NetBSD/sbin/mount/mount.c

777 lines
19 KiB
C

/* $NetBSD: mount.c,v 1.102 2016/10/09 21:03:43 christos Exp $ */
/*
* Copyright (c) 1980, 1989, 1993, 1994
* The Regents of the University of California. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
#include <sys/cdefs.h>
#ifndef lint
__COPYRIGHT("@(#) Copyright (c) 1980, 1989, 1993, 1994\
The Regents of the University of California. All rights reserved.");
#endif /* not lint */
#ifndef lint
#if 0
static char sccsid[] = "@(#)mount.c 8.25 (Berkeley) 5/8/95";
#else
__RCSID("$NetBSD: mount.c,v 1.102 2016/10/09 21:03:43 christos Exp $");
#endif
#endif /* not lint */
#include <sys/param.h>
#include <sys/mount.h>
#include <sys/wait.h>
#include <fs/puffs/puffs_msgif.h>
#include <err.h>
#include <errno.h>
#include <fstab.h>
#include <pwd.h>
#include <signal.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <util.h>
#define MOUNTNAMES
#include <fcntl.h>
#include <sys/disk.h>
#include <sys/disklabel.h>
#include <sys/ioctl.h>
#include "pathnames.h"
#include "mountprog.h"
static int debug, verbose;
static void catopt(char **, const char *);
static const char *
getfslab(const char *str);
static struct statvfs *
getmntpt(const char *);
static int getmntargs(struct statvfs *, char *, size_t);
static int hasopt(const char *, const char *);
static void mangle(char *, int *, const char ** volatile *, int *);
static int mountfs(const char *, const char *, const char *,
int, const char *, const char *, int, char *, size_t);
static void prmount(struct statvfs *);
__dead static void usage(void);
/* Map from mount options to printable formats. */
static const struct opt {
int o_opt;
int o_silent;
const char *o_name;
} optnames[] = {
__MNT_FLAGS
};
static const char ffs_fstype[] = "ffs";
int
main(int argc, char *argv[])
{
const char *mntfromname, *mntonname, **vfslist, *vfstype;
struct fstab *fs;
struct statvfs *mntbuf;
#if 0
FILE *mountdfp;
#endif
int all, ch, forceall, i, init_flags, mntsize, rval;
char *options;
const char *mountopts, *fstypename;
char canonical_path_buf[MAXPATHLEN], buf[MAXPATHLEN];
char *canonical_path;
/* started as "mount" */
all = forceall = init_flags = 0;
options = NULL;
vfslist = NULL;
vfstype = ffs_fstype;
while ((ch = getopt(argc, argv, "Aadfo:rwt:uv")) != -1)
switch (ch) {
case 'A':
all = forceall = 1;
break;
case 'a':
all = 1;
break;
case 'd':
debug = 1;
break;
case 'f':
init_flags |= MNT_FORCE;
break;
case 'o':
if (*optarg)
catopt(&options, optarg);
break;
case 'r':
init_flags |= MNT_RDONLY;
break;
case 't':
if (vfslist != NULL)
errx(EXIT_FAILURE,
"Only one -t option may be specified.");
vfslist = makevfslist(optarg);
vfstype = optarg;
break;
case 'u':
init_flags |= MNT_UPDATE;
break;
case 'v':
verbose++;
break;
case 'w':
init_flags &= ~MNT_RDONLY;
break;
case '?':
default:
usage();
/* NOTREACHED */
}
argc -= optind;
argv += optind;
#define BADTYPE(type) \
(strcmp(type, FSTAB_RO) && \
strcmp(type, FSTAB_RW) && strcmp(type, FSTAB_RQ))
rval = 0;
switch (argc) {
case 0:
if (all)
while ((fs = getfsent()) != NULL) {
if (BADTYPE(fs->fs_type))
continue;
if (checkvfsname(fs->fs_vfstype, vfslist))
continue;
if (hasopt(fs->fs_mntops, "noauto"))
continue;
if (strcmp(fs->fs_spec, "from_mount") == 0) {
if ((mntbuf = getmntpt(fs->fs_file))
== NULL)
errx(EXIT_FAILURE,
"Unknown file system %s",
fs->fs_file);
mntfromname = mntbuf->f_mntfromname;
} else
mntfromname = fs->fs_spec;
mntfromname = getfsspecname(buf, sizeof(buf),
mntfromname);
if (mntfromname == NULL)
err(EXIT_FAILURE, "%s", buf);
if (mountfs(fs->fs_vfstype, mntfromname,
fs->fs_file, init_flags, options,
fs->fs_mntops, !forceall, NULL, 0))
rval = 1;
}
else {
if ((mntsize = getmntinfo(&mntbuf, MNT_NOWAIT)) == 0)
err(EXIT_FAILURE, "getmntinfo");
for (i = 0; i < mntsize; i++) {
if (checkvfsname(mntbuf[i].f_fstypename,
vfslist))
continue;
prmount(&mntbuf[i]);
}
}
return rval;
case 1:
if (vfslist != NULL) {
usage();
/* NOTREACHED */
}
/*
* Create a canonical version of the device or mount path
* passed to us. It's ok for this to fail. It's also ok
* for the result to be exactly the same as the original.
*/
canonical_path = realpath(*argv, canonical_path_buf);
if (init_flags & MNT_UPDATE) {
/*
* Try looking up the canonical path first,
* then try exactly what the user entered.
*/
if ((canonical_path == NULL ||
(mntbuf = getmntpt(canonical_path)) == NULL) &&
(mntbuf = getmntpt(*argv)) == NULL) {
out:
errx(EXIT_FAILURE,
"Unknown special file or file system `%s'",
*argv);
}
mntfromname = mntbuf->f_mntfromname;
if ((fs = getfsfile(mntbuf->f_mntonname)) != NULL) {
if (strcmp(fs->fs_spec, "from_mount") != 0)
mntfromname = fs->fs_spec;
/* ignore the fstab file options. */
fs->fs_mntops = NULL;
}
mntonname = mntbuf->f_mntonname;
fstypename = mntbuf->f_fstypename;
mountopts = NULL;
} else {
/*
* Try looking up the canonical path first,
* then try exactly what the user entered.
*/
if (canonical_path == NULL ||
((fs = getfsfile(canonical_path)) == NULL &&
(fs = getfsspec(canonical_path)) == NULL))
{
if ((fs = getfsfile(*argv)) == NULL &&
(fs = getfsspec(*argv)) == NULL) {
goto out;
}
}
if (BADTYPE(fs->fs_type))
errx(EXIT_FAILURE,
"Unknown file system type for `%s'",
*argv);
if (strcmp(fs->fs_spec, "from_mount") == 0) {
if ((canonical_path == NULL ||
(mntbuf = getmntpt(canonical_path))
== NULL) &&
(mntbuf = getmntpt(*argv)) == NULL)
goto out;
mntfromname = mntbuf->f_mntfromname;
} else
mntfromname = fs->fs_spec;
mntonname = fs->fs_file;
fstypename = fs->fs_vfstype;
mountopts = fs->fs_mntops;
}
mntfromname = getfsspecname(buf, sizeof(buf), mntfromname);
if (mntfromname == NULL)
err(EXIT_FAILURE, "%s", buf);
rval = mountfs(fstypename, mntfromname,
mntonname, init_flags, options, mountopts, 0, NULL, 0);
break;
case 2:
/*
* If -t flag has not been specified, and spec contains either
* a ':' or a '@' then assume that an NFS filesystem is being
* specified ala Sun.
*/
mntfromname = getfsspecname(buf, sizeof(buf), argv[0]);
if (mntfromname == NULL)
err(EXIT_FAILURE, "%s", buf);
if (vfslist == NULL) {
if (strpbrk(argv[0], ":@") != NULL) {
fprintf(stderr, "WARNING: autoselecting nfs "
"based on : or @ in the device name is "
"deprecated!\n"
"WARNING: This behaviour will be removed "
"in a future release\n");
vfstype = "nfs";
} else {
vfstype = getfslab(mntfromname);
if (vfstype == NULL)
vfstype = ffs_fstype;
}
}
rval = mountfs(vfstype, mntfromname, argv[1], init_flags,
options, NULL, 0, NULL, 0);
break;
default:
usage();
}
#if 0 /* disabled because it interferes the service. */
/*
* If the mount was successfully, and done by root, tell mountd the
* good news. Pid checks are probably unnecessary, but don't hurt.
*/
if (rval == 0 && getuid() == 0 &&
(mountdfp = fopen(_PATH_MOUNTDPID, "r")) != NULL) {
int pid;
if (fscanf(mountdfp, "%d", &pid) == 1 &&
pid > 0 && kill(pid, SIGHUP) == -1 && errno != ESRCH)
err(EXIT_FAILURE, "signal mountd");
(void)fclose(mountdfp);
}
#endif
return rval;
}
int
hasopt(const char *mntopts, const char *option)
{
int negative, found;
char *opt, *optbuf;
if (option[0] == 'n' && option[1] == 'o') {
negative = 1;
option += 2;
} else
negative = 0;
optbuf = estrdup(mntopts);
found = 0;
for (opt = optbuf; (opt = strtok(opt, ",")) != NULL; opt = NULL) {
if (opt[0] == 'n' && opt[1] == 'o') {
if (!strcasecmp(opt + 2, option))
found = negative;
} else if (!strcasecmp(opt, option))
found = !negative;
}
free(optbuf);
return (found);
}
static int
mountfs(const char *vfstype, const char *spec, const char *name,
int flags, const char *options, const char *mntopts,
int skipmounted, char *buf, size_t buflen)
{
/* List of directories containing mount_xxx subcommands. */
static const char *edirs[] = {
#ifdef RESCUEDIR
RESCUEDIR,
#endif
_PATH_SBIN,
_PATH_USRSBIN,
NULL
};
const char ** volatile argv, **edir;
struct statvfs *sfp, sf;
pid_t pid;
int pfd[2];
int argc, numfs, i, status, maxargc;
char *optbuf, execname[MAXPATHLEN + 1], execbase[MAXPATHLEN],
mntpath[MAXPATHLEN];
volatile int getargs;
if (realpath(name, mntpath) == NULL) {
warn("realpath `%s'", name);
return (1);
}
name = mntpath;
optbuf = NULL;
if (mntopts)
catopt(&optbuf, mntopts);
if (options) {
catopt(&optbuf, options);
getargs = strstr(options, "getargs") != NULL;
} else
getargs = 0;
if (!mntopts && !options)
catopt(&optbuf, "rw");
if (getargs == 0 && strcmp(name, "/") == 0 && !hasopt(optbuf, "union"))
flags |= MNT_UPDATE;
else if (skipmounted) {
if ((numfs = getmntinfo(&sfp, MNT_WAIT)) == 0) {
warn("getmntinfo");
return (1);
}
for(i = 0; i < numfs; i++) {
const char *mountedtype = sfp[i].f_fstypename;
size_t cmplen = sizeof(sfp[i].f_fstypename);
/* remove "puffs|" from comparisons, if present */
#define TYPESIZE (sizeof(PUFFS_TYPEPREFIX)-1)
if (strncmp(mountedtype,
PUFFS_TYPEPREFIX, TYPESIZE) == 0) {
mountedtype += TYPESIZE;
cmplen -= TYPESIZE;
}
/*
* XXX can't check f_mntfromname,
* thanks to mfs, union, etc.
*/
if (strncmp(name, sfp[i].f_mntonname, MNAMELEN) == 0 &&
strncmp(vfstype, mountedtype, cmplen) == 0) {
if (verbose)
(void)printf("%s on %s type %.*s: "
"%s\n",
sfp[i].f_mntfromname,
sfp[i].f_mntonname,
(int)sizeof(sfp[i].f_fstypename),
sfp[i].f_fstypename,
"already mounted");
return (0);
}
}
}
if (flags & MNT_FORCE)
catopt(&optbuf, "force");
if (flags & MNT_RDONLY)
catopt(&optbuf, "ro");
if (flags & MNT_UPDATE) {
catopt(&optbuf, "update");
/* Figure out the fstype only if we defaulted to ffs */
if (vfstype == ffs_fstype && statvfs(name, &sf) != -1)
vfstype = sf.f_fstypename;
}
maxargc = 64;
argv = ecalloc(maxargc, sizeof(*argv));
if (getargs &&
strncmp(vfstype, PUFFS_TYPEPREFIX, sizeof(PUFFS_TYPEPREFIX)-1) == 0)
(void)snprintf(execbase, sizeof(execbase), "mount_puffs");
else if (hasopt(optbuf, "rump"))
(void)snprintf(execbase, sizeof(execbase), "rump_%s", vfstype);
else
(void)snprintf(execbase, sizeof(execbase), "mount_%s", vfstype);
argc = 0;
argv[argc++] = execbase;
if (optbuf)
mangle(optbuf, &argc, &argv, &maxargc);
argv[argc++] = spec;
argv[argc++] = name;
argv[argc] = NULL;
if ((verbose && buf == NULL) || debug) {
(void)printf("exec:");
for (i = 0; i < argc; i++)
(void)printf(" %s", argv[i]);
(void)printf("\n");
}
if (buf) {
if (pipe(pfd) == -1)
warn("Cannot create pipe");
}
switch (pid = fork()) {
case -1: /* Error. */
warn("fork");
if (optbuf)
free(optbuf);
free(argv);
return (1);
case 0: /* Child. */
if (debug)
_exit(0);
if (buf) {
(void)close(pfd[0]);
(void)close(STDOUT_FILENO);
if (dup2(pfd[1], STDOUT_FILENO) == -1)
warn("Cannot open fd to mount program");
}
/* Go find an executable. */
edir = edirs;
do {
(void)snprintf(execname,
sizeof(execname), "%s/%s", *edir, execbase);
(void)execv(execname, __UNCONST(argv));
if (errno != ENOENT)
warn("exec %s for %s", execname, name);
} while (*++edir != NULL);
if (errno == ENOENT)
warn("exec %s for %s: %s", execbase, name, execbase);
_exit(1);
/* NOTREACHED */
default: /* Parent. */
if (optbuf)
free(optbuf);
free(argv);
if (buf || getargs) {
char tbuf[1024], *ptr;
int nread;
if (buf == NULL) {
ptr = tbuf;
buflen = sizeof(tbuf) - 1;
} else {
ptr = buf;
buflen--;
}
(void)close(pfd[1]);
(void)signal(SIGPIPE, SIG_IGN);
while ((nread = read(pfd[0], ptr, buflen)) > 0) {
buflen -= nread;
ptr += nread;
}
*ptr = '\0';
if (buflen == 0) {
while (read(pfd[0], &nread, sizeof(nread)) > 0)
continue;
}
if (buf == NULL)
(void)fprintf(stdout, "%s", tbuf);
}
if (waitpid(pid, &status, 0) == -1) {
warn("waitpid");
return (1);
}
if (WIFEXITED(status)) {
if (WEXITSTATUS(status) != 0)
return (WEXITSTATUS(status));
} else if (WIFSIGNALED(status)) {
warnx("%s: %s", name, strsignal(WTERMSIG(status)));
return (1);
}
if (buf == NULL) {
if (verbose) {
if (statvfs(name, &sf) == -1) {
warn("statvfs %s", name);
return (1);
}
prmount(&sf);
}
}
break;
}
return (0);
}
static void
prmount(struct statvfs *sfp)
{
int flags;
const struct opt *o;
struct passwd *pw;
int f;
(void)printf("%s on %s type %.*s", sfp->f_mntfromname,
sfp->f_mntonname, (int)sizeof(sfp->f_fstypename),
sfp->f_fstypename);
flags = sfp->f_flag & MNT_VISFLAGMASK;
for (f = 0, o = optnames; flags && o <
&optnames[sizeof(optnames)/sizeof(optnames[0])]; o++)
if (flags & o->o_opt) {
if (!o->o_silent || verbose)
(void)printf("%s%s", !f++ ? " (" : ", ",
o->o_name);
flags &= ~o->o_opt;
}
if (flags)
(void)printf("%sunknown flag%s %#x", !f++ ? " (" : ", ",
flags & (flags - 1) ? "s" : "", flags);
if (sfp->f_owner) {
(void)printf("%smounted by ", !f++ ? " (" : ", ");
if ((pw = getpwuid(sfp->f_owner)) != NULL)
(void)printf("%s", pw->pw_name);
else
(void)printf("%d", sfp->f_owner);
}
if (verbose)
(void)printf("%sfsid: 0x%x/0x%x",
!f++ ? " (" /* ) */: ", ",
sfp->f_fsidx.__fsid_val[0], sfp->f_fsidx.__fsid_val[1]);
if (verbose) {
(void)printf("%s", !f++ ? " (" : ", ");
(void)printf("reads: sync %" PRIu64 " async %" PRIu64 "",
sfp->f_syncreads, sfp->f_asyncreads);
(void)printf(", writes: sync %" PRIu64 " async %" PRIu64 "",
sfp->f_syncwrites, sfp->f_asyncwrites);
if (verbose > 1) {
char buf[2048];
if (getmntargs(sfp, buf, sizeof(buf)))
printf(", [%s: %s]", sfp->f_fstypename, buf);
}
printf(")\n");
} else
(void)printf("%s", f ? ")\n" : "\n");
}
static int
getmntargs(struct statvfs *sfs, char *buf, size_t buflen)
{
if (mountfs(sfs->f_fstypename, sfs->f_mntfromname, sfs->f_mntonname, 0,
"getargs", NULL, 0, buf, buflen))
return (0);
else {
if (*buf == '\0')
return (0);
if ((buf = strchr(buf, '\n')) != NULL)
*buf = '\0';
return (1);
}
}
static struct statvfs *
getmntpt(const char *name)
{
struct statvfs *mntbuf;
int i, mntsize;
mntsize = getmntinfo(&mntbuf, MNT_NOWAIT);
for (i = 0; i < mntsize; i++)
if (strcmp(mntbuf[i].f_mntfromname, name) == 0 ||
strcmp(mntbuf[i].f_mntonname, name) == 0)
return (&mntbuf[i]);
return (NULL);
}
static void
catopt(char **sp, const char *o)
{
char *s, *n;
s = *sp;
if (s) {
easprintf(&n, "%s,%s", s, o);
free(s);
s = n;
} else
s = estrdup(o);
*sp = s;
}
static void
mangle(char *options, int *argcp, const char ** volatile *argvp, int *maxargcp)
{
char *p, *s;
int argc, maxargc;
const char **argv, **nargv;
argc = *argcp;
argv = *argvp;
maxargc = *maxargcp;
for (s = options; (p = strsep(&s, ",")) != NULL;) {
/* Always leave space for one more argument and the NULL. */
if (argc >= maxargc - 4) {
nargv = erealloc(argv, (maxargc << 1) * sizeof(nargv));
argv = nargv;
maxargc <<= 1;
}
if (*p != '\0') {
if (*p == '-') {
argv[argc++] = p;
p = strchr(p, '=');
if (p) {
*p = '\0';
argv[argc++] = p+1;
}
} else if (strcmp(p, "rw") != 0) {
argv[argc++] = "-o";
argv[argc++] = p;
}
}
}
*argcp = argc;
*argvp = argv;
*maxargcp = maxargc;
}
/* Deduce the filesystem type from the disk label. */
static const char *
getfslab(const char *str)
{
static struct dkwedge_info dkw;
struct disklabel dl;
int fd;
int part;
const char *vfstype;
u_char fstype;
char buf[MAXPATHLEN + 1];
char *sp, *ep;
if ((fd = open(str, O_RDONLY)) == -1) {
/*
* Iff we get EBUSY try the raw device. Since mount always uses
* the block device we know we are never passed a raw device.
*/
if (errno != EBUSY)
err(EXIT_FAILURE, "cannot open `%s'", str);
strlcpy(buf, str, MAXPATHLEN);
if ((sp = strrchr(buf, '/')) != NULL)
++sp;
else
sp = buf;
for (ep = sp + strlen(sp) + 1; ep > sp; ep--)
*ep = *(ep - 1);
*sp = 'r';
/* Silently fail here - mount call can display error */
if ((fd = open(buf, O_RDONLY)) == -1)
return (NULL);
}
/* Check to see if this is a wedge. */
if (ioctl(fd, DIOCGWEDGEINFO, &dkw) == 0) {
/* Yup, this is easy. */
(void) close(fd);
return (dkw.dkw_ptype);
}
if (ioctl(fd, DIOCGDINFO, &dl) == -1) {
(void) close(fd);
return (NULL);
}
(void) close(fd);
part = str[strlen(str) - 1] - 'a';
if (part < 0 || part >= dl.d_npartitions)
return (NULL);
/* Return NULL for unknown types - caller can fall back to ffs */
if ((fstype = dl.d_partitions[part].p_fstype) >= FSMAXMOUNTNAMES)
vfstype = NULL;
else
vfstype = mountnames[fstype];
return (vfstype);
}
static void
usage(void)
{
(void)fprintf(stderr, "Usage: %s [-Aadfruvw] [-t type]\n"
"\t%s [-dfruvw] special | node\n"
"\t%s [-dfruvw] [-o options] [-t type] special node\n",
getprogname(), getprogname(), getprogname());
exit(1);
}