__u32 delta_disks; /* change in number of raid_disks */
__u32 new_layout; /* new layout */
__u32 new_chunk; /* new chunk size (bytes) */
- __u8 pad1[128-124]; /* set to 0 when written */
+ __u32 new_offset; /* signed number to add to data_offset in new
+ * layout. 0 == no-change. This can be
+ * different on each device in the array.
+ */
/* constant this-device information - 64 bytes */
__u64 data_offset; /* sector start of data, often 0 */
__u32 dev_number; /* permanent identifier of this device - not role in raid */
__u32 cnt_corrected_read; /* number of read errors that were corrected by re-writing */
__u8 device_uuid[16]; /* user-space setable, ignored by kernel */
- __u8 devflags; /* per-device flags. Only one defined...*/
+ __u8 devflags; /* per-device flags. Only one defined...*/
#define WriteMostly1 1 /* mask for writemostly flag in above */
/* bad block log. If there are any bad blocks the feature flag is set.
* if offset and size are non-zero, that space is reserved and available.
*/
#define MD_FEATURE_RESHAPE_ACTIVE 4
#define MD_FEATURE_BAD_BLOCKS 8 /* badblock list is not empty */
-
-#define MD_FEATURE_ALL (1|2|4|8)
+#define MD_FEATURE_REPLACEMENT 16 /* This device is replacing an
+ * active device with same 'role'.
+ * 'recovery_offset' is also set.
+ */
+#define MD_FEATURE_RESHAPE_BACKWARDS 32 /* Reshape doesn't change number
+ * of devices, but is going
+ * backwards anyway.
+ */
+#define MD_FEATURE_NEW_OFFSET 64 /* new_offset must be honoured */
+#define MD_FEATURE_ALL (MD_FEATURE_BITMAP_OFFSET \
+ |MD_FEATURE_RECOVERY_OFFSET \
+ |MD_FEATURE_RESHAPE_ACTIVE \
+ |MD_FEATURE_BAD_BLOCKS \
+ |MD_FEATURE_REPLACEMENT \
+ |MD_FEATURE_RESHAPE_BACKWARDS \
+ |MD_FEATURE_NEW_OFFSET \
+ )
#ifndef offsetof
#define offsetof(t,f) ((size_t)&(((t*)0)->f))
if (sb->data_offset)
printf(" Data Offset : %llu sectors\n",
(unsigned long long)__le64_to_cpu(sb->data_offset));
+ if (sb->new_offset) {
+ unsigned long long offset = __le64_to_cpu(sb->data_offset);
+ offset += (signed)(int32_t)__le32_to_cpu(sb->new_offset);
+ printf(" New Offset : %llu sectors\n", offset);
+ }
printf(" Super Offset : %llu sectors\n",
(unsigned long long)__le64_to_cpu(sb->super_offset));
if (__le32_to_cpu(sb->feature_map) & MD_FEATURE_RECOVERY_OFFSET)
role = 0xFFFF;
if (role >= 0xFFFE)
printf("spare\n");
+ else if (sb->feature_map & __cpu_to_le32(MD_FEATURE_REPLACEMENT))
+ printf("Replacement device %d\n", role);
else
printf("Active device %d\n", role);
if (role == d)
cnt++;
}
- if (cnt > 1) printf("?");
- else if (cnt == 1) printf("A");
- else printf (".");
+ if (cnt == 2)
+ printf("R");
+ else if (cnt == 1)
+ printf("A");
+ else if (cnt == 0)
+ printf(".");
+ else
+ printf("?");
}
#if 0
/* This is confusing too */
}
if (faulty) printf(" %d failed", faulty);
#endif
- printf(" ('A' == active, '.' == missing)");
+ printf(" ('A' == active, '.' == missing, 'R' == replacing)");
printf("\n");
}
printf("MD_NAME=%.*s\n", len, sb->set_name);
}
+static int examine_badblocks_super1(struct supertype *st, int fd, char *devname)
+{
+ struct mdp_superblock_1 *sb = st->sb;
+ unsigned long long offset;
+ int size;
+ __u64 *bbl, *bbp;
+ int i;
+
+ if (!sb->bblog_size || __le32_to_cpu(sb->bblog_size) > 100
+ || !sb->bblog_offset){
+ printf("No bad-blocks list configured on %s\n", devname);
+ return 0;
+ }
+ if ((sb->feature_map & __cpu_to_le32(MD_FEATURE_BAD_BLOCKS))
+ == 0) {
+ printf("Bad-blocks list is empty in %s\n", devname);
+ return 0;
+ }
+
+ size = __le32_to_cpu(sb->bblog_size)* 512;
+ posix_memalign((void**)&bbl, 4096, size);
+ offset = __le64_to_cpu(sb->super_offset) +
+ (int)__le32_to_cpu(sb->bblog_offset);
+ offset <<= 9;
+ if (lseek64(fd, offset, 0) < 0) {
+ pr_err("Cannot seek to bad-blocks list\n");
+ return 1;
+ }
+ if (read(fd, bbl, size) != size) {
+ pr_err("Cannot read bad-blocks list\n");
+ return 1;
+ }
+ /* 64bits per entry. 10 bits is block-count, 54 bits is block
+ * offset. Blocks are sectors unless bblog->shift makes them bigger
+ */
+ bbp = (__u64*)bbl;
+ printf("Bad-blocks on %s:\n", devname);
+ for (i = 0; i < size/8; i++, bbp++) {
+ __u64 bb = __le64_to_cpu(*bbp);
+ int count = bb & 0x3ff;
+ unsigned long long sector = bb >> 10;
+
+ if (bb + 1 == 0)
+ break;
+
+ sector <<= sb->bblog_shift;
+ count <<= sb->bblog_shift;
+
+ printf("%20llu for %d sectors\n", sector, count);
+ }
+ return 0;
+}
+
#endif
static int match_home1(struct supertype *st, char *homehost)
static void getinfo_super1(struct supertype *st, struct mdinfo *info, char *map)
{
struct mdp_superblock_1 *sb = st->sb;
+ struct bitmap_super_s *bsb = (void*)(((char*)sb)+MAX_SB_SIZE);
+ struct misc_dev_info *misc = (void*)(((char*)sb)+MAX_SB_SIZE+BM_SUPER_SIZE);
int working = 0;
unsigned int i;
unsigned int role;
unsigned int map_disks = info->array.raid_disks;
+ unsigned long long super_offset;
+ unsigned long long data_size;
memset(info, 0, sizeof(*info));
info->array.major_version = 1;
else
role = __le16_to_cpu(sb->dev_roles[__le32_to_cpu(sb->dev_number)]);
+ super_offset = __le64_to_cpu(sb->super_offset);
+ data_size = __le64_to_cpu(sb->size);
+ if (info->data_offset < super_offset) {
+ unsigned long long end;
+ info->space_before = info->data_offset;
+ end = super_offset;
+ if (info->bitmap_offset < 0)
+ end += info->bitmap_offset;
+ if (info->data_offset + data_size < end)
+ info->space_after = end - data_size - info->data_offset;
+ else
+ info->space_after = 0;
+ } else {
+ info->space_before = (info->data_offset -
+ super_offset);
+ if (info->bitmap_offset > 0) {
+ unsigned long long bmend = info->bitmap_offset;
+ unsigned long long size = __le64_to_cpu(bsb->sync_size);
+ size /= __le32_to_cpu(bsb->chunksize) >> 9;
+ size = (size + 7) >> 3;
+ size += sizeof(bitmap_super_t);
+ size = ROUND_UP(size, 4096);
+ size /= 512;
+ size += bmend;
+ if (size < info->space_before)
+ info->space_before -= size;
+ else
+ info->space_before = 0;
+ } else
+ info->space_before -= 8; /* superblock */
+ info->space_after = misc->device_size - data_size - info->data_offset;
+ }
+
info->disk.raid_disk = -1;
switch(role) {
case 0xFFFF:
strncpy(info->name, sb->set_name, 32);
info->name[32] = 0;
+ if ((__le32_to_cpu(sb->feature_map)&MD_FEATURE_REPLACEMENT)) {
+ info->disk.state &= ~(1 << MD_DISK_SYNC);
+ info->disk.state |= 1 << MD_DISK_REPLACEMENT;
+ }
+
+
if (sb->feature_map & __le32_to_cpu(MD_FEATURE_RECOVERY_OFFSET))
info->recovery_start = __le32_to_cpu(sb->recovery_offset);
else
if (sb->feature_map & __le32_to_cpu(MD_FEATURE_RESHAPE_ACTIVE)) {
info->reshape_active = 1;
+ if (info->array.level == 10)
+ info->reshape_active |= RESHAPE_NO_BACKUP;
info->reshape_progress = __le64_to_cpu(sb->reshape_position);
info->new_level = __le32_to_cpu(sb->new_level);
info->delta_disks = __le32_to_cpu(sb->delta_disks);
int rv = 0;
struct mdp_superblock_1 *sb = st->sb;
+ if (strcmp(update, "homehost") == 0 &&
+ homehost) {
+ /* Note that 'homehost' is special as it is really
+ * a "name" update.
+ */
+ char *c;
+ update = "name";
+ c = strchr(sb->set_name, ':');
+ if (c)
+ strncpy(info->name, c+1, 31 - (c-sb->set_name));
+ else
+ strncpy(info->name, sb->set_name, 32);
+ info->name[32] = 0;
+ }
+
if (strcmp(update, "force-one")==0) {
/* Not enough devices for a working array,
* so bring this one up-to-date
} else if (strcmp(update, "assemble")==0) {
int d = info->disk.number;
int want;
- if (info->disk.state == 6)
+ if (info->disk.state & (1<<MD_DISK_ACTIVE))
want = info->disk.raid_disk;
else
want = 0xFFFF;
sb->bblog_shift = 0;
sb->bblog_offset = 0;
}
- } else if (strcmp(update, "homehost") == 0 &&
- homehost) {
- char *c;
- update = "name";
- c = strchr(sb->set_name, ':');
- if (c)
- strncpy(info->name, c+1, 31 - (c-sb->set_name));
- else
- strncpy(info->name, sb->set_name, 32);
- info->name[32] = 0;
} else if (strcmp(update, "name") == 0) {
if (info->name[0] == 0)
sprintf(info->name, "%d", info->array.md_minor);
struct devinfo {
int fd;
char *devname;
+ long long data_offset;
mdu_disk_info_t disk;
struct devinfo *next;
};
#ifndef MDASSEMBLE
/* Add a device to the superblock being created */
static int add_to_super1(struct supertype *st, mdu_disk_info_t *dk,
- int fd, char *devname)
+ int fd, char *devname, unsigned long long data_offset)
{
struct mdp_superblock_1 *sb = st->sb;
__u16 *rp = sb->dev_roles + dk->number;
di->fd = fd;
di->devname = devname;
di->disk = *dk;
+ di->data_offset = data_offset;
di->next = NULL;
*dip = di;
/* We try to leave 0.1% at the start for reshape
* operations, but limit this to 128Meg (0.1% of 10Gig)
* which is plenty for efficient reshapes
+ * However we make it at least 2 chunks as one chunk
+ * is minimum needed for reshape.
*/
headroom = 128 * 1024 * 2;
- while (headroom << 10 > array_size)
+ while (headroom << 10 > array_size &&
+ headroom/2 >= __le32_to_cpu(sb->chunksize) * 2)
headroom >>= 1;
+ data_offset = di->data_offset;
switch(st->minor_version) {
case 0:
sb_offset = dsize;
sb_offset -= 8*2;
sb_offset &= ~(4*2-1);
sb->super_offset = __cpu_to_le64(sb_offset);
- data_offset = __le64_to_cpu(sb->data_offset);
if (data_offset == INVALID_SECTORS)
sb->data_offset = 0;
if (sb_offset < array_size + bm_space)
break;
case 1:
sb->super_offset = __cpu_to_le64(0);
- data_offset = __le64_to_cpu(sb->data_offset);
if (data_offset == INVALID_SECTORS) {
reserved = bm_space + 4*2;
if (reserved < headroom)
case 2:
sb_offset = 4*2;
sb->super_offset = __cpu_to_le64(4*2);
- data_offset = __le64_to_cpu(sb->data_offset);
if (data_offset == INVALID_SECTORS) {
if (4*2 + 4*2 + bm_space + array_size
> dsize)
return -EINVAL;
}
- ioctl(fd, BLKFLSBUF, 0); /* make sure we read current data */
-
-
if (lseek64(fd, sb_offset << 9, 0)< 0LL) {
if (devname)
pr_err("Cannot seek to superblock on %s: %s\n",
{
struct supertype *st = xcalloc(1, sizeof(*st));
- st->container_dev = NoMdDev;
+ st->container_devnm[0] = 0;
st->ss = &super1;
st->max_devs = MAX_DEVS;
st->sb = NULL;
* superblock type st, and reserving 'reserve' sectors for
* a possible bitmap
*/
-static __u64 avail_size1(struct supertype *st, __u64 devsize,
- unsigned long long data_offset)
+static __u64 _avail_size1(struct supertype *st, __u64 devsize,
+ unsigned long long data_offset, int chunksize)
{
struct mdp_superblock_1 *super = st->sb;
int bmspace = 0;
* Limit slack to 128M, but aim for about 0.1%
*/
unsigned long long headroom = 128*1024*2;
- while ((headroom << 10) > devsize)
+ while ((headroom << 10) > devsize &&
+ (chunksize == 0 ||
+ headroom / 2 >= ((unsigned)chunksize*2)*2))
headroom >>= 1;
devsize -= headroom;
}
}
return 0;
}
+static __u64 avail_size1(struct supertype *st, __u64 devsize,
+ unsigned long long data_offset)
+{
+ return _avail_size1(st, devsize, data_offset, 0);
+}
static int
add_internal_bitmap1(struct supertype *st,
}
close(fd);
- *freesize = avail_size1(st, ldsize >> 9, data_offset);
+ *freesize = _avail_size1(st, ldsize >> 9, data_offset, *chunk);
return 1;
}
#endif /* MDASSEMBLE */
.write_init_super = write_init_super1,
.validate_geometry = validate_geometry1,
.add_to_super = add_to_super1,
+ .examine_badblocks = examine_badblocks_super1,
#endif
.match_home = match_home1,
.uuid_from_super = uuid_from_super1,