#ifndef MDASSEMBLE
struct mdinfo *mdi;
#endif
+ int rv = 0;
if (md_get_version(fd) < 9000) {
fprintf(stderr, Name ": need md driver version 0.90.0 or later\n");
vers[9] = mdi->text_version[0];
sysfs_set_str(mdi, NULL, "metadata_version", vers);
- return 1;
+ rv = 1;
+ goto out;
}
} else {
char *cp;
sysfs_set_str(mdi, NULL, "metadata_version", vers);
cp = strchr(vers+10, '/');
- if (*cp)
+ if (cp)
*cp = 0;
ping_monitor(vers+10);
if (mdi->array.level <= 0)
sysfs_set_str(mdi, NULL, "array_state", "active");
}
- return 0;
+ goto out;
}
#endif
if (ioctl(fd, GET_ARRAY_INFO, &array)) {
fprintf(stderr, Name ": %s does not appear to be active.\n",
devname);
- return 1;
+ rv = 1;
+ goto out;
}
if (readonly>0) {
if (ioctl(fd, STOP_ARRAY_RO, NULL)) {
fprintf(stderr, Name ": failed to set readonly for %s: %s\n",
devname, strerror(errno));
- return 1;
+ rv = 1;
+ goto out;
}
} else if (readonly < 0) {
if (ioctl(fd, RESTART_ARRAY_RW, NULL)) {
fprintf(stderr, Name ": failed to set writable for %s: %s\n",
devname, strerror(errno));
- return 1;
+ rv = 1;
+ goto out;
}
}
- return 0;
+out:
+#ifndef MDASSEMBLE
+ if (mdi)
+ sysfs_free(mdi);
+#endif
+ return rv;
}
#ifndef MDASSEMBLE
sprintf(pe, "%d", part);
}
n = readlink(path2, link, sizeof(link));
- if (n && (int)strlen(base) == n &&
+ if (n > 0 && (int)strlen(base) == n &&
strncmp(link, base, n) == 0)
unlink(path2);
}
* quiet < 0 means we will try again if it fails.
*/
mdu_param_t param; /* unused */
+ int rv = 0;
if (runstop == -1 && md_get_version(fd) < 9000) {
if (ioctl(fd, STOP_MD, 0)) {
struct stat stb;
struct mdinfo *mdi;
int devnum;
+ int err;
+ int count;
/* If this is an mdmon managed array, just write 'inactive'
* to the array state and let mdmon clear up.
*/
devnum = fd2devnum(fd);
+ /* Get EXCL access first. If this fails, then attempting
+ * to stop is probably a bad idea.
+ */
+ close(fd);
+ fd = open(devname, O_RDONLY|O_EXCL);
+ if (fd < 0 || fd2devnum(fd) != devnum) {
+ if (fd >= 0)
+ close(fd);
+ fprintf(stderr,
+ Name ": Cannot get exclusive access to %s:"
+ "Perhaps a running "
+ "process, mounted filesystem "
+ "or active volume group?\n",
+ devname);
+ return 1;
+ }
mdi = sysfs_read(fd, -1, GET_LEVEL|GET_VERSION);
if (mdi &&
mdi->array.level > 0 &&
is_subarray(mdi->text_version)) {
+ int err;
/* This is mdmon managed. */
close(fd);
- if (sysfs_set_str(mdi, NULL,
- "array_state", "inactive") < 0) {
- if (quiet == 0)
- fprintf(stderr, Name
- ": failed to stop array %s: %s\n",
- devname, strerror(errno));
- return 1;
+
+ count = 25;
+ while (count &&
+ (err = sysfs_set_str(mdi, NULL,
+ "array_state",
+ "inactive")) < 0
+ && errno == EBUSY) {
+ usleep(200000);
+ count--;
+ }
+ if (err && !quiet) {
+ fprintf(stderr, Name
+ ": failed to stop array %s: %s\n",
+ devname, strerror(errno));
+ rv = 1;
+ goto out;
}
/* Give monitor a chance to act */
ping_monitor(mdi->text_version);
- fd = open(devname, O_RDONLY);
+ fd = open_dev_excl(devnum);
+ if (fd < 0) {
+ fprintf(stderr, Name
+ ": failed to completely stop %s"
+ ": Device is busy\n",
+ devname);
+ rv = 1;
+ goto out;
+ }
} else if (mdi &&
mdi->array.major_version == -1 &&
mdi->array.minor_version == -2 &&
"member %s still active\n",
devname, m->dev);
free_mdstat(mds);
- if (mdi)
- sysfs_free(mdi);
- return 1;
+ rv = 1;
+ goto out;
}
}
- if (fd >= 0 && ioctl(fd, STOP_ARRAY, NULL)) {
+ /* As we have an O_EXCL open, any use of the device
+ * which blocks STOP_ARRAY is probably a transient use,
+ * so it is reasonable to retry for a while - 5 seconds.
+ */
+ count = 25; err = 0;
+ while (count && fd >= 0
+ && (err = ioctl(fd, STOP_ARRAY, NULL)) < 0
+ && errno == EBUSY) {
+ usleep(200000);
+ count --;
+ }
+ if (fd >= 0 && err) {
if (quiet == 0) {
fprintf(stderr, Name
": failed to stop array %s: %s\n",
"process, mounted filesystem "
"or active volume group?\n");
}
- if (mdi)
- sysfs_free(mdi);
- return 1;
+ rv = 1;
+ goto out;
}
/* prior to 2.6.28, KOBJ_CHANGE was not sent when an md array
* was stopped, so We'll do it here just to be sure. Drop any
map_lock(&map);
map_remove(&map, devnum);
map_unlock(&map);
+ out:
+ if (mdi)
+ sysfs_free(mdi);
}
- return 0;
+ return rv;
}
int Manage_resize(char *devname, int fd, long long size, int raid_disks)
int Manage_subdevs(char *devname, int fd,
struct mddev_dev *devlist, int verbose, int test,
- char *update)
+ char *update, int force)
{
/* do something to each dev.
* devmode can be
int lfd = -1;
int sysfd = -1;
int count = 0; /* number of actions taken */
+ struct mdinfo info;
+ int frozen = 0;
if (ioctl(fd, GET_ARRAY_INFO, &array)) {
fprintf(stderr, Name ": cannot get array info for %s\n",
devname);
- return 1;
+ goto abort;
}
+ sysfs_init(&info, fd, 0);
/* array.size is only 32 bit and may be truncated.
* So read from sysfs if possible, and record number of sectors
if (!tst) {
fprintf(stderr, Name ": unsupport array - version %d.%d\n",
array.major_version, array.minor_version);
- return 1;
+ goto abort;
}
stb.st_rdev = 0;
char *dnprintable = dv->devname;
char *add_dev = dv->devname;
int err;
- int re_add_failed = 0;
+ int array_failed;
next = dv->next;
jnext = 0;
if (strcmp(dv->devname, "failed")==0 ||
strcmp(dv->devname, "faulty")==0) {
+ int remaining_disks = array.nr_disks;
if (dv->disposition != 'r') {
fprintf(stderr, Name ": %s only meaningful "
"with -r, not -%c\n",
dv->devname, dv->disposition);
- return 1;
+ goto abort;
}
- for (; j < array.raid_disks + array.nr_disks ; j++) {
+ for (; j < MAX_DISKS && remaining_disks > 0; j++) {
unsigned dev;
disc.number = j;
if (ioctl(fd, GET_DISK_INFO, &disc))
continue;
if (disc.major == 0 && disc.minor == 0)
continue;
+ remaining_disks --;
if ((disc.state & 1) == 0) /* faulty */
continue;
dev = makedev(disc.major, disc.minor);
dnprintable = dvname;
break;
}
- if (jnext == 0)
+ if (next != dv)
continue;
} else if (strcmp(dv->devname, "detached") == 0) {
+ int remaining_disks = array.nr_disks;
if (dv->disposition != 'r' && dv->disposition != 'f') {
fprintf(stderr, Name ": %s only meaningful "
"with -r of -f, not -%c\n",
dv->devname, dv->disposition);
- return 1;
+ goto abort;
}
- for (; j < array.raid_disks + array.nr_disks; j++) {
+ for (; j < MAX_DISKS && remaining_disks > 0; j++) {
int sfd;
unsigned dev;
disc.number = j;
continue;
if (disc.major == 0 && disc.minor == 0)
continue;
+ remaining_disks --;
sprintf(dvname,"%d:%d", disc.major, disc.minor);
sfd = dev_open(dvname, O_RDONLY);
if (sfd >= 0) {
dnprintable = dvname;
break;
}
- if (jnext == 0)
+ if (next != dv)
continue;
} else if (strcmp(dv->devname, "missing") == 0) {
if (dv->disposition != 'a' || dv->re_add == 0) {
fprintf(stderr, Name ": 'missing' only meaningful "
"with --re-add\n");
- return 1;
+ goto abort;
}
if (add_devlist == NULL)
add_devlist = conf_get_devs();
fprintf(stderr, Name ": %s only meaningful "
"with -r or -f, not -%c\n",
dv->devname, dv->disposition);
- return 1;
+ goto abort;
}
sprintf(dname, "dev-%s", dv->devname);
fprintf(stderr, Name ": %s does not appear "
"to be a component of %s\n",
dv->devname, devname);
- return 1;
+ goto abort;
}
}
} else {
dv->devname, strerror(errno));
if (tfd >= 0)
close(tfd);
- return 1;
+ goto abort;
}
close(tfd);
tfd = -1;
fprintf(stderr, Name ": %s is not a "
"block device.\n",
dv->devname);
- return 1;
+ goto abort;
}
}
switch(dv->disposition){
default:
fprintf(stderr, Name ": internal error - devmode[%s]=%d\n",
dv->devname, dv->disposition);
- return 1;
+ goto abort;
case 'a':
/* add the device */
if (subarray) {
fprintf(stderr, Name ": Cannot add disks to a"
" \'member\' array, perform this"
" operation on the parent container\n");
- return 1;
+ goto abort;
}
/* Make sure it isn't in use (in 2.6 or later) */
tfd = dev_open(add_dev, O_RDONLY|O_EXCL|O_DIRECT);
if (tfd < 0) {
fprintf(stderr, Name ": Cannot open %s: %s\n",
dv->devname, strerror(errno));
- return 1;
+ goto abort;
+ }
+ if (!frozen) {
+ if (sysfs_freeze_array(&info) == 1)
+ frozen = 1;
+ else
+ frozen = -1;
}
st = dup_super(tst);
if (add_dev == dv->devname) {
if (!get_dev_size(tfd, dv->devname, &ldsize)) {
+ st->ss->free_super(st);
close(tfd);
- return 1;
+ goto abort;
}
} else if (!get_dev_size(tfd, NULL, &ldsize)) {
+ st->ss->free_super(st);
close(tfd);
tfd = -1;
continue;
}
+ if (tst->ss->validate_geometry(
+ tst, array.level, array.layout,
+ array.raid_disks, NULL,
+ ldsize >> 9, NULL, NULL, 0) == 0) {
+ if (!force) {
+ fprintf(stderr, Name
+ ": %s is larger than %s can "
+ "effectively use.\n"
+ " Add --force is you "
+ "really want to add this device.\n",
+ add_dev, devname);
+ st->ss->free_super(st);
+ close(tfd);
+ goto abort;
+ }
+ fprintf(stderr, Name
+ ": %s is larger than %s can "
+ "effectively use.\n"
+ " Adding anyway as --force "
+ "was given.\n",
+ add_dev, devname);
+ }
if (!tst->ss->external &&
array.major_version == 0 &&
md_get_version(fd)%100 < 2) {
close(tfd);
+ st->ss->free_super(st);
tfd = -1;
if (ioctl(fd, HOT_ADD_DISK,
(unsigned long)stb.st_rdev)==0) {
fprintf(stderr, Name ": hot add failed for %s: %s\n",
add_dev, strerror(errno));
- return 1;
+ goto abort;
}
if (array.not_persistent == 0 || tst->ss->external) {
break;
}
/* FIXME this is a bad test to be using */
- if (!tst->sb) {
+ if (!tst->sb &&
+ dv->re_add) {
+ /* we are re-adding a device to a
+ * completely dead array - have to depend
+ * on kernel to check
+ */
+ } else if (!tst->sb) {
close(tfd);
- fprintf(stderr, Name ": cannot find valid superblock in this array - HELP\n");
- return 1;
+ st->ss->free_super(st);
+ fprintf(stderr, Name ": cannot load array metadata from %s\n", devname);
+ goto abort;
}
/* Make sure device is large enough */
array_size) {
close(tfd);
tfd = -1;
+ st->ss->free_super(st);
if (add_dev != dv->devname)
continue;
fprintf(stderr, Name ": %s not large enough to join array\n",
dv->devname);
- return 1;
+ goto abort;
}
/* Possibly this device was recently part of the array
* and was temporarily removed, and is now being re-added.
* If so, we can simply re-add it.
*/
- tst->ss->uuid_from_super(tst, duuid);
- /* re-add doesn't work for version-1 superblocks
- * before 2.6.18 :-(
- */
- if (array.major_version == 1 &&
- get_linux_version() <= 2006018)
- ;
- else if (st->sb) {
+ if (st->sb) {
struct mdinfo mdi;
st->ss->getinfo_super(st, &mdi, NULL);
st->ss->uuid_from_super(st, ouuid);
+ if (tst->sb)
+ tst->ss->uuid_from_super(tst, duuid);
+ else
+ /* Assume uuid matches: kernel will check */
+ memcpy(duuid, ouuid, sizeof(ouuid));
if ((mdi.disk.state & (1<<MD_DISK_ACTIVE)) &&
!(mdi.disk.state & (1<<MD_DISK_FAULTY)) &&
memcmp(duuid, ouuid, sizeof(ouuid))==0) {
/* look like it is worth a try. Need to
* make sure kernel will accept it though.
*/
+ /* re-add doesn't work for version-1 superblocks
+ * before 2.6.18 :-(
+ */
+ if (array.major_version == 1 &&
+ get_linux_version() <= 2006018)
+ goto skip_re_add;
disc.number = mdi.disk.number;
if (ioctl(fd, GET_DISK_INFO, &disc) != 0
|| disc.major != 0 || disc.minor != 0
- || !enough_fd(fd))
+ )
goto skip_re_add;
disc.major = major(stb.st_rdev);
disc.minor = minor(stb.st_rdev);
remove_partitions(tfd);
close(tfd);
tfd = -1;
- if (update) {
+ if (update || dv->writemostly > 0) {
int rv = -1;
tfd = dev_open(dv->devname, O_RDWR);
+ if (tfd < 0) {
+ fprintf(stderr, Name ": failed to open %s for"
+ " superblock update during re-add\n", dv->devname);
+ st->ss->free_super(st);
+ goto abort;
+ }
- if (tfd >= 0)
+ if (dv->writemostly == 1)
+ rv = st->ss->update_super(
+ st, NULL, "writemostly",
+ devname, verbose, 0, NULL);
+ if (dv->writemostly == 2)
+ rv = st->ss->update_super(
+ st, NULL, "readwrite",
+ devname, verbose, 0, NULL);
+ if (update)
rv = st->ss->update_super(
st, NULL, update,
devname, verbose, 0, NULL);
if (rv == 0)
- rv = tst->ss->store_super(st, tfd);
+ rv = st->ss->store_super(st, tfd);
close(tfd);
tfd = -1;
if (rv != 0) {
fprintf(stderr, Name ": failed to update"
" superblock during re-add\n");
- return 1;
+ st->ss->free_super(st);
+ goto abort;
}
}
/* don't even try if disk is marked as faulty */
if (verbose >= 0)
fprintf(stderr, Name ": re-added %s\n", add_dev);
count++;
+ st->ss->free_super(st);
continue;
}
if (errno == ENOMEM || errno == EROFS) {
fprintf(stderr, Name ": add new device failed for %s: %s\n",
add_dev, strerror(errno));
+ st->ss->free_super(st);
if (add_dev != dv->devname)
continue;
- return 1;
+ goto abort;
}
- skip_re_add:
- re_add_failed = 1;
}
+ skip_re_add:
st->ss->free_super(st);
}
if (add_dev != dv->devname) {
fprintf(stderr, Name
": --re-add for %s to %s is not possible\n",
dv->devname, devname);
- return 1;
+ goto abort;
}
- if (re_add_failed) {
- fprintf(stderr, Name ": %s reports being an active member for %s, but a --re-add fails.\n",
- dv->devname, devname);
- fprintf(stderr, Name ": not performing --add as that would convert %s in to a spare.\n",
- dv->devname);
- fprintf(stderr, Name ": To make this a spare, use \"mdadm --zero-superblock %s\" first.\n",
+ if (array.active_disks < array.raid_disks) {
+ char *avail = calloc(array.raid_disks, 1);
+ int d;
+ int found = 0;
+
+ for (d = 0; d < MAX_DISKS && found < array.active_disks; d++) {
+ disc.number = d;
+ if (ioctl(fd, GET_DISK_INFO, &disc))
+ continue;
+ if (disc.major == 0 && disc.minor == 0)
+ continue;
+ if (!(disc.state & (1<<MD_DISK_SYNC)))
+ continue;
+ avail[disc.raid_disk] = 1;
+ found++;
+ }
+ array_failed = !enough(array.level, array.raid_disks,
+ array.layout, 1, avail);
+ } else
+ array_failed = 0;
+ if (array_failed) {
+ fprintf(stderr, Name ": %s has failed so using --add cannot work and might destroy\n",
+ devname);
+ fprintf(stderr, Name ": data on %s. You should stop the array and re-assemble it.\n",
dv->devname);
if (tfd >= 0)
close(tfd);
- return 1;
+ goto abort;
}
} else {
/* non-persistent. Must ensure that new drive
dv->devname);
if (tfd >= 0)
close(tfd);
- return 1;
+ goto abort;
}
}
/* committed to really trying this device now*/
disc.minor = minor(stb.st_rdev);
disc.number =j;
disc.state = 0;
- if (array.not_persistent==0 || tst->ss->external) {
+ if (array.not_persistent==0) {
int dfd;
if (dv->writemostly == 1)
disc.state |= 1 << MD_DISK_WRITEMOSTLY;
dfd = dev_open(dv->devname, O_RDWR | O_EXCL|O_DIRECT);
- if (tst->ss->external &&
- mdmon_running(tst->container_dev))
- tst->update_tail = &tst->updates;
if (tst->ss->add_to_super(tst, &disc, dfd,
dv->devname)) {
close(dfd);
- return 1;
+ goto abort;
}
if (tst->ss->write_init_super(tst)) {
close(dfd);
- return 1;
+ goto abort;
}
} else if (dv->re_add) {
/* this had better be raid1.
struct mdinfo *sra;
int container_fd;
int devnum = fd2devnum(fd);
+ int dfd;
container_fd = open_dev_excl(devnum);
if (container_fd < 0) {
fprintf(stderr, Name ": add failed for %s:"
" could not get exclusive access to container\n",
dv->devname);
- return 1;
+ tst->ss->free_super(tst);
+ goto abort;
+ }
+
+ dfd = dev_open(dv->devname, O_RDWR | O_EXCL|O_DIRECT);
+ if (mdmon_running(tst->container_dev))
+ tst->update_tail = &tst->updates;
+ if (tst->ss->add_to_super(tst, &disc, dfd,
+ dv->devname)) {
+ close(dfd);
+ close(container_fd);
+ goto abort;
}
+ if (tst->update_tail)
+ flush_metadata_updates(tst);
+ else
+ tst->ss->sync_metadata(tst);
sra = sysfs_read(container_fd, -1, 0);
if (!sra) {
fprintf(stderr, Name ": add failed for %s: sysfs_read failed\n",
dv->devname);
close(container_fd);
- return 1;
+ tst->ss->free_super(tst);
+ goto abort;
}
sra->array.level = LEVEL_CONTAINER;
/* Need to set data_offset and component_size */
" failed for %s\n", dv->devname);
close(container_fd);
sysfs_free(sra);
- return 1;
+ goto abort;
}
- ping_monitor(devnum2devname(devnum));
+ ping_monitor_by_id(devnum);
sysfs_free(sra);
close(container_fd);
- } else if (ioctl(fd, ADD_NEW_DISK, &disc)) {
- fprintf(stderr, Name ": add new device failed for %s as %d: %s\n",
- dv->devname, j, strerror(errno));
- return 1;
+ } else {
+ tst->ss->free_super(tst);
+ if (ioctl(fd, ADD_NEW_DISK, &disc)) {
+ fprintf(stderr, Name ": add new device failed for %s as %d: %s\n",
+ dv->devname, j, strerror(errno));
+ goto abort;
+ }
}
if (verbose >= 0)
fprintf(stderr, Name ": added %s\n", dv->devname);
- tst->ss->free_super(tst);
break;
case 'r':
" operation on the parent container\n");
if (sysfd >= 0)
close(sysfd);
- return 1;
+ goto abort;
}
if (tst->ss->external) {
/* To remove a device from a container, we must
" to container - odd\n");
if (sysfd >= 0)
close(sysfd);
- return 1;
+ goto abort;
}
/* in the detached case it is not possible to
* check if we are the unique holder, so just
errno == EEXIST ? "still in use":
"not a member");
close(lfd);
- return 1;
+ goto abort;
}
}
/* FIXME check that it is a current member */
strerror(errno));
if (lfd >= 0)
close(lfd);
- return 1;
+ goto abort;
}
if (tst->ss->external) {
/*
if (!name) {
fprintf(stderr, Name ": unable to get container name\n");
- return 1;
+ goto abort;
}
ping_manager(name);
dnprintable, strerror(errno));
if (sysfd >= 0)
close(sysfd);
- return 1;
+ goto abort;
}
if (sysfd >= 0)
close(sysfd);
break;
}
}
+ if (frozen > 0)
+ sysfs_set_str(&info, NULL, "sync_action","idle");
if (test && count == 0)
return 2;
return 0;
+
+abort:
+ if (frozen > 0)
+ sysfs_set_str(&info, NULL, "sync_action","idle");
+ return 1;
}
int autodetect(void)
sprintf(devname, "%d:%d", major(devid), minor(devid));
devlist.disposition = 'r';
- if (Manage_subdevs(from_devname, fd2, &devlist, -1, 0, NULL) == 0) {
+ if (Manage_subdevs(from_devname, fd2, &devlist, -1, 0, NULL, 0) == 0) {
devlist.disposition = 'a';
- if (Manage_subdevs(to_devname, fd1, &devlist, -1, 0, NULL) == 0) {
+ if (Manage_subdevs(to_devname, fd1, &devlist, -1, 0, NULL, 0) == 0) {
/* make sure manager is aware of changes */
ping_manager(to_devname);
ping_manager(from_devname);
close(fd2);
return 1;
}
- else Manage_subdevs(from_devname, fd2, &devlist, -1, 0, NULL);
+ else Manage_subdevs(from_devname, fd2, &devlist, -1, 0, NULL, 0);
}
close(fd1);
close(fd2);
projects / thirdparty / mdadm.git / blobdiff