#define MPB_SECTOR_CNT 418
#define IMSM_RESERVED_SECTORS 4096
+#define SECT_PER_MB_SHIFT 11
/* Disk configuration info. */
#define IMSM_MAX_DEVICES 255
__u8 num_members; /* number of member disks */
__u8 num_domains; /* number of parity domains */
__u8 failed_disk_num; /* valid only when state is degraded */
- __u8 reserved[1];
+ __u8 ddf;
__u32 filler[7]; /* expansion area */
#define IMSM_ORD_REBUILD (1 << 24)
__u32 disk_ord_tbl[1]; /* disk_ord_tbl[num_members],
#define MIGR_VERIFY 2 /* analagous to echo check > sync_action */
#define MIGR_GEN_MIGR 3
#define MIGR_STATE_CHANGE 4
+#define MIGR_REPAIR 5
__u8 migr_type; /* Initializing, Rebuilding, ... */
__u8 dirty;
__u8 fs_state; /* fast-sync state for CnG (0xff == disabled) */
static char *map_state_str[] = { "normal", "uninitialized", "degraded", "failed" };
#endif
+static __u8 migr_type(struct imsm_dev *dev)
+{
+ if (dev->vol.migr_type == MIGR_VERIFY &&
+ dev->status & DEV_VERIFY_AND_FIX)
+ return MIGR_REPAIR;
+ else
+ return dev->vol.migr_type;
+}
+
+static void set_migr_type(struct imsm_dev *dev, __u8 migr_type)
+{
+ /* for compatibility with older oroms convert MIGR_REPAIR, into
+ * MIGR_VERIFY w/ DEV_VERIFY_AND_FIX status
+ */
+ if (migr_type == MIGR_REPAIR) {
+ dev->vol.migr_type = MIGR_VERIFY;
+ dev->status |= DEV_VERIFY_AND_FIX;
+ } else {
+ dev->vol.migr_type = migr_type;
+ dev->status &= ~DEV_VERIFY_AND_FIX;
+ }
+}
+
static unsigned int sector_count(__u32 bytes)
{
return ((bytes + (512-1)) & (~(512-1))) / 512;
int fd;
int extent_cnt;
struct extent *e; /* for determining freespace @ create */
+ int raiddisk; /* slot to fill in autolayout */
} *disks;
struct dl *add; /* list of disks to add while mdmon active */
struct dl *missing; /* disks removed while we weren't looking */
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)" : "");
printf(" Chunk Size : %u KiB\n",
__le16_to_cpu(map->blocks_per_strip) / 2);
printf(" Reserved : %d\n", __le32_to_cpu(dev->reserved_blocks));
- printf(" Migrate State : %s", dev->vol.migr_state ? "migrating" : "idle");
- if (dev->vol.migr_state)
- printf(": %s", dev->vol.migr_type ? "rebuilding" : "initializing");
- printf("\n");
+ printf(" Migrate State : %s", dev->vol.migr_state ? "migrating" : "idle\n");
+ if (dev->vol.migr_state) {
+ if (migr_type(dev) == MIGR_INIT)
+ printf(": initializing\n");
+ else if (migr_type(dev) == MIGR_REBUILD)
+ printf(": rebuilding\n");
+ else if (migr_type(dev) == MIGR_VERIFY)
+ printf(": check\n");
+ else if (migr_type(dev) == MIGR_GEN_MIGR)
+ printf(": general migration\n");
+ else if (migr_type(dev) == MIGR_STATE_CHANGE)
+ printf(": state change\n");
+ else if (migr_type(dev) == MIGR_REPAIR)
+ printf(": repair\n");
+ else
+ printf(": <unknown:%d>\n", migr_type(dev));
+ }
printf(" Map State : %s", map_state_str[map->map_state]);
if (dev->vol.migr_state) {
struct imsm_map *map = get_imsm_map(dev, 1);
printf(" Family : %08x\n", __le32_to_cpu(mpb->family_num));
printf(" Generation : %08x\n", __le32_to_cpu(mpb->generation_num));
getinfo_super_imsm(st, &info);
- fname_from_uuid(st, &info, nbuf,'-');
+ fname_from_uuid(st, &info, nbuf, ':');
printf(" UUID : %s\n", nbuf + 5);
sum = __le32_to_cpu(mpb->check_sum);
printf(" Checksum : %08x %s\n", sum,
printf(" Signature : %x\n", __le32_to_cpu(log->signature));
printf(" Entry Count : %d\n", __le32_to_cpu(log->entry_count));
printf(" Spare Blocks : %d\n", __le32_to_cpu(log->reserved_spare_block_count));
- printf(" First Spare : %llx\n", __le64_to_cpu(log->first_spare_lba));
+ printf(" First Spare : %llx\n",
+ (unsigned long long) __le64_to_cpu(log->first_spare_lba));
}
for (i = 0; i < mpb->num_raid_devs; i++) {
struct mdinfo info;
super->current_vol = i;
getinfo_super_imsm(st, &info);
- fname_from_uuid(st, &info, nbuf, '-');
+ fname_from_uuid(st, &info, nbuf, ':');
print_imsm_dev(dev, nbuf + 5, super->disks->index);
}
for (i = 0; i < mpb->num_disks; i++) {
}
}
-static void brief_examine_super_imsm(struct supertype *st)
+static void brief_examine_super_imsm(struct supertype *st, int verbose)
{
/* We just write a generic IMSM ARRAY entry */
struct mdinfo info;
return;
getinfo_super_imsm(st, &info);
- fname_from_uuid(st, &info, nbuf,'-');
- printf("ARRAY metadata=imsm auto=md UUID=%s\n", nbuf + 5);
+ fname_from_uuid(st, &info, nbuf, ':');
for (i = 0; i < super->anchor->num_raid_devs; i++) {
struct imsm_dev *dev = get_imsm_dev(super, i);
super->current_vol = i;
getinfo_super_imsm(st, &info);
- fname_from_uuid(st, &info, nbuf1,'-');
- printf("ARRAY /dev/md/%.16s container=%s\n"
- " member=%d auto=mdp UUID=%s\n",
+ fname_from_uuid(st, &info, nbuf1, ':');
+ printf("ARRAY /dev/md/%.16s container=%s member=%d UUID=%s\n",
dev->volume, nbuf + 5, i, nbuf1 + 5);
}
+ printf("ARRAY metadata=imsm UUID=%s\n", nbuf + 5);
+}
+
+static void export_examine_super_imsm(struct supertype *st)
+{
+ struct intel_super *super = st->sb;
+ struct imsm_super *mpb = super->anchor;
+ struct mdinfo info;
+ char nbuf[64];
+
+ getinfo_super_imsm(st, &info);
+ fname_from_uuid(st, &info, nbuf, ':');
+ printf("MD_METADATA=imsm\n");
+ printf("MD_LEVEL=container\n");
+ printf("MD_UUID=%s\n", nbuf+5);
+ printf("MD_DEVICES=%u\n", mpb->num_disks);
}
static void detail_super_imsm(struct supertype *st, char *homehost)
char nbuf[64];
getinfo_super_imsm(st, &info);
- fname_from_uuid(st, &info, nbuf,'-');
+ fname_from_uuid(st, &info, nbuf, ':');
printf("\n UUID : %s\n", nbuf + 5);
}
struct mdinfo info;
char nbuf[64];
getinfo_super_imsm(st, &info);
- fname_from_uuid(st, &info, nbuf,'-');
+ fname_from_uuid(st, &info, nbuf, ':');
printf(" UUID=%s", nbuf + 5);
}
return err;
}
-static int detail_platform_imsm(int verbose)
+static int detail_platform_imsm(int verbose, int enumerate_only)
{
/* There are two components to imsm platform support, the ahci SATA
* controller and the option-rom. To find the SATA controller we
int host_base = 0;
int port_count = 0;
+ if (enumerate_only) {
+ if (check_env("IMSM_NO_PLATFORM") || find_imsm_orom())
+ return 0;
+ return 2;
+ }
+
list = find_driver_devices("pci", "ahci");
for (hba = list; hba; hba = hba->next)
if (devpath_to_vendor(hba->path) == 0x8086)
imsm_orom_has_raid1e(orom) ? " raid1e" : "",
imsm_orom_has_raid10(orom) ? " raid10" : "",
imsm_orom_has_raid5(orom) ? " raid5" : "");
+ printf(" Chunk Sizes :%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n",
+ imsm_orom_has_chunk(orom, 2) ? " 2k" : "",
+ imsm_orom_has_chunk(orom, 4) ? " 4k" : "",
+ imsm_orom_has_chunk(orom, 8) ? " 8k" : "",
+ imsm_orom_has_chunk(orom, 16) ? " 16k" : "",
+ imsm_orom_has_chunk(orom, 32) ? " 32k" : "",
+ imsm_orom_has_chunk(orom, 64) ? " 64k" : "",
+ imsm_orom_has_chunk(orom, 128) ? " 128k" : "",
+ imsm_orom_has_chunk(orom, 256) ? " 256k" : "",
+ imsm_orom_has_chunk(orom, 512) ? " 512k" : "",
+ imsm_orom_has_chunk(orom, 1024*1) ? " 1M" : "",
+ imsm_orom_has_chunk(orom, 1024*2) ? " 2M" : "",
+ imsm_orom_has_chunk(orom, 1024*4) ? " 4M" : "",
+ imsm_orom_has_chunk(orom, 1024*8) ? " 8M" : "",
+ imsm_orom_has_chunk(orom, 1024*16) ? " 16M" : "",
+ imsm_orom_has_chunk(orom, 1024*32) ? " 32M" : "",
+ imsm_orom_has_chunk(orom, 1024*64) ? " 64M" : "");
printf(" Max Disks : %d\n", orom->tds);
printf(" Max Volumes : %d\n", orom->vpa);
printf(" I/O Controller : %s\n", hba_path);
case 10:
return 0x102;
}
- return -1;
+ return UnSet;
}
static void getinfo_super_imsm_volume(struct supertype *st, struct mdinfo *info)
struct intel_super *super = st->sb;
struct imsm_dev *dev = get_imsm_dev(super, super->current_vol);
struct imsm_map *map = get_imsm_map(dev, 0);
+ struct dl *dl;
+ for (dl = super->disks; dl; dl = dl->next)
+ if (dl->raiddisk == info->disk.raid_disk)
+ break;
info->container_member = super->current_vol;
info->array.raid_disks = map->num_members;
info->array.level = get_imsm_raid_level(map);
info->array.utime = 0;
info->array.chunk_size = __le16_to_cpu(map->blocks_per_strip) << 9;
info->array.state = !dev->vol.dirty;
+ info->custom_array_size = __le32_to_cpu(dev->size_high);
+ info->custom_array_size <<= 32;
+ info->custom_array_size |= __le32_to_cpu(dev->size_low);
info->disk.major = 0;
info->disk.minor = 0;
+ if (dl) {
+ info->disk.major = dl->major;
+ info->disk.minor = dl->minor;
+ }
info->data_offset = __le32_to_cpu(map->pba_of_lba0);
info->component_size = __le32_to_cpu(map->blocks_per_member);
if (map->map_state == IMSM_T_STATE_UNINITIALIZED || dev->vol.dirty)
info->resync_start = 0;
else if (dev->vol.migr_state)
- info->resync_start = __le32_to_cpu(dev->vol.curr_migr_unit);
+ /* FIXME add curr_migr_unit to resync_start conversion */
+ info->resync_start = 0;
else
info->resync_start = ~0ULL;
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 (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,
int rv;
int rsp_len;
int len;
- char *c, *rsp_buf;
+ char *dest;
+ char *src;
+ char *rsp_buf;
+ int i;
memset(scsi_serial, 0, sizeof(scsi_serial));
return rv;
}
- /* trim leading whitespace */
rsp_len = scsi_serial[3];
+ if (!rsp_len) {
+ if (devname)
+ fprintf(stderr,
+ Name ": Failed to retrieve serial for %s\n",
+ devname);
+ return 2;
+ }
rsp_buf = (char *) &scsi_serial[4];
- c = rsp_buf;
- while (isspace(*c))
- c++;
- /* truncate len to the end of rsp_buf if necessary */
- if (c + MAX_RAID_SERIAL_LEN > rsp_buf + rsp_len)
- len = rsp_len - (c - rsp_buf);
- else
+ /* trim all whitespace and non-printable characters and convert
+ * ':' to ';'
+ */
+ for (i = 0, dest = rsp_buf; i < rsp_len; i++) {
+ src = &rsp_buf[i];
+ if (*src > 0x20) {
+ /* ':' is reserved for use in placeholder serial
+ * numbers for missing disks
+ */
+ if (*src == ':')
+ *dest++ = ';';
+ else
+ *dest++ = *src;
+ }
+ }
+ len = dest - rsp_buf;
+ dest = rsp_buf;
+
+ /* truncate leading characters */
+ if (len > MAX_RAID_SERIAL_LEN) {
+ dest += len - MAX_RAID_SERIAL_LEN;
len = MAX_RAID_SERIAL_LEN;
+ }
- /* initialize the buffer and copy rsp_buf characters */
memset(serial, 0, MAX_RAID_SERIAL_LEN);
- memcpy(serial, c, len);
-
- /* trim trailing whitespace starting with the last character copied */
- c = (char *) &serial[len - 1];
- while (isspace(*c) || *c == '\0')
- *c-- = '\0';
+ memcpy(serial, dest, len);
return 0;
}
* 1/ Idle (migr_state=0 map0state=normal||unitialized||degraded||failed)
* 2/ Initialize (migr_state=1 migr_type=MIGR_INIT map0state=normal
* map1state=unitialized)
- * 3/ Verify (Resync) (migr_state=1 migr_type=MIGR_REBUILD map0state=normal
+ * 3/ Repair (Resync) (migr_state=1 migr_type=MIGR_REPAIR map0state=normal
* map1state=normal)
* 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;
+ set_migr_type(dev, 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;
if (posix_memalign(&buf, 512, len) != 0)
return 1;
- memcpy(buf, super->buf, len);
+ memcpy(buf, super->buf, super->len);
+ memset(buf + super->len, 0, len - super->len);
free(super->buf);
super->buf = buf;
super->len = len;
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;
- /* check if this disk is a member of an active array */
- sra = sysfs_read(fd, 0, GET_LEVEL|GET_VERSION|GET_DEVS|GET_STATE);
+ flags = GET_LEVEL|GET_VERSION|GET_DEVS|GET_STATE;
+ if (mdmon_running(devnum))
+ flags |= SKIP_GONE_DEVS;
+
+ /* 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;
}
}
if (st->subarray[0]) {
if (atoi(st->subarray) <= super->anchor->num_raid_devs)
super->current_vol = atoi(st->subarray);
- else
+ else {
+ free_imsm(super);
return 1;
+ }
}
*sbp = super;
- st->container_dev = fd2devnum(fd);
+ st->container_dev = devnum;
if (st->ss == NULL) {
st->ss = &super_imsm;
st->minor_version = 0;
if (load_super_imsm_all(st, fd, &st->sb, devname, 1) == 0)
return 0;
#endif
- if (st->subarray[0])
- return 1; /* FIXME */
+
+ free_super_imsm(st);
super = alloc_super(0);
if (!super) {
return rv;
}
+ if (st->subarray[0]) {
+ if (atoi(st->subarray) <= super->anchor->num_raid_devs)
+ super->current_vol = atoi(st->subarray);
+ else {
+ free_imsm(super);
+ return 1;
+ }
+ }
+
st->sb = super;
if (st->ss == NULL) {
st->ss = &super_imsm;
return info->chunk_size >> 9;
}
-static __u32 info_to_num_data_stripes(mdu_array_info_t *info)
+static __u32 info_to_num_data_stripes(mdu_array_info_t *info, int num_domains)
{
__u32 num_stripes;
num_stripes = (info->size * 2) / info_to_blocks_per_strip(info);
- if (info->level == 1)
- num_stripes /= 2;
+ num_stripes /= num_domains;
return num_stripes;
}
int i;
unsigned long long array_blocks;
size_t size_old, size_new;
+ __u32 num_data_stripes;
if (super->orom && mpb->num_raid_devs >= super->orom->vpa) {
fprintf(stderr, Name": This imsm-container already has the "
*/
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) {
array_blocks = calc_array_size(info->level, info->raid_disks,
info->layout, info->chunk_size,
info->size*2);
+ /* round array size down to closest MB */
+ array_blocks = (array_blocks >> SECT_PER_MB_SHIFT) << SECT_PER_MB_SHIFT;
+
dev->size_low = __cpu_to_le32((__u32) array_blocks);
dev->size_high = __cpu_to_le32((__u32) (array_blocks >> 32));
dev->status = __cpu_to_le32(0);
dev->reserved_blocks = __cpu_to_le32(0);
vol = &dev->vol;
vol->migr_state = 0;
- vol->migr_type = MIGR_INIT;
+ set_migr_type(dev, MIGR_INIT);
vol->dirty = 0;
vol->curr_migr_unit = 0;
map = get_imsm_map(dev, 0);
map->pba_of_lba0 = __cpu_to_le32(super->create_offset);
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;
+ map->ddf = 1;
if (info->level == 1 && info->raid_disks > 2) {
fprintf(stderr, Name": imsm does not support more than 2 disks"
"in a raid1 volume\n");
return 0;
}
+
+ map->raid_level = info->level;
if (info->level == 10) {
map->raid_level = 1;
map->num_domains = info->raid_disks / 2;
- } else {
- map->raid_level = info->level;
- map->num_domains = !!map->raid_level;
- }
+ } else if (info->level == 1)
+ map->num_domains = info->raid_disks;
+ else
+ map->num_domains = 1;
+
+ num_data_stripes = info_to_num_data_stripes(info, map->num_domains);
+ map->num_data_stripes = __cpu_to_le32(num_data_stripes);
map->num_members = info->raid_disks;
for (i = 0; i < map->num_members; i++) {
return 1;
}
- for (dl = super->disks; dl ; dl = dl->next)
- if (dl->major == dk->major &&
- dl->minor == dk->minor)
- break;
+ if (fd == -1) {
+ /* we're doing autolayout so grab the pre-marked (in
+ * validate_geometry) raid_disk
+ */
+ for (dl = super->disks; dl; dl = dl->next)
+ if (dl->raiddisk == dk->raid_disk)
+ break;
+ } else {
+ for (dl = super->disks; dl ; dl = dl->next)
+ if (dl->major == dk->major &&
+ dl->minor == dk->minor)
+ break;
+ }
if (!dl) {
fprintf(stderr, Name ": %s is not a member of the same container\n", devname);
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,
int i, j;
int start_extent;
unsigned long long pos;
- unsigned long long start;
+ unsigned long long start = 0;
unsigned long long maxsize;
unsigned long reserve;
case 1:
if (raiddisks > 2)
return imsm_orom_has_raid1e(orom);
- else
- return imsm_orom_has_raid1(orom);
- case 10: return imsm_orom_has_raid10(orom);
- case 5: return imsm_orom_has_raid5(orom);
+ return imsm_orom_has_raid1(orom) && raiddisks == 2;
+ case 10: return imsm_orom_has_raid10(orom) && raiddisks == 4;
+ case 5: return imsm_orom_has_raid5(orom) && raiddisks > 2;
}
else
return 1; /* not on an Intel RAID platform so anything goes */
return 0;
}
-#define vprintf(fmt, arg...) (void) (verbose && fprintf(stderr, Name fmt, ##arg))
+#define pr_vrb(fmt, arg...) (void) (verbose && fprintf(stderr, Name fmt, ##arg))
/* validate_geometry_imsm_volume - lifted from validate_geometry_ddf_bvd
* FIX ME add ahci details
*/
{
struct stat stb;
struct intel_super *super = st->sb;
+ struct imsm_super *mpb = super->anchor;
struct dl *dl;
unsigned long long pos = 0;
unsigned long long maxsize;
return 0;
if (!is_raid_level_supported(super->orom, level, raiddisks)) {
- vprintf(": platform does not support raid level: %d\n", level);
+ pr_vrb(": platform does not support raid%d with %d disk%s\n",
+ level, raiddisks, raiddisks > 1 ? "s" : "");
return 0;
}
- if (super->orom && !imsm_orom_has_chunk(super->orom, chunk)) {
- vprintf(": platform does not support a chunk size of: %d\n", 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;
}
if (layout != imsm_level_to_layout(level)) {
if (level == 5)
- vprintf(": imsm raid 5 only supports the left-asymmetric layout\n");
+ pr_vrb(": imsm raid 5 only supports the left-asymmetric layout\n");
else if (level == 10)
- vprintf(": imsm raid 10 only supports the n2 layout\n");
+ pr_vrb(": imsm raid 10 only supports the n2 layout\n");
else
- vprintf(": imsm unknown layout %#x for this raid level %d\n",
+ pr_vrb(": imsm unknown layout %#x for this raid level %d\n",
layout, level);
return 0;
}
fprintf(stderr, Name ": %s is not in the "
"same imsm set\n", dev);
return 0;
+ } else if (super->orom && dl->index < 0 && mpb->num_raid_devs) {
+ /* If a volume is present then the current creation attempt
+ * cannot incorporate new spares because the orom may not
+ * understand this configuration (all member disks must be
+ * members of each array in the container).
+ */
+ fprintf(stderr, Name ": %s is a spare and a volume"
+ " is already defined for this container\n", dev);
+ fprintf(stderr, Name ": The option-rom requires all member"
+ " disks to be a member of all volumes\n");
+ return 0;
}
/* retrieve the largest free space block */
return 1;
}
+static int reserve_space(struct supertype *st, int raiddisks,
+ unsigned long long size, int chunk,
+ unsigned long long *freesize)
+{
+ struct intel_super *super = st->sb;
+ struct imsm_super *mpb = super->anchor;
+ struct dl *dl;
+ int i;
+ int extent_cnt;
+ struct extent *e;
+ unsigned long long maxsize;
+ unsigned long long minsize;
+ int cnt;
+ int used;
+
+ /* find the largest common start free region of the possible disks */
+ used = 0;
+ extent_cnt = 0;
+ cnt = 0;
+ for (dl = super->disks; dl; dl = dl->next) {
+ dl->raiddisk = -1;
+
+ if (dl->index >= 0)
+ used++;
+
+ /* don't activate new spares if we are orom constrained
+ * and there is already a volume active in the container
+ */
+ if (super->orom && dl->index < 0 && mpb->num_raid_devs)
+ continue;
+
+ e = get_extents(super, dl);
+ if (!e)
+ continue;
+ for (i = 1; e[i-1].size; i++)
+ ;
+ dl->e = e;
+ dl->extent_cnt = i;
+ extent_cnt += i;
+ cnt++;
+ }
+
+ maxsize = merge_extents(super, extent_cnt);
+ minsize = size;
+ if (size == 0)
+ minsize = chunk;
+
+ if (cnt < raiddisks ||
+ (super->orom && used && used != raiddisks) ||
+ maxsize < minsize) {
+ fprintf(stderr, Name ": not enough devices with space to create array.\n");
+ return 0; /* No enough free spaces large enough */
+ }
+
+ if (size == 0) {
+ size = maxsize;
+ if (chunk) {
+ size /= chunk;
+ size *= chunk;
+ }
+ }
+
+ cnt = 0;
+ for (dl = super->disks; dl; dl = dl->next)
+ if (dl->e)
+ dl->raiddisk = cnt++;
+
+ *freesize = size;
+
+ return 1;
+}
+
static int validate_geometry_imsm(struct supertype *st, int level, int layout,
int raiddisks, int chunk, unsigned long long size,
char *dev, unsigned long long *freesize,
if (!dev) {
if (st->sb && freesize) {
- /* Should do auto-layout here */
- fprintf(stderr, Name ": IMSM does not support auto-layout yet\n");
- return 0;
+ /* we are being asked to automatically layout a
+ * new volume based on the current contents of
+ * the container. If the the parameters can be
+ * satisfied reserve_space will record the disks,
+ * start offset, and size of the volume to be
+ * created. add_to_super and getinfo_super
+ * detect when autolayout is in progress.
+ */
+ return reserve_space(st, raiddisks, size, chunk, freesize);
}
return 1;
}
struct mdinfo *this;
int slot;
+ /* do not publish arrays that are in the middle of an
+ * unsupported migration
+ */
+ if (dev->vol.migr_state &&
+ (migr_type(dev) == MIGR_GEN_MIGR ||
+ migr_type(dev) == MIGR_STATE_CHANGE)) {
+ fprintf(stderr, Name ": cannot assemble volume '%.16s':"
+ " unsupported migration in progress\n",
+ dev->volume);
+ continue;
+ }
+
this = malloc(sizeof(*this));
memset(this, 0, sizeof(*this));
this->next = rest;
info_d = malloc(sizeof(*info_d));
if (!info_d) {
fprintf(stderr, Name ": failed to allocate disk"
- " for volume %s\n", (char *) dev->volume);
+ " for volume %.16s\n", dev->volume);
free(this);
this = rest;
break;
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 ||
if (!dev->vol.migr_state)
return 0;
- if (dev->vol.migr_type == MIGR_INIT)
+ if (migr_type(dev) == MIGR_INIT ||
+ migr_type(dev) == MIGR_REPAIR)
return 1;
migr_map = get_imsm_map(dev, 1);
if (!dev->vol.migr_state)
return 0;
- if (dev->vol.migr_type != MIGR_REBUILD)
+ if (migr_type(dev) != MIGR_REBUILD)
return 0;
migr_map = get_imsm_map(dev, 1);
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");
} else if (!is_resyncing(dev) && !failed) {
/* mark the start of the init process if nothing is failed */
dprintf("imsm: mark resync start (%llu)\n", a->resync_start);
- if (map->map_state == IMSM_T_STATE_NORMAL)
- migrate(dev, IMSM_T_STATE_NORMAL, MIGR_REBUILD);
- else
+ if (map->map_state == IMSM_T_STATE_UNINITIALIZED)
migrate(dev, IMSM_T_STATE_NORMAL, MIGR_INIT);
+ else
+ migrate(dev, IMSM_T_STATE_NORMAL, MIGR_REPAIR);
super->updates_pending++;
}
- /* check if we can update the migration checkpoint */
- if (dev->vol.migr_state &&
- __le32_to_cpu(dev->vol.curr_migr_unit) != a->resync_start) {
- dprintf("imsm: checkpoint migration (%llu)\n", a->resync_start);
- dev->vol.curr_migr_unit = __cpu_to_le32(a->resync_start);
- super->updates_pending++;
- }
+ /* FIXME check if we can update curr_migr_unit from resync_start */
/* mark dirty / clean */
if (dev->vol.dirty != !consistent) {
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 &&
return dl;
}
-static struct dl *imsm_add_spare(struct intel_super *super, int slot, struct active_array *a)
+static struct dl *imsm_add_spare(struct intel_super *super, int slot,
+ struct active_array *a, int activate_new)
{
struct imsm_dev *dev = get_imsm_dev(super, a->info.container_member);
int idx = get_imsm_disk_idx(dev, slot);
- struct imsm_map *map = get_imsm_map(dev, 0);
+ struct imsm_super *mpb = super->anchor;
+ struct imsm_map *map;
unsigned long long esize;
unsigned long long pos;
struct mdinfo *d;
struct extent *ex;
- int j;
+ int i, j;
int found;
__u32 array_start;
+ __u32 blocks;
struct dl *dl;
for (dl = super->disks; dl; dl = dl->next) {
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;
}
+ /* skip pure spares when we are looking for partially
+ * assimilated drives
+ */
+ if (dl->index == -1 && !activate_new)
+ continue;
+
/* Does this unused device have the requisite free space?
- * We need a->info.component_size sectors
+ * It needs to be able to cover all member volumes
*/
ex = get_extents(super, dl);
if (!ex) {
dprintf("cannot get extents\n");
continue;
}
- found = 0;
- j = 0;
- pos = 0;
- array_start = __le32_to_cpu(map->pba_of_lba0);
+ for (i = 0; i < mpb->num_raid_devs; i++) {
+ dev = get_imsm_dev(super, i);
+ map = get_imsm_map(dev, 0);
- do {
- /* check that we can start at pba_of_lba0 with
- * a->info.component_size of space
+ /* check if this disk is already a member of
+ * this array
*/
- esize = ex[j].start - pos;
- if (array_start >= pos &&
- array_start + a->info.component_size < ex[j].start) {
- found = 1;
+ if (get_imsm_disk_slot(map, dl->index) >= 0)
+ continue;
+
+ found = 0;
+ j = 0;
+ pos = 0;
+ array_start = __le32_to_cpu(map->pba_of_lba0);
+ blocks = __le32_to_cpu(map->blocks_per_member);
+
+ do {
+ /* check that we can start at pba_of_lba0 with
+ * blocks_per_member of space
+ */
+ esize = ex[j].start - pos;
+ if (array_start >= pos &&
+ array_start + blocks < ex[j].start) {
+ found = 1;
+ break;
+ }
+ pos = ex[j].start + ex[j].size;
+ j++;
+ } while (ex[j-1].size);
+
+ if (!found)
break;
- }
- pos = ex[j].start + ex[j].size;
- j++;
-
- } while (ex[j-1].size);
+ }
free(ex);
- if (!found) {
- dprintf("%x:%x does not have %llu at %d\n",
+ if (i < mpb->num_raid_devs) {
+ dprintf("%x:%x does not have %u at %u\n",
dl->major, dl->minor,
- a->info.component_size,
- __le32_to_cpu(map->pba_of_lba0));
+ blocks, array_start);
/* No room */
continue;
- } else
- break;
+ }
+ return dl;
}
return dl;
continue;
/*
- * OK, this device needs recovery. Try to re-add the previous
- * occupant of this slot, if this fails add a new spare
+ * OK, this device needs recovery. Try to re-add the
+ * previous occupant of this slot, if this fails see if
+ * we can continue the assimilation of a spare that was
+ * partially assimilated, finally try to activate a new
+ * spare.
*/
dl = imsm_readd(super, i, a);
if (!dl)
- dl = imsm_add_spare(super, i, a);
+ dl = imsm_add_spare(super, i, a, 0);
+ if (!dl)
+ dl = imsm_add_spare(super, i, a, 1);
if (!dl)
continue;
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
free(super->next_buf);
super->next_len = buf_len;
- if (posix_memalign(&super->next_buf, 512, buf_len) != 0)
+ if (posix_memalign(&super->next_buf, 512, buf_len) == 0)
+ memset(super->next_buf, 0, buf_len);
+ else
super->next_buf = NULL;
}
}
#ifndef MDASSEMBLE
.examine_super = examine_super_imsm,
.brief_examine_super = brief_examine_super_imsm,
+ .export_examine_super = export_examine_super_imsm,
.detail_super = detail_super_imsm,
.brief_detail_super = brief_detail_super_imsm,
.write_init_super = write_init_super_imsm,
.free_super = free_super_imsm,
.match_metadata_desc = match_metadata_desc_imsm,
.container_content = container_content_imsm,
+ .default_layout = imsm_level_to_layout,
.external = 1,
.name = "imsm",