#define IMSM_MAX_DEVICES 255
struct imsm_disk {
__u8 serial[MAX_RAID_SERIAL_LEN];/* 0xD8 - 0xE7 ascii serial number */
- __u32 total_blocks; /* 0xE8 - 0xEB total blocks */
+ __u32 total_blocks_lo; /* 0xE8 - 0xEB total blocks lo */
__u32 scsi_id; /* 0xEC - 0xEF scsi ID */
#define SPARE_DISK __cpu_to_le32(0x01) /* Spare */
#define CONFIGURED_DISK __cpu_to_le32(0x02) /* Member of some RaidDev */
#define FAILED_DISK __cpu_to_le32(0x04) /* Permanent failure */
__u32 status; /* 0xF0 - 0xF3 */
__u32 owner_cfg_num; /* which config 0,1,2... owns this disk */
-#define IMSM_DISK_FILLERS 4
- __u32 filler[IMSM_DISK_FILLERS]; /* 0xF4 - 0x107 MPB_DISK_FILLERS for future expansion */
+ __u32 total_blocks_hi; /* 0xF4 - 0xF5 total blocks hi */
+#define IMSM_DISK_FILLERS 3
+ __u32 filler[IMSM_DISK_FILLERS]; /* 0xF5 - 0x107 MPB_DISK_FILLERS for future expansion */
};
/* map selector for map managment
/* RAID map configuration infos. */
struct imsm_map {
- __u32 pba_of_lba0; /* start address of partition */
- __u32 blocks_per_member;/* blocks per member */
- __u32 num_data_stripes; /* number of data stripes */
+ __u32 pba_of_lba0_lo; /* start address of partition */
+ __u32 blocks_per_member_lo;/* blocks per member */
+ __u32 num_data_stripes_lo; /* number of data stripes */
__u16 blocks_per_strip;
__u8 map_state; /* Normal, Uninitialized, Degraded, Failed */
#define IMSM_T_STATE_NORMAL 0
__u8 num_domains; /* number of parity domains */
__u8 failed_disk_num; /* valid only when state is degraded */
__u8 ddf;
- __u32 filler[7]; /* expansion area */
+ __u32 pba_of_lba0_hi;
+ __u32 blocks_per_member_hi;
+ __u32 num_data_stripes_hi;
+ __u32 filler[4]; /* expansion area */
#define IMSM_ORD_REBUILD (1 << 24)
__u32 disk_ord_tbl[1]; /* disk_ord_tbl[num_members],
* top byte contains some flags
* (for recovered migrations) */
} __attribute__ ((__packed__));
+struct md_list {
+ /* usage marker:
+ * 1: load metadata
+ * 2: metadata does not match
+ * 4: already checked
+ */
+ int used;
+ char *devname;
+ int found;
+ int container;
+ dev_t st_rdev;
+ struct md_list *next;
+};
+
+#define pr_vrb(fmt, arg...) (void) (verbose && fprintf(stderr, Name fmt, ##arg))
+
static __u8 migr_type(struct imsm_dev *dev)
{
if (dev->vol.migr_type == MIGR_VERIFY &&
static unsigned int sector_count(__u32 bytes)
{
- return ((bytes + (512-1)) & (~(512-1))) / 512;
+ return ROUND_UP(bytes, 512) / 512;
}
static unsigned int mpb_sectors(struct imsm_super *mpb)
void *migr_rec_buf; /* buffer for I/O operations */
struct migr_record *migr_rec; /* migration record */
};
+ int clean_migration_record_by_mdmon; /* when reshape is switched to next
+ array, it indicates that mdmon is allowed to clean migration
+ record */
size_t len; /* size of the 'buf' allocation */
void *next_buf; /* for realloc'ing buf from the manager */
size_t next_len;
int updates_pending; /* count of pending updates for mdmon */
int current_vol; /* index of raid device undergoing creation */
- __u32 create_offset; /* common start for 'current_vol' */
+ unsigned long long create_offset; /* common start for 'current_vol' */
__u32 random; /* random data for seeding new family numbers */
struct intel_dev *devlist;
struct dl {
static __u32 imsm_min_reserved_sectors(struct intel_super *super);
+static int split_ull(unsigned long long n, __u32 *lo, __u32 *hi)
+{
+ if (lo == 0 || hi == 0)
+ return 1;
+ *lo = __le32_to_cpu((unsigned)n);
+ *hi = __le32_to_cpu((unsigned)(n >> 32));
+ return 0;
+}
+
+static unsigned long long join_u32(__u32 lo, __u32 hi)
+{
+ return (unsigned long long)__le32_to_cpu(lo) |
+ (((unsigned long long)__le32_to_cpu(hi)) << 32);
+}
+
+static unsigned long long total_blocks(struct imsm_disk *disk)
+{
+ if (disk == NULL)
+ return 0;
+ return join_u32(disk->total_blocks_lo, disk->total_blocks_hi);
+}
+
+static unsigned long long pba_of_lba0(struct imsm_map *map)
+{
+ if (map == NULL)
+ return 0;
+ return join_u32(map->pba_of_lba0_lo, map->pba_of_lba0_hi);
+}
+
+static unsigned long long blocks_per_member(struct imsm_map *map)
+{
+ if (map == NULL)
+ return 0;
+ return join_u32(map->blocks_per_member_lo, map->blocks_per_member_hi);
+}
+
+static unsigned long long num_data_stripes(struct imsm_map *map)
+{
+ if (map == NULL)
+ return 0;
+ return join_u32(map->num_data_stripes_lo, map->num_data_stripes_hi);
+}
+
+static void set_total_blocks(struct imsm_disk *disk, unsigned long long n)
+{
+ split_ull(n, &disk->total_blocks_lo, &disk->total_blocks_hi);
+}
+
+static void set_pba_of_lba0(struct imsm_map *map, unsigned long long n)
+{
+ split_ull(n, &map->pba_of_lba0_lo, &map->pba_of_lba0_hi);
+}
+
+static void set_blocks_per_member(struct imsm_map *map, unsigned long long n)
+{
+ split_ull(n, &map->blocks_per_member_lo, &map->blocks_per_member_hi);
+}
+
+static void set_num_data_stripes(struct imsm_map *map, unsigned long long n)
+{
+ split_ull(n, &map->num_data_stripes_lo, &map->num_data_stripes_hi);
+}
+
static struct extent *get_extents(struct intel_super *super, struct dl *dl)
{
/* find a list of used extents on the given physical device */
struct imsm_map *map = get_imsm_map(dev, MAP_0);
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->start = pba_of_lba0(map);
+ e->size = blocks_per_member(map);
e++;
}
}
*/
if (memberships) {
struct extent *last = &rv[memberships - 1];
- __u32 remainder;
+ unsigned long long remainder;
- remainder = __le32_to_cpu(dl->disk.total_blocks) -
- (last->start + last->size);
+ remainder = total_blocks(&dl->disk) - (last->start + last->size);
/* round down to 1k block to satisfy precision of the kernel
* 'size' interface
*/
if (reservation > remainder)
reservation = remainder;
}
- e->start = __le32_to_cpu(dl->disk.total_blocks) - reservation;
+ e->start = total_blocks(&dl->disk) - reservation;
e->size = 0;
return rv;
}
for (i = 0; e[i].size; i++)
continue;
- rv = __le32_to_cpu(dl->disk.total_blocks) - e[i].start;
+ rv = total_blocks(&dl->disk) - e[i].start;
free(e);
{
struct extent *e;
int i;
- __u32 min_active, remainder;
+ unsigned long long min_active;
+ __u32 remainder;
__u32 rv = MPB_SECTOR_CNT + IMSM_RESERVED_SECTORS;
struct dl *dl, *dl_min = NULL;
for (dl = super->disks; dl; dl = dl->next) {
if (dl->index < 0)
continue;
- if (dl->disk.total_blocks < min_active || min_active == 0) {
+ unsigned long long blocks = total_blocks(&dl->disk);
+ if (blocks < min_active || min_active == 0) {
dl_min = dl;
- min_active = dl->disk.total_blocks;
+ min_active = blocks;
}
}
if (!dl_min)
sz += __le32_to_cpu(dev->size_low);
printf(" Array Size : %llu%s\n", (unsigned long long)sz,
human_size(sz * 512));
- sz = __le32_to_cpu(map->blocks_per_member);
+ sz = blocks_per_member(map);
printf(" Per Dev Size : %llu%s\n", (unsigned long long)sz,
human_size(sz * 512));
- printf(" Sector Offset : %u\n",
- __le32_to_cpu(map->pba_of_lba0));
- printf(" Num Stripes : %u\n",
- __le32_to_cpu(map->num_data_stripes));
+ printf(" Sector Offset : %llu\n",
+ pba_of_lba0(map));
+ printf(" Num Stripes : %llu\n",
+ num_data_stripes(map));
printf(" Chunk Size : %u KiB",
__le16_to_cpu(map->blocks_per_strip) / 2);
if (map2)
is_configured(disk) ? " active" : "",
is_failed(disk) ? " failed" : "");
printf(" Id : %08x\n", __le32_to_cpu(disk->scsi_id));
- sz = __le32_to_cpu(disk->total_blocks) - reserved;
+ sz = total_blocks(disk) - reserved;
printf(" Usable Size : %llu%s\n", (unsigned long long)sz,
human_size(sz * 512));
}
* Returns:
* 0 : success
* -1 : fail
+ * -2 : no migration in progress
******************************************************************************/
static int load_imsm_migr_rec(struct intel_super *super, struct mdinfo *info)
{
/* nothing to load,no migration in progress?
*/
if (dev == NULL)
- return 0;
+ return -2;
map = get_imsm_map(dev, MAP_0);
if (info) {
dl->index);
}
- info->data_offset = __le32_to_cpu(map_to_analyse->pba_of_lba0);
- info->component_size =
- __le32_to_cpu(map_to_analyse->blocks_per_member);
+ info->data_offset = pba_of_lba0(map_to_analyse);
+ info->component_size = blocks_per_member(map_to_analyse);
/* check component size aligment
*/
used_disks = imsm_num_data_members(dev, MAP_1);
if (used_disks > 0) {
- array_blocks = map->blocks_per_member *
+ array_blocks = blocks_per_member(map) *
used_disks;
/* round array size down to closest MB
*/
enough = 0;
else /* we're normal, or already degraded */
enough = 1;
-
+ if (is_gen_migration(dev) && missing) {
+ /* during general migration we need all disks
+ * that process is running on.
+ * No new missing disk is allowed.
+ */
+ max_enough = -1;
+ enough = -1;
+ /* no more checks necessary
+ */
+ break;
+ }
/* in the missing/failed disk case check to see
* if at least one array is runnable
*/
__u32 reserved = imsm_reserved_sectors(super, super->disks);
disk = &super->disks->disk;
- info->data_offset = __le32_to_cpu(disk->total_blocks) - reserved;
+ info->data_offset = total_blocks(&super->disks->disk) - reserved;
info->component_size = reserved;
info->disk.state = is_configured(disk) ? (1 << MD_DISK_ACTIVE) : 0;
/* we don't change info->disk.raid_disk here because
mpb = super->anchor;
- if (strcmp(update, "uuid") == 0 && uuid_set && !info->update_private)
- rv = -1;
- else if (strcmp(update, "uuid") == 0 && uuid_set && info->update_private) {
- mpb->orig_family_num = *((__u32 *) info->update_private);
- rv = 0;
- } else if (strcmp(update, "uuid") == 0) {
- __u32 *new_family = malloc(sizeof(*new_family));
-
- /* update orig_family_number with the incoming random
- * data, report the new effective uuid, and store the
- * new orig_family_num for future updates.
+ if (strcmp(update, "uuid") == 0) {
+ /* We take this to mean that the family_num should be updated.
+ * However that is much smaller than the uuid so we cannot really
+ * allow an explicit uuid to be given. And it is hard to reliably
+ * know if one was.
+ * So if !uuid_set we know the current uuid is random and just used
+ * the first 'int' and copy it to the other 3 positions.
+ * Otherwise we require the 4 'int's to be the same as would be the
+ * case if we are using a random uuid. So an explicit uuid will be
+ * accepted as long as all for ints are the same... which shouldn't hurt
*/
- if (new_family) {
- memcpy(&mpb->orig_family_num, info->uuid, sizeof(__u32));
- uuid_from_super_imsm(st, info->uuid);
- *new_family = mpb->orig_family_num;
- info->update_private = new_family;
+ if (!uuid_set) {
+ info->uuid[1] = info->uuid[2] = info->uuid[3] = info->uuid[0];
rv = 0;
+ } else {
+ if (info->uuid[0] != info->uuid[1] ||
+ info->uuid[1] != info->uuid[2] ||
+ info->uuid[2] != info->uuid[3])
+ rv = -1;
+ else
+ rv = 0;
}
+ if (rv == 0)
+ mpb->orig_family_num = info->uuid[0];
} else if (strcmp(update, "assemble") == 0)
rv = 0;
else
rv = readlink(path, dname, sizeof(dname)-1);
if (rv <= 0)
return;
-
+
dname[rv] = '\0';
nm = strrchr(dname, '/');
if (nm) {
len_migr = sizeof_imsm_dev(dev_iter, 1);
if (len_migr > len)
space_needed += len_migr - len;
-
+
dv = malloc(sizeof(*dv));
if (!dv)
return 1;
super->buf = buf;
super->len = len;
}
-
+
return 0;
}
/* This device is migrating */
map0 = get_imsm_map(dev_iter, MAP_0);
map1 = get_imsm_map(dev_iter, MAP_1);
- if (map0->pba_of_lba0 != map1->pba_of_lba0)
+ if (pba_of_lba0(map0) != pba_of_lba0(map1))
/* migration optimization area was used */
return -1;
if (migr_rec->ascending_migr == 0
free(super->buf);
return 2;
}
+ super->clean_migration_record_by_mdmon = 0;
if (!sectors) {
check_sum = __gen_imsm_checksum(super->anchor);
if (super) {
memset(super, 0, sizeof(*super));
super->current_vol = -1;
- super->create_offset = ~((__u32 ) 0);
+ super->create_offset = ~((unsigned long long) 0);
}
return super;
}
return champion;
}
+
+static int
+get_sra_super_block(int fd, struct intel_super **super_list, char *devname, int *max, int keep_fd);
+static int get_super_block(struct intel_super **super_list, int devnum, char *devname,
+ int major, int minor, int keep_fd);
+static int
+get_devlist_super_block(struct md_list *devlist, struct intel_super **super_list,
+ int *max, int keep_fd);
+
+
static int load_super_imsm_all(struct supertype *st, int fd, void **sbp,
- char *devname)
+ char *devname, struct md_list *devlist,
+ int keep_fd)
{
- struct mdinfo *sra;
struct intel_super *super_list = NULL;
struct intel_super *super = NULL;
- int devnum = fd2devnum(fd);
- struct mdinfo *sd;
- int retry;
int err = 0;
- int i;
-
- /* check if 'fd' an opened container */
- sra = sysfs_read(fd, 0, GET_LEVEL|GET_VERSION|GET_DEVS|GET_STATE);
- if (!sra)
- return 1;
+ int i = 0;
- if (sra->array.major_version != -1 ||
- sra->array.minor_version != -2 ||
- strcmp(sra->text_version, "imsm") != 0) {
- err = 1;
+ if (fd >= 0)
+ /* 'fd' is an opened container */
+ err = get_sra_super_block(fd, &super_list, devname, &i, keep_fd);
+ else
+ /* get super block from devlist devices */
+ err = get_devlist_super_block(devlist, &super_list, &i, keep_fd);
+ if (err)
goto error;
- }
- /* load all mpbs */
- for (sd = sra->devs, i = 0; sd; sd = sd->next, i++) {
- struct intel_super *s = alloc_super();
- char nm[32];
- int dfd;
- int rv;
-
- err = 1;
- if (!s)
- goto error;
- s->next = super_list;
- super_list = s;
-
- err = 2;
- sprintf(nm, "%d:%d", sd->disk.major, sd->disk.minor);
- dfd = dev_open(nm, O_RDWR);
- if (dfd < 0)
- goto error;
-
- rv = find_intel_hba_capability(dfd, s, devname);
- /* no orom/efi or non-intel hba of the disk */
- if (rv != 0)
- goto error;
-
- err = load_and_parse_mpb(dfd, s, NULL, 1);
-
- /* retry the load if we might have raced against mdmon */
- if (err == 3 && mdmon_running(devnum))
- for (retry = 0; retry < 3; retry++) {
- usleep(3000);
- err = load_and_parse_mpb(dfd, s, NULL, 1);
- if (err != 3)
- break;
- }
- if (err)
- goto error;
- }
-
/* all mpbs enter, maybe one leaves */
super = imsm_thunderdome(&super_list, i);
if (!super) {
/* load migration record */
err = load_imsm_migr_rec(super, NULL);
- if (err) {
+ if (err == -1) {
+ /* migration is in progress,
+ * but migr_rec cannot be loaded,
+ */
err = 4;
goto error;
}
/* Check migration compatibility */
- if (check_mpb_migr_compatibility(super) != 0) {
+ if ((err == 0) && (check_mpb_migr_compatibility(super) != 0)) {
fprintf(stderr, Name ": Unsupported migration detected");
if (devname)
fprintf(stderr, " on %s\n", devname);
super_list = super_list->next;
free_imsm(s);
}
- sysfs_free(sra);
+
if (err)
return err;
*sbp = super;
- st->container_dev = devnum;
+ if (fd >= 0)
+ st->container_dev = fd2devnum(fd);
+ else
+ st->container_dev = NoMdDev;
if (err == 0 && st->ss == NULL) {
st->ss = &super_imsm;
st->minor_version = 0;
return 0;
}
+
+static int
+get_devlist_super_block(struct md_list *devlist, struct intel_super **super_list,
+ int *max, int keep_fd)
+{
+ struct md_list *tmpdev;
+ int err = 0;
+ int i = 0;
+
+ for (i = 0, tmpdev = devlist; tmpdev; tmpdev = tmpdev->next) {
+ if (tmpdev->used != 1)
+ continue;
+ if (tmpdev->container == 1) {
+ int lmax = 0;
+ int fd = dev_open(tmpdev->devname, O_RDONLY|O_EXCL);
+ if (fd < 0) {
+ fprintf(stderr, Name ": cannot open device %s: %s\n",
+ tmpdev->devname, strerror(errno));
+ err = 8;
+ goto error;
+ }
+ err = get_sra_super_block(fd, super_list,
+ tmpdev->devname, &lmax,
+ keep_fd);
+ i += lmax;
+ close(fd);
+ if (err) {
+ err = 7;
+ goto error;
+ }
+ } else {
+ int major = major(tmpdev->st_rdev);
+ int minor = minor(tmpdev->st_rdev);
+ err = get_super_block(super_list,
+ -1,
+ tmpdev->devname,
+ major, minor,
+ keep_fd);
+ i++;
+ if (err) {
+ err = 6;
+ goto error;
+ }
+ }
+ }
+ error:
+ *max = i;
+ return err;
+}
+
+static int get_super_block(struct intel_super **super_list, int devnum, char *devname,
+ int major, int minor, int keep_fd)
+{
+ struct intel_super*s = NULL;
+ char nm[32];
+ int dfd = -1;
+ int rv;
+ int err = 0;
+ int retry;
+
+ s = alloc_super();
+ if (!s) {
+ err = 1;
+ goto error;
+ }
+
+ sprintf(nm, "%d:%d", major, minor);
+ dfd = dev_open(nm, O_RDWR);
+ if (dfd < 0) {
+ err = 2;
+ goto error;
+ }
+
+ rv = find_intel_hba_capability(dfd, s, devname);
+ /* no orom/efi or non-intel hba of the disk */
+ if (rv != 0) {
+ err = 4;
+ goto error;
+ }
+
+ err = load_and_parse_mpb(dfd, s, NULL, keep_fd);
+
+ /* retry the load if we might have raced against mdmon */
+ if (err == 3 && (devnum != -1) && mdmon_running(devnum))
+ for (retry = 0; retry < 3; retry++) {
+ usleep(3000);
+ err = load_and_parse_mpb(dfd, s, NULL, keep_fd);
+ if (err != 3)
+ break;
+ }
+ error:
+ if (!err) {
+ s->next = *super_list;
+ *super_list = s;
+ } else {
+ if (s)
+ free(s);
+ if (dfd)
+ close(dfd);
+ }
+ if ((dfd >= 0) && (!keep_fd))
+ close(dfd);
+ return err;
+
+}
+
+static int
+get_sra_super_block(int fd, struct intel_super **super_list, char *devname, int *max, int keep_fd)
+{
+ struct mdinfo *sra;
+ int devnum;
+ struct mdinfo *sd;
+ int err = 0;
+ int i = 0;
+ sra = sysfs_read(fd, 0, GET_LEVEL|GET_VERSION|GET_DEVS|GET_STATE);
+ if (!sra)
+ return 1;
+
+ if (sra->array.major_version != -1 ||
+ sra->array.minor_version != -2 ||
+ strcmp(sra->text_version, "imsm") != 0) {
+ err = 1;
+ goto error;
+ }
+ /* load all mpbs */
+ devnum = fd2devnum(fd);
+ for (sd = sra->devs, i = 0; sd; sd = sd->next, i++) {
+ if (get_super_block(super_list, devnum, devname,
+ sd->disk.major, sd->disk.minor, keep_fd) != 0) {
+ err = 7;
+ goto error;
+ }
+ }
+ error:
+ sysfs_free(sra);
+ *max = i;
+ return err;
+}
+
static int load_container_imsm(struct supertype *st, int fd, char *devname)
{
- return load_super_imsm_all(st, fd, &st->sb, devname);
+ return load_super_imsm_all(st, fd, &st->sb, devname, NULL, 1);
}
#endif
return info->chunk_size >> 9;
}
-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);
- num_stripes /= num_domains;
-
- return num_stripes;
-}
-
-static __u32 info_to_blocks_per_member(mdu_array_info_t *info)
+static unsigned long long info_to_blocks_per_member(mdu_array_info_t *info,
+ unsigned long long size)
{
if (info->level == 1)
- return info->size * 2;
+ return size * 2;
else
- return (info->size * 2) & ~(info_to_blocks_per_strip(info) - 1);
+ return (size * 2) & ~(info_to_blocks_per_strip(info) - 1);
}
static void imsm_update_version_info(struct intel_super *super)
int i;
unsigned long long array_blocks;
size_t size_old, size_new;
- __u32 num_data_stripes;
+ unsigned long long num_data_stripes;
if (super->orom && mpb->num_raid_devs >= super->orom->vpa) {
fprintf(stderr, Name": This imsm-container already has the "
strncpy((char *) dev->volume, name, MAX_RAID_SERIAL_LEN);
if (info->level == 1)
- array_blocks = info_to_blocks_per_member(info);
+ array_blocks = info_to_blocks_per_member(info, size);
else
array_blocks = calc_array_size(info->level, info->raid_disks,
info->layout, info->chunk_size,
- info->size*2);
+ size * 2);
/* round array size down to closest MB */
array_blocks = (array_blocks >> SECT_PER_MB_SHIFT) << SECT_PER_MB_SHIFT;
vol->dirty = !info->state;
vol->curr_migr_unit = 0;
map = get_imsm_map(dev, MAP_0);
- map->pba_of_lba0 = __cpu_to_le32(super->create_offset);
- map->blocks_per_member = __cpu_to_le32(info_to_blocks_per_member(info));
+ set_pba_of_lba0(map, super->create_offset);
+ set_blocks_per_member(map, info_to_blocks_per_member(info, size));
map->blocks_per_strip = __cpu_to_le16(info_to_blocks_per_strip(info));
map->failed_disk_num = ~0;
if (info->level > 0)
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);
+ /* info->size is only int so use the 'size' parameter instead */
+ num_data_stripes = (size * 2) / info_to_blocks_per_strip(info);
+ num_data_stripes /= map->num_domains;
+ set_num_data_stripes(map, num_data_stripes);
map->num_members = info->raid_disks;
for (i = 0; i < map->num_members; i++) {
*/
if (super->current_vol == 0) {
for (df = super->missing; df; df = df->next) {
- if (dl->disk.total_blocks > df->disk.total_blocks)
- df->disk.total_blocks = dl->disk.total_blocks;
+ if (total_blocks(&dl->disk) > total_blocks(&df->disk))
+ set_total_blocks(&df->disk, total_blocks(&dl->disk));
_disk = __get_imsm_disk(mpb, df->index);
*_disk = df->disk;
}
get_dev_size(fd, NULL, &size);
size /= 512;
serialcpy(dd->disk.serial, dd->serial);
- dd->disk.total_blocks = __cpu_to_le32(size);
+ set_total_blocks(&dd->disk, size);
+ if (__le32_to_cpu(dd->disk.total_blocks_hi) > 0) {
+ struct imsm_super *mpb = super->anchor;
+ mpb->attributes |= MPB_ATTRIB_2TB_DISK;
+ }
mark_spare(dd);
if (sysfs_disk_to_scsi_id(fd, &id) == 0)
dd->disk.scsi_id = __cpu_to_le32(id);
sum = __gen_imsm_checksum(mpb);
mpb->check_sum = __cpu_to_le32(sum);
+ if (super->clean_migration_record_by_mdmon) {
+ clear_migration_record = 1;
+ super->clean_migration_record_by_mdmon = 0;
+ }
if (clear_migration_record)
memset(super->migr_rec_buf, 0, MIGR_REC_BUF_SIZE);
for (d = super->disks; d ; d = d->next) {
if (d->index < 0 || is_failed(&d->disk))
continue;
- if (store_imsm_mpb(d->fd, mpb))
- fprintf(stderr, "%s: failed for device %d:%d %s\n",
- __func__, d->major, d->minor, strerror(errno));
+
if (clear_migration_record) {
unsigned long long dsize;
perror("Write migr_rec failed");
}
}
+
+ if (store_imsm_mpb(d->fd, mpb))
+ fprintf(stderr,
+ "%s: failed for device %d:%d (fd: %d)%s\n",
+ __func__, d->major, d->minor,
+ d->fd, strerror(errno));
+
if (doclose) {
close(d->fd);
d->fd = -1;
if (maxsize < reserve)
return 0;
- super->create_offset = ~((__u32) 0);
+ super->create_offset = ~((unsigned long long) 0);
if (start + reserve > super->create_offset)
return 0; /* start overflows create_offset */
super->create_offset = start + reserve;
return 0;
}
+
+static int
+active_arrays_by_format(char *name, char* hba, struct md_list **devlist,
+ int dpa, int verbose)
+{
+ struct mdstat_ent *mdstat = mdstat_read(0, 0);
+ struct mdstat_ent *memb = NULL;
+ int count = 0;
+ int num = 0;
+ struct md_list *dv = NULL;
+ int found;
+
+ for (memb = mdstat ; memb ; memb = memb->next) {
+ if (memb->metadata_version &&
+ (strncmp(memb->metadata_version, "external:", 9) == 0) &&
+ (strcmp(&memb->metadata_version[9], name) == 0) &&
+ !is_subarray(memb->metadata_version+9) &&
+ memb->members) {
+ struct dev_member *dev = memb->members;
+ int fd = -1;
+ while(dev && (fd < 0)) {
+ char *path = malloc(strlen(dev->name) + strlen("/dev/") + 1);
+ if (path) {
+ num = sprintf(path, "%s%s", "/dev/", dev->name);
+ if (num > 0)
+ fd = open(path, O_RDONLY, 0);
+ if ((num <= 0) || (fd < 0)) {
+ pr_vrb(": Cannot open %s: %s\n",
+ dev->name, strerror(errno));
+ }
+ free(path);
+ }
+ dev = dev->next;
+ }
+ found = 0;
+ if ((fd >= 0) && disk_attached_to_hba(fd, hba)) {
+ struct mdstat_ent *vol;
+ for (vol = mdstat ; vol ; vol = vol->next) {
+ if ((vol->active > 0) &&
+ vol->metadata_version &&
+ is_container_member(vol, memb->dev)) {
+ found++;
+ count++;
+ }
+ }
+ if (*devlist && (found < dpa)) {
+ dv = calloc(1, sizeof(*dv));
+ if (dv == NULL)
+ fprintf(stderr, Name ": calloc failed\n");
+ else {
+ dv->devname = malloc(strlen(memb->dev) + strlen("/dev/") + 1);
+ if (dv->devname != NULL) {
+ sprintf(dv->devname, "%s%s", "/dev/", memb->dev);
+ dv->found = found;
+ dv->used = 0;
+ dv->next = *devlist;
+ *devlist = dv;
+ } else
+ free(dv);
+ }
+ }
+ }
+ if (fd >= 0)
+ close(fd);
+ }
+ }
+ free_mdstat(mdstat);
+ return count;
+}
+
+#ifdef DEBUG_LOOP
+static struct md_list*
+get_loop_devices(void)
+{
+ int i;
+ struct md_list *devlist = NULL;
+ struct md_list *dv = NULL;
+
+ for(i = 0; i < 12; i++) {
+ dv = calloc(1, sizeof(*dv));
+ if (dv == NULL) {
+ fprintf(stderr, Name ": calloc failed\n");
+ break;
+ }
+ dv->devname = malloc(40);
+ if (dv->devname == NULL) {
+ fprintf(stderr, Name ": malloc failed\n");
+ free(dv);
+ break;
+ }
+ sprintf(dv->devname, "/dev/loop%d", i);
+ dv->next = devlist;
+ devlist = dv;
+ }
+ return devlist;
+}
+#endif
+
+static struct md_list*
+get_devices(const char *hba_path)
+{
+ struct md_list *devlist = NULL;
+ struct md_list *dv = NULL;
+ struct dirent *ent;
+ DIR *dir;
+ int err = 0;
+
+#if DEBUG_LOOP
+ devlist = get_loop_devices();
+ return devlist;
+#endif
+ /* scroll through /sys/dev/block looking for devices attached to
+ * this hba
+ */
+ dir = opendir("/sys/dev/block");
+ for (ent = dir ? readdir(dir) : NULL; ent; ent = readdir(dir)) {
+ int fd;
+ char buf[1024];
+ int major, minor;
+ char *path = NULL;
+ if (sscanf(ent->d_name, "%d:%d", &major, &minor) != 2)
+ continue;
+ path = devt_to_devpath(makedev(major, minor));
+ if (!path)
+ continue;
+ if (!path_attached_to_hba(path, hba_path)) {
+ free(path);
+ path = NULL;
+ continue;
+ }
+ free(path);
+ path = NULL;
+ fd = dev_open(ent->d_name, O_RDONLY);
+ if (fd >= 0) {
+ fd2devname(fd, buf);
+ close(fd);
+ } else {
+ fprintf(stderr, Name ": cannot open device: %s\n",
+ ent->d_name);
+ continue;
+ }
+
+
+ dv = calloc(1, sizeof(*dv));
+ if (dv == NULL) {
+ fprintf(stderr, Name ": malloc failed\n");
+ err = 1;
+ break;
+ }
+ dv->devname = strdup(buf);
+ if (dv->devname == NULL) {
+ fprintf(stderr, Name ": malloc failed\n");
+ err = 1;
+ free(dv);
+ break;
+ }
+ dv->next = devlist;
+ devlist = dv;
+ }
+ if (err) {
+ while(devlist) {
+ dv = devlist;
+ devlist = devlist->next;
+ free(dv->devname);
+ free(dv);
+ }
+ }
+ return devlist;
+}
+
+static int
+count_volumes_list(struct md_list *devlist, char *homehost,
+ int verbose, int *found)
+{
+ struct md_list *tmpdev;
+ int count = 0;
+ struct supertype *st = NULL;
+
+ /* first walk the list of devices to find a consistent set
+ * that match the criterea, if that is possible.
+ * We flag the ones we like with 'used'.
+ */
+ *found = 0;
+ st = match_metadata_desc_imsm("imsm");
+ if (st == NULL) {
+ pr_vrb(": cannot allocate memory for imsm supertype\n");
+ return 0;
+ }
+
+ for (tmpdev = devlist; tmpdev; tmpdev = tmpdev->next) {
+ char *devname = tmpdev->devname;
+ struct stat stb;
+ struct supertype *tst;
+ int dfd;
+ if (tmpdev->used > 1)
+ continue;
+ tst = dup_super(st);
+ if (tst == NULL) {
+ pr_vrb(": cannot allocate memory for imsm supertype\n");
+ goto err_1;
+ }
+ tmpdev->container = 0;
+ dfd = dev_open(devname, O_RDONLY|O_EXCL);
+ if (dfd < 0) {
+ dprintf(": cannot open device %s: %s\n",
+ devname, strerror(errno));
+ tmpdev->used = 2;
+ } else if (fstat(dfd, &stb)< 0) {
+ /* Impossible! */
+ dprintf(": fstat failed for %s: %s\n",
+ devname, strerror(errno));
+ tmpdev->used = 2;
+ } else if ((stb.st_mode & S_IFMT) != S_IFBLK) {
+ dprintf(": %s is not a block device.\n",
+ devname);
+ tmpdev->used = 2;
+ } else if (must_be_container(dfd)) {
+ struct supertype *cst;
+ cst = super_by_fd(dfd, NULL);
+ if (cst == NULL) {
+ dprintf(": cannot recognize container type %s\n",
+ devname);
+ tmpdev->used = 2;
+ } else if (tst->ss != st->ss) {
+ dprintf(": non-imsm container - ignore it: %s\n",
+ devname);
+ tmpdev->used = 2;
+ } else if (!tst->ss->load_container ||
+ tst->ss->load_container(tst, dfd, NULL))
+ tmpdev->used = 2;
+ else {
+ tmpdev->container = 1;
+ }
+ if (cst)
+ cst->ss->free_super(cst);
+ } else {
+ tmpdev->st_rdev = stb.st_rdev;
+ if (tst->ss->load_super(tst,dfd, NULL)) {
+ dprintf(": no RAID superblock on %s\n",
+ devname);
+ tmpdev->used = 2;
+ } else if (tst->ss->compare_super == NULL) {
+ dprintf(": Cannot assemble %s metadata on %s\n",
+ tst->ss->name, devname);
+ tmpdev->used = 2;
+ }
+ }
+ if (dfd >= 0)
+ close(dfd);
+ if (tmpdev->used == 2 || tmpdev->used == 4) {
+ /* Ignore unrecognised devices during auto-assembly */
+ goto loop;
+ }
+ else {
+ struct mdinfo info;
+ tst->ss->getinfo_super(tst, &info, NULL);
+
+ if (st->minor_version == -1)
+ st->minor_version = tst->minor_version;
+
+ if (memcmp(info.uuid, uuid_zero,
+ sizeof(int[4])) == 0) {
+ /* this is a floating spare. It cannot define
+ * an array unless there are no more arrays of
+ * this type to be found. It can be included
+ * in an array of this type though.
+ */
+ tmpdev->used = 3;
+ goto loop;
+ }
+
+ if (st->ss != tst->ss ||
+ st->minor_version != tst->minor_version ||
+ st->ss->compare_super(st, tst) != 0) {
+ /* Some mismatch. If exactly one array matches this host,
+ * we can resolve on that one.
+ * Or, if we are auto assembling, we just ignore the second
+ * for now.
+ */
+ dprintf(": superblock on %s doesn't match others - assembly aborted\n",
+ devname);
+ goto loop;
+ }
+ tmpdev->used = 1;
+ *found = 1;
+ dprintf("found: devname: %s\n", devname);
+ }
+ loop:
+ if (tst)
+ tst->ss->free_super(tst);
+ }
+ if (*found != 0) {
+ int err;
+ if ((err = load_super_imsm_all(st, -1, &st->sb, NULL, devlist, 0)) == 0) {
+ struct mdinfo *iter, *head = st->ss->container_content(st, NULL);
+ for (iter = head; iter; iter = iter->next) {
+ dprintf("content->text_version: %s vol\n",
+ iter->text_version);
+ if (iter->array.state & (1<<MD_SB_BLOCK_VOLUME)) {
+ /* do not assemble arrays with unsupported
+ configurations */
+ dprintf(": Cannot activate member %s.\n",
+ iter->text_version);
+ } else
+ count++;
+ }
+ sysfs_free(head);
+
+ } else {
+ dprintf(" no valid super block on device list: err: %d %p\n",
+ err, st->sb);
+ }
+ } else {
+ dprintf(" no more devices to examin\n");
+ }
+
+ for (tmpdev = devlist; tmpdev; tmpdev = tmpdev->next) {
+ if ((tmpdev->used == 1) && (tmpdev->found)) {
+ if (count) {
+ if (count < tmpdev->found)
+ count = 0;
+ else
+ count -= tmpdev->found;
+ }
+ }
+ if (tmpdev->used == 1)
+ tmpdev->used = 4;
+ }
+ err_1:
+ if (st)
+ st->ss->free_super(st);
+ return count;
+}
+
+
+static int
+count_volumes(char *hba, int dpa, int verbose)
+{
+ struct md_list *devlist = NULL;
+ int count = 0;
+ int found = 0;;
+
+ devlist = get_devices(hba);
+ /* if no intel devices return zero volumes */
+ if (devlist == NULL)
+ return 0;
+
+ count = active_arrays_by_format("imsm", hba, &devlist, dpa, verbose);
+ dprintf(" path: %s active arrays: %d\n", hba, count);
+ if (devlist == NULL)
+ return 0;
+ do {
+ found = 0;
+ count += count_volumes_list(devlist,
+ NULL,
+ verbose,
+ &found);
+ dprintf("found %d count: %d\n", found, count);
+ } while (found);
+
+ dprintf("path: %s total number of volumes: %d\n", hba, count);
+
+ while(devlist) {
+ struct md_list *dv = devlist;
+ devlist = devlist->next;
+ free(dv->devname);
+ free(dv);
+ }
+ return count;
+}
+
static int imsm_default_chunk(const struct imsm_orom *orom)
{
/* up to 512 if the plaform supports it, otherwise the platform max.
return min(512, (1 << fs));
}
-#define pr_vrb(fmt, arg...) (void) (verbose && fprintf(stderr, Name fmt, ##arg))
static int
validate_geometry_imsm_orom(struct intel_super *super, int level, int layout,
int raiddisks, int *chunk, int verbose)
mpb = super->anchor;
- if (mpb->num_raid_devs > 0 && mpb->num_disks != raiddisks) {
- fprintf(stderr, Name ": the option-rom requires all "
- "member disks to be a member of all volumes.\n");
- return 0;
- }
-
if (!validate_geometry_imsm_orom(super, level, layout, raiddisks, chunk, verbose)) {
fprintf(stderr, Name ": RAID gemetry validation failed. "
"Cannot proceed with the action(s).\n");
fprintf(stderr, Name ": The option-rom requires all member"
" disks to be a member of all volumes\n");
return 0;
+ } else if (super->orom && mpb->num_raid_devs > 0 &&
+ mpb->num_disks != raiddisks) {
+ 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 */
}
if (maxsize < size || maxsize == 0) {
- if (verbose)
- fprintf(stderr, Name ": not enough space after merge (%llu < %llu)\n",
- maxsize, size);
+ if (verbose) {
+ if (maxsize == 0)
+ fprintf(stderr, Name ": no free space"
+ " left on device. Aborting...\n");
+ else
+ fprintf(stderr, Name ": not enough space"
+ " to create volume of given size"
+ " (%llu < %llu). Aborting...\n",
+ maxsize, size);
+ }
return 0;
}
*freesize = maxsize;
+ if (super->orom) {
+ int count = count_volumes(super->hba->path,
+ super->orom->dpa, verbose);
+ if (super->orom->vphba <= count) {
+ pr_vrb(": platform does not support more then %d raid volumes.\n",
+ super->orom->vphba);
+ return 0;
+ }
+ }
return 1;
}
size /= 2 * chunk;
size *= 2 * chunk;
}
+ maxsize = size;
+ }
+ if (!check_env("IMSM_NO_PLATFORM") &&
+ mpb->num_raid_devs > 0 && size && size != maxsize) {
+ fprintf(stderr, Name ": attempting to create a second "
+ "volume with size less then remaining space. "
+ "Aborting...\n");
+ return 0;
}
-
cnt = 0;
for (dl = super->disks; dl; dl = dl->next)
if (dl->e)
dev, freesize,
verbose);
}
-
+
if (!dev) {
if (st->sb) {
+ struct intel_super *super = st->sb;
if (!validate_geometry_imsm_orom(st->sb, level, layout,
raiddisks, chunk,
verbose))
* created. add_to_super and getinfo_super
* detect when autolayout is in progress.
*/
+ /* assuming that freesize is always given when array is
+ created */
+ if (super->orom && freesize) {
+ int count;
+ count = count_volumes(super->hba->path,
+ super->orom->dpa, verbose);
+ if (super->orom->vphba <= count) {
+ pr_vrb(": platform does not support more"
+ "then %d raid volumes.\n",
+ super->orom->vphba);
+ return 0;
+ }
+ }
if (freesize)
return reserve_space(st, raiddisks, size,
chunk?*chunk:0, freesize);
*/
struct intel_super *super;
- if (load_super_imsm_all(st, cfd, (void **) &super, NULL) == 0) {
+ if (load_super_imsm_all(st, cfd, (void **) &super, NULL, NULL, 1) == 0) {
st->sb = super;
st->container_dev = fd2devnum(cfd);
close(cfd);
this->array.working_disks++;
info_d->events = __le32_to_cpu(mpb->generation_num);
- info_d->data_offset = __le32_to_cpu(map->pba_of_lba0);
- info_d->component_size = __le32_to_cpu(map->blocks_per_member);
+ info_d->data_offset = pba_of_lba0(map);
+ info_d->component_size = blocks_per_member(map);
}
/* now that the disk list is up-to-date fixup recovery_start */
update_recovery_start(super, dev, this);
{
struct intel_super *super = c->sb;
struct imsm_super *mpb = super->anchor;
-
+
if (atoi(inst) >= mpb->num_raid_devs) {
fprintf(stderr, "%s: subarry index %d, out of range\n",
__func__, atoi(inst));
/* set array size in metadata
*/
map = get_imsm_map(dev, MAP_0);
- array_blocks = map->blocks_per_member * used_disks;
+ array_blocks = blocks_per_member(map) * used_disks;
/* round array size down to closest MB
*/
map2->num_members = prev_num_members;
imsm_set_array_size(dev);
+ super->clean_migration_record_by_mdmon = 1;
super->updates_pending++;
}
}
used_disks = imsm_num_data_members(dev, MAP_0);
if (used_disks > 0) {
array_blocks =
- map->blocks_per_member *
+ blocks_per_member(map) *
used_disks;
/* round array size down to closest MB
*/
found = 0;
j = 0;
pos = 0;
- array_start = __le32_to_cpu(map->pba_of_lba0);
+ array_start = pba_of_lba0(map);
array_end = array_start +
- __le32_to_cpu(map->blocks_per_member) - 1;
+ blocks_per_member(map) - 1;
do {
/* check that we can start at pba_of_lba0 with
di->disk.minor = dl->minor;
di->disk.state = 0;
di->recovery_start = 0;
- di->data_offset = __le32_to_cpu(map->pba_of_lba0);
+ di->data_offset = pba_of_lba0(map);
di->component_size = a->info.component_size;
di->container_member = inst;
super->random = random32();
}
return NULL;
}
-
+
mu->space = NULL;
mu->space_list = NULL;
mu->len = sizeof(struct imsm_update_activate_spare) * num_spares;
}
static int apply_update_activate_spare(struct imsm_update_activate_spare *u,
- struct intel_super *super,
+ struct intel_super *super,
struct active_array *active_array)
{
struct imsm_super *mpb = super->anchor;
}
new_map = get_imsm_map(&u->dev, MAP_0);
- new_start = __le32_to_cpu(new_map->pba_of_lba0);
- new_end = new_start + __le32_to_cpu(new_map->blocks_per_member);
+ new_start = pba_of_lba0(new_map);
+ new_end = new_start + blocks_per_member(new_map);
inf = get_disk_info(u);
/* handle activate_spare versus create race:
for (i = 0; i < mpb->num_raid_devs; i++) {
dev = get_imsm_dev(super, i);
map = get_imsm_map(dev, MAP_0);
- start = __le32_to_cpu(map->pba_of_lba0);
- end = start + __le32_to_cpu(map->blocks_per_member);
+ start = pba_of_lba0(map);
+ end = start + blocks_per_member(map);
if ((new_start >= start && new_start <= end) ||
(start >= new_start && start <= new_end))
/* overlap */;
struct imsm_dev *dev)
{
switch (level) {
+ case 1:
case 10:{
int ret_val = 0;
struct imsm_map *map;
migr_rec->dest_depth_per_unit = GEN_MIGR_AREA_SIZE /
max(map_dest->blocks_per_strip, map_src->blocks_per_strip);
- migr_rec->dest_depth_per_unit *= map_dest->blocks_per_strip;
+ migr_rec->dest_depth_per_unit *=
+ max(map_dest->blocks_per_strip, map_src->blocks_per_strip);
new_data_disks = imsm_num_data_members(dev, MAP_0);
migr_rec->blocks_per_unit =
__cpu_to_le32(migr_rec->dest_depth_per_unit * new_data_disks);
write_offset = ((unsigned long long)
__le32_to_cpu(migr_rec->dest_1st_member_lba) +
- __le32_to_cpu(map_dest->pba_of_lba0)) * 512;
+ pba_of_lba0(map_dest)) * 512;
unit_len = __le32_to_cpu(migr_rec->dest_depth_per_unit) * 512;
if (posix_memalign((void **)&buf, 512, unit_len) != 0)
int change = -1;
int check_devs = 0;
int chunk;
- int devNumChange=0;
- int layout = -1;
+ /* number of added/removed disks in operation result */
+ int devNumChange = 0;
+ /* imsm compatible layout value for array geometry verification */
+ int imsm_layout = -1;
getinfo_super_imsm_volume(st, &info, NULL);
if ((geo->level != info.array.level) &&
change = -1;
goto analyse_change_exit;
}
- layout = geo->layout;
+ imsm_layout = geo->layout;
check_devs = 1;
devNumChange = 1; /* parity disk added */
} else if (geo->level == 10) {
change = CH_TAKEOVER;
check_devs = 1;
devNumChange = 2; /* two mirrors added */
- layout = 0x102; /* imsm supported layout */
+ imsm_layout = 0x102; /* imsm supported layout */
}
break;
case 1:
change = CH_TAKEOVER;
check_devs = 1;
devNumChange = -(geo->raid_disks/2);
- layout = 0; /* imsm raid0 layout */
+ imsm_layout = 0; /* imsm raid0 layout */
}
break;
}
change = -1;
goto analyse_change_exit;
}
- } else
+ } else {
geo->layout = info.array.layout;
+ if (imsm_layout == -1)
+ imsm_layout = info.array.layout;
+ }
if ((geo->chunksize > 0) && (geo->chunksize != UnSet)
&& (geo->chunksize != info.array.chunk_size))
chunk = geo->chunksize / 1024;
if (!validate_geometry_imsm(st,
geo->level,
- layout,
+ imsm_layout,
geo->raid_disks + devNumChange,
&chunk,
geo->size,
"are present in copy area.\n");
goto abort;
}
+ /* Save checkpoint to update migration record for current
+ * reshape position (in md). It can be farther than current
+ * reshape position in metadata.
+ */
+ if (save_checkpoint_imsm(st, sra, UNIT_SRC_NORMAL) == 1) {
+ /* ignore error == 2, this can mean end of reshape here
+ */
+ dprintf("imsm: Cannot write checkpoint to "
+ "migration record (UNIT_SRC_NORMAL, "
+ "initial save)\n");
+ goto abort;
+ }
}
/* size for data */