named/unix/os.c

/*
 * Copyright (C) Internet Systems Consortium, Inc. ("ISC")
 *
 * This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, you can obtain one at https://mozilla.org/MPL/2.0/.
 *
 * See the COPYRIGHT file distributed with this work for additional
 * information regarding copyright ownership.
 */

/*! \file */

#include <config.h>
#include <stdarg.h>
#include <stdbool.h>

#include <sys/types.h>	/* dev_t FreeBSD 2.1 */
#include <sys/stat.h>
#ifdef HAVE_UNAME
#include <sys/utsname.h>
#endif

#include <ctype.h>
#include <errno.h>
#include <fcntl.h>
#include <grp.h>		/* Required for initgroups() on IRIX. */
#include <pwd.h>
#include <stdio.h>
#include <stdlib.h>
#include <signal.h>
#include <syslog.h>
#ifdef HAVE_TZSET
#include <time.h>
#endif
#include <unistd.h>

#include <isc/buffer.h>
#include <isc/file.h>
#include <isc/print.h>
#include <isc/resource.h>
#include <isc/result.h>
#include <isc/strerror.h>
#include <isc/string.h>
#include <isc/util.h>

#include <named/globals.h>
#include <named/main.h>
#include <named/os.h>
#ifdef HAVE_LIBSCF
#include <named/ns_smf_globals.h>
#endif

static char *pidfile = NULL;
static char *lockfile = NULL;
static int devnullfd = -1;
static int singletonfd = -1;

#ifndef ISC_FACILITY
#define ISC_FACILITY LOG_DAEMON
#endif

/*
 * If there's no <linux/capability.h>, we don't care about <sys/prctl.h>
 */
#ifndef HAVE_LINUX_CAPABILITY_H
#undef HAVE_SYS_PRCTL_H
#endif

/*
 * Linux defines:
 *	(T) HAVE_LINUXTHREADS
 *	(C) HAVE_SYS_CAPABILITY_H (or HAVE_LINUX_CAPABILITY_H)
 *	(P) HAVE_SYS_PRCTL_H
 * The possible cases are:
 *	none:	setuid() normally
 *	T:	no setuid()
 *	C:	setuid() normally, drop caps (keep CAP_SETUID)
 *	T+C:	no setuid(), drop caps (don't keep CAP_SETUID)
 *	T+C+P:	setuid() early, drop caps (keep CAP_SETUID)
 *	C+P:	setuid() normally, drop caps (keep CAP_SETUID)
 *	P:	not possible
 *	T+P:	not possible
 *
 * if (C)
 *	caps = BIND_SERVICE + CHROOT + SETGID
 *	if ((T && C && P) || !T)
 *		caps += SETUID
 *	endif
 *	capset(caps)
 * endif
 * if (T && C && P && -u)
 *	setuid()
 * else if (T && -u)
 *	fail
 * --> start threads
 * if (!T && -u)
 *	setuid()
 * if (C && (P || !-u))
 *	caps = BIND_SERVICE
 *	capset(caps)
 * endif
 *
 * It will be nice when Linux threads work properly with setuid().
 */

#ifdef HAVE_LINUXTHREADS
static pid_t mainpid = 0;
#endif

static struct passwd *runas_pw = NULL;
static bool done_setuid = false;
static int dfd[2] = { -1, -1 };

#ifdef HAVE_LINUX_CAPABILITY_H

static bool non_root = false;
static bool non_root_caps = false;

#ifdef HAVE_SYS_CAPABILITY_H
#include <sys/capability.h>
#else
#ifdef HAVE_LINUX_TYPES_H
#include <linux/types.h>
#endif
/*%
 * We define _LINUX_FS_H to prevent it from being included.  We don't need
 * anything from it, and the files it includes cause warnings with 2.2
 * kernels, and compilation failures (due to conflicts between <linux/string.h>
 * and <string.h>) on 2.3 kernels.
 */
#define _LINUX_FS_H
#include <linux/capability.h>
#include <syscall.h>
#ifndef SYS_capset
#ifndef __NR_capset
#include <asm/unistd.h> /* Slackware 4.0 needs this. */
#endif /* __NR_capset */
#define SYS_capset __NR_capset
#endif /* SYS_capset */
#endif /* HAVE_SYS_CAPABILITY_H */

#ifdef HAVE_SYS_PRCTL_H
#include <sys/prctl.h>		/* Required for prctl(). */

/*
 * If the value of PR_SET_KEEPCAPS is not in <sys/prctl.h>, define it
 * here.  This allows setuid() to work on systems running a new enough
 * kernel but with /usr/include/linux pointing to "standard" kernel
 * headers.
 */
#ifndef PR_SET_KEEPCAPS
#define PR_SET_KEEPCAPS 8
#endif

#endif /* HAVE_SYS_PRCTL_H */

#ifdef HAVE_LIBCAP
#define SETCAPS_FUNC "cap_set_proc "
#else
typedef unsigned int cap_t;
#define SETCAPS_FUNC "syscall(capset) "
#endif /* HAVE_LIBCAP */

static void
linux_setcaps(cap_t caps) {
#ifndef HAVE_LIBCAP
	struct __user_cap_header_struct caphead;
	struct __user_cap_data_struct cap;
#endif
	char strbuf[ISC_STRERRORSIZE];

	if ((getuid() != 0 && !non_root_caps) || non_root)
		return;
#ifndef HAVE_LIBCAP
	memset(&caphead, 0, sizeof(caphead));
	caphead.version = _LINUX_CAPABILITY_VERSION;
	caphead.pid = 0;
	memset(&cap, 0, sizeof(cap));
	cap.effective = caps;
	cap.permitted = caps;
	cap.inheritable = 0;
#endif
#ifdef HAVE_LIBCAP
	if (cap_set_proc(caps) < 0) {
#else
	if (syscall(SYS_capset, &caphead, &cap) < 0) {
#endif
		isc__strerror(errno, strbuf, sizeof(strbuf));
		ns_main_earlyfatal(SETCAPS_FUNC "failed: %s:"
				   " please ensure that the capset kernel"
				   " module is loaded.  see insmod(8)",
				   strbuf);
	}
}

#ifdef HAVE_LIBCAP
#define SET_CAP(flag) \
	do { \
		cap_flag_value_t curval; \
		capval = (flag); \
		err = cap_get_flag(curcaps, capval, CAP_PERMITTED, &curval); \
		if (err != -1 && curval) { \
			err = cap_set_flag(caps, CAP_EFFECTIVE, 1, &capval, CAP_SET); \
			if (err == -1) { \
				isc__strerror(errno, strbuf, sizeof(strbuf)); \
				ns_main_earlyfatal("cap_set_proc failed: %s", strbuf); \
			} \
			\
			err = cap_set_flag(caps, CAP_PERMITTED, 1, &capval, CAP_SET); \
			if (err == -1) { \
				isc__strerror(errno, strbuf, sizeof(strbuf)); \
				ns_main_earlyfatal("cap_set_proc failed: %s", strbuf); \
			} \
		} \
	} while (0)
#define INIT_CAP \
	do { \
		caps = cap_init(); \
		if (caps == NULL) { \
			isc__strerror(errno, strbuf, sizeof(strbuf)); \
			ns_main_earlyfatal("cap_init failed: %s", strbuf); \
		} \
		curcaps = cap_get_proc(); \
		if (curcaps == NULL) { \
			isc__strerror(errno, strbuf, sizeof(strbuf)); \
			ns_main_earlyfatal("cap_get_proc failed: %s", strbuf); \
		} \
	} while (0)
#define FREE_CAP \
	{ \
		cap_free(caps); \
		cap_free(curcaps); \
	} while (0)
#else
#define SET_CAP(flag) \
	do { \
		if (curcaps & (1 << (flag))) { \
			caps |= (1 << (flag)); \
		} \
	} while (0)
#define INIT_CAP do { caps = 0; } while (0)
#endif /* HAVE_LIBCAP */

#ifndef HAVE_LIBCAP
/*%
 * Store the bitmask representing the permitted capability set in 'capsp'.  To
 * match libcap-enabled behavior, capget() syscall errors are not reported,
 * they just cause 'capsp' to be set to 0, which effectively prevents any
 * capability from being subsequently requested.
 */
static void
linux_getpermittedcaps(cap_t *capsp) {
	struct __user_cap_header_struct caphead;
	struct __user_cap_data_struct curcaps;

	memset(&caphead, 0, sizeof(caphead));
	caphead.version = _LINUX_CAPABILITY_VERSION;
	caphead.pid = 0;
	memset(&curcaps, 0, sizeof(curcaps));
	syscall(SYS_capget, &caphead, &curcaps);

	*capsp = curcaps.permitted;
}
#endif /* HAVE_LIBCAP */

static void
linux_initialprivs(void) {
	cap_t curcaps;
	cap_t caps;
#ifdef HAVE_LIBCAP
	cap_value_t capval;
	char strbuf[ISC_STRERRORSIZE];
	int err;
#else
	linux_getpermittedcaps(&curcaps);
#endif

	/*%
	 * We don't need most privileges, so we drop them right away.
	 * Later on linux_minprivs() will be called, which will drop our
	 * capabilities to the minimum needed to run the server.
	 */
	INIT_CAP;

	/*
	 * We need to be able to bind() to privileged ports, notably port 53!
	 */
	SET_CAP(CAP_NET_BIND_SERVICE);

	/*
	 * We need chroot() initially too.
	 */
	SET_CAP(CAP_SYS_CHROOT);

#if defined(HAVE_SYS_PRCTL_H) || !defined(HAVE_LINUXTHREADS)
	/*
	 * We can setuid() only if either the kernel supports keeping
	 * capabilities after setuid() (which we don't know until we've
	 * tried) or we're not using threads.  If either of these is
	 * true, we want the setuid capability.
	 */
	SET_CAP(CAP_SETUID);
#endif

	/*
	 * Since we call initgroups, we need this.
	 */
	SET_CAP(CAP_SETGID);

	/*
	 * Without this, we run into problems reading a configuration file
	 * owned by a non-root user and non-world-readable on startup.
	 */
	SET_CAP(CAP_DAC_READ_SEARCH);

	/*
	 * XXX  We might want to add CAP_SYS_RESOURCE, though it's not
	 *      clear it would work right given the way linuxthreads work.
	 * XXXDCL But since we need to be able to set the maximum number
	 * of files, the stack size, data size, and core dump size to
	 * support named.conf options, this is now being added to test.
	 */
	SET_CAP(CAP_SYS_RESOURCE);

	/*
	 * We need to be able to set the ownership of the containing
	 * directory of the pid file when we create it.
	 */
	SET_CAP(CAP_CHOWN);

	linux_setcaps(caps);

#ifdef HAVE_LIBCAP
	FREE_CAP;
#endif
}

static void
linux_minprivs(void) {
	cap_t curcaps;
	cap_t caps;
#ifdef HAVE_LIBCAP
	cap_value_t capval;
	char strbuf[ISC_STRERRORSIZE];
	int err;
#else
	linux_getpermittedcaps(&curcaps);
#endif

	INIT_CAP;
	/*%
	 * Drop all privileges except the ability to bind() to privileged
	 * ports.
	 *
	 * It's important that we drop CAP_SYS_CHROOT.  If we didn't, it
	 * chroot() could be used to escape from the chrooted area.
	 */

	SET_CAP(CAP_NET_BIND_SERVICE);

	/*
	 * XXX  We might want to add CAP_SYS_RESOURCE, though it's not
	 *      clear it would work right given the way linuxthreads work.
	 * XXXDCL But since we need to be able to set the maximum number
	 * of files, the stack size, data size, and core dump size to
	 * support named.conf options, this is now being added to test.
	 */
	SET_CAP(CAP_SYS_RESOURCE);

	linux_setcaps(caps);

#ifdef HAVE_LIBCAP
	FREE_CAP;
#endif
}

#ifdef HAVE_SYS_PRCTL_H
static void
linux_keepcaps(void) {
	char strbuf[ISC_STRERRORSIZE];
	/*%
	 * Ask the kernel to allow us to keep our capabilities after we
	 * setuid().
	 */

	if (prctl(PR_SET_KEEPCAPS, 1, 0, 0, 0) < 0) {
		if (errno != EINVAL) {
			isc__strerror(errno, strbuf, sizeof(strbuf));
			ns_main_earlyfatal("prctl() failed: %s", strbuf);
		}
	} else {
		non_root_caps = true;
		if (getuid() != 0)
			non_root = true;
	}
}
#endif

#endif	/* HAVE_LINUX_CAPABILITY_H */


static void
setup_syslog(const char *progname) {
	int options;

	options = LOG_PID;
#ifdef LOG_NDELAY
	options |= LOG_NDELAY;
#endif
	openlog(isc_file_basename(progname), options, ISC_FACILITY);
}

void
ns_os_init(const char *progname) {
	setup_syslog(progname);
#ifdef HAVE_LINUX_CAPABILITY_H
	linux_initialprivs();
#endif
#ifdef HAVE_LINUXTHREADS
	mainpid = getpid();
#endif
#ifdef SIGXFSZ
	signal(SIGXFSZ, SIG_IGN);
#endif
}

void
ns_os_daemonize(void) {
	pid_t pid;
	char strbuf[ISC_STRERRORSIZE];

	if (pipe(dfd) == -1) {
		isc__strerror(errno, strbuf, sizeof(strbuf));
		ns_main_earlyfatal("pipe(): %s", strbuf);
	}

	pid = fork();
	if (pid == -1) {
		isc__strerror(errno, strbuf, sizeof(strbuf));
		ns_main_earlyfatal("fork(): %s", strbuf);
	}
	if (pid != 0) {
		int n;
		/*
		 * Wait for the child to finish loading for the first time.
		 * This would be so much simpler if fork() worked once we
		 * were multi-threaded.
		 */
		(void)close(dfd[1]);
		do {
			char buf;
			n = read(dfd[0], &buf, 1);
			if (n == 1)
				_exit(0);
		} while (n == -1 && errno == EINTR);
		_exit(1);
	}
	(void)close(dfd[0]);

	/*
	 * We're the child.
	 */

#ifdef HAVE_LINUXTHREADS
	mainpid = getpid();
#endif

	if (setsid() == -1) {
		isc__strerror(errno, strbuf, sizeof(strbuf));
		ns_main_earlyfatal("setsid(): %s", strbuf);
	}

	/*
	 * Try to set stdin, stdout, and stderr to /dev/null, but press
	 * on even if it fails.
	 *
	 * XXXMLG The close() calls here are unneeded on all but NetBSD, but
	 * are harmless to include everywhere.  dup2() is supposed to close
	 * the FD if it is in use, but unproven-pthreads-0.16 is broken
	 * and will end up closing the wrong FD.  This will be fixed eventually,
	 * and these calls will be removed.
	 */
	if (devnullfd != -1) {
		if (devnullfd != STDIN_FILENO) {
			(void)close(STDIN_FILENO);
			(void)dup2(devnullfd, STDIN_FILENO);
		}
		if (devnullfd != STDOUT_FILENO) {
			(void)close(STDOUT_FILENO);
			(void)dup2(devnullfd, STDOUT_FILENO);
		}
		if (devnullfd != STDERR_FILENO && !ns_g_keepstderr) {
			(void)close(STDERR_FILENO);
			(void)dup2(devnullfd, STDERR_FILENO);
		}
	}
}

void
ns_os_started(void) {
	char buf = 0;

	/*
	 * Signal to the parent that we started successfully.
	 */
	if (dfd[0] != -1 && dfd[1] != -1) {
		if (write(dfd[1], &buf, 1) != 1)
			ns_main_earlyfatal("unable to signal parent that we "
					   "otherwise started successfully.");
		close(dfd[1]);
		dfd[0] = dfd[1] = -1;
	}
}

void
ns_os_opendevnull(void) {
	devnullfd = open("/dev/null", O_RDWR, 0);
}

void
ns_os_closedevnull(void) {
	if (devnullfd != STDIN_FILENO &&
	    devnullfd != STDOUT_FILENO &&
	    devnullfd != STDERR_FILENO) {
		close(devnullfd);
		devnullfd = -1;
	}
}

static bool
all_digits(const char *s) {
	if (*s == '\0')
		return (false);
	while (*s != '\0') {
		if (!isdigit((*s)&0xff))
			return (false);
		s++;
	}
	return (true);
}

void
ns_os_chroot(const char *root) {
	char strbuf[ISC_STRERRORSIZE];
#ifdef HAVE_LIBSCF
	ns_smf_chroot = 0;
#endif
	if (root != NULL) {
#ifdef HAVE_CHROOT
		if (chroot(root) < 0) {
			isc__strerror(errno, strbuf, sizeof(strbuf));
			ns_main_earlyfatal("chroot(): %s", strbuf);
		}
#else
		ns_main_earlyfatal("chroot(): disabled");
#endif
		if (chdir("/") < 0) {
			isc__strerror(errno, strbuf, sizeof(strbuf));
			ns_main_earlyfatal("chdir(/): %s", strbuf);
		}
#ifdef HAVE_LIBSCF
		/* Set ns_smf_chroot flag on successful chroot. */
		ns_smf_chroot = 1;
#endif
	}
}

void
ns_os_inituserinfo(const char *username) {
	char strbuf[ISC_STRERRORSIZE];
	if (username == NULL)
		return;

	if (all_digits(username))
		runas_pw = getpwuid((uid_t)atoi(username));
	else
		runas_pw = getpwnam(username);
	endpwent();

	if (runas_pw == NULL)
		ns_main_earlyfatal("user '%s' unknown", username);

	if (getuid() == 0) {
		if (initgroups(runas_pw->pw_name, runas_pw->pw_gid) < 0) {
			isc__strerror(errno, strbuf, sizeof(strbuf));
			ns_main_earlyfatal("initgroups(): %s", strbuf);
		}
	}

}

void
ns_os_changeuser(void) {
	char strbuf[ISC_STRERRORSIZE];
	if (runas_pw == NULL || done_setuid)
		return;

	done_setuid = true;

#ifdef HAVE_LINUXTHREADS
#ifdef HAVE_LINUX_CAPABILITY_H
	if (!non_root_caps)
		ns_main_earlyfatal("-u with Linux threads not supported: "
				   "requires kernel support for "
				   "prctl(PR_SET_KEEPCAPS)");
#else
	ns_main_earlyfatal("-u with Linux threads not supported: "
			   "no capabilities support or capabilities "
			   "disabled at build time");
#endif
#endif

	if (setgid(runas_pw->pw_gid) < 0) {
		isc__strerror(errno, strbuf, sizeof(strbuf));
		ns_main_earlyfatal("setgid(): %s", strbuf);
	}

	if (setuid(runas_pw->pw_uid) < 0) {
		isc__strerror(errno, strbuf, sizeof(strbuf));
		ns_main_earlyfatal("setuid(): %s", strbuf);
	}

#if defined(HAVE_SYS_PRCTL_H) && defined(PR_SET_DUMPABLE)
	/*
	 * Restore the ability of named to drop core after the setuid()
	 * call has disabled it.
	 */
	if (prctl(PR_SET_DUMPABLE,1,0,0,0) < 0) {
		isc__strerror(errno, strbuf, sizeof(strbuf));
		ns_main_earlywarning("prctl(PR_SET_DUMPABLE) failed: %s",
				     strbuf);
	}
#endif
#if defined(HAVE_LINUX_CAPABILITY_H) && !defined(HAVE_LINUXTHREADS)
	linux_minprivs();
#endif
}

uid_t
ns_os_uid(void) {
	if (runas_pw == NULL)
		return (0);
	return (runas_pw->pw_uid);
}

void
ns_os_adjustnofile(void) {
#ifdef HAVE_LINUXTHREADS
	isc_result_t result;
	isc_resourcevalue_t newvalue;

	/*
	 * Linux: max number of open files specified by one thread doesn't seem
	 * to apply to other threads on Linux.
	 */
	newvalue = ISC_RESOURCE_UNLIMITED;

	result = isc_resource_setlimit(isc_resource_openfiles, newvalue);
	if (result != ISC_R_SUCCESS)
		ns_main_earlywarning("couldn't adjust limit on open files");
#endif
}

void
ns_os_minprivs(void) {
#ifdef HAVE_SYS_PRCTL_H
	linux_keepcaps();
#endif

#ifdef HAVE_LINUXTHREADS
	ns_os_changeuser(); /* Call setuid() before threads are started */
#endif

#if defined(HAVE_LINUX_CAPABILITY_H) && defined(HAVE_LINUXTHREADS)
	linux_minprivs();
#endif
}

static int
safe_open(const char *filename, mode_t mode, bool append) {
	int fd;
	struct stat sb;

	if (stat(filename, &sb) == -1) {
		if (errno != ENOENT)
			return (-1);
	} else if ((sb.st_mode & S_IFREG) == 0) {
		errno = EOPNOTSUPP;
		return (-1);
	}

	if (append)
		fd = open(filename, O_WRONLY|O_CREAT|O_APPEND, mode);
	else {
		if (unlink(filename) < 0 && errno != ENOENT)
			return (-1);
		fd = open(filename, O_WRONLY|O_CREAT|O_EXCL, mode);
	}
	return (fd);
}

static void
cleanup_pidfile(void) {
	int n;
	if (pidfile != NULL) {
		n = unlink(pidfile);
		if (n == -1 && errno != ENOENT)
			ns_main_earlywarning("unlink '%s': failed", pidfile);
		free(pidfile);
	}
	pidfile = NULL;
}

static void
cleanup_lockfile(void) {
	if (singletonfd != -1) {
		close(singletonfd);
		singletonfd = -1;
	}

	if (lockfile != NULL) {
		int n = unlink(lockfile);
		if (n == -1 && errno != ENOENT)
			ns_main_earlywarning("unlink '%s': failed", lockfile);
		free(lockfile);
		lockfile = NULL;
	}
}

/*
 * Ensure that a directory exists.
 * NOTE: This function overwrites the '/' characters in 'filename' with
 * nulls. The caller should copy the filename to a fresh buffer first.
 */
static int
mkdirpath(char *filename, void (*report)(const char *, ...)) {
	char *slash = strrchr(filename, '/');
	char strbuf[ISC_STRERRORSIZE];
	unsigned int mode;

	if (slash != NULL && slash != filename) {
		struct stat sb;
		*slash = '\0';

		if (stat(filename, &sb) == -1) {
			if (errno != ENOENT) {
				isc__strerror(errno, strbuf, sizeof(strbuf));
				(*report)("couldn't stat '%s': %s", filename,
					  strbuf);
				goto error;
			}
			if (mkdirpath(filename, report) == -1)
				goto error;
			/*
			 * Handle "//", "/./" and "/../" in path.
			 */
			if (!strcmp(slash + 1, "") ||
			    !strcmp(slash + 1, ".") ||
			    !strcmp(slash + 1, "..")) {
				*slash = '/';
				return (0);
			}
			mode = S_IRUSR | S_IWUSR | S_IXUSR;	/* u=rwx */
			mode |= S_IRGRP | S_IXGRP;		/* g=rx */
			mode |= S_IROTH | S_IXOTH;		/* o=rx */
			if (mkdir(filename, mode) == -1) {
				isc__strerror(errno, strbuf, sizeof(strbuf));
				(*report)("couldn't mkdir '%s': %s", filename,
					  strbuf);
				goto error;
			}
			if (runas_pw != NULL &&
			    chown(filename, runas_pw->pw_uid,
				  runas_pw->pw_gid) == -1) {
				isc__strerror(errno, strbuf, sizeof(strbuf));
				(*report)("couldn't chown '%s': %s", filename,
					  strbuf);
			}
		}
		*slash = '/';
	}
	return (0);

 error:
	*slash = '/';
	return (-1);
}

#ifndef HAVE_LINUXTHREADS
static void
setperms(uid_t uid, gid_t gid) {
#if defined(HAVE_SETEGID) || defined(HAVE_SETRESGID)
	char strbuf[ISC_STRERRORSIZE];
#endif
#if !defined(HAVE_SETEGID) && defined(HAVE_SETRESGID)
	gid_t oldgid, tmpg;
#endif
#if !defined(HAVE_SETEUID) && defined(HAVE_SETRESUID)
	uid_t olduid, tmpu;
#endif
#if defined(HAVE_SETEGID)
	if (getegid() != gid && setegid(gid) == -1) {
		isc__strerror(errno, strbuf, sizeof(strbuf));
		ns_main_earlywarning("unable to set effective gid to %ld: %s",
				     (long)gid, strbuf);
	}
#elif defined(HAVE_SETRESGID)
	if (getresgid(&tmpg, &oldgid, &tmpg) == -1 || oldgid != gid) {
		if (setresgid(-1, gid, -1) == -1) {
			isc__strerror(errno, strbuf, sizeof(strbuf));
			ns_main_earlywarning("unable to set effective "
					     "gid to %d: %s", gid, strbuf);
		}
	}
#endif

#if defined(HAVE_SETEUID)
	if (geteuid() != uid && seteuid(uid) == -1) {
		isc__strerror(errno, strbuf, sizeof(strbuf));
		ns_main_earlywarning("unable to set effective uid to %ld: %s",
				     (long)uid, strbuf);
	}
#elif defined(HAVE_SETRESUID)
	if (getresuid(&tmpu, &olduid, &tmpu) == -1 || olduid != uid) {
		if (setresuid(-1, uid, -1) == -1) {
			isc__strerror(errno, strbuf, sizeof(strbuf));
			ns_main_earlywarning("unable to set effective "
					     "uid to %d: %s", uid, strbuf);
		}
	}
#endif
}
#endif /* !HAVE_LINUXTHREADS */

FILE *
ns_os_openfile(const char *filename, mode_t mode, bool switch_user) {
	char strbuf[ISC_STRERRORSIZE], *f;
	FILE *fp;
	int fd;

	/*
	 * Make the containing directory if it doesn't exist.
	 */
	f = strdup(filename);
	if (f == NULL) {
		isc__strerror(errno, strbuf, sizeof(strbuf));
		ns_main_earlywarning("couldn't strdup() '%s': %s",
				     filename, strbuf);
		return (NULL);
	}
	if (mkdirpath(f, ns_main_earlywarning) == -1) {
		free(f);
		return (NULL);
	}
	free(f);

	if (switch_user && runas_pw != NULL) {
		uid_t olduid = getuid();
		gid_t oldgid = getgid();
#ifdef HAVE_LINUXTHREADS
		REQUIRE(olduid == runas_pw->pw_uid);
		REQUIRE(oldgid == runas_pw->pw_gid);
#else
		/* Set UID/GID to the one we'll be running with eventually */
		setperms(runas_pw->pw_uid, runas_pw->pw_gid);
#endif
		fd = safe_open(filename, mode, false);

#ifndef HAVE_LINUXTHREADS
		/* Restore UID/GID to root */
		setperms(olduid, oldgid);
#endif /* HAVE_LINUXTHREADS */

		if (fd == -1) {
#ifndef HAVE_LINUXTHREADS
			fd = safe_open(filename, mode, false);
			if (fd != -1) {
				ns_main_earlywarning("Required root "
						     "permissions to open "
						     "'%s'.", filename);
			} else {
				ns_main_earlywarning("Could not open "
						     "'%s'.", filename);
			}
			ns_main_earlywarning("Please check file and "
					     "directory permissions "
					     "or reconfigure the filename.");
#else /* HAVE_LINUXTHREADS */
			ns_main_earlywarning("Could not open "
					     "'%s'.", filename);
			ns_main_earlywarning("Please check file and "
					     "directory permissions "
					     "or reconfigure the filename.");
#endif /* HAVE_LINUXTHREADS */
		}
	} else {
		fd = safe_open(filename, mode, false);
	}

	if (fd < 0) {
		isc__strerror(errno, strbuf, sizeof(strbuf));
		ns_main_earlywarning("could not open file '%s': %s",
				     filename, strbuf);
		return (NULL);
	}

	fp = fdopen(fd, "w");
	if (fp == NULL) {
		isc__strerror(errno, strbuf, sizeof(strbuf));
		ns_main_earlywarning("could not fdopen() file '%s': %s",
				     filename, strbuf);
	}

	return (fp);
}

void
ns_os_writepidfile(const char *filename, bool first_time) {
	FILE *fh;
	pid_t pid;
	char strbuf[ISC_STRERRORSIZE];
	void (*report)(const char *, ...);

	/*
	 * The caller must ensure any required synchronization.
	 */

	report = first_time ? ns_main_earlyfatal : ns_main_earlywarning;

	cleanup_pidfile();

	if (filename == NULL)
		return;

	pidfile = strdup(filename);
	if (pidfile == NULL) {
		isc__strerror(errno, strbuf, sizeof(strbuf));
		(*report)("couldn't strdup() '%s': %s", filename, strbuf);
		return;
	}

	fh = ns_os_openfile(filename, S_IRUSR|S_IWUSR|S_IRGRP|S_IROTH,
				  first_time);
	if (fh == NULL) {
		cleanup_pidfile();
		return;
	}
#ifdef HAVE_LINUXTHREADS
	pid = mainpid;
#else
	pid = getpid();
#endif
	if (fprintf(fh, "%ld\n", (long)pid) < 0) {
		(*report)("fprintf() to pid file '%s' failed", filename);
		(void)fclose(fh);
		cleanup_pidfile();
		return;
	}
	if (fflush(fh) == EOF) {
		(*report)("fflush() to pid file '%s' failed", filename);
		(void)fclose(fh);
		cleanup_pidfile();
		return;
	}
	(void)fclose(fh);
}

bool
ns_os_issingleton(const char *filename) {
	char strbuf[ISC_STRERRORSIZE];
	struct flock lock;

	if (singletonfd != -1)
		return (true);

	if (strcasecmp(filename, "none") == 0)
		return (true);

	/*
	 * Make the containing directory if it doesn't exist.
	 */
	lockfile = strdup(filename);
	if (lockfile == NULL) {
		isc__strerror(errno, strbuf, sizeof(strbuf));
		ns_main_earlyfatal("couldn't allocate memory for '%s': %s",
				   filename, strbuf);
	} else {
		int ret = mkdirpath(lockfile, ns_main_earlywarning);
		if (ret == -1) {
			ns_main_earlywarning("couldn't create '%s'", filename);
			cleanup_lockfile();
			return (false);
		}
	}

	/*
	 * ns_os_openfile() uses safeopen() which removes any existing
	 * files. We can't use that here.
	 */
	singletonfd = open(filename, O_WRONLY | O_CREAT,
			   S_IRUSR|S_IWUSR|S_IRGRP|S_IROTH);
	if (singletonfd == -1) {
		cleanup_lockfile();
		return (false);
	}

	memset(&lock, 0, sizeof(lock));
	lock.l_type = F_WRLCK;
	lock.l_whence = SEEK_SET;
	lock.l_start = 0;
	lock.l_len = 1;

	/* Non-blocking (does not wait for lock) */
	if (fcntl(singletonfd, F_SETLK, &lock) == -1) {
		close(singletonfd);
		singletonfd = -1;
		return (false);
	}

	return (true);
}

void
ns_os_shutdown(void) {
	closelog();
	cleanup_pidfile();
	cleanup_lockfile();
}

isc_result_t
ns_os_gethostname(char *buf, size_t len) {
	int n;

	n = gethostname(buf, len);
	return ((n == 0) ? ISC_R_SUCCESS : ISC_R_FAILURE);
}

static char *
next_token(char **stringp, const char *delim) {
	char *res;

	do {
		res = strsep(stringp, delim);
		if (res == NULL)
			break;
	} while (*res == '\0');
	return (res);
}

void
ns_os_shutdownmsg(char *command, isc_buffer_t *text) {
	char *input, *ptr;
	unsigned int n;
	pid_t pid;

	input = command;

	/* Skip the command name. */
	ptr = next_token(&input, " \t");
	if (ptr == NULL)
		return;

	ptr = next_token(&input, " \t");
	if (ptr == NULL)
		return;

	if (strcmp(ptr, "-p") != 0)
		return;

#ifdef HAVE_LINUXTHREADS
	pid = mainpid;
#else
	pid = getpid();
#endif

	n = snprintf((char *)isc_buffer_used(text),
		     isc_buffer_availablelength(text),
		     "pid: %ld", (long)pid);
	/* Only send a message if it is complete. */
	if (n > 0 && n < isc_buffer_availablelength(text))
		isc_buffer_add(text, n);
}

void
ns_os_tzset(void) {
#ifdef HAVE_TZSET
	tzset();
#endif
}

#ifdef HAVE_UNAME
static char unamebuf[sizeof(struct utsname)];
#else
static const char unamebuf[] = { "unknown architecture" };
#endif
static const char *unamep = NULL;

static void
getuname(void) {
#ifdef HAVE_UNAME
	struct utsname uts;

	memset(&uts, 0, sizeof(uts));
	if (uname(&uts) < 0) {
		snprintf(unamebuf, sizeof(unamebuf), "unknown architecture");
		return;
	}

	snprintf(unamebuf, sizeof(unamebuf), "%s %s %s %s", uts.sysname,
		 uts.machine, uts.release, uts.version);
#endif /* ifdef HAVE_UNAME */
	unamep = unamebuf;
}

const char *
ns_os_uname(void) {
	if (unamep == NULL) {
		getuname();
	}
	return (unamep);
}