struct mdinfo *content = NULL;
int report_mismatch = ((inargv && c->verbose >= 0) || c->verbose > 0);
struct domainlist *domains = NULL;
+ dev_t rdev;
tmpdev = devlist; num_devs = 0;
while (tmpdev) {
tmpdev = tmpdev ? tmpdev->next : NULL) {
char *devname = tmpdev->devname;
int dfd;
- struct stat stb;
struct supertype *tst;
struct dev_policy *pol = NULL;
int found_container = 0;
pr_err("cannot open device %s: %s\n",
devname, strerror(errno));
tmpdev->used = 2;
- } else if (fstat(dfd, &stb)< 0) {
- /* Impossible! */
- pr_err("fstat failed for %s: %s\n",
- devname, strerror(errno));
- tmpdev->used = 2;
- } else if ((stb.st_mode & S_IFMT) != S_IFBLK) {
- pr_err("%s is not a block device.\n",
- devname);
+ } else if (!fstat_is_blkdev(dfd, devname, &rdev)) {
tmpdev->used = 2;
} else if (must_be_container(dfd)) {
if (st) {
pr_err("not a recognisable container: %s\n",
devname);
tmpdev->used = 2;
- } else if (!tst->ss->load_container
- || tst->ss->load_container(tst, dfd, NULL)) {
+ } else if (!tst->ss->load_container ||
+ tst->ss->load_container(tst, dfd, NULL)) {
if (report_mismatch)
pr_err("no correct container type: %s\n",
devname);
tmpdev->used = 2;
} else if (auto_assem &&
- !conf_test_metadata(tst->ss->name, (pol = devid_policy(stb.st_rdev)),
+ !conf_test_metadata(tst->ss->name,
+ (pol = devid_policy(rdev)),
tst->ss->match_home(tst, c->homehost) == 1)) {
if (report_mismatch)
pr_err("%s has metadata type %s for which auto-assembly is disabled\n",
tst->ss->name, devname);
tmpdev->used = 2;
} else if (auto_assem && st == NULL &&
- !conf_test_metadata(tst->ss->name, (pol = devid_policy(stb.st_rdev)),
+ !conf_test_metadata(tst->ss->name,
+ (pol = devid_policy(rdev)),
tst->ss->match_home(tst, c->homehost) == 1)) {
if (report_mismatch)
pr_err("%s has metadata type %s for which auto-assembly is disabled\n",
/* Collect domain information from members only */
if (tmpdev && tmpdev->used == 1) {
if (!pol)
- pol = devid_policy(stb.st_rdev);
+ pol = devid_policy(rdev);
domain_merge(&domains, pol, tst?tst->ss->name:NULL);
}
dev_policy_free(pol);
/* Now reject spares that don't match domains of identified members */
for (tmpdev = devlist; tmpdev; tmpdev = tmpdev->next) {
- struct stat stb;
if (tmpdev->used != 3)
continue;
- if (stat(tmpdev->devname, &stb)< 0) {
- pr_err("fstat failed for %s: %s\n",
- tmpdev->devname, strerror(errno));
+ if (!stat_is_blkdev(tmpdev->devname, &rdev)) {
tmpdev->used = 2;
} else {
- struct dev_policy *pol = devid_policy(stb.st_rdev);
+ struct dev_policy *pol = devid_policy(rdev);
int dt = domain_test(domains, pol, NULL);
if (inargv && dt != 0)
/* take this spare as domains match
bestcnt = newbestcnt;
}
if (best[i] >=0 &&
- devices[best[i]].i.events
- == devices[devcnt].i.events
- && (devices[best[i]].i.disk.minor
- != devices[devcnt].i.disk.minor)
- && st->ss == &super0
- && content->array.level != LEVEL_MULTIPATH) {
+ devices[best[i]].i.events ==
+ devices[devcnt].i.events &&
+ (devices[best[i]].i.disk.minor !=
+ devices[devcnt].i.disk.minor) &&
+ st->ss == &super0 &&
+ content->array.level != LEVEL_MULTIPATH) {
/* two different devices with identical superblock.
* Could be a mis-detection caused by overlapping
* partitions. fail-safe.
*stp = st;
return -1;
}
- if (best[i] == -1
- || (devices[best[i]].i.events
- < devices[devcnt].i.events))
+ if (best[i] == -1 || (devices[best[i]].i.events
+ < devices[devcnt].i.events))
best[i] = devcnt;
}
devcnt++;
int okcnt = 0;
while (!enough(content->array.level, content->array.raid_disks,
content->array.layout, 1,
- avail)
- ||
+ avail) ||
(content->reshape_active && content->delta_disks > 0 &&
!enough(content->array.level, (content->array.raid_disks
- content->delta_disks),
- content->new_layout, 1,
- avail)
- )) {
+ content->new_layout, 1, avail))) {
/* Choose the newest best drive which is
* not up-to-date, update the superblock
* and add it.
/* OK */;
else
continue;
- }
+ } else if (devices[j].i.reshape_active !=
+ content->reshape_active ||
+ (devices[j].i.reshape_active &&
+ devices[j].i.reshape_progress !=
+ content->reshape_progress))
+ /* Here, it may be a source of data. If two
+ * devices claim different progresses, it
+ * means that reshape boundaries differ for
+ * their own devices. Kernel will only treat
+ * the first one as reshape progress and
+ * go on. It may cause disaster, so avoid it.
+ */
+ continue;
if (chosen_drive < 0 ||
devices[j].i.events
> devices[chosen_drive].i.events)
if (j >= 0 &&
!devices[j].uptodate &&
devices[j].i.recovery_start == MaxSector &&
- devices[j].i.events == current_events) {
+ devices[j].i.events == current_events &&
+ ((!devices[j].i.reshape_active &&
+ !content->reshape_active) ||
+ (devices[j].i.reshape_active ==
+ content->reshape_active &&
+ devices[j].i.reshape_progress ==
+ content->reshape_progress))) {
chosen_drive = j;
goto add_another;
}
}
/* First, add the raid disks, but add the chosen one last */
- for (i=0; i<= bestcnt; i++) {
+ for (i = 0; i <= bestcnt; i++) {
int j;
if (i < bestcnt) {
j = best[i];
j = chosen_drive;
if (j >= 0 && !devices[j].included) {
- int dfd = dev_open(devices[j].devname,
- O_RDWR|O_EXCL);
+ int dfd;
+
+ dfd = dev_open(devices[j].devname, O_RDWR|O_EXCL);
if (dfd >= 0) {
remove_partitions(dfd);
close(dfd);
if (rv) {
pr_err("failed to add %s to %s: %s\n",
- devices[j].devname,
- mddev,
+ devices[j].devname, mddev,
strerror(errno));
- if (i < content->array.raid_disks * 2
- || i == bestcnt)
+ if (i < content->array.raid_disks * 2 ||
+ i == bestcnt)
okcnt--;
else
sparecnt--;
- } else if (c->verbose > 0)
+ } else if (c->verbose > 0) {
pr_err("added %s to %s as %d%s%s\n",
devices[j].devname, mddev,
devices[j].i.disk.raid_disk,
devices[j].uptodate?"":
" (possibly out of date)",
- (devices[j].i.disk.state & (1<<MD_DISK_REPLACEMENT))?" replacement":"");
+ (devices[j].i.disk.state &
+ (1<<MD_DISK_REPLACEMENT)) ?
+ " replacement":"");
+ }
} else if (j >= 0) {
if (c->verbose > 0)
pr_err("%s is already in %s as %d\n",
devices[j].devname, mddev,
devices[j].i.disk.raid_disk);
- } else if (c->verbose > 0 && i < content->array.raid_disks*2
- && (i&1) == 0)
+ } else if (c->verbose > 0 &&
+ i < content->array.raid_disks * 2 && (i & 1) == 0)
pr_err("no uptodate device for slot %d of %s\n",
i/2, mddev);
}
if (content->array.level == LEVEL_CONTAINER) {
if (c->verbose >= 0) {
pr_err("Container %s has been assembled with %d drive%s",
- mddev, okcnt+sparecnt+journalcnt,
- okcnt+sparecnt+journalcnt==1?"":"s");
+ mddev, okcnt + sparecnt + journalcnt,
+ okcnt + sparecnt + journalcnt == 1 ? "" : "s");
if (okcnt < (unsigned)content->array.raid_disks)
fprintf(stderr, " (out of %d)",
content->array.raid_disks);
if (st->ss->validate_container) {
struct mdinfo *devices_list;
- struct mdinfo *info_devices = xmalloc(sizeof(struct mdinfo)*(okcnt+sparecnt));
+ struct mdinfo *info_devices;
unsigned int count;
+
devices_list = NULL;
- for (count = 0; count < okcnt+sparecnt; count++) {
+ info_devices = xmalloc(sizeof(struct mdinfo) *
+ (okcnt + sparecnt));
+ for (count = 0; count < okcnt + sparecnt; count++) {
info_devices[count] = devices[count].i;
info_devices[count].next = devices_list;
devices_list = &info_devices[count];
if (c->runstop == 1 ||
(c->runstop <= 0 &&
- ( enough(content->array.level, content->array.raid_disks,
- content->array.layout, clean, avail) &&
- (okcnt + rebuilding_cnt >= req_cnt || start_partial_ok)
- ))) {
+ (enough(content->array.level, content->array.raid_disks,
+ content->array.layout, clean, avail) &&
+ (okcnt + rebuilding_cnt >= req_cnt || start_partial_ok)))) {
/* This array is good-to-go.
* If a reshape is in progress then we might need to
* continue monitoring it. In that case we start
* it read-only and let the grow code make it writable.
*/
int rv;
+
if (content->reshape_active &&
!(content->reshape_active & RESHAPE_NO_BACKUP) &&
content->delta_disks <= 0) {
c->backup_file, 0,
c->freeze_reshape);
} else if (c->readonly &&
- sysfs_attribute_available(
- content, NULL, "array_state")) {
+ sysfs_attribute_available(content, NULL,
+ "array_state")) {
rv = sysfs_set_str(content, NULL,
"array_state", "readonly");
} else
pr_err("%s has been started with %d drive%s",
mddev, okcnt, okcnt==1?"":"s");
if (okcnt < (unsigned)content->array.raid_disks)
- fprintf(stderr, " (out of %d)", content->array.raid_disks);
+ fprintf(stderr, " (out of %d)",
+ content->array.raid_disks);
if (rebuilding_cnt)
- fprintf(stderr, "%s %d rebuilding", sparecnt?",":" and", rebuilding_cnt);
+ fprintf(stderr, "%s %d rebuilding",
+ sparecnt?",":" and",
+ rebuilding_cnt);
if (sparecnt)
- fprintf(stderr, " and %d spare%s", sparecnt, sparecnt==1?"":"s");
+ fprintf(stderr, " and %d spare%s",
+ sparecnt,
+ sparecnt == 1 ? "" : "s");
if (content->journal_clean)
- fprintf(stderr, " and %d journal", journalcnt);
+ fprintf(stderr, " and %d journal",
+ journalcnt);
fprintf(stderr, ".\n");
}
if (content->reshape_active &&
* of the stripe cache - default is 256
*/
int chunk_size = content->array.chunk_size;
+
if (content->reshape_active &&
content->new_chunk > chunk_size)
chunk_size = content->new_chunk;
if (256 < 4 * ((chunk_size+4065)/4096)) {
- struct mdinfo *sra = sysfs_read(mdfd, NULL, 0);
+ struct mdinfo *sra;
+
+ sra = sysfs_read(mdfd, NULL, 0);
if (sra)
sysfs_set_num(sra, NULL,
"stripe_cache_size",
if (content->array.level == 6 &&
okcnt + 1 == (unsigned)content->array.raid_disks &&
was_forced) {
- struct mdinfo *sra = sysfs_read(mdfd, NULL, 0);
+ struct mdinfo *sra;
+
+ sra = sysfs_read(mdfd, NULL, 0);
if (sra)
sysfs_set_str(sra, NULL,
"sync_action", "repair");
}
return 0;
}
- pr_err("failed to RUN_ARRAY %s: %s\n",
- mddev, strerror(errno));
+ pr_err("failed to RUN_ARRAY %s: %s\n", mddev, strerror(errno));
if (!enough(content->array.level, content->array.raid_disks,
content->array.layout, 1, avail))
pr_err("Not enough devices to start the array.\n");
else if (!enough(content->array.level,
content->array.raid_disks,
- content->array.layout, clean,
- avail))
+ content->array.layout, clean, avail))
pr_err("Not enough devices to start the array while not clean - consider --force.\n");
return 1;
}
if (c->runstop == -1) {
pr_err("%s assembled from %d drive%s",
- mddev, okcnt, okcnt==1?"":"s");
+ mddev, okcnt, okcnt == 1 ? "" : "s");
if (okcnt != (unsigned)content->array.raid_disks)
- fprintf(stderr, " (out of %d)", content->array.raid_disks);
+ fprintf(stderr, " (out of %d)",
+ content->array.raid_disks);
fprintf(stderr, ", but not started.\n");
return 2;
}
if (c->verbose >= -1) {
- pr_err("%s assembled from %d drive%s", mddev, okcnt, okcnt==1?"":"s");
+ pr_err("%s assembled from %d drive%s",
+ mddev, okcnt, okcnt == 1 ? "" : "s");
if (rebuilding_cnt)
- fprintf(stderr, "%s %d rebuilding", sparecnt?",":" and", rebuilding_cnt);
+ fprintf(stderr, "%s %d rebuilding",
+ sparecnt ? "," : " and", rebuilding_cnt);
if (sparecnt)
- fprintf(stderr, " and %d spare%s", sparecnt, sparecnt==1?"":"s");
+ fprintf(stderr, " and %d spare%s", sparecnt,
+ sparecnt == 1 ? "" : "s");
if (!enough(content->array.level, content->array.raid_disks,
content->array.layout, 1, avail))
fprintf(stderr, " - not enough to start the array.\n");
else if (!enough(content->array.level,
content->array.raid_disks,
- content->array.layout, clean,
- avail))
+ content->array.layout, clean, avail))
fprintf(stderr, " - not enough to start the array while not clean - consider --force.\n");
else {
if (req_cnt == (unsigned)content->array.raid_disks)
- fprintf(stderr, " - need all %d to start it", req_cnt);
+ fprintf(stderr, " - need all %d to start it",
+ req_cnt);
else
fprintf(stderr, " - need %d to start", req_cnt);
fprintf(stderr, " (use --run to insist).\n");
* START_ARRAY
*
*/
- int rv;
- int mdfd;
+ int rv = -1;
+ int mdfd = -1;
int clean;
int auto_assem = (mddev == NULL && !ident->uuid_set &&
- ident->super_minor == UnSet && ident->name[0] == 0
- && (ident->container == NULL || ident->member == NULL));
- struct devs *devices;
+ ident->super_minor == UnSet && ident->name[0] == 0 &&
+ (ident->container == NULL || ident->member == NULL));
+ struct devs *devices = NULL;
char *devmap;
int *best = NULL; /* indexed by raid_disk */
int bestcnt = 0;
char chosen_name[1024];
struct map_ent *map = NULL;
struct map_ent *mp;
+ int locked = 0;
+ struct mdp_superblock_1 *sb;
+ bitmap_super_t *bms;
/*
* If any subdevs are listed, then any that don't
* set of devices failed. Those are now marked as ->used==2 and
* we ignore them and try again
*/
+ if (locked)
+ /*
+ * if come back try_again is called, then need to unlock first,
+ * and lock again since the metadate is re-read.
+ */
+ cluster_release_dlmlock();
if (!st && ident->st)
st = ident->st;
if (c->verbose>0)
if (!st || !st->sb || !content)
return 2;
+ sb = st->sb;
+ bms = (bitmap_super_t*)(((char*)sb) + 4096);
+ if (sb && bms->version == BITMAP_MAJOR_CLUSTERED) {
+ locked = cluster_get_dlmlock();
+ if (locked != 1)
+ return 1;
+ }
+
/* We have a full set of devices - we now need to find the
* array device.
* However there is a risk that we are racing with "mdadm -I"
pr_err("Found some drive for an array that is already active: %s\n",
mp->path);
pr_err("giving up.\n");
- return 1;
+ goto out;
}
for (dv = pre_exist->devs; dv; dv = dv->next) {
/* We want to add this device to our list,
name = strchr(name, ':')+1;
mdfd = create_mddev(mddev, name, ident->autof, trustworthy,
- chosen_name);
+ chosen_name, 0);
}
if (mdfd < 0) {
st->ss->free_super(st);
if (auto_assem)
goto try_again;
- return 1;
+ goto out;
}
mddev = chosen_name;
- if (get_linux_version() < 2004000) {
- pr_err("Assemble requires Linux 2.4 or later, and\n"
- " md driver version 0.90.0 or later.\n"
- " Upgrade your kernel or try --build\n");
- close(mdfd);
- return 1;
- }
if (pre_exist == NULL) {
if (mddev_busy(fd2devnm(mdfd))) {
pr_err("%s already active, cannot restart it!\n",
st->ss->free_super(st);
if (auto_assem)
goto try_again;
- return 1;
+ goto out;
}
/* just incase it was started but has no content */
ioctl(mdfd, STOP_ARRAY, NULL);
err = assemble_container_content(st, mdfd, content, c,
chosen_name, NULL);
close(mdfd);
+ if (locked == 1)
+ cluster_release_dlmlock();
return err;
}
devcnt = load_devices(devices, devmap, ident, &st, devlist,
c, content, mdfd, mddev,
&most_recent, &bestcnt, &best, inargv);
- if (devcnt < 0)
- return 1;
+ if (devcnt < 0) {
+ mdfd = -3;
+ goto out;
+ }
if (devcnt == 0) {
pr_err("no devices found for %s\n",
mddev);
if (st)
st->ss->free_super(st);
- close(mdfd);
- free(devices);
free(devmap);
- return 1;
+ goto out;
}
if (c->update && strcmp(c->update, "byteorder")==0)
: (O_RDONLY|O_EXCL)))< 0) {
pr_err("Cannot open %s: %s\n",
devices[j].devname, strerror(errno));
- close(mdfd);
- free(devices);
- return 1;
+ goto out;
}
if (st->ss->load_super(st,fd, NULL)) {
close(fd);
pr_err("RAID superblock has disappeared from %s\n",
devices[j].devname);
- close(mdfd);
- free(devices);
- return 1;
+ goto out;
}
close(fd);
}
if (st->sb == NULL) {
pr_err("No suitable drives found for %s\n", mddev);
- close(mdfd);
- free(devices);
- return 1;
+ goto out;
}
st->ss->getinfo_super(st, content, NULL);
if (sysfs_init(content, mdfd, NULL)) {
pr_err("Unable to initialize sysfs\n");
- close(mdfd);
- free(devices);
- return 1;
+ goto out;
}
/* after reload context, store journal_clean in context */
int j = best[i];
unsigned int desired_state;
+ if (j < 0)
+ continue;
if (devices[j].i.disk.raid_disk == MD_DISK_ROLE_JOURNAL)
desired_state = (1<<MD_DISK_JOURNAL);
else if (i >= content->array.raid_disks * 2)
else
desired_state = (1<<MD_DISK_ACTIVE) | (1<<MD_DISK_SYNC);
- if (j<0)
- continue;
if (!devices[j].uptodate)
continue;
if (fd < 0) {
pr_err("Could not open %s for write - cannot Assemble array.\n",
devices[chosen_drive].devname);
- close(mdfd);
- free(devices);
- return 1;
+ goto out;
}
if (st->ss->store_super(st, fd)) {
close(fd);
pr_err("Could not re-write superblock on %s\n",
devices[chosen_drive].devname);
- close(mdfd);
- free(devices);
- return 1;
+ goto out;
}
if (c->verbose >= 0)
pr_err("Marking array %s as 'clean'\n",
pr_err("Failed to restore critical section for reshape, sorry.\n");
if (c->backup_file == NULL)
cont_err("Possibly you needed to specify the --backup-file\n");
- close(mdfd);
- free(devices);
- return err;
+ goto out;
}
}
ioctl(mdfd, STOP_ARRAY, NULL);
free(devices);
map_unlock(&map);
+out:
if (rv == 0) {
wait_for(chosen_name, mdfd);
close(mdfd);
usecs <<= 1;
}
}
- } else
+ } else if (mdfd >= 0)
close(mdfd);
/* '2' means 'OK, but not started yet' */
+ if (locked == 1)
+ cluster_release_dlmlock();
+ if (rv == -1) {
+ free(devices);
+ return 1;
+ }
return rv == 2 ? 0 : rv;
}