/* $OpenBSD: getentropy_osx.c,v 1.8 2014/07/21 20:19:47 guenther Exp $ */ /* * Copyright (c) 2014 Theo de Raadt * Copyright (c) 2014 Bob Beck * * 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. * * Emulation of getentropy(2) as documented at: * http://www.openbsd.org/cgi-bin/man.cgi/OpenBSD-current/man2/getentropy.2 */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #define SHA512_Update(a, b, c) (CC_SHA512_Update((a), (b), (c))) #define SHA512_Init(xxx) (CC_SHA512_Init((xxx))) #define SHA512_Final(xxx, yyy) (CC_SHA512_Final((xxx), (yyy))) #define SHA512_CTX CC_SHA512_CTX #define SHA512_DIGEST_LENGTH CC_SHA512_DIGEST_LENGTH #define REPEAT 5 #define min(a, b) (((a) < (b)) ? (a) : (b)) #define HX(a, b) \ do { \ if ((a)) \ HD(errno); \ else \ HD(b); \ } while (0) #define HR(x, l) (SHA512_Update(&ctx, (char *)(x), (l))) #define HD(x) (SHA512_Update(&ctx, (char *)&(x), sizeof (x))) #define HF(x) (SHA512_Update(&ctx, (char *)&(x), sizeof (void*))) int getentropy(void *buf, size_t len); static int gotdata(char *buf, size_t len); static int getentropy_urandom(void *buf, size_t len); static int getentropy_fallback(void *buf, size_t len); int getentropy(void *buf, size_t len) { int ret = -1; if (len > 256) { errno = EIO; return -1; } /* * Try to get entropy with /dev/urandom * * This can fail if the process is inside a chroot or if file * descriptors are exhausted. */ ret = getentropy_urandom(buf, len); if (ret != -1) return (ret); /* * Entropy collection via /dev/urandom and sysctl have failed. * * No other API exists for collecting entropy, and we have * no failsafe way to get it on OSX that is not sensitive * to resource exhaustion. * * We have very few options: * - Even syslog_r is unsafe to call at this low level, so * there is no way to alert the user or program. * - Cannot call abort() because some systems have unsafe * corefiles. * - Could raise(SIGKILL) resulting in silent program termination. * - Return EIO, to hint that arc4random's stir function * should raise(SIGKILL) * - Do the best under the circumstances.... * * This code path exists to bring light to the issue that OSX * does not provide a failsafe API for entropy collection. * * We hope this demonstrates that OSX should consider * providing a new failsafe API which works in a chroot or * when file descriptors are exhausted. */ #undef FAIL_INSTEAD_OF_TRYING_FALLBACK #ifdef FAIL_INSTEAD_OF_TRYING_FALLBACK raise(SIGKILL); #endif ret = getentropy_fallback(buf, len); if (ret != -1) return (ret); errno = EIO; return (ret); } /* * Basic sanity checking; wish we could do better. */ static int gotdata(char *buf, size_t len) { char any_set = 0; size_t i; for (i = 0; i < len; ++i) any_set |= buf[i]; if (any_set == 0) return -1; return 0; } static int getentropy_urandom(void *buf, size_t len) { struct stat st; size_t i; int fd, flags; int save_errno = errno; start: flags = O_RDONLY; #ifdef O_NOFOLLOW flags |= O_NOFOLLOW; #endif #ifdef O_CLOEXEC flags |= O_CLOEXEC; #endif fd = open("/dev/urandom", flags, 0); if (fd == -1) { if (errno == EINTR) goto start; goto nodevrandom; } #ifndef O_CLOEXEC fcntl(fd, F_SETFD, fcntl(fd, F_GETFD) | FD_CLOEXEC); #endif /* Lightly verify that the device node looks sane */ if (fstat(fd, &st) == -1 || !S_ISCHR(st.st_mode)) { close(fd); goto nodevrandom; } for (i = 0; i < len; ) { size_t wanted = len - i; ssize_t ret = read(fd, (char *)buf + i, wanted); if (ret == -1) { if (errno == EAGAIN || errno == EINTR) continue; close(fd); goto nodevrandom; } i += ret; } close(fd); if (gotdata(buf, len) == 0) { errno = save_errno; return 0; /* satisfied */ } nodevrandom: errno = EIO; return -1; } static int tcpmib[] = { CTL_NET, AF_INET, IPPROTO_TCP, TCPCTL_STATS }; static int udpmib[] = { CTL_NET, AF_INET, IPPROTO_UDP, UDPCTL_STATS }; static int ipmib[] = { CTL_NET, AF_INET, IPPROTO_IP, IPCTL_STATS }; static int kmib[] = { CTL_KERN, KERN_USRSTACK }; static int hwmib[] = { CTL_HW, HW_USERMEM }; static int getentropy_fallback(void *buf, size_t len) { uint8_t results[SHA512_DIGEST_LENGTH]; int save_errno = errno, e, pgs = getpagesize(), faster = 0, repeat; static int cnt; struct timespec ts; struct timeval tv; struct rusage ru; sigset_t sigset; struct stat st; SHA512_CTX ctx; static pid_t lastpid; pid_t pid; size_t i, ii, m; char *p; struct tcpstat tcpstat; struct udpstat udpstat; struct ipstat ipstat; u_int64_t mach_time; unsigned int idata; void *addr; pid = getpid(); if (lastpid == pid) { faster = 1; repeat = 2; } else { faster = 0; lastpid = pid; repeat = REPEAT; } for (i = 0; i < len; ) { int j; SHA512_Init(&ctx); for (j = 0; j < repeat; j++) { HX((e = gettimeofday(&tv, NULL)) == -1, tv); if (e != -1) { cnt += (int)tv.tv_sec; cnt += (int)tv.tv_usec; } mach_time = mach_absolute_time(); HD(mach_time); ii = sizeof(addr); HX(sysctl(kmib, sizeof(kmib) / sizeof(kmib[0]), &addr, &ii, NULL, 0) == -1, addr); ii = sizeof(idata); HX(sysctl(hwmib, sizeof(hwmib) / sizeof(hwmib[0]), &idata, &ii, NULL, 0) == -1, idata); ii = sizeof(tcpstat); HX(sysctl(tcpmib, sizeof(tcpmib) / sizeof(tcpmib[0]), &tcpstat, &ii, NULL, 0) == -1, tcpstat); ii = sizeof(udpstat); HX(sysctl(udpmib, sizeof(udpmib) / sizeof(udpmib[0]), &udpstat, &ii, NULL, 0) == -1, udpstat); ii = sizeof(ipstat); HX(sysctl(ipmib, sizeof(ipmib) / sizeof(ipmib[0]), &ipstat, &ii, NULL, 0) == -1, ipstat); HX((pid = getpid()) == -1, pid); HX((pid = getsid(pid)) == -1, pid); HX((pid = getppid()) == -1, pid); HX((pid = getpgid(0)) == -1, pid); HX((e = getpriority(0, 0)) == -1, e); if (!faster) { ts.tv_sec = 0; ts.tv_nsec = 1; (void) nanosleep(&ts, NULL); } HX(sigpending(&sigset) == -1, sigset); HX(sigprocmask(SIG_BLOCK, NULL, &sigset) == -1, sigset); HF(getentropy); /* an addr in this library */ HF(printf); /* an addr in libc */ p = (char *)&p; HD(p); /* an addr on stack */ p = (char *)&errno; HD(p); /* the addr of errno */ if (i == 0) { struct sockaddr_storage ss; struct statvfs stvfs; struct termios tios; struct statfs stfs; socklen_t ssl; off_t off; /* * Prime-sized mappings encourage fragmentation; * thus exposing some address entropy. */ struct mm { size_t npg; void *p; } mm[] = { { 17, MAP_FAILED }, { 3, MAP_FAILED }, { 11, MAP_FAILED }, { 2, MAP_FAILED }, { 5, MAP_FAILED }, { 3, MAP_FAILED }, { 7, MAP_FAILED }, { 1, MAP_FAILED }, { 57, MAP_FAILED }, { 3, MAP_FAILED }, { 131, MAP_FAILED }, { 1, MAP_FAILED }, }; for (m = 0; m < sizeof mm/sizeof(mm[0]); m++) { HX(mm[m].p = mmap(NULL, mm[m].npg * pgs, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANON, -1, (off_t)0), mm[m].p); if (mm[m].p != MAP_FAILED) { size_t mo; /* Touch some memory... */ p = mm[m].p; mo = cnt % (mm[m].npg * pgs - 1); p[mo] = 1; cnt += (int)((long)(mm[m].p) / pgs); } /* Check cnts and times... */ mach_time = mach_absolute_time(); HD(mach_time); cnt += (int)mach_time; HX((e = getrusage(RUSAGE_SELF, &ru)) == -1, ru); if (e != -1) { cnt += (int)ru.ru_utime.tv_sec; cnt += (int)ru.ru_utime.tv_usec; } } for (m = 0; m < sizeof mm/sizeof(mm[0]); m++) { if (mm[m].p != MAP_FAILED) munmap(mm[m].p, mm[m].npg * pgs); mm[m].p = MAP_FAILED; } HX(stat(".", &st) == -1, st); HX(statvfs(".", &stvfs) == -1, stvfs); HX(statfs(".", &stfs) == -1, stfs); HX(stat("/", &st) == -1, st); HX(statvfs("/", &stvfs) == -1, stvfs); HX(statfs("/", &stfs) == -1, stfs); HX((e = fstat(0, &st)) == -1, st); if (e == -1) { if (S_ISREG(st.st_mode) || S_ISFIFO(st.st_mode) || S_ISSOCK(st.st_mode)) { HX(fstatvfs(0, &stvfs) == -1, stvfs); HX(fstatfs(0, &stfs) == -1, stfs); HX((off = lseek(0, (off_t)0, SEEK_CUR)) < 0, off); } if (S_ISCHR(st.st_mode)) { HX(tcgetattr(0, &tios) == -1, tios); } else if (S_ISSOCK(st.st_mode)) { memset(&ss, 0, sizeof ss); ssl = sizeof(ss); HX(getpeername(0, (void *)&ss, &ssl) == -1, ss); } } HX((e = getrusage(RUSAGE_CHILDREN, &ru)) == -1, ru); if (e != -1) { cnt += (int)ru.ru_utime.tv_sec; cnt += (int)ru.ru_utime.tv_usec; } } else { /* Subsequent hashes absorb previous result */ HD(results); } HX((e = gettimeofday(&tv, NULL)) == -1, tv); if (e != -1) { cnt += (int)tv.tv_sec; cnt += (int)tv.tv_usec; } HD(cnt); } SHA512_Final(results, &ctx); memcpy((char *)buf + i, results, min(sizeof(results), len - i)); i += min(sizeof(results), len - i); } explicit_bzero(&ctx, sizeof ctx); explicit_bzero(results, sizeof results); if (gotdata(buf, len) == 0) { errno = save_errno; return 0; /* satisfied */ } errno = EIO; return -1; }