's the first (0) argument for x86_64 */ EXILE_SYSCALL_FILTER_LOAD_ARG(0), { EXILE_SYSCALL_VOW_CLONE, EXILE_BPF_RETURN_MATCHING, 1 }, { EXILE_SYSCALL_VOW_THREAD, EXILE_BPF_CMP_SET(CLONE_VM, 0, EXILE_SYSCALL_EXIT_BPF_NO_MATCH), 1}, { EXILE_SYSCALL_VOW_THREAD, EXILE_BPF_CMP_SET(CLONE_THREAD, 0, EXILE_SYSCALL_EXIT_BPF_NO_MATCH), 1}, { 0, EXILE_BPF_NO_MATCH_SET(CLONE_NEWCGROUP), 0}, { 0, EXILE_BPF_NO_MATCH_SET(CLONE_NEWIPC),0}, { 0, EXILE_BPF_NO_MATCH_SET(CLONE_NEWNET),0}, { 0, EXILE_BPF_NO_MATCH_SET(CLONE_NEWNS),0}, { 0, EXILE_BPF_NO_MATCH_SET(CLONE_NEWPID),0}, { 0, EXILE_BPF_NO_MATCH_SET(CLONE_NEWUSER),0}, { 0, EXILE_BPF_NO_MATCH_SET(CLONE_NEWUTS),0}, }; struct exile_syscall_filter prctl_filter[] ={ EXILE_SYSCALL_FILTER_LOAD_ARG(0), { EXILE_SYSCALL_VOW_PRCTL, EXILE_BPF_RETURN_MATCHING, 1}, { EXILE_SYSCALL_VOW_SECCOMP_INSTALL, EXILE_BPF_MATCH(PR_SET_SECCOMP), 1 }, { EXILE_SYSCALL_VOW_STDIO, EXILE_BPF_MATCH(PR_SET_NO_NEW_PRIVS),1}, { EXILE_SYSCALL_VOW_STDIO, EXILE_BPF_MATCH(PR_GET_NO_NEW_PRIVS),1}, { EXILE_SYSCALL_VOW_STDIO, EXILE_BPF_MATCH(PR_GET_NAME),1}, { EXILE_SYSCALL_VOW_STDIO, EXILE_BPF_MATCH(PR_SET_NAME),1}, { EXILE_SYSCALL_VOW_STDIO, EXILE_BPF_MATCH(PR_CAPBSET_READ), 1}, { 0, EXILE_BPF_RETURN_NOT_MATCHING, 0} }; struct exile_syscall_filter *current_filter = NULL; size_t current_count = 0; *policy = EXILE_SYSCALL_ALLOW; switch(syscall) { case EXILE_SYS(mmap): case EXILE_SYS(mprotect): current_filter = mmap_filter; current_count = COUNT_EXILE_SYSCALL_FILTER(mmap_filter); break; case EXILE_SYS(ioctl): current_filter = ioctl_filter; current_count = COUNT_EXILE_SYSCALL_FILTER(ioctl_filter); break; case EXILE_SYS(open): case EXILE_SYS(openat): case EXILE_SYS(open_by_handle_at): if(syscall == EXILE_SYS(openat) || syscall == EXILE_SYS(open_by_handle_at)) { /* for openat, it's the third arg */ open_filter[0] = (struct exile_syscall_filter) EXILE_SYSCALL_FILTER_LOAD_ARG(2); } current_filter = open_filter; current_count = COUNT_EXILE_SYSCALL_FILTER(open_filter); break; case EXILE_SYS(openat2): *policy = EXILE_SYSCALL_DENY_RET_NOSYS; return 0; break; case EXILE_SYS(socket): current_filter = socket_filter; current_count = COUNT_EXILE_SYSCALL_FILTER(socket_filter); break; case EXILE_SYS(setsockopt): current_filter = setsockopt_filter; current_count = COUNT_EXILE_SYSCALL_FILTER(setsockopt_filter); break; case EXILE_SYS(clone): current_filter = clone_filter; current_count = COUNT_EXILE_SYSCALL_FILTER(clone_filter); break; case EXILE_SYS(clone3): if((vow_promises & EXILE_SYSCALL_VOW_CLONE) == 0) { *policy = EXILE_SYSCALL_DENY_RET_NOSYS; return 0; } break; case EXILE_SYS(prctl): current_filter = prctl_filter; current_count = COUNT_EXILE_SYSCALL_FILTER(prctl_filter); break; } int out_filter_index = 0; for(size_t i = 0; i < current_count; i++) { struct exile_syscall_filter *c = ¤t_filter[i]; int set = 0; if(c->vowmask & vow_promises) { set = 1; } if(c->whenset == set || c->vowmask == 0) { filter[out_filter_index++] = c->filter; } } return out_filter_index; } int exile_append_vow_promises(struct exile_policy *policy, uint64_t vow_promises) { for(unsigned int i = 0; i < sizeof(exile_vow_map)/sizeof(exile_vow_map[0]); i++) { struct syscall_vow_map *current_map = &exile_vow_map[i]; if(current_map->vowmask & vow_promises) { struct sock_filter filter[EXILE_ARGFILTERS_COUNT]; long syscall = current_map->syscall; int syscall_policy = EXILE_SYSCALL_ALLOW; int argfilters = get_vow_argfilter(syscall, vow_promises, filter, &syscall_policy); int ret = exile_append_syscall_policy(policy, syscall, syscall_policy, filter, argfilters); if(ret != 0) { EXILE_LOG_ERROR("Failed adding syscall policy from vow while processing %li\n", syscall); return ret; } } } int vow_policy = (vow_promises & EXILE_SYSCALL_VOW_DENY_ERROR) ? EXILE_SYSCALL_DENY_RET_ERROR : EXILE_SYSCALL_DENY_KILL_PROCESS; return exile_append_syscall_default_policy(policy, vow_policy); } /* Creates an empty policy struct without opinionated defaults. * * Must be freed using exile_free_policy() * @returns: empty policy */ struct exile_policy *exile_create_policy() { struct exile_policy *result = (struct exile_policy *) calloc(1, sizeof(struct exile_policy)); if(result == NULL) { EXILE_LOG_ERROR("Failed to allocate memory for policy\n"); return NULL; } result->path_policies_tail = &(result->path_policies); result->syscall_policies_tail = &(result->syscall_policies); return result; } /* Creates the default policy * Must be freed using exile_free_policy() * * @returns: default policy */ struct exile_policy *exile_init_policy() { struct exile_policy *result = exile_create_policy(); if(result == NULL) { return NULL; } result->drop_caps = 0; result->not_dumpable = 1; result->no_new_privs = 1; result->namespace_options = EXILE_UNSHARE_AUTOMATIC; result->namespace_uid = 0; result->namespace_gid = 0; return result; } /* Appends path policies to the exile_policy object * The last paramater must be NULL * * This function does not copy parameters. All passed paths * MUST NOT be freed until exile_enable_policy() is called! * * @returns: 0 on success, -1 on failure */ int (exile_append_path_policies)(struct exile_policy *exile_policy, unsigned int path_policy, ...) { va_list args; const char *path; va_start(args, path_policy); path = va_arg(args, char*); while(path != NULL) { int fd = open(path, O_PATH); if(fd == -1) { EXILE_LOG_ERROR("Failed to open %s: %s\n", path, strerror(errno)); exile_policy->exile_flags |= EXILE_FLAG_ADD_PATH_POLICY_FAIL; return -1; } close(fd); struct exile_path_policy *newpolicy = (struct exile_path_policy *) calloc(1, sizeof(struct exile_path_policy)); if(newpolicy == NULL) { EXILE_LOG_ERROR("Failed to allocate memory for path policy\n"); exile_policy->exile_flags |= EXILE_FLAG_ADD_PATH_POLICY_FAIL; return -1; } newpolicy->path = path; newpolicy->policy = path_policy; newpolicy->next = NULL; *(exile_policy->path_policies_tail) = newpolicy; exile_policy->path_policies_tail = &(newpolicy->next); path = va_arg(args, char*); } va_end(args); return 0; } /* * Fills buffer with random characters a-z. * The string will be null terminated. * * @returns: number of written chars (excluding terminating null byte) on success */ int random_string(char *buffer, size_t buffer_length) { int r = getrandom(buffer, buffer_length-1, GRND_NONBLOCK); if(r != -1 && (size_t) r == buffer_length-1) { int i = 0; while(i < r) { buffer[i] = 'a' + ((unsigned int)buffer[i] % 26); ++i; } buffer[buffer_length-1] = '\0'; return i; } return 0; } /* Creates a directory/file and all necessary parent directories * @returns: 0 on success, -ERRNO on failure */ int mkpath(const char *p, mode_t mode, int baseisfile) { char path[PATH_MAX + 1] = {0}; int ret = snprintf(path, sizeof(path), "%s%c", p, (baseisfile) ? '\0' : '/'); if(ret < 0) { EXILE_LOG_ERROR("error during path concatination\n"); return -EINVAL; } if((size_t)ret >= sizeof(path)) { EXILE_LOG_ERROR("path concatination truncated\n"); return -EINVAL; } char *begin = path; char *end = begin + 1; while(*end) { if(*end == '/') { *end = 0; if(mkdir(begin, mode) < 0) { if(errno != EEXIST) { EXILE_LOG_ERROR("Failed to create directory: %s\n", begin); return -1; } } *end = '/'; while(*end == '/') { ++end; } } else { ++end; } } if(baseisfile) { ret = creat(p, mode); if(ret == -1) { EXILE_LOG_ERROR("Failed to create file: %s\n", begin); return ret; } close(ret); return 0; } return 0; } /* @returns: argument for mount(2) flags */ static int get_policy_mount_flags(struct exile_path_policy *policy) { int result = 0; if( (policy->policy & EXILE_FS_ALLOW_DEV) == 0) { result |= MS_NODEV; } if( (policy->policy & EXILE_FS_ALLOW_EXEC) == 0) { result |= MS_NOEXEC; } if( (policy->policy & EXILE_FS_ALLOW_SETUID) == 0) { result |= MS_NOSUID; } if( (policy->policy & EXILE_FS_ALLOW_ALL_WRITE) == 0) { result |= MS_RDONLY; } if( (policy->policy & EXILE_MOUNT_NOT_REC) == 0) { result |= MS_REC; } return result; } int path_policy_needs_landlock(struct exile_path_policy *path_policy) { unsigned int policy = path_policy->policy; #if HAVE_LANDLOCK == 1 if(policy >= EXILE_FS_ALLOW_REMOVE_DIR) { return 1; } #endif //Can't need it if we don't have support at compile time return 0; } /* TODO: we can do va_args */ char *concat_path(const char *first, const char *second) { char *result = (char *) calloc(1, PATH_MAX); if(result == NULL) { EXILE_LOG_ERROR("calloc failed\n"); return NULL; } //TODO: We can strip multiple redundant slashes int written = snprintf(result, PATH_MAX, "%s/%s", first, second); if(written < 0) { EXILE_LOG_ERROR("Error during path concatination\n"); free(result); return NULL; } if(written >= PATH_MAX) { EXILE_LOG_ERROR("path concatination truncated\n"); free(result); return NULL; } return result; } /* Creates the file system hierarchy for the chroot * @returns: 0 on sucess, -ERRNO on failure */ static int create_chroot_dirs(const char *chroot_target_path, struct exile_path_policy *path_policy) { while(path_policy != NULL) { struct stat sb; int ret = stat(path_policy->path, &sb); if(ret < 0) { EXILE_LOG_ERROR("stat failed\n"); return ret; } int baseisfile = 0; if(S_ISREG(sb.st_mode)) { baseisfile = 1; } char *path_inside_chroot = concat_path(chroot_target_path, path_policy->path); if(path_inside_chroot == NULL) { return 1; } ret = mkpath(path_inside_chroot, 0700, baseisfile); if(ret < 0) { EXILE_LOG_ERROR("Error creating directory structure %s while mounting paths to chroot: %s\n", path_inside_chroot, strerror(errno)); free(path_inside_chroot); return ret; } path_policy = path_policy->next; free(path_inside_chroot); } return 0; } static int perform_mounts(const char *chroot_target_path, struct exile_path_policy *path_policy) { while(path_policy != NULL) { if(path_policy->policy & EXILE_FS_ALLOW_ALL_READ || path_policy->policy & EXILE_FS_ALLOW_ALL_WRITE) { int mount_flags = get_policy_mount_flags(path_policy); char *path_inside_chroot = concat_path(chroot_target_path, path_policy->path); if(path_inside_chroot == NULL) { return 1; } //all we do is bind mounts mount_flags |= MS_BIND; int ret = mount(path_policy->path, path_inside_chroot, NULL, mount_flags, NULL); if(ret < 0 ) { EXILE_LOG_ERROR("Failed to mount %s to %s: %s\n", path_policy->path, path_inside_chroot, strerror(errno)); free(path_inside_chroot); return ret; } //remount so noexec, readonly etc. take effect ret = mount(NULL, path_inside_chroot, NULL, mount_flags | MS_REMOUNT, NULL); if(ret < 0 ) { EXILE_LOG_ERROR("Failed to remount %s: %s\n", path_inside_chroot, strerror(errno)); free(path_inside_chroot); return ret; } free(path_inside_chroot); } path_policy = path_policy->next; } return 0; } /* * Frees the memory taken by a exile_policy object */ void exile_free_policy(struct exile_policy *ctxt) { if(ctxt != NULL) { struct exile_path_policy *current = ctxt->path_policies; while(current != NULL) { struct exile_path_policy *tmp = current; current = current->next; free(tmp); } struct exile_syscall_policy *sc_policy = ctxt->syscall_policies; while(sc_policy != NULL) { struct exile_syscall_policy *tmp = sc_policy; sc_policy = sc_policy->next; free(tmp); } free(ctxt); } } /* Enters the specified namespaces */ static int enter_namespaces(int namespace_options, uid_t namespace_uid, gid_t namespace_gid) { if(namespace_options & EXILE_UNSHARE_USER) { uid_t current_uid = getuid(); gid_t current_gid = getgid(); char buf[1024] = {0}; int ret = unshare(CLONE_NEWUSER); if(ret == -1) { EXILE_LOG_ERROR("Failed to unshare user namespaces: %s\n", strerror(errno)); return ret; } int fd = open("/proc/self/setgroups", O_WRONLY); if(fd == -1) { EXILE_LOG_ERROR("Failed to open /proc/self/setgroups for writing"); return -1; } int writesize = snprintf(buf, sizeof(buf), "deny"); int writeret = write(fd, buf, writesize); if(writeret < 0 || writeret < writesize) { EXILE_LOG_ERROR("Failed to write to /proc/self/setgroups: %i (%s)\n", writeret, strerror(errno)); return -1; } close(fd); fd = open("/proc/self/uid_map", O_WRONLY); if(fd == -1) { EXILE_LOG_ERROR("Failed to open /proc/self/uid_map for writing"); return -1; } writesize = snprintf(buf, sizeof(buf), "%u %u 1\n", namespace_uid, current_uid); writeret = write(fd, buf, writesize); if(writeret < 0 || writeret < writesize) { EXILE_LOG_ERROR("Failed to write to /proc/self/uid_map: %i (%s)\n", writeret, strerror(errno)); return -1; } close(fd); fd = open("/proc/self/gid_map", O_WRONLY); if(fd == -1) { EXILE_LOG_ERROR("Failed to open /proc/self/gid_map for writing"); return -1; } writesize = snprintf(buf, sizeof(buf), "%u %u 1\n", namespace_gid, current_gid); writeret = write(fd, buf, writesize); if(writeret < 0 || writeret < writesize) { EXILE_LOG_ERROR("Failed to write to /proc/self/gid_map: %i (%s)\n", writeret, strerror(errno)); return -1; } close(fd); } if(namespace_options & EXILE_UNSHARE_MOUNT) { int ret = unshare(CLONE_NEWNS); if(ret == -1) { EXILE_LOG_ERROR("Failed to unshare mount namespaces: %s\n", strerror(errno)); return ret; } } if(namespace_options & EXILE_UNSHARE_NETWORK) { int ret = unshare(CLONE_NEWNET); if(ret == -1) { EXILE_LOG_ERROR("Failed to unshare network namespace: %s\n", strerror(errno)); return ret; } } return 0; } /* Drops all capabiltiies held by the process * * @returns: 0 on sucess, -1 on error */ static int drop_caps() { int cap = 0; int res = 0; while((res = prctl(PR_CAPBSET_DROP, cap, 0, 0, 0)) == 0) { ++cap; } if(res == -1 && errno != EINVAL) { EXILE_LOG_ERROR("Failed to drop the capability bounding set!\n"); return -errno; } //TODO: systems that are not 64 bit struct __user_cap_header_struct h = { 0 }; h.pid = 0; h.version = _LINUX_CAPABILITY_VERSION_3; struct __user_cap_data_struct drop[2]; drop[0].effective = 0; drop[0].permitted = 0; drop[0].inheritable = 0; drop[1].effective = 0; drop[1].permitted = 0; drop[1].inheritable = 0; if(capset(&h, drop) == -1) { EXILE_LOG_ERROR("Failed to drop capabilities: %s\n", strerror(errno)); return -errno; } return 0; } static void assign_filter(struct sock_filter *left, struct sock_filter *right, struct sock_filter *endfilter) { if(left <= endfilter) { *left = *right; return; } EXILE_LOG_ERROR("Too many syscall filters installed! Aborting.\n"); abort(); } static struct sock_filter *append_syscall_to_bpf(struct exile_syscall_policy *syscallpolicy, struct sock_filter *filter, struct sock_filter *endfilter) { unsigned int action = syscallpolicy->policy; if(action == EXILE_SYSCALL_ALLOW) { action = SECCOMP_RET_ALLOW; } if(action == EXILE_SYSCALL_DENY_KILL_PROCESS) { action = SECCOMP_RET_KILL_PROCESS; } if(action == EXILE_SYSCALL_DENY_RET_ERROR) { action = SECCOMP_RET_ERRNO|EACCES; } if(action == EXILE_SYSCALL_DENY_RET_NOSYS) { action = SECCOMP_RET_ERRNO|ENOSYS; } long syscall = syscallpolicy->syscall; struct sock_filter syscall_load = BPF_STMT(BPF_LD+BPF_W+BPF_ABS, offsetof(struct seccomp_data, nr)); assign_filter(filter, &syscall_load, endfilter); ++filter; if(syscall != EXILE_SYSCALL_MATCH_ALL) { /* How many steps forward to jump when we don't match. This is either the last statement, * i. e. the default action or the next syscall policy */ __u8 next_syscall_pc = 1; if(__builtin_add_overflow(next_syscall_pc, syscallpolicy->argfilterscount, &next_syscall_pc)) { EXILE_LOG_ERROR("Overflow while trying to calculate jump offset\n"); abort(); } struct sock_filter syscall_check = EXILE_BPF_CMP_EQ((unsigned int) syscall, 0, next_syscall_pc); assign_filter(filter, &syscall_check, endfilter); ++filter; --next_syscall_pc; struct sock_filter return_matching = EXILE_BPF_RETURN_MATCHING; struct sock_filter return_not_matching = EXILE_BPF_RETURN_NOT_MATCHING; for(size_t i = 0; i < syscallpolicy->argfilterscount; i++) { assign_filter(filter, &syscallpolicy->argfilters[i], endfilter); struct sock_filter *current = filter; __u8 jump_count_next_syscall = next_syscall_pc; __u8 jump_count_return = jump_count_next_syscall - 1; if(current->jt == EXILE_SYSCALL_EXIT_BPF_NO_MATCH) { current->jt = jump_count_next_syscall; } if(current->jt == EXILE_SYSCALL_EXIT_BPF_RETURN) { current->jt = jump_count_return; } if(current->jf == EXILE_SYSCALL_EXIT_BPF_NO_MATCH) { current->jf = jump_count_next_syscall; } if(current->jf == EXILE_SYSCALL_EXIT_BPF_RETURN) { current->jf = jump_count_return; } if(current->code == return_matching.code && current->k == return_matching.k) { current->k = jump_count_return; } if(current->code == return_not_matching.code && current->k == return_not_matching.k) { current->k = jump_count_next_syscall; } --next_syscall_pc; ++filter; } } struct sock_filter syscall_action = BPF_STMT(BPF_RET+BPF_K, action); /* TODO: we can do better than adding this below every jump */ assign_filter(filter, &syscall_action, endfilter); return ++filter; } static int is_valid_syscall_policy(unsigned int policy) { return policy == EXILE_SYSCALL_ALLOW || policy == EXILE_SYSCALL_DENY_RET_ERROR || policy == EXILE_SYSCALL_DENY_KILL_PROCESS || policy == EXILE_SYSCALL_DENY_RET_NOSYS; } /* * Enables the seccomp policy * * policy: exile policy object * * @returns: 0 on success, -1 on error */ int exile_enable_syscall_policy(struct exile_policy *policy) { struct sock_filter filter[1024] = { BPF_STMT(BPF_LD+BPF_W+BPF_ABS,offsetof(struct seccomp_data, arch)), BPF_JUMP (BPF_JMP+BPF_JEQ+BPF_K, SECCOMP_AUDIT_ARCH, 1, 0), BPF_STMT(BPF_RET+BPF_K, SECCOMP_RET_KILL_PROCESS), BPF_STMT(BPF_LD+BPF_W+BPF_ABS, offsetof(struct seccomp_data, nr)), BPF_JUMP(BPF_JMP+BPF_JGE+BPF_K, __X32_SYSCALL_BIT, 0, 1), BPF_STMT(BPF_RET+BPF_K, SECCOMP_RET_KILL_PROCESS), }; struct sock_filter *current_target = &filter[6]; struct sock_filter *end = &filter[1023]; struct exile_syscall_policy *current_policy = policy->syscall_policies; while(current_policy) { if(!is_valid_syscall_policy(current_policy->policy)) { EXILE_LOG_ERROR("invalid syscall policy specified\n"); return -1; } current_target = append_syscall_to_bpf(current_policy, current_target, end); current_policy = current_policy->next; } unsigned short len = (current_target - &filter[0]); struct sock_fprog prog = { .len = len , .filter = filter, }; if (prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, &prog) == -1) { EXILE_LOG_ERROR("prctl SET_SECCOMP %s\n", strerror(errno)); return -1; } return 0; } #if HAVE_LANDLOCK == 1 static unsigned int exile_flags_to_landlock(unsigned int flags, int statmode) { unsigned int result = 0; if(flags & EXILE_FS_ALLOW_ALL_READ) { result |= LANDLOCK_ACCESS_FS_READ_FILE; if(S_ISDIR(statmode)) { result |= LANDLOCK_ACCESS_FS_READ_DIR; } } if(flags & EXILE_FS_ALLOW_ALL_WRITE) { result |= LANDLOCK_ACCESS_FS_WRITE_FILE; if(S_ISDIR(statmode)) { result |= LANDLOCK_ACCESS_FS_REMOVE_DIR; result |= LANDLOCK_ACCESS_FS_REMOVE_FILE; result |= LANDLOCK_ACCESS_FS_MAKE_DIR; result |= LANDLOCK_ACCESS_FS_MAKE_FIFO; result |= LANDLOCK_ACCESS_FS_MAKE_REG; result |= LANDLOCK_ACCESS_FS_MAKE_SOCK; result |= LANDLOCK_ACCESS_FS_MAKE_SYM; } } if(flags & EXILE_FS_ALLOW_EXEC) { result |= LANDLOCK_ACCESS_FS_EXECUTE; } if(flags & EXILE_FS_ALLOW_WRITE_FILE) { result |= LANDLOCK_ACCESS_FS_WRITE_FILE; } if(S_ISDIR(statmode)) { if(flags & EXILE_FS_ALLOW_DEV) { result |= LANDLOCK_ACCESS_FS_MAKE_BLOCK; result |= LANDLOCK_ACCESS_FS_MAKE_CHAR; } if(flags & EXILE_FS_ALLOW_MAKE_BLOCK) { result |= LANDLOCK_ACCESS_FS_MAKE_BLOCK; } if(flags & EXILE_FS_ALLOW_MAKE_CHAR) { result |= LANDLOCK_ACCESS_FS_MAKE_CHAR; } if(flags & EXILE_FS_ALLOW_MAKE_DIR) { result |= LANDLOCK_ACCESS_FS_MAKE_DIR; } if(flags & EXILE_FS_ALLOW_MAKE_FIFO) { result |= LANDLOCK_ACCESS_FS_MAKE_FIFO; } if(flags & EXILE_FS_ALLOW_MAKE_REG) { result |= LANDLOCK_ACCESS_FS_MAKE_REG; } if(flags & EXILE_FS_ALLOW_MAKE_SOCK) { result |= LANDLOCK_ACCESS_FS_MAKE_SOCK; } if(flags & EXILE_FS_ALLOW_MAKE_SYM) { result |= LANDLOCK_ACCESS_FS_MAKE_SYM; } if(flags & EXILE_FS_ALLOW_REMOVE) { result |= LANDLOCK_ACCESS_FS_REMOVE_DIR; result |= LANDLOCK_ACCESS_FS_REMOVE_FILE; } if(flags & EXILE_FS_ALLOW_REMOVE_DIR) { result |= LANDLOCK_ACCESS_FS_REMOVE_DIR; } if(flags & EXILE_FS_ALLOW_REMOVE_FILE) { result |= LANDLOCK_ACCESS_FS_REMOVE_FILE; } if(flags & EXILE_FS_ALLOW_READ_DIR) { result |= LANDLOCK_ACCESS_FS_READ_DIR; } } return result; } static int landlock_prepare_ruleset(struct exile_path_policy *policies) { int ruleset_fd = -1; struct landlock_ruleset_attr ruleset_attr = {0}; /* We here want the maximum possible ruleset, so set the var to the max possible bitmask. Stolen/Adapted from: [linux src]/security/landlock/limits.h */ ruleset_attr.handled_access_fs = ((LANDLOCK_ACCESS_FS_MAKE_SYM << 1) - 1); ruleset_fd = landlock_create_ruleset(&ruleset_attr, sizeof(ruleset_attr), 0); if (ruleset_fd < 0) { EXILE_LOG_ERROR("Failed to create landlock ruleset\n"); return -1; } struct exile_path_policy *policy = policies; while(policy != NULL) { struct landlock_path_beneath_attr path_beneath = {0}; path_beneath.parent_fd = open(policy->path, O_PATH | O_CLOEXEC); if(path_beneath.parent_fd < 0) { EXILE_LOG_ERROR("Failed to open policy path %s while preparing landlock ruleset\n", policy->path); close(ruleset_fd); return path_beneath.parent_fd; } struct stat sb; int ret = fstat(path_beneath.parent_fd, &sb); if(ret) { EXILE_LOG_ERROR("fstat failed %s\n", strerror(errno)); close(ruleset_fd); return ret; } path_beneath.allowed_access = exile_flags_to_landlock(policy->policy, sb.st_mode); ret = landlock_add_rule(ruleset_fd, LANDLOCK_RULE_PATH_BENEATH, &path_beneath, 0); if(ret) { EXILE_LOG_ERROR("Failed to update ruleset while processsing policy path %s\n", policy->path); close(ruleset_fd); return ret; } policy = policy->next; } return ruleset_fd; } #endif /* Checks for illogical or dangerous combinations */ static int check_policy_sanity(struct exile_policy *policy) { if(policy->no_new_privs != 1) { if(policy->syscall_policies != NULL) { EXILE_LOG_ERROR("no_new_privs = 1 is required for seccomp filtering!\n"); return -1; } } int can_use_landlock = exile_landlock_is_available(); if(!can_use_landlock) { struct exile_path_policy *path_policy = policy->path_policies; while(path_policy) { if(path_policy_needs_landlock(path_policy)) { EXILE_LOG_ERROR("A path policy (%s) needs landlock, but landlock is not available. Fallback not possible\n", path_policy->path); return -1; } path_policy = path_policy->next; } } /* TODO: check if we have ALLOWED, but no default deny */ if(policy->mount_path_policies_to_chroot == 1) { if(policy->path_policies == NULL) { EXILE_LOG_ERROR("Cannot mount path policies to chroot if none are given\n"); return -1; } if(!(policy->namespace_options & EXILE_UNSHARE_MOUNT)) { EXILE_LOG_ERROR("mount_path_policies_to_chroot = 1 requires unsharing mount namespace\n"); return -1; } } if(policy->path_policies != NULL) { if(policy->mount_path_policies_to_chroot != 1) { #if HAVE_LANDLOCK != 1 EXILE_LOG_ERROR("Path policies cannot be enforced! System needs landlock support or set mount_path_policies_to_chroot = 1\n"); return -1; #endif } if(policy->no_fs == 1) { EXILE_LOG_ERROR("If path_policies are specified, no_fs cannot be set to 1\n"); return -1; } } struct exile_syscall_policy *syscall_policy = policy->syscall_policies; if(syscall_policy != NULL) { /* A few sanitiy checks... but we cannot check overall whether it's reasonable */ int i = 0; int last_match_all = -1; int match_all_policy = 0; int last_policy = 0; while(syscall_policy) { if(syscall_policy->syscall == EXILE_SYSCALL_MATCH_ALL) { if(policy->vow_promises != 0) { EXILE_LOG_ERROR("It's not possible to specify a default, all matching syscall policy while also using vows\n"); return -1; } last_match_all = i; match_all_policy = syscall_policy->policy; } else { last_policy = syscall_policy->policy; } syscall_policy = syscall_policy->next; ++i; } if(policy->vow_promises == 0 && (last_match_all == -1 || i - last_match_all != 1)) { EXILE_LOG_ERROR("The last entry in the syscall policy list must match all syscalls (default rule)\n"); return -1; } /* Most likely a mistake and not intended */ if(last_policy == match_all_policy) { EXILE_LOG_ERROR("Last policy for a syscall matches default policy\n"); return -1; } } return 0; } static void close_file_fds() { long max_files = sysconf(_SC_OPEN_MAX); for(long i = 3; i <= max_files; i++) { struct stat statbuf; int fd = (int) max_files; int result = fstat(i, &statbuf); if(result == -1 && errno != EBADF && errno != EACCES) { EXILE_LOG_ERROR("Could not fstat %i: %s\n", fd, strerror(errno)); abort(); } int type = statbuf.st_mode & S_IFMT; if(type != S_IFIFO && type != S_IFSOCK) { /* No error check, retrying not recommended */ close(fd); } } } /* Takes away file system access from the process * * We use this when "no_fs" is given in the policy. * * This is useful for restricted subprocesses that do some computational work * and do not require filesystem access * * @returns: 0 on success, < 0 on error */ static int enable_no_fs(struct exile_policy *policy) { close_file_fds(); if(chdir("/proc/self/fdinfo") != 0) { EXILE_LOG_ERROR("Failed to change to safe directory: %s\n", strerror(errno)); return -1; } if(chroot(".") != 0) { EXILE_LOG_ERROR("Failed to chroot into safe directory: %s\n", strerror(errno)); return -1; } if(chdir("/") != 0) { EXILE_LOG_ERROR("Failed to chdir into safe directory inside chroot: %s\n", strerror(errno)); return -1; } return 0; } /* Enables the specified exile_policy. * * This function is not atomic (and can't be). This means some * policies can apply, while others may fail. * * This function returns success only if all policies applied. * * The state is undefined if this function fails. The process generally * should exit. * * @returns: 0 on success (all policies applied), < 0 on error (none or some policies dit not apply) */ int exile_enable_policy(struct exile_policy *policy) { if((policy->exile_flags & EXILE_FLAG_ADD_PATH_POLICY_FAIL) || (policy->exile_flags & EXILE_FLAG_ADD_SYSCALL_POLICY_FAIL)) { EXILE_LOG_ERROR("At least one syscall or path policy was not successfully added!\n"); return -1; } if(check_policy_sanity(policy) != 0) { EXILE_LOG_ERROR("Policy sanity check failed. Cannot apply policy!\n"); return -EINVAL; } if(policy->keep_fds_open != 1) { close_file_fds(); } if(enter_namespaces(policy->namespace_options, policy->namespace_uid, policy->namespace_gid) < 0) { EXILE_LOG_ERROR("Error while trying to enter namespaces\n"); return -1; } int can_use_landlock = exile_landlock_is_available(); /* Fallback to chroot mechanism to enforce policies. Ignore mount_path_policies_to_chroot * if we have no other option (so no landlock) */ if((policy->mount_path_policies_to_chroot || !can_use_landlock) && policy->path_policies != NULL) { if(*policy->chroot_target_path == '\0') { char random_str[17]; if(random_string(random_str, sizeof(random_str)) == 16) { int res = snprintf(policy->chroot_target_path, sizeof(policy->chroot_target_path), "%s/.sandbox_%" PRIdMAX "_%s", EXILE_TEMP_DIR, (intmax_t)getpid(), random_str); if(res < 0) { EXILE_LOG_ERROR("error during path concatination\n"); return -EINVAL; } if(res >= PATH_MAX) { EXILE_LOG_ERROR("path concatination truncated\n"); return -EINVAL; } } else { EXILE_LOG_ERROR("Error creating random sandbox directory name\n"); return -1; } } if(create_chroot_dirs(policy->chroot_target_path, policy->path_policies) < 0) { EXILE_LOG_ERROR("bind mounting of path policies failed\n"); return -1; } if(perform_mounts(policy->chroot_target_path, policy->path_policies) < 0) { EXILE_LOG_ERROR("Failed to remount\n"); return -1; } } if(*policy->chroot_target_path != '\0') { if(chroot(policy->chroot_target_path) < 0) { EXILE_LOG_ERROR("failed to enter %s\n", policy->chroot_target_path); return -1; } const char *chdir_target_path = policy->chdir_path; if(chdir_target_path == NULL) { chdir_target_path = "/"; } if(chdir(chdir_target_path) < 0) { EXILE_LOG_ERROR("chdir to %s failed\n", policy->chdir_path); return -1; } } #if HAVE_LANDLOCK == 1 int landlock_ruleset_fd = -1; if(can_use_landlock && policy->path_policies != NULL) { landlock_ruleset_fd = landlock_prepare_ruleset(policy->path_policies); if(landlock_ruleset_fd < 0) { EXILE_LOG_ERROR("Failed to prepare landlock ruleset: %s\n", strerror(errno)); return -1; } } #endif if(policy->no_fs) { if(enable_no_fs(policy) != 0) { EXILE_LOG_ERROR("Failed to take away filesystem access of process\n"); return -1; } } if(policy->no_new_fds) { const struct rlimit nofile = {0, 0}; if (setrlimit(RLIMIT_NOFILE, &nofile) == -1) { EXILE_LOG_ERROR("setrlimit: Failed to set rlimit: %s\n", strerror(errno)); return -1; } } if(policy->drop_caps) { if(drop_caps() < 0) { EXILE_LOG_ERROR("failed to drop capabilities\n"); return -1; } } if(policy->not_dumpable) { if(prctl(PR_SET_DUMPABLE, 0) == -1) { EXILE_LOG_ERROR("prctl: PR_SET_DUMPABLE failed\n"); return -1; } } if(policy->no_new_privs) { if(prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0) == -1) { EXILE_LOG_ERROR("prctl: PR_SET_NO_NEW_PRIVS failed: %s\n", strerror(errno)); return -1; } } #if HAVE_LANDLOCK == 1 if (can_use_landlock && policy->path_policies != NULL && landlock_restrict_self(landlock_ruleset_fd, 0) != 0) { perror("Failed to enforce ruleset"); close(landlock_ruleset_fd); return -1; } close(landlock_ruleset_fd); #endif if(policy->vow_promises != 0) { int ret = exile_append_vow_promises(policy, policy->vow_promises); if(ret != 0) { EXILE_LOG_ERROR("exile_append_vow_promises() failed: %i\n", ret); return ret; } } if(policy->syscall_policies != NULL) { return exile_enable_syscall_policy(policy); } return 0; } /* Convenience wrapper for the vow-related subset of exile.h * * Only installs seccomp filters for the specified vow promises. * * Useful if only vow is required from exile.h, but nothing else * * Comparable with OpenBSD's pledge(), subsequent calls can only reduce allowed syscalls. * * Here, adding more promises than a previous call set may return success, but * won't be allowed during execution. * * Due to the nature of seccomp, it's furthermore required the EXILE_SYSCALL_VOW_SECCOMP_INSTALL promise * is set if further calls are expected. Generally, it's reasonable for the last call to * exile_vow() a program makes to not set EXILE_SYSCALL_VOW_SECCOMP_INSTALL. * * There are no seperate exec_promises. All children of the process inherit the filter. * . * Return value: 0 on success, any other value on failure. */ int exile_vow(uint64_t promises) { struct __user_cap_header_struct h = { 0 }; h.pid = 0; h.version = _LINUX_CAPABILITY_VERSION_3; struct __user_cap_data_struct cap[2]; cap[0].effective = 0; cap[0].permitted = 0; cap[0].inheritable = 0; cap[1].effective = 0; cap[1].permitted = 0; cap[1].inheritable = 0; if(capget(&h, cap) == -1) { EXILE_LOG_ERROR("Failed to get capabilities: %s\n", strerror(errno)); return -errno; } struct exile_policy *policy = exile_create_policy(); if(policy == NULL) { EXILE_LOG_ERROR("Failed to create policy\n"); return 1; } policy->vow_promises = promises; if((cap[0].effective & (1<no_new_privs = 1; } int ret = exile_enable_policy(policy); exile_free_policy(policy); return ret; } int exile_clone_handle(void *arg) { struct exile_launch_params *params = (struct exile_launch_params *) arg; struct exile_policy *policy = (struct exile_policy *) params->policy; int ret = exile_enable_policy(policy); if(ret != 0) { EXILE_LOG_ERROR("Failed to enable policy\n"); close(params->child_read_pipe[1]); close(params->child_write_pipe[0]); return 1; } ret = dup2(params->child_read_pipe[1], 1); if(ret == -1) { EXILE_LOG_ERROR("Failed to redirect stdout to pipe\n"); return 1; } ret = params->func(params->funcarg); fclose(stdout); close(params->child_read_pipe[1]); close(params->child_write_pipe[0]); return ret; } /* Helper to easily execute a single function sandboxed. * * Creates a child-process, then activates the policy contained in launch_params, * and jumps to the specified function, passing the specified argument to it. * Returns a fd connected to stdout in the child process, as well as a fd allowing to write * to the child. * * if cloneflags is 0, the default ones are passed to clone(), otherwise the value of cloneflags * * Return value: Negative on error, otherwise the file descriptor to read from*/ int exile_launch(struct exile_launch_params *launch_params, struct exile_launch_result *launch_result) { int ret = pipe(launch_params->child_read_pipe); if(ret != 0) { EXILE_LOG_ERROR("read pipe creation failed\n"); return ret; } ret = pipe(launch_params->child_write_pipe); if(ret != 0) { EXILE_LOG_ERROR("write pipe creation failed\n"); return ret; } struct rlimit rlimit; ret = getrlimit(RLIMIT_STACK, &rlimit); if(ret != 0) { EXILE_LOG_ERROR("Failed to get stack size: %s\n", strerror(errno)); return ret; } size_t size = rlimit.rlim_cur; char *stack = (char *) calloc(1, size); if(stack == NULL) { EXILE_LOG_ERROR("Failed to allocate stack memory for child\n"); return 1; } stack += size; ret = clone(&exile_clone_handle, stack, 17 /* SIGCHLD */, launch_params); if(ret == -1) { EXILE_LOG_ERROR("clone failed(): %s\n", strerror(errno)); return ret; } close(launch_params->child_read_pipe[1]); close(launch_params->child_write_pipe[0]); launch_result->tid = ret; launch_result->read_fd = launch_params->child_read_pipe[0]; launch_result->write_fd = launch_params->child_write_pipe[1]; return 0; } /* Helper for exile_launch, to easily read all output from a function * This function will read all output from a sandboxed function. It's up to the caller to ensure * that enough memory will be available. * * The result is \0 terminated. The "n" parameter contains the size of the result, not including the \0. * * Return value: All data written by the function. The result should be passed to free() once not needed. NULL will * be returned on error. */ char *exile_launch_get(struct exile_launch_params *launch_params, size_t *n) { *n = 0; struct exile_launch_result launch_result; int launch = exile_launch(launch_params, &launch_result); if(launch < 0) { return NULL; } char *result = NULL; size_t size = 0; FILE *stream = open_memstream(&result, &size); while(1) { char buffer[4096]; int ret = read(launch_result.read_fd, buffer, sizeof(buffer)); if(ret == 0) { break; } if(ret == -1) { if(errno == EINTR) { continue; } EXILE_LOG_ERROR("Failed to read from read file descriptor\n"); close(launch_result.read_fd); fclose(stream); return NULL; } size_t written = fwrite(buffer, 1, ret, stream); if(written != (size_t) ret) { EXILE_LOG_ERROR("Short item write"); /* TODO: can we seek and free? */ close(launch_result.read_fd); fclose(stream); return NULL; } } fclose(stream); close(launch_result.read_fd); *n = size; return result; }