blob: 9426976e0860d6386dad898a20c53db97a3d3444 [file] [log] [blame]
/*
md.c : Multiple Devices driver for Linux
Copyright (C) 1998, 1999, 2000 Ingo Molnar
completely rewritten, based on the MD driver code from Marc Zyngier
Changes:
- RAID-1/RAID-5 extensions by Miguel de Icaza, Gadi Oxman, Ingo Molnar
- RAID-6 extensions by H. Peter Anvin <hpa@zytor.com>
- boot support for linear and striped mode by Harald Hoyer <HarryH@Royal.Net>
- kerneld support by Boris Tobotras <boris@xtalk.msk.su>
- kmod support by: Cyrus Durgin
- RAID0 bugfixes: Mark Anthony Lisher <markal@iname.com>
- Devfs support by Richard Gooch <rgooch@atnf.csiro.au>
- lots of fixes and improvements to the RAID1/RAID5 and generic
RAID code (such as request based resynchronization):
Neil Brown <neilb@cse.unsw.edu.au>.
- persistent bitmap code
Copyright (C) 2003-2004, Paul Clements, SteelEye Technology, Inc.
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2, or (at your option)
any later version.
You should have received a copy of the GNU General Public License
(for example /usr/src/linux/COPYING); if not, write to the Free
Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
Errors, Warnings, etc.
Please use:
pr_crit() for error conditions that risk data loss
pr_err() for error conditions that are unexpected, like an IO error
or internal inconsistency
pr_warn() for error conditions that could have been predicated, like
adding a device to an array when it has incompatible metadata
pr_info() for every interesting, very rare events, like an array starting
or stopping, or resync starting or stopping
pr_debug() for everything else.
*/
#include <linux/sched/signal.h>
#include <linux/kthread.h>
#include <linux/blkdev.h>
#include <linux/badblocks.h>
#include <linux/sysctl.h>
#include <linux/seq_file.h>
#include <linux/fs.h>
#include <linux/poll.h>
#include <linux/ctype.h>
#include <linux/string.h>
#include <linux/hdreg.h>
#include <linux/proc_fs.h>
#include <linux/random.h>
#include <linux/module.h>
#include <linux/reboot.h>
#include <linux/file.h>
#include <linux/compat.h>
#include <linux/delay.h>
#include <linux/raid/md_p.h>
#include <linux/raid/md_u.h>
#include <linux/slab.h>
#include <linux/percpu-refcount.h>
#include <trace/events/block.h>
#include "md.h"
#include "md-bitmap.h"
#include "md-cluster.h"
#ifndef MODULE
static void autostart_arrays(int part);
#endif
/* pers_list is a list of registered personalities protected
* by pers_lock.
* pers_lock does extra service to protect accesses to
* mddev->thread when the mutex cannot be held.
*/
static LIST_HEAD(pers_list);
static DEFINE_SPINLOCK(pers_lock);
static struct kobj_type md_ktype;
struct md_cluster_operations *md_cluster_ops;
EXPORT_SYMBOL(md_cluster_ops);
struct module *md_cluster_mod;
EXPORT_SYMBOL(md_cluster_mod);
static DECLARE_WAIT_QUEUE_HEAD(resync_wait);
static struct workqueue_struct *md_wq;
static struct workqueue_struct *md_misc_wq;
static int remove_and_add_spares(struct mddev *mddev,
struct md_rdev *this);
static void mddev_detach(struct mddev *mddev);
/*
* Default number of read corrections we'll attempt on an rdev
* before ejecting it from the array. We divide the read error
* count by 2 for every hour elapsed between read errors.
*/
#define MD_DEFAULT_MAX_CORRECTED_READ_ERRORS 20
/*
* Current RAID-1,4,5 parallel reconstruction 'guaranteed speed limit'
* is 1000 KB/sec, so the extra system load does not show up that much.
* Increase it if you want to have more _guaranteed_ speed. Note that
* the RAID driver will use the maximum available bandwidth if the IO
* subsystem is idle. There is also an 'absolute maximum' reconstruction
* speed limit - in case reconstruction slows down your system despite
* idle IO detection.
*
* you can change it via /proc/sys/dev/raid/speed_limit_min and _max.
* or /sys/block/mdX/md/sync_speed_{min,max}
*/
static int sysctl_speed_limit_min = 1000;
static int sysctl_speed_limit_max = 200000;
static inline int speed_min(struct mddev *mddev)
{
return mddev->sync_speed_min ?
mddev->sync_speed_min : sysctl_speed_limit_min;
}
static inline int speed_max(struct mddev *mddev)
{
return mddev->sync_speed_max ?
mddev->sync_speed_max : sysctl_speed_limit_max;
}
static struct ctl_table_header *raid_table_header;
static struct ctl_table raid_table[] = {
{
.procname = "speed_limit_min",
.data = &sysctl_speed_limit_min,
.maxlen = sizeof(int),
.mode = S_IRUGO|S_IWUSR,
.proc_handler = proc_dointvec,
},
{
.procname = "speed_limit_max",
.data = &sysctl_speed_limit_max,
.maxlen = sizeof(int),
.mode = S_IRUGO|S_IWUSR,
.proc_handler = proc_dointvec,
},
{ }
};
static struct ctl_table raid_dir_table[] = {
{
.procname = "raid",
.maxlen = 0,
.mode = S_IRUGO|S_IXUGO,
.child = raid_table,
},
{ }
};
static struct ctl_table raid_root_table[] = {
{
.procname = "dev",
.maxlen = 0,
.mode = 0555,
.child = raid_dir_table,
},
{ }
};
static const struct block_device_operations md_fops;
static int start_readonly;
/*
* The original mechanism for creating an md device is to create
* a device node in /dev and to open it. This causes races with device-close.
* The preferred method is to write to the "new_array" module parameter.
* This can avoid races.
* Setting create_on_open to false disables the original mechanism
* so all the races disappear.
*/
static bool create_on_open = true;
struct bio *bio_alloc_mddev(gfp_t gfp_mask, int nr_iovecs,
struct mddev *mddev)
{
struct bio *b;
if (!mddev || !bioset_initialized(&mddev->bio_set))
return bio_alloc(gfp_mask, nr_iovecs);
b = bio_alloc_bioset(gfp_mask, nr_iovecs, &mddev->bio_set);
if (!b)
return NULL;
return b;
}
EXPORT_SYMBOL_GPL(bio_alloc_mddev);
static struct bio *md_bio_alloc_sync(struct mddev *mddev)
{
if (!mddev || !bioset_initialized(&mddev->sync_set))
return bio_alloc(GFP_NOIO, 1);
return bio_alloc_bioset(GFP_NOIO, 1, &mddev->sync_set);
}
/*
* We have a system wide 'event count' that is incremented
* on any 'interesting' event, and readers of /proc/mdstat
* can use 'poll' or 'select' to find out when the event
* count increases.
*
* Events are:
* start array, stop array, error, add device, remove device,
* start build, activate spare
*/
static DECLARE_WAIT_QUEUE_HEAD(md_event_waiters);
static atomic_t md_event_count;
void md_new_event(struct mddev *mddev)
{
atomic_inc(&md_event_count);
wake_up(&md_event_waiters);
}
EXPORT_SYMBOL_GPL(md_new_event);
/*
* Enables to iterate over all existing md arrays
* all_mddevs_lock protects this list.
*/
static LIST_HEAD(all_mddevs);
static DEFINE_SPINLOCK(all_mddevs_lock);
/*
* iterates through all used mddevs in the system.
* We take care to grab the all_mddevs_lock whenever navigating
* the list, and to always hold a refcount when unlocked.
* Any code which breaks out of this loop while own
* a reference to the current mddev and must mddev_put it.
*/
#define for_each_mddev(_mddev,_tmp) \
\
for (({ spin_lock(&all_mddevs_lock); \
_tmp = all_mddevs.next; \
_mddev = NULL;}); \
({ if (_tmp != &all_mddevs) \
mddev_get(list_entry(_tmp, struct mddev, all_mddevs));\
spin_unlock(&all_mddevs_lock); \
if (_mddev) mddev_put(_mddev); \
_mddev = list_entry(_tmp, struct mddev, all_mddevs); \
_tmp != &all_mddevs;}); \
({ spin_lock(&all_mddevs_lock); \
_tmp = _tmp->next;}) \
)
/* Rather than calling directly into the personality make_request function,
* IO requests come here first so that we can check if the device is
* being suspended pending a reconfiguration.
* We hold a refcount over the call to ->make_request. By the time that
* call has finished, the bio has been linked into some internal structure
* and so is visible to ->quiesce(), so we don't need the refcount any more.
*/
static bool is_suspended(struct mddev *mddev, struct bio *bio)
{
if (mddev->suspended)
return true;
if (bio_data_dir(bio) != WRITE)
return false;
if (mddev->suspend_lo >= mddev->suspend_hi)
return false;
if (bio->bi_iter.bi_sector >= mddev->suspend_hi)
return false;
if (bio_end_sector(bio) < mddev->suspend_lo)
return false;
return true;
}
void md_handle_request(struct mddev *mddev, struct bio *bio)
{
check_suspended:
rcu_read_lock();
if (is_suspended(mddev, bio)) {
DEFINE_WAIT(__wait);
for (;;) {
prepare_to_wait(&mddev->sb_wait, &__wait,
TASK_UNINTERRUPTIBLE);
if (!is_suspended(mddev, bio))
break;
rcu_read_unlock();
schedule();
rcu_read_lock();
}
finish_wait(&mddev->sb_wait, &__wait);
}
atomic_inc(&mddev->active_io);
rcu_read_unlock();
if (!mddev->pers->make_request(mddev, bio)) {
atomic_dec(&mddev->active_io);
wake_up(&mddev->sb_wait);
goto check_suspended;
}
if (atomic_dec_and_test(&mddev->active_io) && mddev->suspended)
wake_up(&mddev->sb_wait);
}
EXPORT_SYMBOL(md_handle_request);
static blk_qc_t md_make_request(struct request_queue *q, struct bio *bio)
{
const int rw = bio_data_dir(bio);
const int sgrp = op_stat_group(bio_op(bio));
struct mddev *mddev = q->queuedata;
unsigned int sectors;
int cpu;
blk_queue_split(q, &bio);
if (mddev == NULL || mddev->pers == NULL) {
bio_io_error(bio);
return BLK_QC_T_NONE;
}
if (mddev->ro == 1 && unlikely(rw == WRITE)) {
if (bio_sectors(bio) != 0)
bio->bi_status = BLK_STS_IOERR;
bio_endio(bio);
return BLK_QC_T_NONE;
}
/*
* save the sectors now since our bio can
* go away inside make_request
*/
sectors = bio_sectors(bio);
/* bio could be mergeable after passing to underlayer */
bio->bi_opf &= ~REQ_NOMERGE;
md_handle_request(mddev, bio);
cpu = part_stat_lock();
part_stat_inc(cpu, &mddev->gendisk->part0, ios[sgrp]);
part_stat_add(cpu, &mddev->gendisk->part0, sectors[sgrp], sectors);
part_stat_unlock();
return BLK_QC_T_NONE;
}
/* mddev_suspend makes sure no new requests are submitted
* to the device, and that any requests that have been submitted
* are completely handled.
* Once mddev_detach() is called and completes, the module will be
* completely unused.
*/
void mddev_suspend(struct mddev *mddev)
{
WARN_ON_ONCE(mddev->thread && current == mddev->thread->tsk);
lockdep_assert_held(&mddev->reconfig_mutex);
if (mddev->suspended++)
return;
synchronize_rcu();
wake_up(&mddev->sb_wait);
set_bit(MD_ALLOW_SB_UPDATE, &mddev->flags);
smp_mb__after_atomic();
wait_event(mddev->sb_wait, atomic_read(&mddev->active_io) == 0);
mddev->pers->quiesce(mddev, 1);
clear_bit_unlock(MD_ALLOW_SB_UPDATE, &mddev->flags);
wait_event(mddev->sb_wait, !test_bit(MD_UPDATING_SB, &mddev->flags));
del_timer_sync(&mddev->safemode_timer);
}
EXPORT_SYMBOL_GPL(mddev_suspend);
void mddev_resume(struct mddev *mddev)
{
lockdep_assert_held(&mddev->reconfig_mutex);
if (--mddev->suspended)
return;
wake_up(&mddev->sb_wait);
mddev->pers->quiesce(mddev, 0);
set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
md_wakeup_thread(mddev->thread);
md_wakeup_thread(mddev->sync_thread); /* possibly kick off a reshape */
}
EXPORT_SYMBOL_GPL(mddev_resume);
int mddev_congested(struct mddev *mddev, int bits)
{
struct md_personality *pers = mddev->pers;
int ret = 0;
rcu_read_lock();
if (mddev->suspended)
ret = 1;
else if (pers && pers->congested)
ret = pers->congested(mddev, bits);
rcu_read_unlock();
return ret;
}
EXPORT_SYMBOL_GPL(mddev_congested);
static int md_congested(void *data, int bits)
{
struct mddev *mddev = data;
return mddev_congested(mddev, bits);
}
/*
* Generic flush handling for md
*/
static void md_end_flush(struct bio *bio)
{
struct md_rdev *rdev = bio->bi_private;
struct mddev *mddev = rdev->mddev;
rdev_dec_pending(rdev, mddev);
if (atomic_dec_and_test(&mddev->flush_pending)) {
/* The pre-request flush has finished */
queue_work(md_wq, &mddev->flush_work);
}
bio_put(bio);
}
static void md_submit_flush_data(struct work_struct *ws);
static void submit_flushes(struct work_struct *ws)
{
struct mddev *mddev = container_of(ws, struct mddev, flush_work);
struct md_rdev *rdev;
mddev->start_flush = ktime_get_boottime();
INIT_WORK(&mddev->flush_work, md_submit_flush_data);
atomic_set(&mddev->flush_pending, 1);
rcu_read_lock();
rdev_for_each_rcu(rdev, mddev)
if (rdev->raid_disk >= 0 &&
!test_bit(Faulty, &rdev->flags)) {
/* Take two references, one is dropped
* when request finishes, one after
* we reclaim rcu_read_lock
*/
struct bio *bi;
atomic_inc(&rdev->nr_pending);
atomic_inc(&rdev->nr_pending);
rcu_read_unlock();
bi = bio_alloc_mddev(GFP_NOIO, 0, mddev);
bi->bi_end_io = md_end_flush;
bi->bi_private = rdev;
bio_set_dev(bi, rdev->bdev);
bi->bi_opf = REQ_OP_WRITE | REQ_PREFLUSH;
atomic_inc(&mddev->flush_pending);
submit_bio(bi);
rcu_read_lock();
rdev_dec_pending(rdev, mddev);
}
rcu_read_unlock();
if (atomic_dec_and_test(&mddev->flush_pending))
queue_work(md_wq, &mddev->flush_work);
}
static void md_submit_flush_data(struct work_struct *ws)
{
struct mddev *mddev = container_of(ws, struct mddev, flush_work);
struct bio *bio = mddev->flush_bio;
/*
* must reset flush_bio before calling into md_handle_request to avoid a
* deadlock, because other bios passed md_handle_request suspend check
* could wait for this and below md_handle_request could wait for those
* bios because of suspend check
*/
mddev->last_flush = mddev->start_flush;
mddev->flush_bio = NULL;
wake_up(&mddev->sb_wait);
if (bio->bi_iter.bi_size == 0) {
/* an empty barrier - all done */
bio_endio(bio);
} else {
bio->bi_opf &= ~REQ_PREFLUSH;
md_handle_request(mddev, bio);
}
}
/*
* Manages consolidation of flushes and submitting any flushes needed for
* a bio with REQ_PREFLUSH. Returns true if the bio is finished or is
* being finished in another context. Returns false if the flushing is
* complete but still needs the I/O portion of the bio to be processed.
*/
bool md_flush_request(struct mddev *mddev, struct bio *bio)
{
ktime_t start = ktime_get_boottime();
spin_lock_irq(&mddev->lock);
wait_event_lock_irq(mddev->sb_wait,
!mddev->flush_bio ||
ktime_after(mddev->last_flush, start),
mddev->lock);
if (!ktime_after(mddev->last_flush, start)) {
WARN_ON(mddev->flush_bio);
mddev->flush_bio = bio;
bio = NULL;
}
spin_unlock_irq(&mddev->lock);
if (!bio) {
INIT_WORK(&mddev->flush_work, submit_flushes);
queue_work(md_wq, &mddev->flush_work);
} else {
/* flush was performed for some other bio while we waited. */
if (bio->bi_iter.bi_size == 0)
/* an empty barrier - all done */
bio_endio(bio);
else {
bio->bi_opf &= ~REQ_PREFLUSH;
return false;
}
}
return true;
}
EXPORT_SYMBOL(md_flush_request);
static inline struct mddev *mddev_get(struct mddev *mddev)
{
atomic_inc(&mddev->active);
return mddev;
}
static void mddev_delayed_delete(struct work_struct *ws);
static void mddev_put(struct mddev *mddev)
{
if (!atomic_dec_and_lock(&mddev->active, &all_mddevs_lock))
return;
if (!mddev->raid_disks && list_empty(&mddev->disks) &&
mddev->ctime == 0 && !mddev->hold_active) {
/* Array is not configured at all, and not held active,
* so destroy it */
list_del_init(&mddev->all_mddevs);
/*
* Call queue_work inside the spinlock so that
* flush_workqueue() after mddev_find will succeed in waiting
* for the work to be done.
*/
INIT_WORK(&mddev->del_work, mddev_delayed_delete);
queue_work(md_misc_wq, &mddev->del_work);
}
spin_unlock(&all_mddevs_lock);
}
static void md_safemode_timeout(struct timer_list *t);
void mddev_init(struct mddev *mddev)
{
kobject_init(&mddev->kobj, &md_ktype);
mutex_init(&mddev->open_mutex);
mutex_init(&mddev->reconfig_mutex);
mutex_init(&mddev->bitmap_info.mutex);
INIT_LIST_HEAD(&mddev->disks);
INIT_LIST_HEAD(&mddev->all_mddevs);
timer_setup(&mddev->safemode_timer, md_safemode_timeout, 0);
atomic_set(&mddev->active, 1);
atomic_set(&mddev->openers, 0);
atomic_set(&mddev->active_io, 0);
spin_lock_init(&mddev->lock);
atomic_set(&mddev->flush_pending, 0);
init_waitqueue_head(&mddev->sb_wait);
init_waitqueue_head(&mddev->recovery_wait);
mddev->reshape_position = MaxSector;
mddev->reshape_backwards = 0;
mddev->last_sync_action = "none";
mddev->resync_min = 0;
mddev->resync_max = MaxSector;
mddev->level = LEVEL_NONE;
}
EXPORT_SYMBOL_GPL(mddev_init);
static struct mddev *mddev_find(dev_t unit)
{
struct mddev *mddev, *new = NULL;
if (unit && MAJOR(unit) != MD_MAJOR)
unit &= ~((1<<MdpMinorShift)-1);
retry:
spin_lock(&all_mddevs_lock);
if (unit) {
list_for_each_entry(mddev, &all_mddevs, all_mddevs)
if (mddev->unit == unit) {
mddev_get(mddev);
spin_unlock(&all_mddevs_lock);
kfree(new);
return mddev;
}
if (new) {
list_add(&new->all_mddevs, &all_mddevs);
spin_unlock(&all_mddevs_lock);
new->hold_active = UNTIL_IOCTL;
return new;
}
} else if (new) {
/* find an unused unit number */
static int next_minor = 512;
int start = next_minor;
int is_free = 0;
int dev = 0;
while (!is_free) {
dev = MKDEV(MD_MAJOR, next_minor);
next_minor++;
if (next_minor > MINORMASK)
next_minor = 0;
if (next_minor == start) {
/* Oh dear, all in use. */
spin_unlock(&all_mddevs_lock);
kfree(new);
return NULL;
}
is_free = 1;
list_for_each_entry(mddev, &all_mddevs, all_mddevs)
if (mddev->unit == dev) {
is_free = 0;
break;
}
}
new->unit = dev;
new->md_minor = MINOR(dev);
new->hold_active = UNTIL_STOP;
list_add(&new->all_mddevs, &all_mddevs);
spin_unlock(&all_mddevs_lock);
return new;
}
spin_unlock(&all_mddevs_lock);
new = kzalloc(sizeof(*new), GFP_KERNEL);
if (!new)
return NULL;
new->unit = unit;
if (MAJOR(unit) == MD_MAJOR)
new->md_minor = MINOR(unit);
else
new->md_minor = MINOR(unit) >> MdpMinorShift;
mddev_init(new);
goto retry;
}
static struct attribute_group md_redundancy_group;
void mddev_unlock(struct mddev *mddev)
{
if (mddev->to_remove) {
/* These cannot be removed under reconfig_mutex as
* an access to the files will try to take reconfig_mutex
* while holding the file unremovable, which leads to
* a deadlock.
* So hold set sysfs_active while the remove in happeing,
* and anything else which might set ->to_remove or my
* otherwise change the sysfs namespace will fail with
* -EBUSY if sysfs_active is still set.
* We set sysfs_active under reconfig_mutex and elsewhere
* test it under the same mutex to ensure its correct value
* is seen.
*/
struct attribute_group *to_remove = mddev->to_remove;
mddev->to_remove = NULL;
mddev->sysfs_active = 1;
mutex_unlock(&mddev->reconfig_mutex);
if (mddev->kobj.sd) {
if (to_remove != &md_redundancy_group)
sysfs_remove_group(&mddev->kobj, to_remove);
if (mddev->pers == NULL ||
mddev->pers->sync_request == NULL) {
sysfs_remove_group(&mddev->kobj, &md_redundancy_group);
if (mddev->sysfs_action)
sysfs_put(mddev->sysfs_action);
mddev->sysfs_action = NULL;
}
}
mddev->sysfs_active = 0;
} else
mutex_unlock(&mddev->reconfig_mutex);
/* As we've dropped the mutex we need a spinlock to
* make sure the thread doesn't disappear
*/
spin_lock(&pers_lock);
md_wakeup_thread(mddev->thread);
wake_up(&mddev->sb_wait);
spin_unlock(&pers_lock);
}
EXPORT_SYMBOL_GPL(mddev_unlock);
struct md_rdev *md_find_rdev_nr_rcu(struct mddev *mddev, int nr)
{
struct md_rdev *rdev;
rdev_for_each_rcu(rdev, mddev)
if (rdev->desc_nr == nr)
return rdev;
return NULL;
}
EXPORT_SYMBOL_GPL(md_find_rdev_nr_rcu);
static struct md_rdev *find_rdev(struct mddev *mddev, dev_t dev)
{
struct md_rdev *rdev;
rdev_for_each(rdev, mddev)
if (rdev->bdev->bd_dev == dev)
return rdev;
return NULL;
}
struct md_rdev *md_find_rdev_rcu(struct mddev *mddev, dev_t dev)
{
struct md_rdev *rdev;
rdev_for_each_rcu(rdev, mddev)
if (rdev->bdev->bd_dev == dev)
return rdev;
return NULL;
}
EXPORT_SYMBOL_GPL(md_find_rdev_rcu);
static struct md_personality *find_pers(int level, char *clevel)
{
struct md_personality *pers;
list_for_each_entry(pers, &pers_list, list) {
if (level != LEVEL_NONE && pers->level == level)
return pers;
if (strcmp(pers->name, clevel)==0)
return pers;
}
return NULL;
}
/* return the offset of the super block in 512byte sectors */
static inline sector_t calc_dev_sboffset(struct md_rdev *rdev)
{
sector_t num_sectors = i_size_read(rdev->bdev->bd_inode) / 512;
return MD_NEW_SIZE_SECTORS(num_sectors);
}
static int alloc_disk_sb(struct md_rdev *rdev)
{
rdev->sb_page = alloc_page(GFP_KERNEL);
if (!rdev->sb_page)
return -ENOMEM;
return 0;
}
void md_rdev_clear(struct md_rdev *rdev)
{
if (rdev->sb_page) {
put_page(rdev->sb_page);
rdev->sb_loaded = 0;
rdev->sb_page = NULL;
rdev->sb_start = 0;
rdev->sectors = 0;
}
if (rdev->bb_page) {
put_page(rdev->bb_page);
rdev->bb_page = NULL;
}
badblocks_exit(&rdev->badblocks);
}
EXPORT_SYMBOL_GPL(md_rdev_clear);
static void super_written(struct bio *bio)
{
struct md_rdev *rdev = bio->bi_private;
struct mddev *mddev = rdev->mddev;
if (bio->bi_status) {
pr_err("md: super_written gets error=%d\n", bio->bi_status);
md_error(mddev, rdev);
if (!test_bit(Faulty, &rdev->flags)
&& (bio->bi_opf & MD_FAILFAST)) {
set_bit(MD_SB_NEED_REWRITE, &mddev->sb_flags);
set_bit(LastDev, &rdev->flags);
}
} else
clear_bit(LastDev, &rdev->flags);
if (atomic_dec_and_test(&mddev->pending_writes))
wake_up(&mddev->sb_wait);
rdev_dec_pending(rdev, mddev);
bio_put(bio);
}
void md_super_write(struct mddev *mddev, struct md_rdev *rdev,
sector_t sector, int size, struct page *page)
{
/* write first size bytes of page to sector of rdev
* Increment mddev->pending_writes before returning
* and decrement it on completion, waking up sb_wait
* if zero is reached.
* If an error occurred, call md_error
*/
struct bio *bio;
int ff = 0;
if (!page)
return;
if (test_bit(Faulty, &rdev->flags))
return;
bio = md_bio_alloc_sync(mddev);
atomic_inc(&rdev->nr_pending);
bio_set_dev(bio, rdev->meta_bdev ? rdev->meta_bdev : rdev->bdev);
bio->bi_iter.bi_sector = sector;
bio_add_page(bio, page, size, 0);
bio->bi_private = rdev;
bio->bi_end_io = super_written;
if (test_bit(MD_FAILFAST_SUPPORTED, &mddev->flags) &&
test_bit(FailFast, &rdev->flags) &&
!test_bit(LastDev, &rdev->flags))
ff = MD_FAILFAST;
bio->bi_opf = REQ_OP_WRITE | REQ_SYNC | REQ_PREFLUSH | REQ_FUA | ff;
atomic_inc(&mddev->pending_writes);
submit_bio(bio);
}
int md_super_wait(struct mddev *mddev)
{
/* wait for all superblock writes that were scheduled to complete */
wait_event(mddev->sb_wait, atomic_read(&mddev->pending_writes)==0);
if (test_and_clear_bit(MD_SB_NEED_REWRITE, &mddev->sb_flags))
return -EAGAIN;
return 0;
}
int sync_page_io(struct md_rdev *rdev, sector_t sector, int size,
struct page *page, int op, int op_flags, bool metadata_op)
{
struct bio *bio = md_bio_alloc_sync(rdev->mddev);
int ret;
if (metadata_op && rdev->meta_bdev)
bio_set_dev(bio, rdev->meta_bdev);
else
bio_set_dev(bio, rdev->bdev);
bio_set_op_attrs(bio, op, op_flags);
if (metadata_op)
bio->bi_iter.bi_sector = sector + rdev->sb_start;
else if (rdev->mddev->reshape_position != MaxSector &&
(rdev->mddev->reshape_backwards ==
(sector >= rdev->mddev->reshape_position)))
bio->bi_iter.bi_sector = sector + rdev->new_data_offset;
else
bio->bi_iter.bi_sector = sector + rdev->data_offset;
bio_add_page(bio, page, size, 0);
submit_bio_wait(bio);
ret = !bio->bi_status;
bio_put(bio);
return ret;
}
EXPORT_SYMBOL_GPL(sync_page_io);
static int read_disk_sb(struct md_rdev *rdev, int size)
{
char b[BDEVNAME_SIZE];
if (rdev->sb_loaded)
return 0;
if (!sync_page_io(rdev, 0, size, rdev->sb_page, REQ_OP_READ, 0, true))
goto fail;
rdev->sb_loaded = 1;
return 0;
fail:
pr_err("md: disabled device %s, could not read superblock.\n",
bdevname(rdev->bdev,b));
return -EINVAL;
}
static int md_uuid_equal(mdp_super_t *sb1, mdp_super_t *sb2)
{
return sb1->set_uuid0 == sb2->set_uuid0 &&
sb1->set_uuid1 == sb2->set_uuid1 &&
sb1->set_uuid2 == sb2->set_uuid2 &&
sb1->set_uuid3 == sb2->set_uuid3;
}
static int md_sb_equal(mdp_super_t *sb1, mdp_super_t *sb2)
{
int ret;
mdp_super_t *tmp1, *tmp2;
tmp1 = kmalloc(sizeof(*tmp1),GFP_KERNEL);
tmp2 = kmalloc(sizeof(*tmp2),GFP_KERNEL);
if (!tmp1 || !tmp2) {
ret = 0;
goto abort;
}
*tmp1 = *sb1;
*tmp2 = *sb2;
/*
* nr_disks is not constant
*/
tmp1->nr_disks = 0;
tmp2->nr_disks = 0;
ret = (memcmp(tmp1, tmp2, MD_SB_GENERIC_CONSTANT_WORDS * 4) == 0);
abort:
kfree(tmp1);
kfree(tmp2);
return ret;
}
static u32 md_csum_fold(u32 csum)
{
csum = (csum & 0xffff) + (csum >> 16);
return (csum & 0xffff) + (csum >> 16);
}
static unsigned int calc_sb_csum(mdp_super_t *sb)
{
u64 newcsum = 0;
u32 *sb32 = (u32*)sb;
int i;
unsigned int disk_csum, csum;
disk_csum = sb->sb_csum;
sb->sb_csum = 0;
for (i = 0; i < MD_SB_BYTES/4 ; i++)
newcsum += sb32[i];
csum = (newcsum & 0xffffffff) + (newcsum>>32);
#ifdef CONFIG_ALPHA
/* This used to use csum_partial, which was wrong for several
* reasons including that different results are returned on
* different architectures. It isn't critical that we get exactly
* the same return value as before (we always csum_fold before
* testing, and that removes any differences). However as we
* know that csum_partial always returned a 16bit value on
* alphas, do a fold to maximise conformity to previous behaviour.
*/
sb->sb_csum = md_csum_fold(disk_csum);
#else
sb->sb_csum = disk_csum;
#endif
return csum;
}
/*
* Handle superblock details.
* We want to be able to handle multiple superblock formats
* so we have a common interface to them all, and an array of
* different handlers.
* We rely on user-space to write the initial superblock, and support
* reading and updating of superblocks.
* Interface methods are:
* int load_super(struct md_rdev *dev, struct md_rdev *refdev, int minor_version)
* loads and validates a superblock on dev.
* if refdev != NULL, compare superblocks on both devices
* Return:
* 0 - dev has a superblock that is compatible with refdev
* 1 - dev has a superblock that is compatible and newer than refdev
* so dev should be used as the refdev in future
* -EINVAL superblock incompatible or invalid
* -othererror e.g. -EIO
*
* int validate_super(struct mddev *mddev, struct md_rdev *dev)
* Verify that dev is acceptable into mddev.
* The first time, mddev->raid_disks will be 0, and data from
* dev should be merged in. Subsequent calls check that dev
* is new enough. Return 0 or -EINVAL
*
* void sync_super(struct mddev *mddev, struct md_rdev *dev)
* Update the superblock for rdev with data in mddev
* This does not write to disc.
*
*/
struct super_type {
char *name;
struct module *owner;
int (*load_super)(struct md_rdev *rdev,
struct md_rdev *refdev,
int minor_version);
int (*validate_super)(struct mddev *mddev,
struct md_rdev *rdev);
void (*sync_super)(struct mddev *mddev,
struct md_rdev *rdev);
unsigned long long (*rdev_size_change)(struct md_rdev *rdev,
sector_t num_sectors);
int (*allow_new_offset)(struct md_rdev *rdev,
unsigned long long new_offset);
};
/*
* Check that the given mddev has no bitmap.
*
* This function is called from the run method of all personalities that do not
* support bitmaps. It prints an error message and returns non-zero if mddev
* has a bitmap. Otherwise, it returns 0.
*
*/
int md_check_no_bitmap(struct mddev *mddev)
{
if (!mddev->bitmap_info.file && !mddev->bitmap_info.offset)
return 0;
pr_warn("%s: bitmaps are not supported for %s\n",
mdname(mddev), mddev->pers->name);
return 1;
}
EXPORT_SYMBOL(md_check_no_bitmap);
/*
* load_super for 0.90.0
*/
static int super_90_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor_version)
{
char b[BDEVNAME_SIZE], b2[BDEVNAME_SIZE];
mdp_super_t *sb;
int ret;
/*
* Calculate the position of the superblock (512byte sectors),
* it's at the end of the disk.
*
* It also happens to be a multiple of 4Kb.
*/
rdev->sb_start = calc_dev_sboffset(rdev);
ret = read_disk_sb(rdev, MD_SB_BYTES);
if (ret)
return ret;
ret = -EINVAL;
bdevname(rdev->bdev, b);
sb = page_address(rdev->sb_page);
if (sb->md_magic != MD_SB_MAGIC) {
pr_warn("md: invalid raid superblock magic on %s\n", b);
goto abort;
}
if (sb->major_version != 0 ||
sb->minor_version < 90 ||
sb->minor_version > 91) {
pr_warn("Bad version number %d.%d on %s\n",
sb->major_version, sb->minor_version, b);
goto abort;
}
if (sb->raid_disks <= 0)
goto abort;
if (md_csum_fold(calc_sb_csum(sb)) != md_csum_fold(sb->sb_csum)) {
pr_warn("md: invalid superblock checksum on %s\n", b);
goto abort;
}
rdev->preferred_minor = sb->md_minor;
rdev->data_offset = 0;
rdev->new_data_offset = 0;
rdev->sb_size = MD_SB_BYTES;
rdev->badblocks.shift = -1;
if (sb->level == LEVEL_MULTIPATH)
rdev->desc_nr = -1;
else
rdev->desc_nr = sb->this_disk.number;
if (!refdev) {
ret = 1;
} else {
__u64 ev1, ev2;
mdp_super_t *refsb = page_address(refdev->sb_page);
if (!md_uuid_equal(refsb, sb)) {
pr_warn("md: %s has different UUID to %s\n",
b, bdevname(refdev->bdev,b2));
goto abort;
}
if (!md_sb_equal(refsb, sb)) {
pr_warn("md: %s has same UUID but different superblock to %s\n",
b, bdevname(refdev->bdev, b2));
goto abort;
}
ev1 = md_event(sb);
ev2 = md_event(refsb);
if (ev1 > ev2)
ret = 1;
else
ret = 0;
}
rdev->sectors = rdev->sb_start;
/* Limit to 4TB as metadata cannot record more than that.
* (not needed for Linear and RAID0 as metadata doesn't
* record this size)
*/
if (IS_ENABLED(CONFIG_LBDAF) && (u64)rdev->sectors >= (2ULL << 32) &&
sb->level >= 1)
rdev->sectors = (sector_t)(2ULL << 32) - 2;
if (rdev->sectors < ((sector_t)sb->size) * 2 && sb->level >= 1)
/* "this cannot possibly happen" ... */
ret = -EINVAL;
abort:
return ret;
}
/*
* validate_super for 0.90.0
*/
static int super_90_validate(struct mddev *mddev, struct md_rdev *rdev)
{
mdp_disk_t *desc;
mdp_super_t *sb = page_address(rdev->sb_page);
__u64 ev1 = md_event(sb);
rdev->raid_disk = -1;
clear_bit(Faulty, &rdev->flags);
clear_bit(In_sync, &rdev->flags);
clear_bit(Bitmap_sync, &rdev->flags);
clear_bit(WriteMostly, &rdev->flags);
if (mddev->raid_disks == 0) {
mddev->major_version = 0;
mddev->minor_version = sb->minor_version;
mddev->patch_version = sb->patch_version;
mddev->external = 0;
mddev->chunk_sectors = sb->chunk_size >> 9;
mddev->ctime = sb->ctime;
mddev->utime = sb->utime;
mddev->level = sb->level;
mddev->clevel[0] = 0;
mddev->layout = sb->layout;
mddev->raid_disks = sb->raid_disks;
mddev->dev_sectors = ((sector_t)sb->size) * 2;
mddev->events = ev1;
mddev->bitmap_info.offset = 0;
mddev->bitmap_info.space = 0;
/* bitmap can use 60 K after the 4K superblocks */
mddev->bitmap_info.default_offset = MD_SB_BYTES >> 9;
mddev->bitmap_info.default_space = 64*2 - (MD_SB_BYTES >> 9);
mddev->reshape_backwards = 0;
if (mddev->minor_version >= 91) {
mddev->reshape_position = sb->reshape_position;
mddev->delta_disks = sb->delta_disks;
mddev->new_level = sb->new_level;
mddev->new_layout = sb->new_layout;
mddev->new_chunk_sectors = sb->new_chunk >> 9;
if (mddev->delta_disks < 0)
mddev->reshape_backwards = 1;
} else {
mddev->reshape_position = MaxSector;
mddev->delta_disks = 0;
mddev->new_level = mddev->level;
mddev->new_layout = mddev->layout;
mddev->new_chunk_sectors = mddev->chunk_sectors;
}
if (sb->state & (1<<MD_SB_CLEAN))
mddev->recovery_cp = MaxSector;
else {
if (sb->events_hi == sb->cp_events_hi &&
sb->events_lo == sb->cp_events_lo) {
mddev->recovery_cp = sb->recovery_cp;
} else
mddev->recovery_cp = 0;
}
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);
mddev->max_disks = MD_SB_DISKS;
if (sb->state & (1<<MD_SB_BITMAP_PRESENT) &&
mddev->bitmap_info.file == NULL) {
mddev->bitmap_info.offset =
mddev->bitmap_info.default_offset;
mddev->bitmap_info.space =
mddev->bitmap_info.default_space;
}
} else if (mddev->pers == NULL) {
/* Insist on good event counter while assembling, except
* for spares (which don't need an event count) */
++ev1;
if (sb->disks[rdev->desc_nr].state & (
(1<<MD_DISK_SYNC) | (1 << MD_DISK_ACTIVE)))
if (ev1 < mddev->events)
return -EINVAL;
} else if (mddev->bitmap) {
/* if adding to array with a bitmap, then we can accept an
* older device ... but not too old.
*/
if (ev1 < mddev->bitmap->events_cleared)
return 0;
if (ev1 < mddev->events)
set_bit(Bitmap_sync, &rdev->flags);
} else {
if (ev1 < mddev->events)
/* just a hot-add of a new device, leave raid_disk at -1 */
return 0;
}
if (mddev->level != LEVEL_MULTIPATH) {
desc = sb->disks + rdev->desc_nr;
if (desc->state & (1<<MD_DISK_FAULTY))
set_bit(Faulty, &rdev->flags);
else if (desc->state & (1<<MD_DISK_SYNC) /* &&
desc->raid_disk < mddev->raid_disks */) {
set_bit(In_sync, &rdev->flags);
rdev->raid_disk = desc->raid_disk;
rdev->saved_raid_disk = desc->raid_disk;
} else if (desc->state & (1<<MD_DISK_ACTIVE)) {
/* active but not in sync implies recovery up to
* reshape position. We don't know exactly where
* that is, so set to zero for now */
if (mddev->minor_version >= 91) {
rdev->recovery_offset = 0;
rdev->raid_disk = desc->raid_disk;
}
}
if (desc->state & (1<<MD_DISK_WRITEMOSTLY))
set_bit(WriteMostly, &rdev->flags);
if (desc->state & (1<<MD_DISK_FAILFAST))
set_bit(FailFast, &rdev->flags);
} else /* MULTIPATH are always insync */
set_bit(In_sync, &rdev->flags);
return 0;
}
/*
* sync_super for 0.90.0
*/
static void super_90_sync(struct mddev *mddev, struct md_rdev *rdev)
{
mdp_super_t *sb;
struct md_rdev *rdev2;
int next_spare = mddev->raid_disks;
/* make rdev->sb match mddev data..
*
* 1/ zero out disks
* 2/ Add info for each disk, keeping track of highest desc_nr (next_spare);
* 3/ any empty disks < next_spare become removed
*
* disks[0] gets initialised to REMOVED because
* we cannot be sure from other fields if it has
* been initialised or not.
*/
int i;
int active=0, working=0,failed=0,spare=0,nr_disks=0;
rdev->sb_size = MD_SB_BYTES;
sb = page_address(rdev->sb_page);
memset(sb, 0, sizeof(*sb));
sb->md_magic = MD_SB_MAGIC;
sb->major_version = mddev->major_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 = clamp_t(time64_t, mddev->ctime, 0, U32_MAX);
sb->level = mddev->level;
sb->size = mddev->dev_sectors / 2;
sb->raid_disks = mddev->raid_disks;
sb->md_minor = mddev->md_minor;
sb->not_persistent = 0;
sb->utime = clamp_t(time64_t, mddev->utime, 0, U32_MAX);
sb->state = 0;
sb->events_hi = (mddev->events>>32);
sb->events_lo = (u32)mddev->events;
if (mddev->reshape_position == MaxSector)
sb->minor_version = 90;
else {
sb->minor_version = 91;
sb->reshape_position = mddev->reshape_position;
sb->new_level = mddev->new_level;
sb->delta_disks = mddev->delta_disks;
sb->new_layout = mddev->new_layout;
sb->new_chunk = mddev->new_chunk_sectors << 9;
}
mddev->minor_version = sb->minor_version;
if (mddev->in_sync)
{
sb->recovery_cp = mddev->recovery_cp;
sb->cp_events_hi = (mddev->events>>32);
sb->cp_events_lo = (u32)mddev->events;
if (mddev->recovery_cp == MaxSector)
sb->state = (1<< MD_SB_CLEAN);
} else
sb->recovery_cp = 0;
sb->layout = mddev->layout;
sb->chunk_size = mddev->chunk_sectors << 9;
if (mddev->bitmap && mddev->bitmap_info.file == NULL)
sb->state |= (1<<MD_SB_BITMAP_PRESENT);
sb->disks[0].state = (1<<MD_DISK_REMOVED);
rdev_for_each(rdev2, mddev) {
mdp_disk_t *d;
int desc_nr;
int is_active = test_bit(In_sync, &rdev2->flags);
if (rdev2->raid_disk >= 0 &&
sb->minor_version >= 91)
/* we have nowhere to store the recovery_offset,
* but if it is not below the reshape_position,
* we can piggy-back on that.
*/
is_active = 1;
if (rdev2->raid_disk < 0 ||
test_bit(Faulty, &rdev2->flags))
is_active = 0;
if (is_active)
desc_nr = rdev2->raid_disk;
else
desc_nr = next_spare++;
rdev2->desc_nr = desc_nr;
d = &sb->disks[rdev2->desc_nr];
nr_disks++;
d->number = rdev2->desc_nr;
d->major = MAJOR(rdev2->bdev->bd_dev);
d->minor = MINOR(rdev2->bdev->bd_dev);
if (is_active)
d->raid_disk = rdev2->raid_disk;
else
d->raid_disk = rdev2->desc_nr; /* compatibility */
if (test_bit(Faulty, &rdev2->flags))
d->state = (1<<MD_DISK_FAULTY);
else if (is_active) {
d->state = (1<<MD_DISK_ACTIVE);
if (test_bit(In_sync, &rdev2->flags))
d->state |= (1<<MD_DISK_SYNC);
active++;
working++;
} else {
d->state = 0;
spare++;
working++;
}
if (test_bit(WriteMostly, &rdev2->flags))
d->state |= (1<<MD_DISK_WRITEMOSTLY);
if (test_bit(FailFast, &rdev2->flags))
d->state |= (1<<MD_DISK_FAILFAST);
}
/* now set the "removed" and "faulty" bits on any missing devices */
for (i=0 ; i < mddev->raid_disks ; i++) {
mdp_disk_t *d = &sb->disks[i];
if (d->state == 0 && d->number == 0) {
d->number = i;
d->raid_disk = i;
d->state = (1<<MD_DISK_REMOVED);
d->state |= (1<<MD_DISK_FAULTY);
failed++;
}
}
sb->nr_disks = nr_disks;
sb->active_disks = active;
sb->working_disks = working;
sb->failed_disks = failed;
sb->spare_disks = spare;
sb->this_disk = sb->disks[rdev->desc_nr];
sb->sb_csum = calc_sb_csum(sb);
}
/*
* rdev_size_change for 0.90.0
*/
static unsigned long long
super_90_rdev_size_change(struct md_rdev *rdev, sector_t num_sectors)
{
if (num_sectors && num_sectors < rdev->mddev->dev_sectors)
return 0; /* component must fit device */
if (rdev->mddev->bitmap_info.offset)
return 0; /* can't move bitmap */
rdev->sb_start = calc_dev_sboffset(rdev);
if (!num_sectors || num_sectors > rdev->sb_start)
num_sectors = rdev->sb_start;
/* Limit to 4TB as metadata cannot record more than that.
* 4TB == 2^32 KB, or 2*2^32 sectors.
*/
if (IS_ENABLED(CONFIG_LBDAF) && (u64)num_sectors >= (2ULL << 32) &&
rdev->mddev->level >= 1)
num_sectors = (sector_t)(2ULL << 32) - 2;
do {
md_super_write(rdev->mddev, rdev, rdev->sb_start, rdev->sb_size,
rdev->sb_page);
} while (md_super_wait(rdev->mddev) < 0);
return num_sectors;
}
static int
super_90_allow_new_offset(struct md_rdev *rdev, unsigned long long new_offset)
{
/* non-zero offset changes not possible with v0.90 */
return new_offset == 0;
}
/*
* version 1 superblock
*/
static __le32 calc_sb_1_csum(struct mdp_superblock_1 *sb)
{
__le32 disk_csum;
u32 csum;
unsigned long long newcsum;
int size = 256 + le32_to_cpu(sb->max_dev)*2;
__le32 *isuper = (__le32*)sb;
disk_csum = sb->sb_csum;
sb->sb_csum = 0;
newcsum = 0;
for (; size >= 4; size -= 4)
newcsum += le32_to_cpu(*isuper++);
if (size == 2)
newcsum += le16_to_cpu(*(__le16*) isuper);
csum = (newcsum & 0xffffffff) + (newcsum >> 32);
sb->sb_csum = disk_csum;
return cpu_to_le32(csum);
}
static int super_1_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor_version)
{
struct mdp_superblock_1 *sb;
int ret;
sector_t sb_start;
sector_t sectors;
char b[BDEVNAME_SIZE], b2[BDEVNAME_SIZE];
int bmask;
/*
* Calculate the position of the superblock in 512byte sectors.
* It is always aligned to a 4K boundary and
* depeding on minor_version, it can be:
* 0: At least 8K, but less than 12K, from end of device
* 1: At start of device
* 2: 4K from start of device.
*/
switch(minor_version) {
case 0:
sb_start = i_size_read(rdev->bdev->bd_inode) >> 9;
sb_start -= 8*2;
sb_start &= ~(sector_t)(4*2-1);
break;
case 1:
sb_start = 0;
break;
case 2:
sb_start = 8;
break;
default:
return -EINVAL;
}
rdev->sb_start = sb_start;
/* superblock is rarely larger than 1K, but it can be larger,
* and it is safe to read 4k, so we do that
*/
ret = read_disk_sb(rdev, 4096);
if (ret) return ret;
sb = page_address(rdev->sb_page);
if (sb->magic != cpu_to_le32(MD_SB_MAGIC) ||
sb->major_version != cpu_to_le32(1) ||
le32_to_cpu(sb->max_dev) > (4096-256)/2 ||
le64_to_cpu(sb->super_offset) != rdev->sb_start ||
(le32_to_cpu(sb->feature_map) & ~MD_FEATURE_ALL) != 0)
return -EINVAL;
if (calc_sb_1_csum(sb) != sb->sb_csum) {
pr_warn("md: invalid superblock checksum on %s\n",
bdevname(rdev->bdev,b));
return -EINVAL;
}
if (le64_to_cpu(sb->data_size) < 10) {
pr_warn("md: data_size too small on %s\n",
bdevname(rdev->bdev,b));
return -EINVAL;
}
if (sb->pad0 ||
sb->pad3[0] ||
memcmp(sb->pad3, sb->pad3+1, sizeof(sb->pad3) - sizeof(sb->pad3[1])))
/* Some padding is non-zero, might be a new feature */
return -EINVAL;
rdev->preferred_minor = 0xffff;
rdev->data_offset = le64_to_cpu(sb->data_offset);
rdev->new_data_offset = rdev->data_offset;
if ((le32_to_cpu(sb->feature_map) & MD_FEATURE_RESHAPE_ACTIVE) &&
(le32_to_cpu(sb->feature_map) & MD_FEATURE_NEW_OFFSET))
rdev->new_data_offset += (s32)le32_to_cpu(sb->new_offset);
atomic_set(&rdev->corrected_errors, le32_to_cpu(sb->cnt_corrected_read));
rdev->sb_size = le32_to_cpu(sb->max_dev) * 2 + 256;
bmask = queue_logical_block_size(rdev->bdev->bd_disk->queue)-1;
if (rdev->sb_size & bmask)
rdev->sb_size = (rdev->sb_size | bmask) + 1;
if (minor_version
&& rdev->data_offset < sb_start + (rdev->sb_size/512))
return -EINVAL;
if (minor_version
&& rdev->new_data_offset < sb_start + (rdev->sb_size/512))
return -EINVAL;
if (sb->level == cpu_to_le32(LEVEL_MULTIPATH))
rdev->desc_nr = -1;
else
rdev->desc_nr = le32_to_cpu(sb->dev_number);
if (!rdev->bb_page) {
rdev->bb_page = alloc_page(GFP_KERNEL);
if (!rdev->bb_page)
return -ENOMEM;
}
if ((le32_to_cpu(sb->feature_map) & MD_FEATURE_BAD_BLOCKS) &&
rdev->badblocks.count == 0) {
/* need to load the bad block list.
* Currently we limit it to one page.
*/
s32 offset;
sector_t bb_sector;
u64 *bbp;
int i;
int sectors = le16_to_cpu(sb->bblog_size);
if (sectors > (PAGE_SIZE / 512))
return -EINVAL;
offset = le32_to_cpu(sb->bblog_offset);
if (offset == 0)
return -EINVAL;
bb_sector = (long long)offset;
if (!sync_page_io(rdev, bb_sector, sectors << 9,
rdev->bb_page, REQ_OP_READ, 0, true))
return -EIO;
bbp = (u64 *)page_address(rdev->bb_page);
rdev->badblocks.shift = sb->bblog_shift;
for (i = 0 ; i < (sectors << (9-3)) ; i++, bbp++) {
u64 bb = le64_to_cpu(*bbp);
int count = bb & (0x3ff);
u64 sector = bb >> 10;
sector <<= sb->bblog_shift;
count <<= sb->bblog_shift;
if (bb + 1 == 0)
break;
if (badblocks_set(&rdev->badblocks, sector, count, 1))
return -EINVAL;
}
} else if (sb->bblog_offset != 0)
rdev->badblocks.shift = 0;
if ((le32_to_cpu(sb->feature_map) &
(MD_FEATURE_PPL | MD_FEATURE_MULTIPLE_PPLS))) {
rdev->ppl.offset = (__s16)le16_to_cpu(sb->ppl.offset);
rdev->ppl.size = le16_to_cpu(sb->ppl.size);
rdev->ppl.sector = rdev->sb_start + rdev->ppl.offset;
}
if (!refdev) {
ret = 1;
} else {
__u64 ev1, ev2;
struct mdp_superblock_1 *refsb = page_address(refdev->sb_page);
if (memcmp(sb->set_uuid, refsb->set_uuid, 16) != 0 ||
sb->level != refsb->level ||
sb->layout != refsb->layout ||
sb->chunksize != refsb->chunksize) {
pr_warn("md: %s has strangely different superblock to %s\n",
bdevname(rdev->bdev,b),
bdevname(refdev->bdev,b2));
return -EINVAL;
}
ev1 = le64_to_cpu(sb->events);
ev2 = le64_to_cpu(refsb->events);
if (ev1 > ev2)
ret = 1;
else
ret = 0;
}
if (minor_version) {
sectors = (i_size_read(rdev->bdev->bd_inode) >> 9);
sectors -= rdev->data_offset;
} else
sectors = rdev->sb_start;
if (sectors < le64_to_cpu(sb->data_size))
return -EINVAL;
rdev->sectors = le64_to_cpu(sb->data_size);
return ret;
}
static int super_1_validate(struct mddev *mddev, struct md_rdev *rdev)
{
struct mdp_superblock_1 *sb = page_address(rdev->sb_page);
__u64 ev1 = le64_to_cpu(sb->events);
rdev->raid_disk = -1;
clear_bit(Faulty, &rdev->flags);
clear_bit(In_sync, &rdev->flags);
clear_bit(Bitmap_sync, &rdev->flags);
clear_bit(WriteMostly, &rdev->flags);
if (mddev->raid_disks == 0) {
mddev->major_version = 1;
mddev->patch_version = 0;
mddev->external = 0;
mddev->chunk_sectors = le32_to_cpu(sb->chunksize);
mddev->ctime = le64_to_cpu(sb->ctime);
mddev->utime = le64_to_cpu(sb->utime);
mddev->level = le32_to_cpu(sb->level);
mddev->clevel[0] = 0;
mddev->layout = le32_to_cpu(sb->layout);
mddev->raid_disks = le32_to_cpu(sb->raid_disks);
mddev->dev_sectors = le64_to_cpu(sb->size);
mddev->events = ev1;
mddev->bitmap_info.offset = 0;
mddev->bitmap_info.space = 0;
/* Default location for bitmap is 1K after superblock
* using 3K - total of 4K
*/
mddev->bitmap_info.default_offset = 1024 >> 9;
mddev->bitmap_info.default_space = (4096-1024) >> 9;
mddev->reshape_backwards = 0;
mddev->recovery_cp = le64_to_cpu(sb->resync_offset);
memcpy(mddev->uuid, sb->set_uuid, 16);
mddev->max_disks = (4096-256)/2;
if ((le32_to_cpu(sb->feature_map) & MD_FEATURE_BITMAP_OFFSET) &&
mddev->bitmap_info.file == NULL) {
mddev->bitmap_info.offset =
(__s32)le32_to_cpu(sb->bitmap_offset);
/* Metadata doesn't record how much space is available.
* For 1.0, we assume we can use up to the superblock
* if before, else to 4K beyond superblock.
* For others, assume no change is possible.
*/
if (mddev->minor_version > 0)
mddev->bitmap_info.space = 0;
else if (mddev->bitmap_info.offset > 0)
mddev->bitmap_info.space =
8 - mddev->bitmap_info.offset;
else
mddev->bitmap_info.space =
-mddev->bitmap_info.offset;
}
if ((le32_to_cpu(sb->feature_map) & MD_FEATURE_RESHAPE_ACTIVE)) {
mddev->reshape_position = le64_to_cpu(sb->reshape_position);
mddev->delta_disks = le32_to_cpu(sb->delta_disks);
mddev->new_level = le32_to_cpu(sb->new_level);
mddev->new_layout = le32_to_cpu(sb->new_layout);
mddev->new_chunk_sectors = le32_to_cpu(sb->new_chunk);
if (mddev->delta_disks < 0 ||
(mddev->delta_disks == 0 &&
(le32_to_cpu(sb->feature_map)
& MD_FEATURE_RESHAPE_BACKWARDS)))
mddev->reshape_backwards = 1;
} else {
mddev->reshape_position = MaxSector;
mddev->delta_disks = 0;
mddev->new_level = mddev->level;
mddev->new_layout = mddev->layout;
mddev->new_chunk_sectors = mddev->chunk_sectors;
}
if (le32_to_cpu(sb->feature_map) & MD_FEATURE_JOURNAL)
set_bit(MD_HAS_JOURNAL, &mddev->flags);
if (le32_to_cpu(sb->feature_map) &
(MD_FEATURE_PPL | MD_FEATURE_MULTIPLE_PPLS)) {
if (le32_to_cpu(sb->feature_map) &
(MD_FEATURE_BITMAP_OFFSET | MD_FEATURE_JOURNAL))
return -EINVAL;
if ((le32_to_cpu(sb->feature_map) & MD_FEATURE_PPL) &&
(le32_to_cpu(sb->feature_map) &
MD_FEATURE_MULTIPLE_PPLS))
return -EINVAL;
set_bit(MD_HAS_PPL, &mddev->flags);
}
} else if (mddev->pers == NULL) {
/* Insist of good event counter while assembling, except for
* spares (which don't need an event count) */
++ev1;
if (rdev->desc_nr >= 0 &&
rdev->desc_nr < le32_to_cpu(sb->max_dev) &&
(le16_to_cpu(sb->dev_roles[rdev->desc_nr]) < MD_DISK_ROLE_MAX ||
le16_to_cpu(sb->dev_roles[rdev->desc_nr]) == MD_DISK_ROLE_JOURNAL))
if (ev1 < mddev->events)
return -EINVAL;
} else if (mddev->bitmap) {
/* If adding to array with a bitmap, then we can accept an
* older device, but not too old.
*/
if (ev1 < mddev->bitmap->events_cleared)
return 0;
if (ev1 < mddev->events)
set_bit(Bitmap_sync, &rdev->flags);
} else {
if (ev1 < mddev->events)
/* just a hot-add of a new device, leave raid_disk at -1 */
return 0;
}
if (mddev->level != LEVEL_MULTIPATH) {
int role;
if (rdev->desc_nr < 0 ||
rdev->desc_nr >= le32_to_cpu(sb->max_dev)) {
role = MD_DISK_ROLE_SPARE;
rdev->desc_nr = -1;
} else
role = le16_to_cpu(sb->dev_roles[rdev->desc_nr]);
switch(role) {
case MD_DISK_ROLE_SPARE: /* spare */
break;
case MD_DISK_ROLE_FAULTY: /* faulty */
set_bit(Faulty, &rdev->flags);
break;
case MD_DISK_ROLE_JOURNAL: /* journal device */
if (!(le32_to_cpu(sb->feature_map) & MD_FEATURE_JOURNAL)) {
/* journal device without journal feature */
pr_warn("md: journal device provided without journal feature, ignoring the device\n");
return -EINVAL;
}
set_bit(Journal, &rdev->flags);
rdev->journal_tail = le64_to_cpu(sb->journal_tail);
rdev->raid_disk = 0;
break;
default:
rdev->saved_raid_disk = role;
if ((le32_to_cpu(sb->feature_map) &
MD_FEATURE_RECOVERY_OFFSET)) {
rdev->recovery_offset = le64_to_cpu(sb->recovery_offset);
if (!(le32_to_cpu(sb->feature_map) &
MD_FEATURE_RECOVERY_BITMAP))
rdev->saved_raid_disk = -1;
} else {
/*
* If the array is FROZEN, then the device can't
* be in_sync with rest of array.
*/
if (!test_bit(MD_RECOVERY_FROZEN,
&mddev->recovery))
set_bit(In_sync, &rdev->flags);
}
rdev->raid_disk = role;
break;
}
if (sb->devflags & WriteMostly1)
set_bit(WriteMostly, &rdev->flags);
if (sb->devflags & FailFast1)
set_bit(FailFast, &rdev->flags);
if (le32_to_cpu(sb->feature_map) & MD_FEATURE_REPLACEMENT)
set_bit(Replacement, &rdev->flags);
} else /* MULTIPATH are always insync */
set_bit(In_sync, &rdev->flags);
return 0;
}
static void super_1_sync(struct mddev *mddev, struct md_rdev *rdev)
{
struct mdp_superblock_1 *sb;
struct md_rdev *rdev2;
int max_dev, i;
/* make rdev->sb match mddev and rdev data. */
sb = page_address(rdev->sb_page);
sb->feature_map = 0;
sb->pad0 = 0;
sb->recovery_offset = cpu_to_le64(0);
memset(sb->pad3, 0, sizeof(sb->pad3));
sb->utime = cpu_to_le64((__u64)mddev->utime);
sb->events = cpu_to_le64(mddev->events);
if (mddev->in_sync)
sb->resync_offset = cpu_to_le64(mddev->recovery_cp);
else if (test_bit(MD_JOURNAL_CLEAN, &mddev->flags))
sb->resync_offset = cpu_to_le64(MaxSector);
else
sb->resync_offset = cpu_to_le64(0);
sb->cnt_corrected_read = cpu_to_le32(atomic_read(&rdev->corrected_errors));
sb->raid_disks = cpu_to_le32(mddev->raid_disks);
sb->size = cpu_to_le64(mddev->dev_sectors);
sb->chunksize = cpu_to_le32(mddev->chunk_sectors);
sb->level = cpu_to_le32(mddev->level);
sb->layout = cpu_to_le32(mddev->layout);
if (test_bit(FailFast, &rdev->flags))
sb->devflags |= FailFast1;
else
sb->devflags &= ~FailFast1;
if (test_bit(WriteMostly, &rdev->flags))
sb->devflags |= WriteMostly1;
else
sb->devflags &= ~WriteMostly1;
sb->data_offset = cpu_to_le64(rdev->data_offset);
sb->data_size = cpu_to_le64(rdev->sectors);
if (mddev->bitmap && mddev->bitmap_info.file == NULL) {
sb->bitmap_offset = cpu_to_le32((__u32)mddev->bitmap_info.offset);
sb->feature_map = cpu_to_le32(MD_FEATURE_BITMAP_OFFSET);
}
if (rdev->raid_disk >= 0 && !test_bit(Journal, &rdev->flags) &&
!test_bit(In_sync, &rdev->flags)) {
sb->feature_map |=
cpu_to_le32(MD_FEATURE_RECOVERY_OFFSET);
sb->recovery_offset =
cpu_to_le64(rdev->recovery_offset);
if (rdev->saved_raid_disk >= 0 && mddev->bitmap)
sb->feature_map |=
cpu_to_le32(MD_FEATURE_RECOVERY_BITMAP);
}
/* Note: recovery_offset and journal_tail share space */
if (test_bit(Journal, &rdev->flags))
sb->journal_tail = cpu_to_le64(rdev->journal_tail);
if (test_bit(Replacement, &rdev->flags))
sb->feature_map |=
cpu_to_le32(MD_FEATURE_REPLACEMENT);
if (mddev->reshape_position != MaxSector) {
sb->feature_map |= cpu_to_le32(MD_FEATURE_RESHAPE_ACTIVE);
sb->reshape_position = cpu_to_le64(mddev->reshape_position);
sb->new_layout = cpu_to_le32(mddev->new_layout);
sb->delta_disks = cpu_to_le32(mddev->delta_disks);
sb->new_level = cpu_to_le32(mddev->new_level);
sb->new_chunk = cpu_to_le32(mddev->new_chunk_sectors);
if (mddev->delta_disks == 0 &&
mddev->reshape_backwards)
sb->feature_map
|= cpu_to_le32(MD_FEATURE_RESHAPE_BACKWARDS);
if (rdev->new_data_offset != rdev->data_offset) {
sb->feature_map
|= cpu_to_le32(MD_FEATURE_NEW_OFFSET);
sb->new_offset = cpu_to_le32((__u32)(rdev->new_data_offset
- rdev->data_offset));
}
}
if (mddev_is_clustered(mddev))
sb->feature_map |= cpu_to_le32(MD_FEATURE_CLUSTERED);
if (rdev->badblocks.count == 0)
/* Nothing to do for bad blocks*/ ;
else if (sb->bblog_offset == 0)
/* Cannot record bad blocks on this device */
md_error(mddev, rdev);
else {
struct badblocks *bb = &rdev->badblocks;
u64 *bbp = (u64 *)page_address(rdev->bb_page);
u64 *p = bb->page;
sb->feature_map |= cpu_to_le32(MD_FEATURE_BAD_BLOCKS);
if (bb->changed) {
unsigned seq;
retry:
seq = read_seqbegin(&bb->lock);
memset(bbp, 0xff, PAGE_SIZE);
for (i = 0 ; i < bb->count ; i++) {
u64 internal_bb = p[i];
u64 store_bb = ((BB_OFFSET(internal_bb) << 10)
| BB_LEN(internal_bb));
bbp[i] = cpu_to_le64(store_bb);
}
bb->changed = 0;
if (read_seqretry(&bb->lock, seq))
goto retry;
bb->sector = (rdev->sb_start +
(int)le32_to_cpu(sb->bblog_offset));
bb->size = le16_to_cpu(sb->bblog_size);
}
}
max_dev = 0;
rdev_for_each(rdev2, mddev)
if (rdev2->desc_nr+1 > max_dev)
max_dev = rdev2->desc_nr+1;
if (max_dev > le32_to_cpu(sb->max_dev)) {
int bmask;
sb->max_dev = cpu_to_le32(max_dev);
rdev->sb_size = max_dev * 2 + 256;
bmask = queue_logical_block_size(rdev->bdev->bd_disk->queue)-1;
if (rdev->sb_size & bmask)
rdev->sb_size = (rdev->sb_size | bmask) + 1;
} else
max_dev = le32_to_cpu(sb->max_dev);
for (i=0; i<max_dev;i++)
sb->dev_roles[i] = cpu_to_le16(MD_DISK_ROLE_SPARE);
if (test_bit(MD_HAS_JOURNAL, &mddev->flags))
sb->feature_map |= cpu_to_le32(MD_FEATURE_JOURNAL);
if (test_bit(MD_HAS_PPL, &mddev->flags)) {
if (test_bit(MD_HAS_MULTIPLE_PPLS, &mddev->flags))
sb->feature_map |=
cpu_to_le32(MD_FEATURE_MULTIPLE_PPLS);
else
sb->feature_map |= cpu_to_le32(MD_FEATURE_PPL);
sb->ppl.offset = cpu_to_le16(rdev->ppl.offset);
sb->ppl.size = cpu_to_le16(rdev->ppl.size);
}
rdev_for_each(rdev2, mddev) {
i = rdev2->desc_nr;
if (test_bit(Faulty, &rdev2->flags))
sb->dev_roles[i] = cpu_to_le16(MD_DISK_ROLE_FAULTY);
else if (test_bit(In_sync, &rdev2->flags))
sb->dev_roles[i] = cpu_to_le16(rdev2->raid_disk);
else if (test_bit(Journal, &rdev2->flags))
sb->dev_roles[i] = cpu_to_le16(MD_DISK_ROLE_JOURNAL);
else if (rdev2->raid_disk >= 0)
sb->dev_roles[i] = cpu_to_le16(rdev2->raid_disk);
else
sb->dev_roles[i] = cpu_to_le16(MD_DISK_ROLE_SPARE);
}
sb->sb_csum = calc_sb_1_csum(sb);
}
static unsigned long long
super_1_rdev_size_change(struct md_rdev *rdev, sector_t num_sectors)
{
struct mdp_superblock_1 *sb;
sector_t max_sectors;
if (num_sectors && num_sectors < rdev->mddev->dev_sectors)
return 0; /* component must fit device */
if (rdev->data_offset != rdev->new_data_offset)
return 0; /* too confusing */
if (rdev->sb_start < rdev->data_offset) {
/* minor versions 1 and 2; superblock before data */
max_sectors = i_size_read(rdev->bdev->bd_inode) >> 9;
max_sectors -= rdev->data_offset;
if (!num_sectors || num_sectors > max_sectors)
num_sectors = max_sectors;
} else if (rdev->mddev->bitmap_info.offset) {
/* minor version 0 with bitmap we can't move */
return 0;
} else {
/* minor version 0; superblock after data */
sector_t sb_start;
sb_start = (i_size_read(rdev->bdev->bd_inode) >> 9) - 8*2;
sb_start &= ~(sector_t)(4*2 - 1);
max_sectors = rdev->sectors + sb_start - rdev->sb_start;
if (!num_sectors || num_sectors > max_sectors)
num_sectors = max_sectors;
rdev->sb_start = sb_start;
}
sb = page_address(rdev->sb_page);
sb->data_size = cpu_to_le64(num_sectors);
sb->super_offset = cpu_to_le64(rdev->sb_start);
sb->sb_csum = calc_sb_1_csum(sb);
do {
md_super_write(rdev->mddev, rdev, rdev->sb_start, rdev->sb_size,
rdev->sb_page);
} while (md_super_wait(rdev->mddev) < 0);
return num_sectors;
}
static int
super_1_allow_new_offset(struct md_rdev *rdev,
unsigned long long new_offset)
{
/* All necessary checks on new >= old have been done */
struct bitmap *bitmap;
if (new_offset >= rdev->data_offset)
return 1;
/* with 1.0 metadata, there is no metadata to tread on
* so we can always move back */
if (rdev->mddev->minor_version == 0)
return 1;
/* otherwise we must be sure not to step on
* any metadata, so stay:
* 36K beyond start of superblock
* beyond end of badblocks
* beyond write-intent bitmap
*/
if (rdev->sb_start + (32+4)*2 > new_offset)
return 0;
bitmap = rdev->mddev->bitmap;
if (bitmap && !rdev->mddev->bitmap_info.file &&
rdev->sb_start + rdev->mddev->bitmap_info.offset +
bitmap->storage.file_pages * (PAGE_SIZE>>9) > new_offset)
return 0;
if (rdev->badblocks.sector + rdev->badblocks.size > new_offset)
return 0;
return 1;
}
static struct super_type super_types[] = {
[0] = {
.name = "0.90.0",
.owner = THIS_MODULE,
.load_super = super_90_load,
.validate_super = super_90_validate,
.sync_super = super_90_sync,
.rdev_size_change = super_90_rdev_size_change,
.allow_new_offset = super_90_allow_new_offset,
},
[1] = {
.name = "md-1",
.owner = THIS_MODULE,
.load_super = super_1_load,
.validate_super = super_1_validate,
.sync_super = super_1_sync,
.rdev_size_change = super_1_rdev_size_change,
.allow_new_offset = super_1_allow_new_offset,
},
};
static void sync_super(struct mddev *mddev, struct md_rdev *rdev)
{
if (mddev->sync_super) {
mddev->sync_super(mddev, rdev);
return;
}
BUG_ON(mddev->major_version >= ARRAY_SIZE(super_types));
super_types[mddev->major_version].sync_super(mddev, rdev);
}
static int match_mddev_units(struct mddev *mddev1, struct mddev *mddev2)
{
struct md_rdev *rdev, *rdev2;
rcu_read_lock();
rdev_for_each_rcu(rdev, mddev1) {
if (test_bit(Faulty, &rdev->flags) ||
test_bit(Journal, &rdev->flags) ||
rdev->raid_disk == -1)
continue;
rdev_for_each_rcu(rdev2, mddev2) {
if (test_bit(Faulty, &rdev2->flags) ||
test_bit(Journal, &rdev2->flags) ||
rdev2->raid_disk == -1)
continue;
if (rdev->bdev->bd_contains ==
rdev2->bdev->bd_contains) {
rcu_read_unlock();
return 1;
}
}
}
rcu_read_unlock();
return 0;
}
static LIST_HEAD(pending_raid_disks);
/*
* Try to register data integrity profile for an mddev
*
* This is called when an array is started and after a disk has been kicked
* from the array. It only succeeds if all working and active component devices
* are integrity capable with matching profiles.
*/
int md_integrity_register(struct mddev *mddev)
{
struct md_rdev *rdev, *reference = NULL;
if (list_empty(&mddev->disks))
return 0; /* nothing to do */
if (!mddev->gendisk || blk_get_integrity(mddev->gendisk))
return 0; /* shouldn't register, or already is */
rdev_for_each(rdev, mddev) {
/* skip spares and non-functional disks */
if (test_bit(Faulty, &rdev->flags))
continue;
if (rdev->raid_disk < 0)
continue;
if (!reference) {
/* Use the first rdev as the reference */
reference = rdev;
continue;
}
/* does this rdev's profile match the reference profile? */
if (blk_integrity_compare(reference->bdev->bd_disk,
rdev->bdev->bd_disk) < 0)
return -EINVAL;
}
if (!reference || !bdev_get_integrity(reference->bdev))
return 0;
/*
* All component devices are integrity capable and have matching
* profiles, register the common profile for the md device.
*/
blk_integrity_register(mddev->gendisk,
bdev_get_integrity(reference->bdev));
pr_debug("md: data integrity enabled on %s\n", mdname(mddev));
if (bioset_integrity_create(&mddev->bio_set, BIO_POOL_SIZE)) {
pr_err("md: failed to create integrity pool for %s\n",
mdname(mddev));
return -EINVAL;
}
return 0;
}
EXPORT_SYMBOL(md_integrity_register);
/*
* Attempt to add an rdev, but only if it is consistent with the current
* integrity profile
*/
int md_integrity_add_rdev(struct md_rdev *rdev, struct mddev *mddev)
{
struct blk_integrity *bi_rdev;
struct blk_integrity *bi_mddev;
char name[BDEVNAME_SIZE];
if (!mddev->gendisk)
return 0;
bi_rdev = bdev_get_integrity(rdev->bdev);
bi_mddev = blk_get_integrity(mddev->gendisk);
if (!bi_mddev) /* nothing to do */
return 0;
if (blk_integrity_compare(mddev->gendisk, rdev->bdev->bd_disk) != 0) {
pr_err("%s: incompatible integrity profile for %s\n",
mdname(mddev), bdevname(rdev->bdev, name));
return -ENXIO;
}
return 0;
}
EXPORT_SYMBOL(md_integrity_add_rdev);
static int bind_rdev_to_array(struct md_rdev *rdev, struct mddev *mddev)
{
char b[BDEVNAME_SIZE];
struct kobject *ko;
int err;
/* prevent duplicates */
if (find_rdev(mddev, rdev->bdev->bd_dev))
return -EEXIST;
if ((bdev_read_only(rdev->bdev) || bdev_read_only(rdev->meta_bdev)) &&
mddev->pers)
return -EROFS;
/* make sure rdev->sectors exceeds mddev->dev_sectors */
if (!test_bit(Journal, &rdev->flags) &&
rdev->sectors &&
(mddev->dev_sectors == 0 || rdev->sectors < mddev->dev_sectors)) {
if (mddev->pers) {
/* Cannot change size, so fail
* If mddev->level <= 0, then we don't care
* about aligning sizes (e.g. linear)
*/
if (mddev->level > 0)
return -ENOSPC;
} else
mddev->dev_sectors = rdev->sectors;
}
/* Verify rdev->desc_nr is unique.
* If it is -1, assign a free number, else
* check number is not in use
*/
rcu_read_lock();
if (rdev->desc_nr < 0) {
int choice = 0;
if (mddev->pers)
choice = mddev->raid_disks;
while (md_find_rdev_nr_rcu(mddev, choice))
choice++;
rdev->desc_nr = choice;
} else {
if (md_find_rdev_nr_rcu(mddev, rdev->desc_nr)) {
rcu_read_unlock();
return -EBUSY;
}
}
rcu_read_unlock();
if (!test_bit(Journal, &rdev->flags) &&
mddev->max_disks && rdev->desc_nr >= mddev->max_disks) {
pr_warn("md: %s: array is limited to %d devices\n",
mdname(mddev), mddev->max_disks);
return -EBUSY;
}
bdevname(rdev->bdev,b);
strreplace(b, '/', '!');
rdev->mddev = mddev;
pr_debug("md: bind<%s>\n", b);
if ((err = kobject_add(&rdev->kobj, &mddev->kobj, "dev-%s", b)))
goto fail;
ko = &part_to_dev(rdev->bdev->bd_part)->kobj;
if (sysfs_create_link(&rdev->kobj, ko, "block"))
/* failure here is OK */;
rdev->sysfs_state = sysfs_get_dirent_safe(rdev->kobj.sd, "state");
list_add_rcu(&rdev->same_set, &mddev->disks);
bd_link_disk_holder(rdev->bdev, mddev->gendisk);
/* May as well allow recovery to be retried once */
mddev->recovery_disabled++;
return 0;
fail:
pr_warn("md: failed to register dev-%s for %s\n",
b, mdname(mddev));
return err;
}
static void md_delayed_delete(struct work_struct *ws)
{
struct md_rdev *rdev = container_of(ws, struct md_rdev, del_work);
kobject_del(&rdev->kobj);
kobject_put(&rdev->kobj);
}
static void unbind_rdev_from_array(struct md_rdev *rdev)
{
char b[BDEVNAME_SIZE];
bd_unlink_disk_holder(rdev->bdev, rdev->mddev->gendisk);
list_del_rcu(&rdev->same_set);
pr_debug("md: unbind<%s>\n", bdevname(rdev->bdev,b));
rdev->mddev = NULL;
sysfs_remove_link(&rdev->kobj, "block");
sysfs_put(rdev->sysfs_state);
rdev->sysfs_state = NULL;
rdev->badblocks.count = 0;
/* We need to delay this, otherwise we can deadlock when
* writing to 'remove' to "dev/state". We also need
* to delay it due to rcu usage.
*/
synchronize_rcu();
INIT_WORK(&rdev->del_work, md_delayed_delete);
kobject_get(&rdev->kobj);
queue_work(md_misc_wq, &rdev->del_work);
}
/*
* prevent the device from being mounted, repartitioned or
* otherwise reused by a RAID array (or any other kernel
* subsystem), by bd_claiming the device.
*/
static int lock_rdev(struct md_rdev *rdev, dev_t dev, int shared)
{
int err = 0;
struct block_device *bdev;
char b[BDEVNAME_SIZE];
bdev = blkdev_get_by_dev(dev, FMODE_READ|FMODE_WRITE|FMODE_EXCL,
shared ? (struct md_rdev *)lock_rdev : rdev);
if (IS_ERR(bdev)) {
pr_warn("md: could not open %s.\n", __bdevname(dev, b));
return PTR_ERR(bdev);
}
rdev->bdev = bdev;
return err;
}
static void unlock_rdev(struct md_rdev *rdev)
{
struct block_device *bdev = rdev->bdev;
rdev->bdev = NULL;
blkdev_put(bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
}
void md_autodetect_dev(dev_t dev);
static void export_rdev(struct md_rdev *rdev)
{
char b[BDEVNAME_SIZE];
pr_debug("md: export_rdev(%s)\n", bdevname(rdev->bdev,b));
md_rdev_clear(rdev);
#ifndef MODULE
if (test_bit(AutoDetected, &rdev->flags))
md_autodetect_dev(rdev->bdev->bd_dev);
#endif
unlock_rdev(rdev);
kobject_put(&rdev->kobj);
}
void md_kick_rdev_from_array(struct md_rdev *rdev)
{
unbind_rdev_from_array(rdev);
export_rdev(rdev);
}
EXPORT_SYMBOL_GPL(md_kick_rdev_from_array);
static void export_array(struct mddev *mddev)
{
struct md_rdev *rdev;
while (!list_empty(&mddev->disks)) {
rdev = list_first_entry(&mddev->disks, struct md_rdev,
same_set);
md_kick_rdev_from_array(rdev);
}
mddev->raid_disks = 0;
mddev->major_version = 0;
}
static bool set_in_sync(struct mddev *mddev)
{
lockdep_assert_held(&mddev->lock);
if (!mddev->in_sync) {
mddev->sync_checkers++;
spin_unlock(&mddev->lock);
percpu_ref_switch_to_atomic_sync(&mddev->writes_pending);
spin_lock(&mddev->lock);
if (!mddev->in_sync &&
percpu_ref_is_zero(&mddev->writes_pending)) {
mddev->in_sync = 1;
/*
* Ensure ->in_sync is visible before we clear
* ->sync_checkers.
*/
smp_mb();
set_bit(MD_SB_CHANGE_CLEAN, &mddev->sb_flags);
sysfs_notify_dirent_safe(mddev->sysfs_state);
}
if (--mddev->sync_checkers == 0)
percpu_ref_switch_to_percpu(&mddev->writes_pending);
}
if (mddev->safemode == 1)
mddev->safemode = 0;
return mddev->in_sync;
}
static void sync_sbs(struct mddev *mddev, int nospares)
{
/* Update each superblock (in-memory image), but
* if we are allowed to, skip spares which already
* have the right event counter, or have one earlier
* (which would mean they aren't being marked as dirty
* with the rest of the array)
*/
struct md_rdev *rdev;
rdev_for_each(rdev, mddev) {
if (rdev->sb_events == mddev->events ||
(nospares &&
rdev->raid_disk < 0 &&
rdev->sb_events+1 == mddev->events)) {
/* Don't update this superblock */
rdev->sb_loaded = 2;
} else {
sync_super(mddev, rdev);
rdev->sb_loaded = 1;
}
}
}
static bool does_sb_need_changing(struct mddev *mddev)
{
struct md_rdev *rdev;
struct mdp_superblock_1 *sb;
int role;
/* Find a good rdev */
rdev_for_each(rdev, mddev)
if ((rdev->raid_disk >= 0) && !test_bit(Faulty, &rdev->flags))
break;
/* No good device found. */
if (!rdev)
return false;
sb = page_address(rdev->sb_page);
/* Check if a device has become faulty or a spare become active */
rdev_for_each(rdev, mddev) {
role = le16_to_cpu(sb->dev_roles[rdev->desc_nr]);
/* Device activated? */
if (role == 0xffff && rdev->raid_disk >=0 &&
!test_bit(Faulty, &rdev->flags))
return true;
/* Device turned faulty? */
if (test_bit(Faulty, &rdev->flags) && (role < 0xfffd))
return true;
}
/* Check if any mddev parameters have changed */
if ((mddev->dev_sectors != le64_to_cpu(sb->size)) ||
(mddev->reshape_position != le64_to_cpu(sb->reshape_position)) ||
(mddev->layout != le32_to_cpu(sb->layout)) ||
(mddev->raid_disks != le32_to_cpu(sb->raid_disks)) ||
(mddev->chunk_sectors != le32_to_cpu(sb->chunksize)))
return true;
return false;
}
void md_update_sb(struct mddev *mddev, int force_change)
{
struct md_rdev *rdev;
int sync_req;
int nospares = 0;
int any_badblocks_changed = 0;
int ret = -1;
if (mddev->ro) {
if (force_change)
set_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags);
return;
}
repeat:
if (mddev_is_clustered(mddev)) {
if (test_and_clear_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags))
force_change = 1;
if (test_and_clear_bit(MD_SB_CHANGE_CLEAN, &mddev->sb_flags))
nospares = 1;
ret = md_cluster_ops->metadata_update_start(mddev);
/* Has someone else has updated the sb */
if (!does_sb_need_changing(mddev)) {
if (ret == 0)
md_cluster_ops->metadata_update_cancel(mddev);
bit_clear_unless(&mddev->sb_flags, BIT(MD_SB_CHANGE_PENDING),
BIT(MD_SB_CHANGE_DEVS) |
BIT(MD_SB_CHANGE_CLEAN));
return;
}
}
/*
* First make sure individual recovery_offsets are correct
* curr_resync_completed can only be used during recovery.
* During reshape/resync it might use array-addresses rather
* that device addresses.
*/
rdev_for_each(rdev, mddev) {
if (rdev->raid_disk >= 0 &&
mddev->delta_disks >= 0 &&
test_bit(MD_RECOVERY_RUNNING, &mddev->recovery) &&
test_bit(MD_RECOVERY_RECOVER, &mddev->recovery) &&
!test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery) &&
!test_bit(Journal, &rdev->flags) &&
!test_bit(In_sync, &rdev->flags) &&
mddev->curr_resync_completed > rdev->recovery_offset)
rdev->recovery_offset = mddev->curr_resync_completed;
}
if (!mddev->persistent) {
clear_bit(MD_SB_CHANGE_CLEAN, &mddev->sb_flags);
clear_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags);
if (!mddev->external) {
clear_bit(MD_SB_CHANGE_PENDING, &mddev->sb_flags);
rdev_for_each(rdev, mddev) {
if (rdev->badblocks.changed) {
rdev->badblocks.changed = 0;
ack_all_badblocks(&rdev->badblocks);
md_error(mddev, rdev);
}
clear_bit(Blocked, &rdev->flags);
clear_bit(BlockedBadBlocks, &rdev->flags);
wake_up(&rdev->blocked_wait);
}
}
wake_up(&mddev->sb_wait);
return;
}
spin_lock(&mddev->lock);
mddev->utime = ktime_get_real_seconds();
if (test_and_clear_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags))
force_change = 1;
if (test_and_clear_bit(MD_SB_CHANGE_CLEAN, &mddev->sb_flags))
/* just a clean<-> dirty transition, possibly leave spares alone,
* though if events isn't the right even/odd, we will have to do
* spares after all
*/
nospares = 1;
if (force_change)
nospares = 0;
if (mddev->degraded)
/* If the array is degraded, then skipping spares is both
* dangerous and fairly pointless.
* Dangerous because a device that was removed from the array
* might have a event_count that still looks up-to-date,
* so it can be re-added without a resync.
* Pointless because if there are any spares to skip,
* then a recovery will happen and soon that array won't
* be degraded any more and the spare can go back to sleep then.
*/
nospares = 0;
sync_req = mddev->in_sync;
/* If this is just a dirty<->clean transition, and the array is clean
* and 'events' is odd, we can roll back to the previous clean state */
if (nospares
&& (mddev->in_sync && mddev->recovery_cp == MaxSector)
&& mddev->can_decrease_events
&& mddev->events != 1) {
mddev->events--;
mddev->can_decrease_events = 0;
} else {
/* otherwise we have to go forward and ... */
mddev->events ++;
mddev->can_decrease_events = nospares;
}
/*
* This 64-bit counter should never wrap.
* Either we are in around ~1 trillion A.C., assuming
* 1 reboot per second, or we have a bug...
*/
WARN_ON(mddev->events == 0);
rdev_for_each(rdev, mddev) {
if (rdev->badblocks.changed)
any_badblocks_changed++;
if (test_bit(Faulty, &rdev->flags))
set_bit(FaultRecorded, &rdev->flags);
}
sync_sbs(mddev, nospares);
spin_unlock(&mddev->lock);
pr_debug("md: updating %s RAID superblock on device (in sync %d)\n",
mdname(mddev), mddev->in_sync);
if (mddev->queue)
blk_add_trace_msg(mddev->queue, "md md_update_sb");
rewrite:
md_bitmap_update_sb(mddev->bitmap);
rdev_for_each(rdev, mddev) {
char b[BDEVNAME_SIZE];
if (rdev->sb_loaded != 1)
continue; /* no noise on spare devices */
if (!test_bit(Faulty, &rdev->flags)) {
md_super_write(mddev,rdev,
rdev->sb_start, rdev->sb_size,
rdev->sb_page);
pr_debug("md: (write) %s's sb offset: %llu\n",
bdevname(rdev->bdev, b),
(unsigned long long)rdev->sb_start);
rdev->sb_events = mddev->events;
if (rdev->badblocks.size) {
md_super_write(mddev, rdev,
rdev->badblocks.sector,
rdev->badblocks.size << 9,
rdev->bb_page);
rdev->badblocks.size = 0;
}
} else
pr_debug("md: %s (skipping faulty)\n",
bdevname(rdev->bdev, b));
if (mddev->level == LEVEL_MULTIPATH)
/* only need to write one superblock... */
break;
}
if (md_super_wait(mddev) < 0)
goto rewrite;
/* if there was a failure, MD_SB_CHANGE_DEVS was set, and we re-write super */
if (mddev_is_clustered(mddev) && ret == 0)
md_cluster_ops->metadata_update_finish(mddev);
if (mddev->in_sync != sync_req ||
!bit_clear_unless(&mddev->sb_flags, BIT(MD_SB_CHANGE_PENDING),
BIT(MD_SB_CHANGE_DEVS) | BIT(MD_SB_CHANGE_CLEAN)))
/* have to write it out again */
goto repeat;
wake_up(&mddev->sb_wait);
if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery))
sysfs_notify(&mddev->kobj, NULL, "sync_completed");
rdev_for_each(rdev, mddev) {
if (test_and_clear_bit(FaultRecorded, &rdev->flags))
clear_bit(Blocked, &rdev->flags);
if (any_badblocks_changed)
ack_all_badblocks(&rdev->badblocks);
clear_bit(BlockedBadBlocks, &rdev->flags);
wake_up(&rdev->blocked_wait);
}
}
EXPORT_SYMBOL(md_update_sb);
static int add_bound_rdev(struct md_rdev *rdev)
{
struct mddev *mddev = rdev->mddev;
int err = 0;
bool add_journal = test_bit(Journal, &rdev->flags);
if (!mddev->pers->hot_remove_disk || add_journal) {
/* If there is hot_add_disk but no hot_remove_disk
* then added disks for geometry changes,
* and should be added immediately.
*/
super_types[mddev->major_version].
validate_super(mddev, rdev);
if (add_journal)
mddev_suspend(mddev);
err = mddev->pers->hot_add_disk(mddev, rdev);
if (add_journal)
mddev_resume(mddev);
if (err) {
md_kick_rdev_from_array(rdev);
return err;
}
}
sysfs_notify_dirent_safe(rdev->sysfs_state);
set_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags);
if (mddev->degraded)
set_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
md_new_event(mddev);
md_wakeup_thread(mddev->thread);
return 0;
}
/* words written to sysfs files may, or may not, be \n terminated.
* We want to accept with case. For this we use cmd_match.
*/
static int cmd_match(const char *cmd, const char *str)
{
/* See if cmd, written into a sysfs file, matches
* str. They must either be the same, or cmd can
* have a trailing newline
*/
while (*cmd && *str && *cmd == *str) {
cmd++;
str++;
}
if (*cmd == '\n')
cmd++;
if (*str || *cmd)
return 0;
return 1;
}
struct rdev_sysfs_entry {
struct attribute attr;
ssize_t (*show)(struct md_rdev *, char *);
ssize_t (*store)(struct md_rdev *, const char *, size_t);
};
static ssize_t
state_show(struct md_rdev *rdev, char *page)
{
char *sep = ",";
size_t len = 0;
unsigned long flags = READ_ONCE(rdev->flags);
if (test_bit(Faulty, &flags) ||
(!test_bit(ExternalBbl, &flags) &&
rdev->badblocks.unacked_exist))
len += sprintf(page+len, "faulty%s", sep);
if (test_bit(In_sync, &flags))
len += sprintf(page+len, "in_sync%s", sep);
if (test_bit(Journal, &flags))
len += sprintf(page+len, "journal%s", sep);
if (test_bit(WriteMostly, &flags))
len += sprintf(page+len, "write_mostly%s", sep);
if (test_bit(Blocked, &flags) ||
(rdev->badblocks.unacked_exist
&& !test_bit(Faulty, &flags)))
len += sprintf(page+len, "blocked%s", sep);
if (!test_bit(Faulty, &flags) &&
!test_bit(Journal, &flags) &&
!test_bit(In_sync, &flags))
len += sprintf(page+len, "spare%s", sep);
if (test_bit(WriteErrorSeen, &flags))
len += sprintf(page+len, "write_error%s", sep);
if (test_bit(WantReplacement, &flags))
len += sprintf(page+len, "want_replacement%s", sep);
if (test_bit(Replacement, &flags))
len += sprintf(page+len, "replacement%s", sep);
if (test_bit(ExternalBbl, &flags))
len += sprintf(page+len, "external_bbl%s", sep);
if (test_bit(FailFast, &flags))
len += sprintf(page+len, "failfast%s", sep);
if (len)
len -= strlen(sep);
return len+sprintf(page+len, "\n");
}
static ssize_t
state_store(struct md_rdev *rdev, const char *buf, size_t len)
{
/* can write
* faulty - simulates an error
* remove - disconnects the device
* writemostly - sets write_mostly
* -writemostly - clears write_mostly
* blocked - sets the Blocked flags
* -blocked - clears the Blocked and possibly simulates an error
* insync - sets Insync providing device isn't active
* -insync - clear Insync for a device with a slot assigned,
* so that it gets rebuilt based on bitmap
* write_error - sets WriteErrorSeen
* -write_error - clears WriteErrorSeen
* {,-}failfast - set/clear FailFast
*/
int err = -EINVAL;
if (cmd_match(buf, "faulty") && rdev->mddev->pers) {
md_error(rdev->mddev, rdev);
if (test_bit(Faulty, &rdev->flags))
err = 0;
else
err = -EBUSY;
} else if (cmd_match(buf, "remove")) {
if (rdev->mddev->pers) {
clear_bit(Blocked, &rdev->flags);
remove_and_add_spares(rdev->mddev, rdev);
}
if (rdev->raid_disk >= 0)
err = -EBUSY;
else {
struct mddev *mddev = rdev->mddev;
err = 0;
if (mddev_is_clustered(mddev))
err = md_cluster_ops->remove_disk(mddev, rdev);
if (err == 0) {
md_kick_rdev_from_array(rdev);
if (mddev->pers) {
set_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags);
md_wakeup_thread(mddev->thread);
}
md_new_event(mddev);
}
}
} else if (cmd_match(buf, "writemostly")) {
set_bit(WriteMostly, &rdev->flags);
err = 0;
} else if (cmd_match(buf, "-writemostly")) {
clear_bit(WriteMostly, &rdev->flags);
err = 0;
} else if (cmd_match(buf, "blocked")) {
set_bit(Blocked, &rdev->flags);
err = 0;
} else if (cmd_match(buf, "-blocked")) {
if (!test_bit(Faulty, &rdev->flags) &&
!test_bit(ExternalBbl, &rdev->flags) &&
rdev->badblocks.unacked_exist) {
/* metadata handler doesn't understand badblocks,
* so we need to fail the device
*/
md_error(rdev->mddev, rdev);
}
clear_bit(Blocked, &rdev->flags);
clear_bit(BlockedBadBlocks, &rdev->flags);
wake_up(&rdev->blocked_wait);
set_bit(MD_RECOVERY_NEEDED, &rdev->mddev->recovery);
md_wakeup_thread(rdev->mddev->thread);
err = 0;
} else if (cmd_match(buf, "insync") && rdev->raid_disk == -1) {
set_bit(In_sync, &rdev->flags);
err = 0;
} else if (cmd_match(buf, "failfast")) {
set_bit(FailFast, &rdev->flags);
err = 0;
} else if (cmd_match(buf, "-failfast")) {
clear_bit(FailFast, &rdev->flags);
err = 0;
} else if (cmd_match(buf, "-insync") && rdev->raid_disk >= 0 &&
!test_bit(Journal, &rdev->flags)) {
if (rdev->mddev->pers == NULL) {
clear_bit(In_sync, &rdev->flags);
rdev->saved_raid_disk = rdev->raid_disk;
rdev->raid_disk = -1;
err = 0;
}
} else if (cmd_match(buf, "write_error")) {
set_bit(WriteErrorSeen, &rdev->flags);
err = 0;
} else if (cmd_match(buf, "-write_error")) {
clear_bit(WriteErrorSeen, &rdev->flags);
err = 0;
} else if (cmd_match(buf, "want_replacement")) {
/* Any non-spare device that is not a replacement can
* become want_replacement at any time, but we then need to
* check if recovery is needed.
*/
if (rdev->raid_disk >= 0 &&
!test_bit(Journal, &rdev->flags) &&
!test_bit(Replacement, &rdev->flags))
set_bit(WantReplacement, &rdev->flags);
set_bit(MD_RECOVERY_NEEDED, &rdev->mddev->recovery);
md_wakeup_thread(rdev->mddev->thread);
err = 0;
} else if (cmd_match(buf, "-want_replacement")) {
/* Clearing 'want_replacement' is always allowed.
* Once replacements starts it is too late though.
*/
err = 0;
clear_bit(WantReplacement, &rdev->flags);
} else if (cmd_match(buf, "replacement")) {
/* Can only set a device as a replacement when array has not
* yet been started. Once running, replacement is automatic
* from spares, or by assigning 'slot'.
*/
if (rdev->mddev->pers)
err = -EBUSY;
else {
set_bit(Replacement, &rdev->flags);
err = 0;
}
} else if (cmd_match(buf, "-replacement")) {
/* Similarly, can only clear Replacement before start */
if (rdev->mddev->pers)
err = -EBUSY;
else {
clear_bit(Replacement, &rdev->flags);
err = 0;
}
} else if (cmd_match(buf, "re-add")) {
if (!rdev->mddev->pers)
err = -EINVAL;