map->disk_ord_tbl[slot] = __cpu_to_le32(ord);
}
+static int get_imsm_disk_slot(struct imsm_map *map, int idx)
+{
+ int slot;
+ __u32 ord;
+
+ for (slot = 0; slot < map->num_members; slot++) {
+ ord = __le32_to_cpu(map->disk_ord_tbl[slot]);
+ if (ord_to_idx(ord) == idx)
+ return slot;
+ }
+
+ return -1;
+}
+
static int get_imsm_raid_level(struct imsm_map *map)
{
if (map->raid_level == 1) {
static int count_memberships(struct dl *dl, struct intel_super *super)
{
int memberships = 0;
- int i, j;
+ int i;
for (i = 0; i < super->anchor->num_raid_devs; i++) {
struct imsm_dev *dev = get_imsm_dev(super, i);
struct imsm_map *map = get_imsm_map(dev, 0);
- for (j = 0; j < map->num_members; j++) {
- __u32 index = get_imsm_disk_idx(dev, j);
-
- if (index == dl->index)
- memberships++;
- }
+ if (get_imsm_disk_slot(map, dl->index) >= 0)
+ memberships++;
}
return memberships;
{
/* find a list of used extents on the given physical device */
struct extent *rv, *e;
- int i, j;
+ int i;
int memberships = count_memberships(dl, super);
__u32 reservation = MPB_SECTOR_CNT + IMSM_RESERVED_SECTORS;
struct imsm_dev *dev = get_imsm_dev(super, i);
struct imsm_map *map = get_imsm_map(dev, 0);
- for (j = 0; j < map->num_members; j++) {
- __u32 index = get_imsm_disk_idx(dev, j);
-
- if (index == dl->index) {
- e->start = __le32_to_cpu(map->pba_of_lba0);
- e->size = __le32_to_cpu(map->blocks_per_member);
- e++;
- }
+ if (get_imsm_disk_slot(map, dl->index) >= 0) {
+ e->start = __le32_to_cpu(map->pba_of_lba0);
+ e->size = __le32_to_cpu(map->blocks_per_member);
+ e++;
}
}
qsort(rv, memberships, sizeof(*rv), cmp_extent);
printf(" UUID : %s\n", uuid);
printf(" RAID Level : %d\n", get_imsm_raid_level(map));
printf(" Members : %d\n", map->num_members);
- for (slot = 0; slot < map->num_members; slot++)
- if (disk_idx== get_imsm_disk_idx(dev, slot))
- break;
- if (slot < map->num_members) {
+ slot = get_imsm_disk_slot(map, disk_idx);
+ if (slot >= 0) {
ord = get_imsm_ord_tbl_ent(dev, slot);
printf(" This Slot : %d%s\n", slot,
ord & IMSM_ORD_REBUILD ? " (out-of-sync)" : "");
uuid_from_super_imsm(st, info->uuid);
}
+/* check the config file to see if we can return a real uuid for this spare */
+static void fixup_container_spare_uuid(struct mdinfo *inf)
+{
+ struct mddev_ident_s *array_list;
+
+ if (inf->array.level != LEVEL_CONTAINER ||
+ memcmp(inf->uuid, uuid_match_any, sizeof(int[4])) != 0)
+ return;
+
+ array_list = conf_get_ident(NULL);
+
+ for (; array_list; array_list = array_list->next) {
+ if (array_list->uuid_set) {
+ struct supertype *_sst; /* spare supertype */
+ struct supertype *_cst; /* container supertype */
+
+ _cst = array_list->st;
+ _sst = _cst->ss->match_metadata_desc(inf->text_version);
+ if (_sst) {
+ memcpy(inf->uuid, array_list->uuid, sizeof(int[4]));
+ free(_sst);
+ break;
+ }
+ }
+ }
+}
static void getinfo_super_imsm(struct supertype *st, struct mdinfo *info)
{
info->component_size = reserved;
s = disk->status;
info->disk.state = s & CONFIGURED_DISK ? (1 << MD_DISK_ACTIVE) : 0;
+ /* we don't change info->disk.raid_disk here because
+ * this state will be finalized in mdmon after we have
+ * found the 'most fresh' version of the metadata
+ */
+ info->disk.state |= s & FAILED_DISK ? (1 << MD_DISK_FAULTY) : 0;
info->disk.state |= s & SPARE_DISK ? 0 : (1 << MD_DISK_SYNC);
- if (s & FAILED_DISK || super->disks->index == -2) {
- info->disk.state |= 1 << MD_DISK_FAULTY;
- info->disk.raid_disk = -2;
- }
}
/* only call uuid_from_super_imsm when this disk is part of a populated container,
*/
if (info->disk.state & (1 << MD_DISK_SYNC) || super->anchor->num_raid_devs)
uuid_from_super_imsm(st, info->uuid);
- else
+ else {
memcpy(info->uuid, uuid_match_any, sizeof(int[4]));
+ fixup_container_spare_uuid(info);
+ }
}
static int update_super_imsm(struct supertype *st, struct mdinfo *info,
* 4/ Rebuild (migr_state=1 migr_type=MIGR_REBUILD map0state=normal
* map1state=degraded)
*/
-static void migrate(struct imsm_dev *dev, __u8 to_state, int rebuild_resync)
+static void migrate(struct imsm_dev *dev, __u8 to_state, int migr_type)
{
struct imsm_map *dest;
struct imsm_map *src = get_imsm_map(dev, 0);
dev->vol.migr_state = 1;
- dev->vol.migr_type = rebuild_resync;
+ dev->vol.migr_type = migr_type;
dev->vol.curr_migr_unit = 0;
dest = get_imsm_map(dev, 1);
+ /* duplicate and then set the target end state in map[0] */
memcpy(dest, src, sizeof_imsm_map(src));
+ if (migr_type == MIGR_REBUILD) {
+ __u32 ord;
+ int i;
+
+ for (i = 0; i < src->num_members; i++) {
+ ord = __le32_to_cpu(src->disk_ord_tbl[i]);
+ set_imsm_ord_tbl_ent(src, i, ord_to_idx(ord));
+ }
+ }
+
src->map_state = to_state;
}
static void end_migration(struct imsm_dev *dev, __u8 map_state)
{
struct imsm_map *map = get_imsm_map(dev, 0);
+ struct imsm_map *prev = get_imsm_map(dev, dev->vol.migr_state);
+ int i;
+
+ /* merge any IMSM_ORD_REBUILD bits that were not successfully
+ * completed in the last migration.
+ *
+ * FIXME add support for online capacity expansion and
+ * raid-level-migration
+ */
+ for (i = 0; i < prev->num_members; i++)
+ map->disk_ord_tbl[i] |= prev->disk_ord_tbl[i];
dev->vol.migr_state = 0;
dev->vol.curr_migr_unit = 0;
sectors = mpb_sectors(anchor) - 1;
free(anchor);
if (!sectors) {
+ check_sum = __gen_imsm_checksum(super->anchor);
+ if (check_sum != __le32_to_cpu(super->anchor->check_sum)) {
+ if (devname)
+ fprintf(stderr,
+ Name ": IMSM checksum %x != %x on %s\n",
+ check_sum,
+ __le32_to_cpu(super->anchor->check_sum),
+ devname);
+ return 2;
+ }
+
rc = load_imsm_disk(fd, super, devname, 0);
if (rc == 0)
rc = parse_raid_devices(super);
Name ": IMSM checksum %x != %x on %s\n",
check_sum, __le32_to_cpu(super->anchor->check_sum),
devname);
- return 2;
+ return 3;
}
/* FIXME the BBM log is disk specific so we cannot use this global
super->create_offset = ~((__u32 ) 0);
if (!check_env("IMSM_NO_PLATFORM"))
super->orom = find_imsm_orom();
- if (super->orom) {
+ if (super->orom && !check_env("IMSM_TEST_OROM")) {
struct sys_dev *list, *ent;
/* find the first intel ahci controller */
dl = serial_to_dl(disk->serial, super);
if (dl)
continue;
- /* ok we have a 'disk' without a live entry in
- * super->disks
- */
- if (disk->status & FAILED_DISK || !(disk->status & USABLE_DISK))
- continue; /* never mind, already marked */
dl = malloc(sizeof(*dl));
if (!dl)
dl->index = i;
serialcpy(dl->serial, disk->serial);
dl->disk = *disk;
+ dl->e = NULL;
dl->next = super->missing;
super->missing = dl;
}
char nm[20];
int dfd;
int rv;
+ int devnum = fd2devnum(fd);
+ int retry;
+ enum sysfs_read_flags flags;
+
+ flags = GET_LEVEL|GET_VERSION|GET_DEVS|GET_STATE;
+ if (mdmon_running(devnum))
+ flags |= SKIP_GONE_DEVS;
- /* check if this disk is a member of an active array */
- sra = sysfs_read(fd, 0, GET_LEVEL|GET_VERSION|GET_DEVS|GET_STATE);
+ /* check if 'fd' an opened container */
+ sra = sysfs_read(fd, 0, flags);
if (!sra)
return 1;
return 2;
}
rv = load_imsm_mpb(dfd, super, NULL);
+
+ /* retry the load if we might have raced against mdmon */
+ if (rv == 3 && mdmon_running(devnum))
+ for (retry = 0; retry < 3; retry++) {
+ usleep(3000);
+ rv = load_imsm_mpb(dfd, super, NULL);
+ if (rv != 3)
+ break;
+ }
if (!keep_fd)
close(dfd);
if (rv == 0) {
}
} else {
free_imsm(super);
- return 2;
+ return rv;
}
}
}
*sbp = super;
- st->container_dev = fd2devnum(fd);
+ st->container_dev = devnum;
if (st->ss == NULL) {
st->ss = &super_imsm;
st->minor_version = 0;
*/
if (super->current_vol == 0)
mpb->num_disks = 0;
+
+ for (i = 0; i < super->current_vol; i++) {
+ dev = get_imsm_dev(super, i);
+ if (strncmp((char *) dev->volume, name,
+ MAX_RAID_SERIAL_LEN) == 0) {
+ fprintf(stderr, Name": '%s' is already defined for this container\n",
+ name);
+ return 0;
+ }
+ }
+
sprintf(st->subarray, "%d", idx);
dv = malloc(sizeof(*dv));
if (!dv) {
map->blocks_per_member = __cpu_to_le32(info_to_blocks_per_member(info));
map->blocks_per_strip = __cpu_to_le16(info_to_blocks_per_strip(info));
map->num_data_stripes = __cpu_to_le32(info_to_num_data_stripes(info));
+ map->failed_disk_num = ~0;
map->map_state = info->level ? IMSM_T_STATE_UNINITIALIZED :
IMSM_T_STATE_NORMAL;
dd->index = -1;
dd->devname = devname ? strdup(devname) : NULL;
dd->fd = fd;
+ dd->e = NULL;
rv = imsm_read_serial(fd, devname, dd->serial);
if (rv) {
fprintf(stderr,
level, raiddisks, raiddisks > 1 ? "s" : "");
return 0;
}
- if (super->orom && !imsm_orom_has_chunk(super->orom, chunk)) {
+ if (super->orom && level != 1 &&
+ !imsm_orom_has_chunk(super->orom, chunk)) {
pr_vrb(": platform does not support a chunk size of: %d\n", chunk);
return 0;
}
int failed = 0;
struct imsm_disk *disk;
struct imsm_map *map = get_imsm_map(dev, 0);
+ struct imsm_map *prev = get_imsm_map(dev, dev->vol.migr_state);
+ __u32 ord;
+ int idx;
- for (i = 0; i < map->num_members; i++) {
- __u32 ord = get_imsm_ord_tbl_ent(dev, i);
- int idx = ord_to_idx(ord);
+ /* at the beginning of migration we set IMSM_ORD_REBUILD on
+ * disks that are being rebuilt. New failures are recorded to
+ * map[0]. So we look through all the disks we started with and
+ * see if any failures are still present, or if any new ones
+ * have arrived
+ *
+ * FIXME add support for online capacity expansion and
+ * raid-level-migration
+ */
+ for (i = 0; i < prev->num_members; i++) {
+ ord = __le32_to_cpu(prev->disk_ord_tbl[i]);
+ ord |= __le32_to_cpu(map->disk_ord_tbl[i]);
+ idx = ord_to_idx(ord);
disk = get_imsm_disk(super, idx);
if (!disk || disk->status & FAILED_DISK ||
return 0;
}
-static void mark_failure(struct imsm_disk *disk)
+/* return true if we recorded new information */
+static int mark_failure(struct imsm_dev *dev, struct imsm_disk *disk, int idx)
{
- if (disk->status & FAILED_DISK)
- return;
+ __u32 ord;
+ int slot;
+ struct imsm_map *map;
+
+ /* new failures are always set in map[0] */
+ map = get_imsm_map(dev, 0);
+
+ slot = get_imsm_disk_slot(map, idx);
+ if (slot < 0)
+ return 0;
+
+ ord = __le32_to_cpu(map->disk_ord_tbl[slot]);
+ if ((disk->status & FAILED_DISK) && (ord & IMSM_ORD_REBUILD))
+ return 0;
+
disk->status |= FAILED_DISK;
+ set_imsm_ord_tbl_ent(map, slot, idx | IMSM_ORD_REBUILD);
+ if (map->failed_disk_num == ~0)
+ map->failed_disk_num = slot;
+ return 1;
+}
+
+static void mark_missing(struct imsm_dev *dev, struct imsm_disk *disk, int idx)
+{
+ mark_failure(dev, disk, idx);
+
+ if (disk->scsi_id == __cpu_to_le32(~(__u32)0))
+ return;
+
disk->scsi_id = __cpu_to_le32(~(__u32)0);
memmove(&disk->serial[0], &disk->serial[1], MAX_RAID_SERIAL_LEN - 1);
}
dprintf("imsm: mark missing\n");
end_migration(dev, map_state);
for (dl = super->missing; dl; dl = dl->next)
- mark_failure(&dl->disk);
+ mark_missing(dev, &dl->disk, dl->index);
super->updates_pending++;
}
if (is_resync_complete(a)) {
/* complete intialization / resync,
- * recovery is completed in ->set_disk
+ * recovery and interrupted recovery is completed in
+ * ->set_disk
*/
if (is_resyncing(dev)) {
dprintf("imsm: mark resync done\n");
disk = get_imsm_disk(super, ord_to_idx(ord));
/* check for new failures */
- if ((state & DS_FAULTY) && !(disk->status & FAILED_DISK)) {
- mark_failure(disk);
- super->updates_pending++;
+ if (state & DS_FAULTY) {
+ if (mark_failure(dev, disk, ord_to_idx(ord)))
+ super->updates_pending++;
}
/* check if in_sync */
- if (state & DS_INSYNC && ord & IMSM_ORD_REBUILD) {
+ if (state & DS_INSYNC && ord & IMSM_ORD_REBUILD && is_rebuilding(dev)) {
struct imsm_map *migr_map = get_imsm_map(dev, 1);
set_imsm_ord_tbl_ent(migr_map, n, ord_to_idx(ord));
/* check if recovery complete, newly degraded, or failed */
if (map_state == IMSM_T_STATE_NORMAL && is_rebuilding(dev)) {
end_migration(dev, map_state);
+ map = get_imsm_map(dev, 0);
+ map->failed_disk_num = ~0;
super->updates_pending++;
} else if (map_state == IMSM_T_STATE_DEGRADED &&
map->map_state != map_state &&
continue;
/* skip in use or failed drives */
- if (dl->disk.status & FAILED_DISK || idx == dl->index) {
- dprintf("%x:%x status ( %s%s)\n",
- dl->major, dl->minor,
- dl->disk.status & FAILED_DISK ? "failed " : "",
- idx == dl->index ? "in use " : "");
+ if (dl->disk.status & FAILED_DISK || idx == dl->index ||
+ dl->index == -2) {
+ dprintf("%x:%x status (failed: %d index: %d)\n",
+ dl->major, dl->minor,
+ (dl->disk.status & FAILED_DISK) == FAILED_DISK, idx);
continue;
}
/* check if this disk is already a member of
* this array
*/
- for (j = 0; j < map->num_members; j++)
- if (get_imsm_disk_idx(dev, j) == dl->index)
- break;
- if (j < map->num_members)
+ if (get_imsm_disk_slot(map, dl->index) >= 0)
continue;
found = 0;
found = 0;
for (a = st->arrays; a ; a = a->next) {
dev = get_imsm_dev(super, a->info.container_member);
- for (i = 0; i < map->num_members; i++)
- if (victim == get_imsm_disk_idx(dev, i))
- found++;
+ map = get_imsm_map(dev, 0);
+
+ if (get_imsm_disk_slot(map, victim) >= 0)
+ found++;
}
/* delete the victim if it is no longer being
projects / thirdparty / mdadm.git / blobdiff