Skip to content

L
linux
Commit bab5d712 authored by Neil Brown's avatar Neil Brown Committed by Trond Myklebust
Browse files
Options

[PATCH] MD - Remove dependance on superblock 

Remove dependance on superblock

All the remaining field of interest in the superblock
get duplicated in the mddev struture and this is treated as
authoritative.  The superblock gets completely generated at
write time, and all useful information extracted at read time.

This means that we can slot in different superblock formats
without affecting the bulk of the code.
parent 5e601b35
Hide whitespace changes
Inline Side-by-side
Showing with 160 additions and 126 deletions
drivers/md/linear.c
View file @ bab5d712
...@@ -55,7 +55,7 @@ static int linear_run (mddev_t *mddev) ...@@ -55,7 +55,7 @@ static int linear_run (mddev_t *mddev)
int j = rdev->raid_disk; int j = rdev->raid_disk;
dev_info_t *disk = conf->disks + j; dev_info_t *disk = conf->disks + j;
if (j < 0 || j > mddev->sb->raid_disks || disk->bdev) { if (j < 0 || j > mddev->raid_disks || disk->bdev) {
printk("linear: disk numbering problem. Aborting!\n"); printk("linear: disk numbering problem. Aborting!\n");
goto out; goto out;
} }
...@@ -67,7 +67,7 @@ static int linear_run (mddev_t *mddev) ...@@ -67,7 +67,7 @@ static int linear_run (mddev_t *mddev)
conf->smallest = disk; conf->smallest = disk;
cnt++; cnt++;
} }
if (cnt != mddev->sb->raid_disks) { if (cnt != mddev->raid_disks) {
printk("linear: not enough drives present. Aborting!\n"); printk("linear: not enough drives present. Aborting!\n");
goto out; goto out;
} }
...@@ -186,7 +186,7 @@ static int linear_status (char *page, mddev_t *mddev) ...@@ -186,7 +186,7 @@ static int linear_status (char *page, mddev_t *mddev)
} }
sz += sprintf(page+sz, "\n"); sz += sprintf(page+sz, "\n");
#endif #endif
sz += sprintf(page+sz, " %dk rounding", mddev->sb->chunk_size/1024); sz += sprintf(page+sz, " %dk rounding", mddev->chunk_size/1024);
return sz; return sz;
} }
......
drivers/md/md.c
View file @ bab5d712
...@@ -311,8 +311,8 @@ static unsigned int calc_dev_size(struct block_device *bdev, mddev_t *mddev) ...@@ -311,8 +311,8 @@ static unsigned int calc_dev_size(struct block_device *bdev, mddev_t *mddev)
size = calc_dev_sboffset(bdev); size = calc_dev_sboffset(bdev);
else else
size = bdev->bd_inode->i_size >> BLOCK_SIZE_BITS; size = bdev->bd_inode->i_size >> BLOCK_SIZE_BITS;
if (mddev->sb->chunk_size) if (mddev->chunk_size)
size &= ~(mddev->sb->chunk_size/1024 - 1); size &= ~(mddev->chunk_size/1024 - 1);
return size; return size;
} }
...@@ -322,14 +322,10 @@ static unsigned int zoned_raid_size(mddev_t *mddev) ...@@ -322,14 +322,10 @@ static unsigned int zoned_raid_size(mddev_t *mddev)
mdk_rdev_t * rdev; mdk_rdev_t * rdev;
struct list_head *tmp; struct list_head *tmp;
if (!mddev->sb) {
MD_BUG();
return -EINVAL;
}
/* /*
* do size and offset calculations. * do size and offset calculations.
*/ */
mask = ~(mddev->sb->chunk_size/1024 - 1); mask = ~(mddev->chunk_size/1024 - 1);
ITERATE_RDEV(mddev,rdev,tmp) { ITERATE_RDEV(mddev,rdev,tmp) {
rdev->size &= mask; rdev->size &= mask;
...@@ -859,7 +855,31 @@ static int sync_sbs(mddev_t * mddev) ...@@ -859,7 +855,31 @@ static int sync_sbs(mddev_t * mddev)
int active=0, working=0,failed=0,spare=0,nr_disks=0; int active=0, working=0,failed=0,spare=0,nr_disks=0;
sb = mddev->sb; sb = mddev->sb;
memset(sb, 0, sizeof(*sb));
sb->md_magic = MD_SB_MAGIC;
sb->major_version = mddev->major_version;
sb->minor_version = mddev->minor_version;
sb->patch_version = mddev->patch_version;
sb->gvalid_words = 0; /* ignored */
memcpy(&sb->set_uuid0, mddev->uuid+0, 4);
memcpy(&sb->set_uuid1, mddev->uuid+4, 4);
memcpy(&sb->set_uuid2, mddev->uuid+8, 4);
memcpy(&sb->set_uuid3, mddev->uuid+12,4);
sb->ctime = mddev->ctime;
sb->level = mddev->level;
sb->size = mddev->size;
sb->raid_disks = mddev->raid_disks;
sb->md_minor = mddev->__minor;
sb->not_persistent = !mddev->persistent; sb->not_persistent = !mddev->persistent;
sb->utime = mddev->utime;
sb->state = mddev->state;
sb->events_hi = (mddev->events>>32);
sb->events_lo = (u32)mddev->events;
sb->layout = mddev->layout;
sb->chunk_size = mddev->chunk_size;
sb->disks[0].state = (1<<MD_DISK_REMOVED); sb->disks[0].state = (1<<MD_DISK_REMOVED);
ITERATE_RDEV(mddev,rdev,tmp) { ITERATE_RDEV(mddev,rdev,tmp) {
...@@ -920,18 +940,17 @@ static void md_update_sb(mddev_t * mddev) ...@@ -920,18 +940,17 @@ static void md_update_sb(mddev_t * mddev)
mddev->sb_dirty = 0; mddev->sb_dirty = 0;
repeat: repeat:
mddev->sb->utime = CURRENT_TIME; mddev->utime = CURRENT_TIME;
if (!(++mddev->sb->events_lo)) mddev->events ++;
++mddev->sb->events_hi;
if (!(mddev->sb->events_lo | mddev->sb->events_hi)) { if (!mddev->events) {
/* /*
* oops, this 64-bit counter should never wrap. * oops, this 64-bit counter should never wrap.
* Either we are in around ~1 trillion A.C., assuming * Either we are in around ~1 trillion A.C., assuming
* 1 reboot per second, or we have a bug: * 1 reboot per second, or we have a bug:
*/ */
MD_BUG(); MD_BUG();
mddev->sb->events_lo = mddev->sb->events_hi = 0xffffffff; mddev->events --;
} }
sync_sbs(mddev); sync_sbs(mddev);
...@@ -1150,7 +1169,25 @@ static int analyze_sbs(mddev_t * mddev) ...@@ -1150,7 +1169,25 @@ static int analyze_sbs(mddev_t * mddev)
} }
memcpy (sb, freshest->sb, sizeof(*sb)); memcpy (sb, freshest->sb, sizeof(*sb));
mddev->major_version = sb->major_version;
mddev->minor_version = sb->minor_version;
mddev->patch_version = sb->patch_version;
mddev->persistent = ! sb->not_persistent; mddev->persistent = ! sb->not_persistent;
mddev->chunk_size = sb->chunk_size;
mddev->ctime = sb->ctime;
mddev->utime = sb->utime;
mddev->level = sb->level;
mddev->layout = sb->layout;
mddev->raid_disks = sb->raid_disks;
mddev->state = sb->state;
mddev->size = sb->size;
mddev->events = md_event(sb);
memcpy(mddev->uuid+0, &sb->set_uuid0, 4);
memcpy(mddev->uuid+4, &sb->set_uuid1, 4);
memcpy(mddev->uuid+8, &sb->set_uuid2, 4);
memcpy(mddev->uuid+12,&sb->set_uuid3, 4);
/* /*
* at this point we have picked the 'best' superblock * at this point we have picked the 'best' superblock
* from all available superblocks. * from all available superblocks.
...@@ -1161,11 +1198,10 @@ static int analyze_sbs(mddev_t * mddev) ...@@ -1161,11 +1198,10 @@ static int analyze_sbs(mddev_t * mddev)
/* /*
* Kick all non-fresh devices * Kick all non-fresh devices
*/ */
__u64 ev1, ev2; __u64 ev1;
ev1 = md_event(rdev->sb); ev1 = md_event(rdev->sb);
ev2 = md_event(sb);
++ev1; ++ev1;
if (ev1 < ev2) { if (ev1 < mddev->events) {
printk(KERN_WARNING "md: kicking non-fresh %s from array!\n", printk(KERN_WARNING "md: kicking non-fresh %s from array!\n",
partition_name(rdev->dev)); partition_name(rdev->dev));
kick_rdev_from_array(rdev); kick_rdev_from_array(rdev);
...@@ -1179,7 +1215,7 @@ static int analyze_sbs(mddev_t * mddev) ...@@ -1179,7 +1215,7 @@ static int analyze_sbs(mddev_t * mddev)
*/ */
i = 0; i = 0;
ITERATE_RDEV(mddev,rdev,tmp) { ITERATE_RDEV(mddev,rdev,tmp) {
if (sb->level == LEVEL_MULTIPATH) { if (mddev->level == LEVEL_MULTIPATH) {
rdev->alias_device = !!i; rdev->alias_device = !!i;
rdev->desc_nr = i++; rdev->desc_nr = i++;
rdev->raid_disk = rdev->desc_nr; rdev->raid_disk = rdev->desc_nr;
...@@ -1195,7 +1231,7 @@ static int analyze_sbs(mddev_t * mddev) ...@@ -1195,7 +1231,7 @@ static int analyze_sbs(mddev_t * mddev)
rdev->faulty = 1; rdev->faulty = 1;
kick_rdev_from_array(rdev); kick_rdev_from_array(rdev);
} else if (desc->state & (1<<MD_DISK_SYNC) && } else if (desc->state & (1<<MD_DISK_SYNC) &&
rdev->raid_disk < mddev->sb->raid_disks) rdev->raid_disk < mddev->raid_disks)
rdev->in_sync = 1; rdev->in_sync = 1;
} }
} }
...@@ -1204,16 +1240,16 @@ static int analyze_sbs(mddev_t * mddev) ...@@ -1204,16 +1240,16 @@ static int analyze_sbs(mddev_t * mddev)
/* /*
* Check if we can support this RAID array * Check if we can support this RAID array
*/ */
if (sb->major_version != MD_MAJOR_VERSION || if (mddev->major_version != MD_MAJOR_VERSION ||
sb->minor_version > MD_MINOR_VERSION) { mddev->minor_version > MD_MINOR_VERSION) {
printk(OLD_VERSION, mdidx(mddev), sb->major_version, printk(OLD_VERSION, mdidx(mddev), mddev->major_version,
sb->minor_version, sb->patch_version); mddev->minor_version, mddev->patch_version);
goto abort; goto abort;
} }
if ((sb->state != (1 << MD_SB_CLEAN)) && ((sb->level == 1) || if ((mddev->state != (1 << MD_SB_CLEAN)) && ((mddev->level == 1) ||
(sb->level == 4) || (sb->level == 5))) (mddev->level == 4) || (mddev->level == 5)))
printk(NOT_CLEAN_IGNORE, mdidx(mddev)); printk(NOT_CLEAN_IGNORE, mdidx(mddev));
return 0; return 0;
...@@ -1230,7 +1266,6 @@ static int device_size_calculation(mddev_t * mddev) ...@@ -1230,7 +1266,6 @@ static int device_size_calculation(mddev_t * mddev)
{ {
int data_disks = 0; int data_disks = 0;
unsigned int readahead; unsigned int readahead;
mdp_super_t *sb = mddev->sb;
struct list_head *tmp; struct list_head *tmp;
mdk_rdev_t *rdev; mdk_rdev_t *rdev;
...@@ -1248,16 +1283,16 @@ static int device_size_calculation(mddev_t * mddev) ...@@ -1248,16 +1283,16 @@ static int device_size_calculation(mddev_t * mddev)
continue; continue;
} }
rdev->size = calc_dev_size(rdev->bdev, mddev); rdev->size = calc_dev_size(rdev->bdev, mddev);
if (rdev->size < sb->chunk_size / 1024) { if (rdev->size < mddev->chunk_size / 1024) {
printk(KERN_WARNING printk(KERN_WARNING
"md: Dev %s smaller than chunk_size: %ldk < %dk\n", "md: Dev %s smaller than chunk_size: %ldk < %dk\n",
partition_name(rdev->dev), partition_name(rdev->dev),
rdev->size, sb->chunk_size / 1024); rdev->size, mddev->chunk_size / 1024);
return -EINVAL; return -EINVAL;
} }
} }
switch (sb->level) { switch (mddev->level) {
case LEVEL_MULTIPATH: case LEVEL_MULTIPATH:
data_disks = 1; data_disks = 1;
break; break;
...@@ -1273,30 +1308,30 @@ static int device_size_calculation(mddev_t * mddev) ...@@ -1273,30 +1308,30 @@ static int device_size_calculation(mddev_t * mddev)
break; break;
case 0: case 0:
zoned_raid_size(mddev); zoned_raid_size(mddev);
data_disks = sb->raid_disks; data_disks = mddev->raid_disks;
break; break;
case 1: case 1:
data_disks = 1; data_disks = 1;
break; break;
case 4: case 4:
case 5: case 5:
data_disks = sb->raid_disks-1; data_disks = mddev->raid_disks-1;
break; break;
default: default:
printk(UNKNOWN_LEVEL, mdidx(mddev), sb->level); printk(UNKNOWN_LEVEL, mdidx(mddev), mddev->level);
goto abort; goto abort;
} }
if (!md_size[mdidx(mddev)]) if (!md_size[mdidx(mddev)])
md_size[mdidx(mddev)] = sb->size * data_disks; md_size[mdidx(mddev)] = mddev->size * data_disks;
readahead = (VM_MAX_READAHEAD * 1024) / PAGE_SIZE; readahead = (VM_MAX_READAHEAD * 1024) / PAGE_SIZE;
if (!sb->level || (sb->level == 4) || (sb->level == 5)) { if (!mddev->level || (mddev->level == 4) || (mddev->level == 5)) {
readahead = (mddev->sb->chunk_size>>PAGE_SHIFT) * 4 * data_disks; readahead = (mddev->chunk_size>>PAGE_SHIFT) * 4 * data_disks;
if (readahead < data_disks * (MAX_SECTORS>>(PAGE_SHIFT-9))*2) if (readahead < data_disks * (MAX_SECTORS>>(PAGE_SHIFT-9))*2)
readahead = data_disks * (MAX_SECTORS>>(PAGE_SHIFT-9))*2; readahead = data_disks * (MAX_SECTORS>>(PAGE_SHIFT-9))*2;
} else { } else {
// (no multipath branch - it uses the default setting) // (no multipath branch - it uses the default setting)
if (sb->level == -3) if (mddev->level == -3)
readahead = 0; readahead = 0;
} }
...@@ -1351,8 +1386,8 @@ static int do_md_run(mddev_t * mddev) ...@@ -1351,8 +1386,8 @@ static int do_md_run(mddev_t * mddev)
return -EINVAL; return -EINVAL;
} }
chunk_size = mddev->sb->chunk_size; chunk_size = mddev->chunk_size;
pnum = level_to_pers(mddev->sb->level); pnum = level_to_pers(mddev->level);
if ((pnum != MULTIPATH) && (pnum != RAID1)) { if ((pnum != MULTIPATH) && (pnum != RAID1)) {
if (!chunk_size) { if (!chunk_size) {
...@@ -1383,7 +1418,7 @@ static int do_md_run(mddev_t * mddev) ...@@ -1383,7 +1418,7 @@ static int do_md_run(mddev_t * mddev)
} else } else
if (chunk_size) if (chunk_size)
printk(KERN_INFO "md: RAID level %d does not need chunksize! Continuing anyway.\n", printk(KERN_INFO "md: RAID level %d does not need chunksize! Continuing anyway.\n",
mddev->sb->level); mddev->level);
if (pnum >= MAX_PERSONALITY) { if (pnum >= MAX_PERSONALITY) {
MD_BUG(); MD_BUG();
...@@ -1443,12 +1478,12 @@ static int do_md_run(mddev_t * mddev) ...@@ -1443,12 +1478,12 @@ static int do_md_run(mddev_t * mddev)
return -EINVAL; return -EINVAL;
} }
mddev->in_sync = (mddev->sb->state & (1<<MD_SB_CLEAN)); mddev->in_sync = (mddev->state & (1<<MD_SB_CLEAN));
/* if personality doesn't have "sync_request", then /* if personality doesn't have "sync_request", then
* a dirty array doesn't mean anything * a dirty array doesn't mean anything
*/ */
if (mddev->pers->sync_request) if (mddev->pers->sync_request)
mddev->sb->state &= ~(1 << MD_SB_CLEAN); mddev->state &= ~(1 << MD_SB_CLEAN);
md_update_sb(mddev); md_update_sb(mddev);
md_recover_arrays(); md_recover_arrays();
...@@ -1556,7 +1591,7 @@ static int do_md_stop(mddev_t * mddev, int ro) ...@@ -1556,7 +1591,7 @@ static int do_md_stop(mddev_t * mddev, int ro)
*/ */
if (mddev->in_sync) { if (mddev->in_sync) {
printk(KERN_INFO "md: marking sb clean...\n"); printk(KERN_INFO "md: marking sb clean...\n");
mddev->sb->state |= 1 << MD_SB_CLEAN; mddev->state |= 1 << MD_SB_CLEAN;
} }
md_update_sb(mddev); md_update_sb(mddev);
} }
...@@ -1812,7 +1847,6 @@ static int get_version(void * arg) ...@@ -1812,7 +1847,6 @@ static int get_version(void * arg)
return 0; return 0;
} }
#define SET_FROM_SB(x) info.x = mddev->sb->x
static int get_array_info(mddev_t * mddev, void * arg) static int get_array_info(mddev_t * mddev, void * arg)
{ {
mdu_array_info_t info; mdu_array_info_t info;
...@@ -1820,10 +1854,6 @@ static int get_array_info(mddev_t * mddev, void * arg) ...@@ -1820,10 +1854,6 @@ static int get_array_info(mddev_t * mddev, void * arg)
mdk_rdev_t *rdev; mdk_rdev_t *rdev;
struct list_head *tmp; struct list_head *tmp;
if (!mddev->sb) {
MD_BUG();
return -EINVAL;
}
nr=working=active=failed=spare=0; nr=working=active=failed=spare=0;
ITERATE_RDEV(mddev,rdev,tmp) { ITERATE_RDEV(mddev,rdev,tmp) {
nr++; nr++;
...@@ -1838,26 +1868,27 @@ static int get_array_info(mddev_t * mddev, void * arg) ...@@ -1838,26 +1868,27 @@ static int get_array_info(mddev_t * mddev, void * arg)
} }
} }
SET_FROM_SB(major_version); info.major_version = mddev->major_version;
SET_FROM_SB(minor_version); info.major_version = mddev->major_version;
SET_FROM_SB(patch_version); info.minor_version = mddev->minor_version;
SET_FROM_SB(ctime); info.patch_version = mddev->patch_version;
SET_FROM_SB(level); info.ctime = mddev->ctime;
SET_FROM_SB(size); info.level = mddev->level;
SET_FROM_SB(nr_disks); info.size = mddev->size;
SET_FROM_SB(raid_disks); info.nr_disks = nr;
SET_FROM_SB(md_minor); info.raid_disks = mddev->raid_disks;
info.md_minor = mddev->__minor;
info.not_persistent= !mddev->persistent; info.not_persistent= !mddev->persistent;
SET_FROM_SB(utime); info.utime = mddev->utime;
SET_FROM_SB(state); info.state = mddev->state;
info.active_disks = active; info.active_disks = active;
info.working_disks = working; info.working_disks = working;
info.failed_disks = failed; info.failed_disks = failed;
info.spare_disks = spare; info.spare_disks = spare;
SET_FROM_SB(layout); info.layout = mddev->layout;
SET_FROM_SB(chunk_size); info.chunk_size = mddev->chunk_size;
if (copy_to_user(arg, &info, sizeof(info))) if (copy_to_user(arg, &info, sizeof(info)))
return -EFAULT; return -EFAULT;
...@@ -1873,9 +1904,6 @@ static int get_disk_info(mddev_t * mddev, void * arg) ...@@ -1873,9 +1904,6 @@ static int get_disk_info(mddev_t * mddev, void * arg)
unsigned int nr; unsigned int nr;
mdk_rdev_t *rdev; mdk_rdev_t *rdev;
if (!mddev->sb)
return -EINVAL;
if (copy_from_user(&info, arg, sizeof(info))) if (copy_from_user(&info, arg, sizeof(info)))
return -EFAULT; return -EFAULT;
...@@ -1949,7 +1977,7 @@ static int add_new_disk(mddev_t * mddev, mdu_disk_info_t *info) ...@@ -1949,7 +1977,7 @@ static int add_new_disk(mddev_t * mddev, mdu_disk_info_t *info)
rdev->desc_nr = info->number; rdev->desc_nr = info->number;
rdev->raid_disk = info->raid_disk; rdev->raid_disk = info->raid_disk;
rdev->faulty = 0; rdev->faulty = 0;
if (rdev->raid_disk < mddev->sb->raid_disks) if (rdev->raid_disk < mddev->raid_disks)
rdev->in_sync = (info->state & (1<<MD_DISK_SYNC)); rdev->in_sync = (info->state & (1<<MD_DISK_SYNC));
else else
rdev->in_sync = 0; rdev->in_sync = 0;
...@@ -1962,8 +1990,8 @@ static int add_new_disk(mddev_t * mddev, mdu_disk_info_t *info) ...@@ -1962,8 +1990,8 @@ static int add_new_disk(mddev_t * mddev, mdu_disk_info_t *info)
size = calc_dev_size(rdev->bdev, mddev); size = calc_dev_size(rdev->bdev, mddev);
rdev->sb_offset = calc_dev_sboffset(rdev->bdev); rdev->sb_offset = calc_dev_sboffset(rdev->bdev);
if (!mddev->sb->size || (mddev->sb->size > size)) if (!mddev->size || (mddev->size > size))
mddev->sb->size = size; mddev->size = size;
} }
return 0; return 0;
...@@ -2074,9 +2102,9 @@ static int hot_add_disk(mddev_t * mddev, kdev_t dev) ...@@ -2074,9 +2102,9 @@ static int hot_add_disk(mddev_t * mddev, kdev_t dev)
size = calc_dev_size(rdev->bdev, mddev); size = calc_dev_size(rdev->bdev, mddev);
if (size < mddev->sb->size) { if (size < mddev->size) {
printk(KERN_WARNING "md%d: disk size %d blocks < array size %d\n", printk(KERN_WARNING "md%d: disk size %d blocks < array size %ld\n",
mdidx(mddev), size, mddev->sb->size); mdidx(mddev), size, mddev->size);
err = -ENOSPC; err = -ENOSPC;
goto abort_export; goto abort_export;
} }
...@@ -2097,7 +2125,7 @@ static int hot_add_disk(mddev_t * mddev, kdev_t dev) ...@@ -2097,7 +2125,7 @@ static int hot_add_disk(mddev_t * mddev, kdev_t dev)
rdev->size = size; rdev->size = size;
rdev->sb_offset = calc_dev_sboffset(rdev->bdev); rdev->sb_offset = calc_dev_sboffset(rdev->bdev);
for (i = mddev->sb->raid_disks; i < MD_SB_DISKS; i++) for (i = mddev->raid_disks; i < MD_SB_DISKS; i++)
if (find_rdev_nr(mddev,i)==NULL) if (find_rdev_nr(mddev,i)==NULL)
break; break;
...@@ -2135,43 +2163,38 @@ static int hot_add_disk(mddev_t * mddev, kdev_t dev) ...@@ -2135,43 +2163,38 @@ static int hot_add_disk(mddev_t * mddev, kdev_t dev)
return err; return err;
} }
#define SET_SB(x) mddev->sb->x = info->x
static int set_array_info(mddev_t * mddev, mdu_array_info_t *info) static int set_array_info(mddev_t * mddev, mdu_array_info_t *info)
{ {
if (alloc_array_sb(mddev)) if (alloc_array_sb(mddev))
return -ENOMEM; return -ENOMEM;
mddev->sb->major_version = MD_MAJOR_VERSION; mddev->major_version = MD_MAJOR_VERSION;
mddev->sb->minor_version = MD_MINOR_VERSION; mddev->minor_version = MD_MINOR_VERSION;
mddev->sb->patch_version = MD_PATCHLEVEL_VERSION; mddev->patch_version = MD_PATCHLEVEL_VERSION;
mddev->sb->ctime = CURRENT_TIME; mddev->ctime = CURRENT_TIME;
SET_SB(level); mddev->level = info->level;
SET_SB(size); mddev->size = info->size;
SET_SB(nr_disks); mddev->raid_disks = info->raid_disks;
SET_SB(raid_disks); /* don't set __minor, it is determined by which /dev/md* was
SET_SB(md_minor); * openned
*/
mddev->state = info->state;
mddev->persistent = ! info->not_persistent; mddev->persistent = ! info->not_persistent;
SET_SB(state); mddev->layout = info->layout;
mddev->chunk_size = info->chunk_size;
SET_SB(layout);
SET_SB(chunk_size);
mddev->sb->md_magic = MD_SB_MAGIC;
/* /*
* Generate a 128 bit UUID * Generate a 128 bit UUID
*/ */
get_random_bytes(&mddev->sb->set_uuid0, 4); get_random_bytes(mddev->uuid, 16);
get_random_bytes(&mddev->sb->set_uuid1, 4);
get_random_bytes(&mddev->sb->set_uuid2, 4);
get_random_bytes(&mddev->sb->set_uuid3, 4);
return 0; return 0;
} }
#undef SET_SB
static int set_disk_faulty(mddev_t *mddev, kdev_t dev) static int set_disk_faulty(mddev_t *mddev, kdev_t dev)
{ {
...@@ -2680,7 +2703,7 @@ static int status_resync(char * page, mddev_t * mddev) ...@@ -2680,7 +2703,7 @@ static int status_resync(char * page, mddev_t * mddev)
unsigned long max_blocks, resync, res, dt, db, rt; unsigned long max_blocks, resync, res, dt, db, rt;
resync = (mddev->curr_resync - atomic_read(&mddev->recovery_active))/2; resync = (mddev->curr_resync - atomic_read(&mddev->recovery_active))/2;
max_blocks = mddev->sb->size; max_blocks = mddev->size;
/* /*
* Should not happen. * Should not happen.
...@@ -2939,7 +2962,7 @@ static void md_do_sync(void *data) ...@@ -2939,7 +2962,7 @@ static void md_do_sync(void *data)
} }
} while (mddev->curr_resync < 2); } while (mddev->curr_resync < 2);
max_sectors = mddev->sb->size << 1; max_sectors = mddev->size << 1;
printk(KERN_INFO "md: syncing RAID array md%d\n", mdidx(mddev)); printk(KERN_INFO "md: syncing RAID array md%d\n", mdidx(mddev));
printk(KERN_INFO "md: minimum _guaranteed_ reconstruction speed: %d KB/sec/disc.\n", sysctl_speed_limit_min); printk(KERN_INFO "md: minimum _guaranteed_ reconstruction speed: %d KB/sec/disc.\n", sysctl_speed_limit_min);
...@@ -3491,7 +3514,7 @@ void __init md_setup_drive(void) ...@@ -3491,7 +3514,7 @@ void __init md_setup_drive(void)
dinfo.state = (1<<MD_DISK_ACTIVE)|(1<<MD_DISK_SYNC); dinfo.state = (1<<MD_DISK_ACTIVE)|(1<<MD_DISK_SYNC);
dinfo.major = major(dev); dinfo.major = major(dev);
dinfo.minor = minor(dev); dinfo.minor = minor(dev);
mddev->sb->raid_disks++; mddev->raid_disks++;
err = add_new_disk (mddev, &dinfo); err = add_new_disk (mddev, &dinfo);
} }
} else { } else {
......
drivers/md/multipath.c
View file @ bab5d712
...@@ -424,7 +424,7 @@ static void multipathd (void *data) ...@@ -424,7 +424,7 @@ static void multipathd (void *data)
"multipath: detected IO path differences!\n" "multipath: detected IO path differences!\n"
#define ARRAY_IS_ACTIVE KERN_INFO \ #define ARRAY_IS_ACTIVE KERN_INFO \
"multipath: array md%d active with %d out of %d IO paths (%d spare IO paths)\n" "multipath: array md%d active with %d out of %d IO paths\n"
#define THREAD_ERROR KERN_ERR \ #define THREAD_ERROR KERN_ERR \
"multipath: couldn't allocate thread for md%d\n" "multipath: couldn't allocate thread for md%d\n"
...@@ -434,15 +434,14 @@ static int multipath_run (mddev_t *mddev) ...@@ -434,15 +434,14 @@ static int multipath_run (mddev_t *mddev)
multipath_conf_t *conf; multipath_conf_t *conf;
int disk_idx; int disk_idx;
struct multipath_info *disk; struct multipath_info *disk;
mdp_super_t *sb = mddev->sb;
mdk_rdev_t *rdev; mdk_rdev_t *rdev;
struct list_head *tmp; struct list_head *tmp;
int num_rdevs = 0; int num_rdevs = 0;
MOD_INC_USE_COUNT; MOD_INC_USE_COUNT;
if (sb->level != LEVEL_MULTIPATH) { if (mddev->level != LEVEL_MULTIPATH) {
printk(INVALID_LEVEL, mdidx(mddev), sb->level); printk(INVALID_LEVEL, mdidx(mddev), mddev->level);
goto out; goto out;
} }
/* /*
...@@ -492,7 +491,7 @@ static int multipath_run (mddev_t *mddev) ...@@ -492,7 +491,7 @@ static int multipath_run (mddev_t *mddev)
num_rdevs++; num_rdevs++;
} }
conf->raid_disks = sb->raid_disks = num_rdevs; conf->raid_disks = mddev->raid_disks = num_rdevs;
mddev->sb_dirty = 1; mddev->sb_dirty = 1;
conf->mddev = mddev; conf->mddev = mddev;
conf->device_lock = SPIN_LOCK_UNLOCKED; conf->device_lock = SPIN_LOCK_UNLOCKED;
...@@ -520,8 +519,8 @@ static int multipath_run (mddev_t *mddev) ...@@ -520,8 +519,8 @@ static int multipath_run (mddev_t *mddev)
} }
} }
printk(ARRAY_IS_ACTIVE, mdidx(mddev), sb->active_disks, printk(ARRAY_IS_ACTIVE, mdidx(mddev), conf->working_disks,
sb->raid_disks, sb->spare_disks); mddev->raid_disks);
/* /*
* Ok, everything is just fine now * Ok, everything is just fine now
*/ */
......
drivers/md/raid0.c
View file @ bab5d712
...@@ -89,7 +89,7 @@ static int create_strip_zones (mddev_t *mddev) ...@@ -89,7 +89,7 @@ static int create_strip_zones (mddev_t *mddev)
ITERATE_RDEV(mddev, rdev1, tmp1) { ITERATE_RDEV(mddev, rdev1, tmp1) {
int j = rdev1->sb->this_disk.raid_disk; int j = rdev1->sb->this_disk.raid_disk;
if (j < 0 || j >= mddev->sb->raid_disks) { if (j < 0 || j >= mddev->raid_disks) {
printk("raid0: bad disk number %d - aborting!\n", j); printk("raid0: bad disk number %d - aborting!\n", j);
goto abort; goto abort;
} }
...@@ -102,9 +102,9 @@ static int create_strip_zones (mddev_t *mddev) ...@@ -102,9 +102,9 @@ static int create_strip_zones (mddev_t *mddev)
smallest = rdev1; smallest = rdev1;
cnt++; cnt++;
} }
if (cnt != mddev->sb->raid_disks) { if (cnt != mddev->raid_disks) {
printk("raid0: too few disks (%d of %d) - aborting!\n", cnt, printk("raid0: too few disks (%d of %d) - aborting!\n", cnt,
mddev->sb->raid_disks); mddev->raid_disks);
goto abort; goto abort;
} }
zone->nb_dev = cnt; zone->nb_dev = cnt;
...@@ -271,7 +271,7 @@ static int raid0_make_request (request_queue_t *q, struct bio *bio) ...@@ -271,7 +271,7 @@ static int raid0_make_request (request_queue_t *q, struct bio *bio)
mdk_rdev_t *tmp_dev; mdk_rdev_t *tmp_dev;
unsigned long chunk, block, rsect; unsigned long chunk, block, rsect;
chunk_size = mddev->sb->chunk_size >> 10; chunk_size = mddev->chunk_size >> 10;
chunksize_bits = ffz(~chunk_size); chunksize_bits = ffz(~chunk_size);
block = bio->bi_sector >> 1; block = bio->bi_sector >> 1;
hash = conf->hash_table + block / conf->smallest->size; hash = conf->hash_table + block / conf->smallest->size;
...@@ -360,7 +360,7 @@ static int raid0_status (char *page, mddev_t *mddev) ...@@ -360,7 +360,7 @@ static int raid0_status (char *page, mddev_t *mddev)
conf->strip_zone[j].size); conf->strip_zone[j].size);
} }
#endif #endif
sz += sprintf(page + sz, " %dk chunks", mddev->sb->chunk_size/1024); sz += sprintf(page + sz, " %dk chunks", mddev->chunk_size/1024);
return sz; return sz;
} }
......
drivers/md/raid1.c
View file @ bab5d712
...@@ -1060,7 +1060,7 @@ static int sync_request(mddev_t *mddev, sector_t sector_nr, int go_faster) ...@@ -1060,7 +1060,7 @@ static int sync_request(mddev_t *mddev, sector_t sector_nr, int go_faster)
if (init_resync(conf)) if (init_resync(conf))
return -ENOMEM; return -ENOMEM;
max_sector = mddev->sb->size << 1; max_sector = mddev->size << 1;
if (sector_nr >= max_sector) { if (sector_nr >= max_sector) {
close_sync(conf); close_sync(conf);
return 0; return 0;
...@@ -1180,14 +1180,13 @@ static int run(mddev_t *mddev) ...@@ -1180,14 +1180,13 @@ static int run(mddev_t *mddev)
conf_t *conf; conf_t *conf;
int i, j, disk_idx; int i, j, disk_idx;
mirror_info_t *disk; mirror_info_t *disk;
mdp_super_t *sb = mddev->sb;
mdk_rdev_t *rdev; mdk_rdev_t *rdev;
struct list_head *tmp; struct list_head *tmp;
MOD_INC_USE_COUNT; MOD_INC_USE_COUNT;
if (sb->level != 1) { if (mddev->level != 1) {
printk(INVALID_LEVEL, mdidx(mddev), sb->level); printk(INVALID_LEVEL, mdidx(mddev), mddev->level);
goto out; goto out;
} }
/* /*
...@@ -1266,7 +1265,7 @@ static int run(mddev_t *mddev) ...@@ -1266,7 +1265,7 @@ static int run(mddev_t *mddev)
disk->head_position = 0; disk->head_position = 0;
} }
} }
conf->raid_disks = sb->raid_disks; conf->raid_disks = mddev->raid_disks;
conf->mddev = mddev; conf->mddev = mddev;
conf->device_lock = SPIN_LOCK_UNLOCKED; conf->device_lock = SPIN_LOCK_UNLOCKED;
...@@ -1316,7 +1315,7 @@ static int run(mddev_t *mddev) ...@@ -1316,7 +1315,7 @@ static int run(mddev_t *mddev)
} }
} }
printk(ARRAY_IS_ACTIVE, mdidx(mddev), sb->raid_disks - mddev->degraded, sb->raid_disks); printk(ARRAY_IS_ACTIVE, mdidx(mddev), mddev->raid_disks - mddev->degraded, mddev->raid_disks);
/* /*
* Ok, everything is just fine now * Ok, everything is just fine now
*/ */
......
drivers/md/raid5.c
View file @ bab5d712
...@@ -1290,7 +1290,7 @@ static int sync_request (mddev_t *mddev, sector_t sector_nr, int go_faster) ...@@ -1290,7 +1290,7 @@ static int sync_request (mddev_t *mddev, sector_t sector_nr, int go_faster)
int raid_disks = conf->raid_disks; int raid_disks = conf->raid_disks;
int data_disks = raid_disks-1; int data_disks = raid_disks-1;
if (sector_nr >= mddev->sb->size <<1) if (sector_nr >= mddev->size <<1)
/* just being told to finish up .. nothing to do */ /* just being told to finish up .. nothing to do */
return 0; return 0;
...@@ -1364,15 +1364,14 @@ static int run (mddev_t *mddev) ...@@ -1364,15 +1364,14 @@ static int run (mddev_t *mddev)
{ {
raid5_conf_t *conf; raid5_conf_t *conf;
int i, raid_disk, memory; int i, raid_disk, memory;
mdp_super_t *sb = mddev->sb;
mdk_rdev_t *rdev; mdk_rdev_t *rdev;
struct disk_info *disk; struct disk_info *disk;
struct list_head *tmp; struct list_head *tmp;
MOD_INC_USE_COUNT; MOD_INC_USE_COUNT;
if (sb->level != 5 && sb->level != 4) { if (mddev->level != 5 && mddev->level != 4) {
printk("raid5: md%d: raid level not set to 4/5 (%d)\n", mdidx(mddev), sb->level); printk("raid5: md%d: raid level not set to 4/5 (%d)\n", mdidx(mddev), mddev->level);
MOD_DEC_USE_COUNT; MOD_DEC_USE_COUNT;
return -EIO; return -EIO;
} }
...@@ -1444,7 +1443,7 @@ static int run (mddev_t *mddev) ...@@ -1444,7 +1443,7 @@ static int run (mddev_t *mddev)
} }
} }
for (i = 0; i < sb->raid_disks; i++) { for (i = 0; i < conf->raid_disks; i++) {
disk = conf->disks + i; disk = conf->disks + i;
if (!disk->used_slot) { if (!disk->used_slot) {
...@@ -1457,15 +1456,15 @@ static int run (mddev_t *mddev) ...@@ -1457,15 +1456,15 @@ static int run (mddev_t *mddev)
} }
} }
conf->raid_disks = sb->raid_disks; conf->raid_disks = mddev->raid_disks;
/* /*
* 0 for a fully functional array, 1 for a degraded array. * 0 for a fully functional array, 1 for a degraded array.
*/ */
mddev->degraded = conf->failed_disks = conf->raid_disks - conf->working_disks; mddev->degraded = conf->failed_disks = conf->raid_disks - conf->working_disks;
conf->mddev = mddev; conf->mddev = mddev;
conf->chunk_size = sb->chunk_size; conf->chunk_size = mddev->chunk_size;
conf->level = sb->level; conf->level = mddev->level;
conf->algorithm = sb->layout; conf->algorithm = mddev->layout;
conf->max_nr_stripes = NR_STRIPES; conf->max_nr_stripes = NR_STRIPES;
#if 0 #if 0
...@@ -1490,7 +1489,7 @@ static int run (mddev_t *mddev) ...@@ -1490,7 +1489,7 @@ static int run (mddev_t *mddev)
} }
if (mddev->degraded == 1 && if (mddev->degraded == 1 &&
!(sb->state & (1<<MD_SB_CLEAN))) { !(mddev->state & (1<<MD_SB_CLEAN))) {
printk(KERN_ERR "raid5: cannot start dirty degraded array for md%d\n", mdidx(mddev)); printk(KERN_ERR "raid5: cannot start dirty degraded array for md%d\n", mdidx(mddev));
goto abort; goto abort;
} }
...@@ -1515,10 +1514,12 @@ static int run (mddev_t *mddev) ...@@ -1515,10 +1514,12 @@ static int run (mddev_t *mddev)
} else } else
printk(KERN_INFO "raid5: allocated %dkB for md%d\n", memory, mdidx(mddev)); printk(KERN_INFO "raid5: allocated %dkB for md%d\n", memory, mdidx(mddev));
if (conf->working_disks == conf->raid_disks) if (mddev->degraded == 0)
printk("raid5: raid level %d set md%d active with %d out of %d devices, algorithm %d\n", conf->level, mdidx(mddev), sb->active_disks, sb->raid_disks, conf->algorithm); printk("raid5: raid level %d set md%d active with %d out of %d devices, algorithm %d\n", conf->level, mdidx(mddev),
mddev->raid_disks-mddev->degraded, mddev->raid_disks, conf->algorithm);
else else
printk(KERN_ALERT "raid5: raid level %d set md%d active with %d out of %d devices, algorithm %d\n", conf->level, mdidx(mddev), sb->active_disks, sb->raid_disks, conf->algorithm); printk(KERN_ALERT "raid5: raid level %d set md%d active with %d out of %d devices, algorithm %d\n", conf->level, mdidx(mddev),
mddev->raid_disks = mddev->degraded, mddev->raid_disks, conf->algorithm);
print_raid5_conf(conf); print_raid5_conf(conf);
...@@ -1590,10 +1591,9 @@ static void printall (raid5_conf_t *conf) ...@@ -1590,10 +1591,9 @@ static void printall (raid5_conf_t *conf)
static int status (char *page, mddev_t *mddev) static int status (char *page, mddev_t *mddev)
{ {
raid5_conf_t *conf = (raid5_conf_t *) mddev->private; raid5_conf_t *conf = (raid5_conf_t *) mddev->private;
mdp_super_t *sb = mddev->sb;
int sz = 0, i; int sz = 0, i;
sz += sprintf (page+sz, " level %d, %dk chunk, algorithm %d", sb->level, sb->chunk_size >> 10, sb->layout); sz += sprintf (page+sz, " level %d, %dk chunk, algorithm %d", mddev->level, mddev->chunk_size >> 10, mddev->layout);
sz += sprintf (page+sz, " [%d/%d] [", conf->raid_disks, conf->working_disks); sz += sprintf (page+sz, " [%d/%d] [", conf->raid_disks, conf->working_disks);
for (i = 0; i < conf->raid_disks; i++) for (i = 0; i < conf->raid_disks; i++)
sz += sprintf (page+sz, "%s", conf->disks[i].operational ? "U" : "_"); sz += sprintf (page+sz, "%s", conf->disks[i].operational ? "U" : "_");
......
include/linux/raid/md_k.h
View file @ bab5d712
...@@ -175,7 +175,20 @@ struct mddev_s ...@@ -175,7 +175,20 @@ struct mddev_s
int sb_dirty; int sb_dirty;
int ro; int ro;
/* Superblock information */
int major_version,
minor_version,
patch_version;
int persistent; int persistent;
int chunk_size;
time_t ctime, utime;
int level, layout;
int raid_disks;
unsigned long state;
sector_t size; /* used size of component devices */
__u64 events;
char uuid[16];
struct mdk_thread_s *sync_thread; /* doing resync or reconstruct */ struct mdk_thread_s *sync_thread; /* doing resync or reconstruct */
unsigned long curr_resync; /* blocks scheduled */ unsigned long curr_resync; /* blocks scheduled */
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7