summaryrefslogtreecommitdiffstats
path: root/drivers/md
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/md')
-rw-r--r--drivers/md/bitmap.c8
-rw-r--r--drivers/md/dm-bio-list.h14
-rw-r--r--drivers/md/dm-crypt.c4
-rw-r--r--drivers/md/dm-emc.c10
-rw-r--r--drivers/md/dm-hw-handler.h2
-rw-r--r--drivers/md/dm-io.c15
-rw-r--r--drivers/md/dm-ioctl.c16
-rw-r--r--drivers/md/dm-linear.c4
-rw-r--r--drivers/md/dm-log.c24
-rw-r--r--drivers/md/dm-log.h10
-rw-r--r--drivers/md/dm-mpath.c52
-rw-r--r--drivers/md/dm-mpath.h4
-rw-r--r--drivers/md/dm-path-selector.h12
-rw-r--r--drivers/md/dm-raid1.c25
-rw-r--r--drivers/md/dm-round-robin.c12
-rw-r--r--drivers/md/dm-snap.c33
-rw-r--r--drivers/md/dm-stripe.c2
-rw-r--r--drivers/md/dm-zero.c2
-rw-r--r--drivers/md/dm.c130
-rw-r--r--drivers/md/dm.h19
-rw-r--r--drivers/md/md.c12
21 files changed, 294 insertions, 116 deletions
diff --git a/drivers/md/bitmap.c b/drivers/md/bitmap.c
index d6f614738bb..5432d07c074 100644
--- a/drivers/md/bitmap.c
+++ b/drivers/md/bitmap.c
@@ -212,8 +212,8 @@ char *file_path(struct file *file, char *buf, int count)
if (!buf)
return NULL;
- d = file->f_dentry;
- v = file->f_vfsmnt;
+ d = file->f_path.dentry;
+ v = file->f_path.mnt;
buf = d_path(d, v, buf, count);
@@ -349,7 +349,7 @@ static struct page *read_page(struct file *file, unsigned long index,
unsigned long count)
{
struct page *page = NULL;
- struct inode *inode = file->f_dentry->d_inode;
+ struct inode *inode = file->f_path.dentry->d_inode;
struct buffer_head *bh;
sector_t block;
@@ -662,7 +662,7 @@ static void bitmap_file_put(struct bitmap *bitmap)
bitmap_file_unmap(bitmap);
if (file) {
- struct inode *inode = file->f_dentry->d_inode;
+ struct inode *inode = file->f_path.dentry->d_inode;
invalidate_inode_pages(inode->i_mapping);
fput(file);
}
diff --git a/drivers/md/dm-bio-list.h b/drivers/md/dm-bio-list.h
index bbf4615f0e3..da4349649f7 100644
--- a/drivers/md/dm-bio-list.h
+++ b/drivers/md/dm-bio-list.h
@@ -44,6 +44,20 @@ static inline void bio_list_merge(struct bio_list *bl, struct bio_list *bl2)
bl->tail = bl2->tail;
}
+static inline void bio_list_merge_head(struct bio_list *bl,
+ struct bio_list *bl2)
+{
+ if (!bl2->head)
+ return;
+
+ if (bl->head)
+ bl2->tail->bi_next = bl->head;
+ else
+ bl->tail = bl2->tail;
+
+ bl->head = bl2->head;
+}
+
static inline struct bio *bio_list_pop(struct bio_list *bl)
{
struct bio *bio = bl->head;
diff --git a/drivers/md/dm-crypt.c b/drivers/md/dm-crypt.c
index a1086ee8ccc..4c2471ee054 100644
--- a/drivers/md/dm-crypt.c
+++ b/drivers/md/dm-crypt.c
@@ -220,7 +220,7 @@ static int crypt_iv_benbi_ctr(struct crypt_config *cc, struct dm_target *ti,
const char *opts)
{
unsigned int bs = crypto_blkcipher_blocksize(cc->tfm);
- int log = long_log2(bs);
+ int log = ilog2(bs);
/* we need to calculate how far we must shift the sector count
* to get the cipher block count, we use this shift in _gen */
@@ -962,7 +962,7 @@ static int crypt_map(struct dm_target *ti, struct bio *bio,
atomic_set(&io->pending, 0);
kcryptd_queue_io(io);
- return 0;
+ return DM_MAPIO_SUBMITTED;
}
static int crypt_status(struct dm_target *ti, status_type_t type,
diff --git a/drivers/md/dm-emc.c b/drivers/md/dm-emc.c
index 2b2d45d7baa..265c467854d 100644
--- a/drivers/md/dm-emc.c
+++ b/drivers/md/dm-emc.c
@@ -40,7 +40,7 @@ static inline void free_bio(struct bio *bio)
static int emc_endio(struct bio *bio, unsigned int bytes_done, int error)
{
- struct path *path = bio->bi_private;
+ struct dm_path *path = bio->bi_private;
if (bio->bi_size)
return 1;
@@ -61,7 +61,7 @@ static int emc_endio(struct bio *bio, unsigned int bytes_done, int error)
return 0;
}
-static struct bio *get_failover_bio(struct path *path, unsigned data_size)
+static struct bio *get_failover_bio(struct dm_path *path, unsigned data_size)
{
struct bio *bio;
struct page *page;
@@ -96,7 +96,7 @@ static struct bio *get_failover_bio(struct path *path, unsigned data_size)
}
static struct request *get_failover_req(struct emc_handler *h,
- struct bio *bio, struct path *path)
+ struct bio *bio, struct dm_path *path)
{
struct request *rq;
struct block_device *bdev = bio->bi_bdev;
@@ -133,7 +133,7 @@ static struct request *get_failover_req(struct emc_handler *h,
}
static struct request *emc_trespass_get(struct emc_handler *h,
- struct path *path)
+ struct dm_path *path)
{
struct bio *bio;
struct request *rq;
@@ -191,7 +191,7 @@ static struct request *emc_trespass_get(struct emc_handler *h,
}
static void emc_pg_init(struct hw_handler *hwh, unsigned bypassed,
- struct path *path)
+ struct dm_path *path)
{
struct request *rq;
struct request_queue *q = bdev_get_queue(path->dev->bdev);
diff --git a/drivers/md/dm-hw-handler.h b/drivers/md/dm-hw-handler.h
index 15f5629e231..32eff28e4ad 100644
--- a/drivers/md/dm-hw-handler.h
+++ b/drivers/md/dm-hw-handler.h
@@ -32,7 +32,7 @@ struct hw_handler_type {
void (*destroy) (struct hw_handler *hwh);
void (*pg_init) (struct hw_handler *hwh, unsigned bypassed,
- struct path *path);
+ struct dm_path *path);
unsigned (*error) (struct hw_handler *hwh, struct bio *bio);
int (*status) (struct hw_handler *hwh, status_type_t type,
char *result, unsigned int maxlen);
diff --git a/drivers/md/dm-io.c b/drivers/md/dm-io.c
index da663d2ff55..4eb73d39521 100644
--- a/drivers/md/dm-io.c
+++ b/drivers/md/dm-io.c
@@ -92,12 +92,12 @@ void dm_io_put(unsigned int num_pages)
*---------------------------------------------------------------*/
static inline void bio_set_region(struct bio *bio, unsigned region)
{
- bio->bi_io_vec[bio->bi_max_vecs - 1].bv_len = region;
+ bio->bi_io_vec[bio->bi_max_vecs].bv_len = region;
}
static inline unsigned bio_get_region(struct bio *bio)
{
- return bio->bi_io_vec[bio->bi_max_vecs - 1].bv_len;
+ return bio->bi_io_vec[bio->bi_max_vecs].bv_len;
}
/*-----------------------------------------------------------------
@@ -136,6 +136,7 @@ static int endio(struct bio *bio, unsigned int done, int error)
zero_fill_bio(bio);
dec_count(io, bio_get_region(bio), error);
+ bio->bi_max_vecs++;
bio_put(bio);
return 0;
@@ -250,16 +251,18 @@ static void do_region(int rw, unsigned int region, struct io_region *where,
while (remaining) {
/*
- * Allocate a suitably sized bio, we add an extra
- * bvec for bio_get/set_region().
+ * Allocate a suitably sized-bio: we add an extra
+ * bvec for bio_get/set_region() and decrement bi_max_vecs
+ * to hide it from bio_add_page().
*/
- num_bvecs = (remaining / (PAGE_SIZE >> 9)) + 2;
+ num_bvecs = (remaining / (PAGE_SIZE >> SECTOR_SHIFT)) + 2;
bio = bio_alloc_bioset(GFP_NOIO, num_bvecs, _bios);
bio->bi_sector = where->sector + (where->count - remaining);
bio->bi_bdev = where->bdev;
bio->bi_end_io = endio;
bio->bi_private = io;
bio->bi_destructor = dm_bio_destructor;
+ bio->bi_max_vecs--;
bio_set_region(bio, region);
/*
@@ -302,7 +305,7 @@ static void dispatch_io(int rw, unsigned int num_regions,
}
/*
- * Drop the extra refence that we were holding to avoid
+ * Drop the extra reference that we were holding to avoid
* the io being completed too early.
*/
dec_count(io, 0, 0);
diff --git a/drivers/md/dm-ioctl.c b/drivers/md/dm-ioctl.c
index 4510ad8f971..cd6a184536a 100644
--- a/drivers/md/dm-ioctl.c
+++ b/drivers/md/dm-ioctl.c
@@ -765,7 +765,7 @@ out:
static int do_suspend(struct dm_ioctl *param)
{
int r = 0;
- int do_lockfs = 1;
+ unsigned suspend_flags = DM_SUSPEND_LOCKFS_FLAG;
struct mapped_device *md;
md = find_device(param);
@@ -773,10 +773,12 @@ static int do_suspend(struct dm_ioctl *param)
return -ENXIO;
if (param->flags & DM_SKIP_LOCKFS_FLAG)
- do_lockfs = 0;
+ suspend_flags &= ~DM_SUSPEND_LOCKFS_FLAG;
+ if (param->flags & DM_NOFLUSH_FLAG)
+ suspend_flags |= DM_SUSPEND_NOFLUSH_FLAG;
if (!dm_suspended(md))
- r = dm_suspend(md, do_lockfs);
+ r = dm_suspend(md, suspend_flags);
if (!r)
r = __dev_status(md, param);
@@ -788,7 +790,7 @@ static int do_suspend(struct dm_ioctl *param)
static int do_resume(struct dm_ioctl *param)
{
int r = 0;
- int do_lockfs = 1;
+ unsigned suspend_flags = DM_SUSPEND_LOCKFS_FLAG;
struct hash_cell *hc;
struct mapped_device *md;
struct dm_table *new_map;
@@ -814,9 +816,11 @@ static int do_resume(struct dm_ioctl *param)
if (new_map) {
/* Suspend if it isn't already suspended */
if (param->flags & DM_SKIP_LOCKFS_FLAG)
- do_lockfs = 0;
+ suspend_flags &= ~DM_SUSPEND_LOCKFS_FLAG;
+ if (param->flags & DM_NOFLUSH_FLAG)
+ suspend_flags |= DM_SUSPEND_NOFLUSH_FLAG;
if (!dm_suspended(md))
- dm_suspend(md, do_lockfs);
+ dm_suspend(md, suspend_flags);
r = dm_swap_table(md, new_map);
if (r) {
diff --git a/drivers/md/dm-linear.c b/drivers/md/dm-linear.c
index 00234909b3d..17753d80ad2 100644
--- a/drivers/md/dm-linear.c
+++ b/drivers/md/dm-linear.c
@@ -77,7 +77,7 @@ static int linear_map(struct dm_target *ti, struct bio *bio,
bio->bi_bdev = lc->dev->bdev;
bio->bi_sector = lc->start + (bio->bi_sector - ti->begin);
- return 1;
+ return DM_MAPIO_REMAPPED;
}
static int linear_status(struct dm_target *ti, status_type_t type,
@@ -108,7 +108,7 @@ static int linear_ioctl(struct dm_target *ti, struct inode *inode,
struct dentry fake_dentry = {};
fake_file.f_mode = lc->dev->mode;
- fake_file.f_dentry = &fake_dentry;
+ fake_file.f_path.dentry = &fake_dentry;
fake_dentry.d_inode = bdev->bd_inode;
return blkdev_driver_ioctl(bdev->bd_inode, &fake_file, bdev->bd_disk, cmd, arg);
diff --git a/drivers/md/dm-log.c b/drivers/md/dm-log.c
index 64b764bd02c..6a926135184 100644
--- a/drivers/md/dm-log.c
+++ b/drivers/md/dm-log.c
@@ -466,6 +466,7 @@ static int disk_resume(struct dirty_log *log)
/* copy clean across to sync */
memcpy(lc->sync_bits, lc->clean_bits, size);
lc->sync_count = count_bits32(lc->clean_bits, lc->bitset_uint32_count);
+ lc->sync_search = 0;
/* set the correct number of regions in the header */
lc->header.nr_regions = lc->region_count;
@@ -480,6 +481,13 @@ static uint32_t core_get_region_size(struct dirty_log *log)
return lc->region_size;
}
+static int core_resume(struct dirty_log *log)
+{
+ struct log_c *lc = (struct log_c *) log->context;
+ lc->sync_search = 0;
+ return 0;
+}
+
static int core_is_clean(struct dirty_log *log, region_t region)
{
struct log_c *lc = (struct log_c *) log->context;
@@ -549,16 +557,19 @@ static int core_get_resync_work(struct dirty_log *log, region_t *region)
return 1;
}
-static void core_complete_resync_work(struct dirty_log *log, region_t region,
- int success)
+static void core_set_region_sync(struct dirty_log *log, region_t region,
+ int in_sync)
{
struct log_c *lc = (struct log_c *) log->context;
log_clear_bit(lc, lc->recovering_bits, region);
- if (success) {
+ if (in_sync) {
log_set_bit(lc, lc->sync_bits, region);
lc->sync_count++;
- }
+ } else if (log_test_bit(lc->sync_bits, region)) {
+ lc->sync_count--;
+ log_clear_bit(lc, lc->sync_bits, region);
+ }
}
static region_t core_get_sync_count(struct dirty_log *log)
@@ -618,6 +629,7 @@ static struct dirty_log_type _core_type = {
.module = THIS_MODULE,
.ctr = core_ctr,
.dtr = core_dtr,
+ .resume = core_resume,
.get_region_size = core_get_region_size,
.is_clean = core_is_clean,
.in_sync = core_in_sync,
@@ -625,7 +637,7 @@ static struct dirty_log_type _core_type = {
.mark_region = core_mark_region,
.clear_region = core_clear_region,
.get_resync_work = core_get_resync_work,
- .complete_resync_work = core_complete_resync_work,
+ .set_region_sync = core_set_region_sync,
.get_sync_count = core_get_sync_count,
.status = core_status,
};
@@ -644,7 +656,7 @@ static struct dirty_log_type _disk_type = {
.mark_region = core_mark_region,
.clear_region = core_clear_region,
.get_resync_work = core_get_resync_work,
- .complete_resync_work = core_complete_resync_work,
+ .set_region_sync = core_set_region_sync,
.get_sync_count = core_get_sync_count,
.status = disk_status,
};
diff --git a/drivers/md/dm-log.h b/drivers/md/dm-log.h
index 5ae5309ebf2..86a301c8daf 100644
--- a/drivers/md/dm-log.h
+++ b/drivers/md/dm-log.h
@@ -90,12 +90,12 @@ struct dirty_log_type {
int (*get_resync_work)(struct dirty_log *log, region_t *region);
/*
- * This notifies the log that the resync of an area has
- * been completed. The log should then mark this region
- * as CLEAN.
+ * This notifies the log that the resync status of a region
+ * has changed. It also clears the region from the recovering
+ * list (if present).
*/
- void (*complete_resync_work)(struct dirty_log *log,
- region_t region, int success);
+ void (*set_region_sync)(struct dirty_log *log,
+ region_t region, int in_sync);
/*
* Returns the number of regions that are in sync.
diff --git a/drivers/md/dm-mpath.c b/drivers/md/dm-mpath.c
index cf8bf052138..3aa01350696 100644
--- a/drivers/md/dm-mpath.c
+++ b/drivers/md/dm-mpath.c
@@ -31,7 +31,7 @@ struct pgpath {
struct priority_group *pg; /* Owning PG */
unsigned fail_count; /* Cumulative failure count */
- struct path path;
+ struct dm_path path;
};
#define path_to_pgpath(__pgp) container_of((__pgp), struct pgpath, path)
@@ -229,7 +229,7 @@ static void __switch_pg(struct multipath *m, struct pgpath *pgpath)
static int __choose_path_in_pg(struct multipath *m, struct priority_group *pg)
{
- struct path *path;
+ struct dm_path *path;
path = pg->ps.type->select_path(&pg->ps, &m->repeat_count);
if (!path)
@@ -282,10 +282,27 @@ failed:
m->current_pg = NULL;
}
+/*
+ * Check whether bios must be queued in the device-mapper core rather
+ * than here in the target.
+ *
+ * m->lock must be held on entry.
+ *
+ * If m->queue_if_no_path and m->saved_queue_if_no_path hold the
+ * same value then we are not between multipath_presuspend()
+ * and multipath_resume() calls and we have no need to check
+ * for the DMF_NOFLUSH_SUSPENDING flag.
+ */
+static int __must_push_back(struct multipath *m)
+{
+ return (m->queue_if_no_path != m->saved_queue_if_no_path &&
+ dm_noflush_suspending(m->ti));
+}
+
static int map_io(struct multipath *m, struct bio *bio, struct mpath_io *mpio,
unsigned was_queued)
{
- int r = 1;
+ int r = DM_MAPIO_REMAPPED;
unsigned long flags;
struct pgpath *pgpath;
@@ -310,11 +327,13 @@ static int map_io(struct multipath *m, struct bio *bio, struct mpath_io *mpio,
!m->queue_io)
queue_work(kmultipathd, &m->process_queued_ios);
pgpath = NULL;
- r = 0;
- } else if (!pgpath)
- r = -EIO; /* Failed */
- else
+ r = DM_MAPIO_SUBMITTED;
+ } else if (pgpath)
bio->bi_bdev = pgpath->path.dev->bdev;
+ else if (__must_push_back(m))
+ r = DM_MAPIO_REQUEUE;
+ else
+ r = -EIO; /* Failed */
mpio->pgpath = pgpath;
@@ -372,8 +391,10 @@ static void dispatch_queued_ios(struct multipath *m)
r = map_io(m, bio, mpio, 1);
if (r < 0)
bio_endio(bio, bio->bi_size, r);
- else if (r == 1)
+ else if (r == DM_MAPIO_REMAPPED)
generic_make_request(bio);
+ else if (r == DM_MAPIO_REQUEUE)
+ bio_endio(bio, bio->bi_size, -EIO);
bio = next;
}
@@ -783,7 +804,7 @@ static int multipath_map(struct dm_target *ti, struct bio *bio,
map_context->ptr = mpio;
bio->bi_rw |= (1 << BIO_RW_FAILFAST);
r = map_io(m, bio, mpio, 0);
- if (r < 0)
+ if (r < 0 || r == DM_MAPIO_REQUEUE)
mempool_free(mpio, m->mpio_pool);
return r;
@@ -957,7 +978,7 @@ static int bypass_pg_num(struct multipath *m, const char *pgstr, int bypassed)
/*
* pg_init must call this when it has completed its initialisation
*/
-void dm_pg_init_complete(struct path *path, unsigned err_flags)
+void dm_pg_init_complete(struct dm_path *path, unsigned err_flags)
{
struct pgpath *pgpath = path_to_pgpath(path);
struct priority_group *pg = pgpath->pg;
@@ -1007,7 +1028,10 @@ static int do_end_io(struct multipath *m, struct bio *bio,
spin_lock_irqsave(&m->lock, flags);
if (!m->nr_valid_paths) {
- if (!m->queue_if_no_path) {
+ if (__must_push_back(m)) {
+ spin_unlock_irqrestore(&m->lock, flags);
+ return DM_ENDIO_REQUEUE;
+ } else if (!m->queue_if_no_path) {
spin_unlock_irqrestore(&m->lock, flags);
return -EIO;
} else {
@@ -1042,7 +1066,7 @@ static int do_end_io(struct multipath *m, struct bio *bio,
queue_work(kmultipathd, &m->process_queued_ios);
spin_unlock_irqrestore(&m->lock, flags);
- return 1; /* io not complete */
+ return DM_ENDIO_INCOMPLETE; /* io not complete */
}
static int multipath_end_io(struct dm_target *ti, struct bio *bio,
@@ -1060,7 +1084,7 @@ static int multipath_end_io(struct dm_target *ti, struct bio *bio,
if (ps->type->end_io)
ps->type->end_io(ps, &pgpath->path);
}
- if (r <= 0)
+ if (r != DM_ENDIO_INCOMPLETE)
mempool_free(mpio, m->mpio_pool);
return r;
@@ -1272,7 +1296,7 @@ static int multipath_ioctl(struct dm_target *ti, struct inode *inode,
struct dentry fake_dentry = {};
int r = 0;
- fake_file.f_dentry = &fake_dentry;
+ fake_file.f_path.dentry = &fake_dentry;
spin_lock_irqsave(&m->lock, flags);
diff --git a/drivers/md/dm-mpath.h b/drivers/md/dm-mpath.h
index 8a4bf2b6d52..b9cdcbb3ed5 100644
--- a/drivers/md/dm-mpath.h
+++ b/drivers/md/dm-mpath.h
@@ -11,7 +11,7 @@
struct dm_dev;
-struct path {
+struct dm_path {
struct dm_dev *dev; /* Read-only */
unsigned is_active; /* Read-only */
@@ -20,6 +20,6 @@ struct path {
};
/* Callback for hwh_pg_init_fn to use when complete */
-void dm_pg_init_complete(struct path *path, unsigned err_flags);
+void dm_pg_init_complete(struct dm_path *path, unsigned err_flags);
#endif
diff --git a/drivers/md/dm-path-selector.h b/drivers/md/dm-path-selector.h
index 732d06a84f8..27357b85d73 100644
--- a/drivers/md/dm-path-selector.h
+++ b/drivers/md/dm-path-selector.h
@@ -44,7 +44,7 @@ struct path_selector_type {
* Add an opaque path object, along with some selector specific
* path args (eg, path priority).
*/
- int (*add_path) (struct path_selector *ps, struct path *path,
+ int (*add_path) (struct path_selector *ps, struct dm_path *path,
int argc, char **argv, char **error);
/*
@@ -55,27 +55,27 @@ struct path_selector_type {
* calling the function again. 0 means don't call it again unless
* the path fails.
*/
- struct path *(*select_path) (struct path_selector *ps,
+ struct dm_path *(*select_path) (struct path_selector *ps,
unsigned *repeat_count);
/*
* Notify the selector that a path has failed.
*/
- void (*fail_path) (struct path_selector *ps, struct path *p);
+ void (*fail_path) (struct path_selector *ps, struct dm_path *p);
/*
* Ask selector to reinstate a path.
*/
- int (*reinstate_path) (struct path_selector *ps, struct path *p);
+ int (*reinstate_path) (struct path_selector *ps, struct dm_path *p);
/*
* Table content based on parameters added in ps_add_path_fn
* or path selector status
*/
- int (*status) (struct path_selector *ps, struct path *path,
+ int (*status) (struct path_selector *ps, struct dm_path *path,
status_type_t type, char *result, unsigned int maxlen);
- int (*end_io) (struct path_selector *ps, struct path *path);
+ int (*end_io) (struct path_selector *ps, struct dm_path *path);
};
/* Register a path selector */
diff --git a/drivers/md/dm-raid1.c b/drivers/md/dm-raid1.c
index fc8cbb168e3..23a642619be 100644
--- a/drivers/md/dm-raid1.c
+++ b/drivers/md/dm-raid1.c
@@ -344,6 +344,17 @@ static void dispatch_bios(struct mirror_set *ms, struct bio_list *bio_list)
}
}
+static void complete_resync_work(struct region *reg, int success)
+{
+ struct region_hash *rh = reg->rh;
+
+ rh->log->type->set_region_sync(rh->log, reg->key, success);
+ dispatch_bios(rh->ms, &reg->delayed_bios);
+ if (atomic_dec_and_test(&rh->recovery_in_flight))
+ wake_up_all(&_kmirrord_recovery_stopped);
+ up(&rh->recovery_count);
+}
+
static void rh_update_states(struct region_hash *rh)
{
struct region *reg, *next;
@@ -383,11 +394,7 @@ static void rh_update_states(struct region_hash *rh)
*/
list_for_each_entry_safe (reg, next, &recovered, list) {
rh->log->type->clear_region(rh->log, reg->key);
- rh->log->type->complete_resync_work(rh->log, reg->key, 1);
- dispatch_bios(rh->ms, &reg->delayed_bios);
- if (atomic_dec_and_test(&rh->recovery_in_flight))
- wake_up_all(&_kmirrord_recovery_stopped);
- up(&rh->recovery_count);
+ complete_resync_work(reg, 1);
mempool_free(reg, rh->region_pool);
}
@@ -1137,7 +1144,7 @@ static int mirror_map(struct dm_target *ti, struct bio *bio,
if (rw == WRITE) {
queue_bio(ms, bio, rw);
- return 0;
+ return DM_MAPIO_SUBMITTED;
}
r = ms->rh.log->type->in_sync(ms->rh.log,
@@ -1146,7 +1153,7 @@ static int mirror_map(struct dm_target *ti, struct bio *bio,
return r;
if (r == -EWOULDBLOCK) /* FIXME: ugly */
- r = 0;
+ r = DM_MAPIO_SUBMITTED;
/*
* We don't want to fast track a recovery just for a read
@@ -1159,7 +1166,7 @@ static int mirror_map(struct dm_target *ti, struct bio *bio,
if (!r) {
/* Pass this io over to the daemon */
queue_bio(ms, bio, rw);
- return 0;
+ return DM_MAPIO_SUBMITTED;
}
m = choose_mirror(ms, bio->bi_sector);
@@ -1167,7 +1174,7 @@ static int mirror_map(struct dm_target *ti, struct bio *bio,
return -EIO;
map_bio(ms, m, bio);
- return 1;
+ return DM_MAPIO_REMAPPED;
}
static int mirror_end_io(struct dm_target *ti, struct bio *bio,
diff --git a/drivers/md/dm-round-robin.c b/drivers/md/dm-round-robin.c
index 6f9fcd4db9b..a348a97b65a 100644
--- a/drivers/md/dm-round-robin.c
+++ b/drivers/md/dm-round-robin.c
@@ -21,7 +21,7 @@
*---------------------------------------------------------------*/
struct path_info {
struct list_head list;
- struct path *path;
+ struct dm_path *path;
unsigned repeat_count;
};
@@ -80,7 +80,7 @@ static void rr_destroy(struct path_selector *ps)
ps->context = NULL;
}
-static int rr_status(struct path_selector *ps, struct path *path,
+static int rr_status(struct path_selector *ps, struct dm_path *path,
status_type_t type, char *result, unsigned int maxlen)
{
struct path_info *pi;
@@ -106,7 +106,7 @@ static int rr_status(struct path_selector *ps, struct path *path,
* Called during initialisation to register each path with an
* optional repeat_count.
*/
-static int rr_add_path(struct path_selector *ps, struct path *path,
+static int rr_add_path(struct path_selector *ps, struct dm_path *path,
int argc, char **argv, char **error)
{
struct selector *s = (struct selector *) ps->context;
@@ -141,7 +141,7 @@ static int rr_add_path(struct path_selector *ps, struct path *path,
return 0;
}
-static void rr_fail_path(struct path_selector *ps, struct path *p)
+static void rr_fail_path(struct path_selector *ps, struct dm_path *p)
{
struct selector *s = (struct selector *) ps->context;
struct path_info *pi = p->pscontext;
@@ -149,7 +149,7 @@ static void rr_fail_path(struct path_selector *ps, struct path *p)
list_move(&pi->list, &s->invalid_paths);
}
-static int rr_reinstate_path(struct path_selector *ps, struct path *p)
+static int rr_reinstate_path(struct path_selector *ps, struct dm_path *p)
{
struct selector *s = (struct selector *) ps->context;
struct path_info *pi = p->pscontext;
@@ -159,7 +159,7 @@ static int rr_reinstate_path(struct path_selector *ps, struct path *p)
return 0;
}
-static struct path *rr_select_path(struct path_selector *ps,
+static struct dm_path *rr_select_path(struct path_selector *ps,
unsigned *repeat_count)
{
struct selector *s = (struct selector *) ps->context;
diff --git a/drivers/md/dm-snap.c b/drivers/md/dm-snap.c
index b0ce2ce8227..0821a2b68a7 100644
--- a/drivers/md/dm-snap.c
+++ b/drivers/md/dm-snap.c
@@ -39,7 +39,7 @@
*/
#define SNAPSHOT_PAGES 256
-struct workqueue_struct *ksnapd;
+static struct workqueue_struct *ksnapd;
static void flush_queued_bios(struct work_struct *work);
struct pending_exception {
@@ -564,6 +564,17 @@ static int snapshot_ctr(struct dm_target *ti, unsigned int argc, char **argv)
return r;
}
+static void __free_exceptions(struct dm_snapshot *s)
+{
+ kcopyd_client_destroy(s->kcopyd_client);
+ s->kcopyd_client = NULL;
+
+ exit_exception_table(&s->pending, pending_cache);
+ exit_exception_table(&s->complete, exception_cache);
+
+ s->store.destroy(&s->store);
+}
+
static void snapshot_dtr(struct dm_target *ti)
{
struct dm_snapshot *s = (struct dm_snapshot *) ti->private;
@@ -574,13 +585,7 @@ static void snapshot_dtr(struct dm_target *ti)
/* After this returns there can be no new kcopyd jobs. */
unregister_snapshot(s);
- kcopyd_client_destroy(s->kcopyd_client);
-
- exit_exception_table(&s->pending, pending_cache);
- exit_exception_table(&s->complete, exception_cache);
-
- /* Deallocate memory used */
- s->store.destroy(&s->store);
+ __free_exceptions(s);
dm_put_device(ti, s->origin);
dm_put_device(ti, s->cow);
@@ -868,7 +873,7 @@ static int snapshot_map(struct dm_target *ti, struct bio *bio,
{
struct exception *e;
struct dm_snapshot *s = (struct dm_snapshot *) ti->private;
- int r = 1;
+ int r = DM_MAPIO_REMAPPED;
chunk_t chunk;
struct pending_exception *pe = NULL;
@@ -914,7 +919,7 @@ static int snapshot_map(struct dm_target *ti, struct bio *bio,
remap_exception(s, &pe->e, bio);
bio_list_add(&pe->snapshot_bios, bio);
- r = 0;
+ r = DM_MAPIO_SUBMITTED;
if (!pe->started) {
/* this is protected by snap->lock */
@@ -992,7 +997,7 @@ static int snapshot_status(struct dm_target *ti, status_type_t type,
*---------------------------------------------------------------*/
static int __origin_write(struct list_head *snapshots, struct bio *bio)
{
- int r = 1, first = 0;
+ int r = DM_MAPIO_REMAPPED, first = 0;
struct dm_snapshot *snap;
struct exception *e;
struct pending_exception *pe, *next_pe, *primary_pe = NULL;
@@ -1050,7 +1055,7 @@ static int __origin_write(struct list_head *snapshots, struct bio *bio)
bio_list_add(&primary_pe->origin_bios, bio);
- r = 0;
+ r = DM_MAPIO_SUBMITTED;
}
if (!pe->primary_pe) {
@@ -1099,7 +1104,7 @@ static int __origin_write(struct list_head *snapshots, struct bio *bio)
static int do_origin(struct dm_dev *origin, struct bio *bio)
{
struct origin *o;
- int r = 1;
+ int r = DM_MAPIO_REMAPPED;
down_read(&_origins_lock);
o = __lookup_origin(origin->bdev);
@@ -1156,7 +1161,7 @@ static int origin_map(struct dm_target *ti, struct bio *bio,
return -EOPNOTSUPP;
/* Only tell snapshots if this is a write */
- return (bio_rw(bio) == WRITE) ? do_origin(dev, bio) : 1;
+ return (bio_rw(bio) == WRITE) ? do_origin(dev, bio) : DM_MAPIO_REMAPPED;
}
#define min_not_zero(l, r) (l == 0) ? r : ((r == 0) ? l : min(l, r))
diff --git a/drivers/md/dm-stripe.c b/drivers/md/dm-stripe.c
index 6c29fcecd89..51f5e076001 100644
--- a/drivers/md/dm-stripe.c
+++ b/drivers/md/dm-stripe.c
@@ -186,7 +186,7 @@ static int stripe_map(struct dm_target *ti, struct bio *bio,
bio->bi_bdev = sc->stripe[stripe].dev->bdev;
bio->bi_sector = sc->stripe[stripe].physical_start +
(chunk << sc->chunk_shift) + (offset & sc->chunk_mask);
- return 1;
+ return DM_MAPIO_REMAPPED;
}
static int stripe_status(struct dm_target *ti,
diff --git a/drivers/md/dm-zero.c b/drivers/md/dm-zero.c
index ea569f7348d..f314d7dc9c2 100644
--- a/drivers/md/dm-zero.c
+++ b/drivers/md/dm-zero.c
@@ -46,7 +46,7 @@ static int zero_map(struct dm_target *ti, struct bio *bio,
bio_endio(bio, bio->bi_size, 0);
/* accepted bio, don't make new request */
- return 0;
+ return DM_MAPIO_SUBMITTED;
}
static struct target_type zero_target = {
diff --git a/drivers/md/dm.c b/drivers/md/dm.c
index 7ec1b112a6d..fe7c56e1043 100644
--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@ -68,10 +68,12 @@ union map_info *dm_get_mapinfo(struct bio *bio)
#define DMF_FROZEN 2
#define DMF_FREEING 3
#define DMF_DELETING 4
+#define DMF_NOFLUSH_SUSPENDING 5
struct mapped_device {
struct rw_semaphore io_lock;
struct semaphore suspend_lock;
+ spinlock_t pushback_lock;
rwlock_t map_lock;
atomic_t holders;
atomic_t open_count;
@@ -89,7 +91,8 @@ struct mapped_device {
*/
atomic_t pending;
wait_queue_head_t wait;
- struct bio_list deferred;
+ struct bio_list deferred;
+ struct bio_list pushback;
/*
* The current mapping.
@@ -444,23 +447,50 @@ int dm_set_geometry(struct mapped_device *md, struct hd_geometry *geo)
* you this clearly demarcated crap.
*---------------------------------------------------------------*/
+static int __noflush_suspending(struct mapped_device *md)
+{
+ return test_bit(DMF_NOFLUSH_SUSPENDING, &md->flags);
+}
+
/*
* Decrements the number of outstanding ios that a bio has been
* cloned into, completing the original io if necc.
*/
static void dec_pending(struct dm_io *io, int error)
{
- if (error)
+ unsigned long flags;
+
+ /* Push-back supersedes any I/O errors */
+ if (error && !(io->error > 0 && __noflush_suspending(io->md)))
io->error = error;
if (atomic_dec_and_test(&io->io_count)) {
+ if (io->error == DM_ENDIO_REQUEUE) {
+ /*
+ * Target requested pushing back the I/O.
+ * This must be handled before the sleeper on
+ * suspend queue merges the pushback list.
+ */
+ spin_lock_irqsave(&io->md->pushback_lock, flags);
+ if (__noflush_suspending(io->md))
+ bio_list_add(&io->md->pushback, io->bio);
+ else
+ /* noflush suspend was interrupted. */
+ io->error = -EIO;
+ spin_unlock_irqrestore(&io->md->pushback_lock, flags);
+ }
+
if (end_io_acct(io))
/* nudge anyone waiting on suspend queue */
wake_up(&io->md->wait);
- blk_add_trace_bio(io->md->queue, io->bio, BLK_TA_COMPLETE);
+ if (io->error != DM_ENDIO_REQUEUE) {
+ blk_add_trace_bio(io->md->queue, io->bio,
+ BLK_TA_COMPLETE);
+
+ bio_endio(io->bio, io->bio->bi_size, io->error);
+ }
- bio_endio(io->bio, io->bio->bi_size, io->error);
free_io(io->md, io);
}
}
@@ -480,12 +510,19 @@ static int clone_endio(struct bio *bio, unsigned int done, int error)
if (endio) {
r = endio(tio->ti, bio, error, &tio->info);
- if (r < 0)
+ if (r < 0 || r == DM_ENDIO_REQUEUE)
+ /*
+ * error and requeue request are handled
+ * in dec_pending().
+ */
error = r;
-
- else if (r > 0)
- /* the target wants another shot at the io */
+ else if (r == DM_ENDIO_INCOMPLETE)
+ /* The target will handle the io */
return 1;
+ else if (r) {
+ DMWARN("unimplemented target endio return value: %d", r);
+ BUG();
+ }
}
dec_pending(tio->io, error);
@@ -543,7 +580,7 @@ static void __map_bio(struct dm_target *ti, struct bio *clone,
atomic_inc(&tio->io->io_count);
sector = clone->bi_sector;
r = ti->type->map(ti, clone, &tio->info);
- if (r > 0) {
+ if (r == DM_MAPIO_REMAPPED) {
/* the bio has been remapped so dispatch it */
blk_add_trace_remap(bdev_get_queue(clone->bi_bdev), clone,
@@ -551,10 +588,8 @@ static void __map_bio(struct dm_target *ti, struct bio *clone,
clone->bi_sector);
generic_make_request(clone);
- }
-
- else if (r < 0) {
- /* error the io and bail out */
+ } else if (r < 0 || r == DM_MAPIO_REQUEUE) {
+ /* error the io and bail out, or requeue it if needed */
md = tio->io->md;
dec_pending(tio->io, r);
/*
@@ -563,6 +598,9 @@ static void __map_bio(struct dm_target *ti, struct bio *clone,
clone->bi_private = md->bs;
bio_put(clone);
free_tio(md, tio);
+ } else if (r) {
+ DMWARN("unimplemented target map return value: %d", r);
+ BUG();
}
}
@@ -948,6 +986,7 @@ static struct mapped_device *alloc_dev(int minor)
memset(md, 0, sizeof(*md));
init_rwsem(&md->io_lock);
init_MUTEX(&md->suspend_lock);
+ spin_lock_init(&md->pushback_lock);
rwlock_init(&md->map_lock);
atomic_set(&md->holders, 1);
atomic_set(&md->open_count, 0);
@@ -966,8 +1005,8 @@ static struct mapped_device *alloc_dev(int minor)
md->queue->issue_flush_fn = dm_flush_all;
md->io_pool = mempool_create_slab_pool(MIN_IOS, _io_cache);
- if (!md->io_pool)
- goto bad2;
+ if (!md->io_pool)
+ goto bad2;
md->tio_pool = mempool_create_slab_pool(MIN_IOS, _tio_cache);
if (!md->tio_pool)
@@ -1275,12 +1314,15 @@ static void unlock_fs(struct mapped_device *md)
* dm_bind_table, dm_suspend must be called to flush any in
* flight bios and ensure that any further io gets deferred.
*/
-int dm_suspend(struct mapped_device *md, int do_lockfs)
+int dm_suspend(struct mapped_device *md, unsigned suspend_flags)
{
struct dm_table *map = NULL;
+ unsigned long flags;
DECLARE_WAITQUEUE(wait, current);
struct bio *def;
int r = -EINVAL;
+ int do_lockfs = suspend_flags & DM_SUSPEND_LOCKFS_FLAG ? 1 : 0;
+ int noflush = suspend_flags & DM_SUSPEND_NOFLUSH_FLAG ? 1 : 0;
down(&md->suspend_lock);
@@ -1289,6 +1331,13 @@ int dm_suspend(struct mapped_device *md, int do_lockfs)
map = dm_get_table(md);
+ /*
+ * DMF_NOFLUSH_SUSPENDING must be set before presuspend.
+ * This flag is cleared before dm_suspend returns.
+ */
+ if (noflush)
+ set_bit(DMF_NOFLUSH_SUSPENDING, &md->flags);
+
/* This does not get reverted if there's an error later. */
dm_table_presuspend_targets(map);
@@ -1296,11 +1345,14 @@ int dm_suspend(struct mapped_device *md, int do_lockfs)
if (!md->suspended_bdev) {
DMWARN("bdget failed in dm_suspend");
r = -ENOMEM;
- goto out;
+ goto flush_and_out;
}
- /* Flush I/O to the device. */
- if (do_lockfs) {
+ /*
+ * Flush I/O to the device.
+ * noflush supersedes do_lockfs, because lock_fs() needs to flush I/Os.
+ */
+ if (do_lockfs && !noflush) {
r = lock_fs(md);
if (r)
goto out;
@@ -1336,6 +1388,14 @@ int dm_suspend(struct mapped_device *md, int do_lockfs)
down_write(&md->io_lock);
remove_wait_queue(&md->wait, &wait);
+ if (noflush) {
+ spin_lock_irqsave(&md->pushback_lock, flags);
+ clear_bit(DMF_NOFLUSH_SUSPENDING, &md->flags);
+ bio_list_merge_head(&md->deferred, &md->pushback);
+ bio_list_init(&md->pushback);
+ spin_unlock_irqrestore(&md->pushback_lock, flags);
+ }
+
/* were we interrupted ? */
r = -EINTR;
if (atomic_read(&md->pending)) {
@@ -1344,7 +1404,7 @@ int dm_suspend(struct mapped_device *md, int do_lockfs)
__flush_deferred_io(md, def);
up_write(&md->io_lock);
unlock_fs(md);
- goto out;
+ goto out; /* pushback list is already flushed, so skip flush */
}
up_write(&md->io_lock);
@@ -1354,6 +1414,25 @@ int dm_suspend(struct mapped_device *md, int do_lockfs)
r = 0;
+flush_and_out:
+ if (r && noflush) {
+ /*
+ * Because there may be already I/Os in the pushback list,
+ * flush them before return.
+ */
+ down_write(&md->io_lock);
+
+ spin_lock_irqsave(&md->pushback_lock, flags);
+ clear_bit(DMF_NOFLUSH_SUSPENDING, &md->flags);
+ bio_list_merge_head(&md->deferred, &md->pushback);
+ bio_list_init(&md->pushback);
+ spin_unlock_irqrestore(&md->pushback_lock, flags);
+
+ def = bio_list_get(&md->deferred);
+ __flush_deferred_io(md, def);
+ up_write(&md->io_lock);
+ }
+
out:
if (r && md->suspended_bdev) {
bdput(md->suspended_bdev);
@@ -1440,6 +1519,17 @@ int dm_suspended(struct mapped_device *md)
return test_bit(DMF_SUSPENDED, &md->flags);
}
+int dm_noflush_suspending(struct dm_target *ti)
+{
+ struct mapped_device *md = dm_table_get_md(ti->table);
+ int r = __noflush_suspending(md);
+
+ dm_put(md);
+
+ return r;
+}
+EXPORT_SYMBOL_GPL(dm_noflush_suspending);
+
static struct block_device_operations dm_blk_dops = {
.open = dm_blk_open,
.release = dm_blk_close,
diff --git a/drivers/md/dm.h b/drivers/md/dm.h
index a48ec5e3c1f..2f796b1436b 100644
--- a/drivers/md/dm.h
+++ b/drivers/md/dm.h
@@ -33,6 +33,25 @@
#define SECTOR_SHIFT 9
/*
+ * Definitions of return values from target end_io function.
+ */
+#define DM_ENDIO_INCOMPLETE 1
+#define DM_ENDIO_REQUEUE 2
+
+/*
+ * Definitions of return values from target map function.
+ */
+#define DM_MAPIO_SUBMITTED 0
+#define DM_MAPIO_REMAPPED 1
+#define DM_MAPIO_REQUEUE DM_ENDIO_REQUEUE
+
+/*
+ * Suspend feature flags
+ */
+#define DM_SUSPEND_LOCKFS_FLAG (1 << 0)
+#define DM_SUSPEND_NOFLUSH_FLAG (1 << 1)
+
+/*
* List of devices that a metadevice uses and should open/close.
*/
struct dm_dev {
diff --git a/drivers/md/md.c b/drivers/md/md.c
index 6c4345bde07..53bd46dba0c 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -1413,7 +1413,7 @@ static int lock_rdev(mdk_rdev_t *rdev, dev_t dev)
struct block_device *bdev;
char b[BDEVNAME_SIZE];
- bdev = open_partition_by_devnum(dev, FMODE_READ|FMODE_WRITE);
+ bdev = open_by_devnum(dev, FMODE_READ|FMODE_WRITE);
if (IS_ERR(bdev)) {
printk(KERN_ERR "md: could not open %s.\n",
__bdevname(dev, b));
@@ -1423,7 +1423,7 @@ static int lock_rdev(mdk_rdev_t *rdev, dev_t dev)
if (err) {
printk(KERN_ERR "md: could not bd_claim %s.\n",
bdevname(bdev, b));
- blkdev_put_partition(bdev);
+ blkdev_put(bdev);
return err;
}
rdev->bdev = bdev;
@@ -1437,7 +1437,7 @@ static void unlock_rdev(mdk_rdev_t *rdev)
if (!bdev)
MD_BUG();
bd_release(bdev);
- blkdev_put_partition(bdev);
+ blkdev_put(bdev);
}
void md_autodetect_dev(dev_t dev);
@@ -4423,7 +4423,7 @@ static int md_open(struct inode *inode, struct file *file)
mddev_t *mddev = inode->i_bdev->bd_disk->private_data;
int err;
- if ((err = mddev_lock(mddev)))
+ if ((err = mutex_lock_interruptible_nested(&mddev->reconfig_mutex, 1)))
goto out;
err = 0;
@@ -4846,8 +4846,8 @@ static int md_seq_show(struct seq_file *seq, void *v)
chunk_kb ? "KB" : "B");
if (bitmap->file) {
seq_printf(seq, ", file: ");
- seq_path(seq, bitmap->file->f_vfsmnt,
- bitmap->file->f_dentry," \t\n");
+ seq_path(seq, bitmap->file->f_path.mnt,
+ bitmap->file->f_path.dentry," \t\n");
}
seq_printf(seq, "\n");
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2024-10-22 20:49:37 +0000