X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=drivers%2Fmd%2Fdm.c;h=08a4654cc02d82e69d55ef72e5e2bc062e2de76e;hb=97bf2856c6014879bd04983a3e9dfcdac1e7fe85;hp=03381e0036073f7daf4586a2ae71363f5da43944;hpb=9213980e6a70d8473e0ffd4b39ab5b6caaba9ff5;p=linux-2.6.git diff --git a/drivers/md/dm.c b/drivers/md/dm.c index 03381e003..08a4654cc 100644 --- a/drivers/md/dm.c +++ b/drivers/md/dm.c @@ -1,5 +1,6 @@ /* * Copyright (C) 2001, 2002 Sistina Software (UK) Limited. + * Copyright (C) 2004-2006 Red Hat, Inc. All rights reserved. * * This file is released under the GPL. */ @@ -9,19 +10,27 @@ #include #include +#include #include #include #include #include #include #include +#include +#include +#include +#include +#include + +#define DM_MSG_PREFIX "core" static const char *_name = DM_NAME; -#define MAX_DEVICES 1024 static unsigned int major = 0; static unsigned int _major = 0; +static DEFINE_SPINLOCK(_minor_lock); /* * One of these is allocated per bio. */ @@ -30,6 +39,7 @@ struct dm_io { int error; struct bio *bio; atomic_t io_count; + unsigned long start_time; }; /* @@ -42,29 +52,49 @@ struct target_io { union map_info info; }; +union map_info *dm_get_mapinfo(struct bio *bio) +{ + if (bio && bio->bi_private) + return &((struct target_io *)bio->bi_private)->info; + return NULL; +} + +#define MINOR_ALLOCED ((void *)-1) + /* * Bits for the md->flags field. */ #define DMF_BLOCK_IO 0 #define DMF_SUSPENDED 1 -#define DMF_FS_LOCKED 2 +#define DMF_FROZEN 2 +#define DMF_FREEING 3 +#define DMF_DELETING 4 +#define DMF_NOFLUSH_SUSPENDING 5 struct mapped_device { - struct rw_semaphore lock; + struct rw_semaphore io_lock; + struct semaphore suspend_lock; + spinlock_t pushback_lock; rwlock_t map_lock; atomic_t holders; + atomic_t open_count; + xid_t xid; unsigned long flags; request_queue_t *queue; struct gendisk *disk; + char name[16]; + + void *interface_ptr; /* * A list of ios that arrived while we were suspended. */ atomic_t pending; wait_queue_head_t wait; - struct bio_list deferred; + struct bio_list deferred; + struct bio_list pushback; /* * The current mapping. @@ -77,6 +107,8 @@ struct mapped_device { mempool_t *io_pool; mempool_t *tio_pool; + struct bio_set *bs; + /* * Event handling. */ @@ -87,11 +119,15 @@ struct mapped_device { * freeze/thaw support require holding onto a super block */ struct super_block *frozen_sb; + struct block_device *suspended_bdev; + + /* forced geometry settings */ + struct hd_geometry geometry; }; #define MIN_IOS 256 -static kmem_cache_t *_io_cache; -static kmem_cache_t *_tio_cache; +static struct kmem_cache *_io_cache; +static struct kmem_cache *_tio_cache; static int __init local_init(void) { @@ -131,30 +167,27 @@ static void local_exit(void) kmem_cache_destroy(_io_cache); if (unregister_blkdev(_major, _name) < 0) - DMERR("devfs_unregister_blkdev failed"); + DMERR("unregister_blkdev failed"); _major = 0; DMINFO("cleaned up"); } -/* - * We have a lot of init/exit functions, so it seems easier to - * store them in an array. The disposable macro 'xx' - * expands a prefix into a pair of function names. - */ -static struct { - int (*init) (void); - void (*exit) (void); - -} _inits[] = { -#define xx(n) {n ## _init, n ## _exit}, - xx(local) - xx(dm_target) - xx(dm_linear) - xx(dm_stripe) - xx(dm_interface) -#undef xx +int (*_inits[])(void) __initdata = { + local_init, + dm_target_init, + dm_linear_init, + dm_stripe_init, + dm_interface_init, +}; + +void (*_exits[])(void) = { + local_exit, + dm_target_exit, + dm_linear_exit, + dm_stripe_exit, + dm_interface_exit, }; static int __init dm_init(void) @@ -164,7 +197,7 @@ static int __init dm_init(void) int r, i; for (i = 0; i < count; i++) { - r = _inits[i].init(); + r = _inits[i](); if (r) goto bad; } @@ -173,17 +206,17 @@ static int __init dm_init(void) bad: while (i--) - _inits[i].exit(); + _exits[i](); return r; } static void __exit dm_exit(void) { - int i = ARRAY_SIZE(_inits); + int i = ARRAY_SIZE(_exits); while (i--) - _inits[i].exit(); + _exits[i](); } /* @@ -192,10 +225,28 @@ static void __exit dm_exit(void) static int dm_blk_open(struct inode *inode, struct file *file) { struct mapped_device *md; + int ret = -ENXIO; + + spin_lock(&_minor_lock); md = inode->i_bdev->bd_disk->private_data; + if (!md) + goto out; + + if (test_bit(DMF_FREEING, &md->flags) || + test_bit(DMF_DELETING, &md->flags)) + goto out; + + ret = -EACCES; + if (!vx_check(md->xid, VS_IDENT|VS_HOSTID)) + goto out; + dm_get(md); - return 0; + atomic_inc(&md->open_count); + ret = 0; +out: + spin_unlock(&_minor_lock); + return ret; } static int dm_blk_close(struct inode *inode, struct file *file) @@ -203,10 +254,81 @@ static int dm_blk_close(struct inode *inode, struct file *file) struct mapped_device *md; md = inode->i_bdev->bd_disk->private_data; + atomic_dec(&md->open_count); dm_put(md); return 0; } +int dm_open_count(struct mapped_device *md) +{ + return atomic_read(&md->open_count); +} + +/* + * Guarantees nothing is using the device before it's deleted. + */ +int dm_lock_for_deletion(struct mapped_device *md) +{ + int r = 0; + + spin_lock(&_minor_lock); + + if (dm_open_count(md)) + r = -EBUSY; + else + set_bit(DMF_DELETING, &md->flags); + + spin_unlock(&_minor_lock); + + return r; +} + +static int dm_blk_getgeo(struct block_device *bdev, struct hd_geometry *geo) +{ + struct mapped_device *md = bdev->bd_disk->private_data; + + return dm_get_geometry(md, geo); +} + +static int dm_blk_ioctl(struct inode *inode, struct file *file, + unsigned int cmd, unsigned long arg) +{ + struct mapped_device *md; + struct dm_table *map; + struct dm_target *tgt; + int r = -ENOTTY; + + /* We don't really need this lock, but we do need 'inode'. */ + unlock_kernel(); + + md = inode->i_bdev->bd_disk->private_data; + + map = dm_get_table(md); + + if (!map || !dm_table_get_size(map)) + goto out; + + /* We only support devices that have a single target */ + if (dm_table_get_num_targets(map) != 1) + goto out; + + tgt = dm_table_get_target(map, 0); + + if (dm_suspended(md)) { + r = -EAGAIN; + goto out; + } + + if (tgt->type->ioctl) + r = tgt->type->ioctl(tgt, inode, file, cmd, arg); + +out: + dm_table_put(map); + + lock_kernel(); + return r; +} + static inline struct dm_io *alloc_io(struct mapped_device *md) { return mempool_alloc(md->io_pool, GFP_NOIO); @@ -227,21 +349,51 @@ static inline void free_tio(struct mapped_device *md, struct target_io *tio) mempool_free(tio, md->tio_pool); } +static void start_io_acct(struct dm_io *io) +{ + struct mapped_device *md = io->md; + + io->start_time = jiffies; + + preempt_disable(); + disk_round_stats(dm_disk(md)); + preempt_enable(); + dm_disk(md)->in_flight = atomic_inc_return(&md->pending); +} + +static int end_io_acct(struct dm_io *io) +{ + struct mapped_device *md = io->md; + struct bio *bio = io->bio; + unsigned long duration = jiffies - io->start_time; + int pending; + int rw = bio_data_dir(bio); + + preempt_disable(); + disk_round_stats(dm_disk(md)); + preempt_enable(); + dm_disk(md)->in_flight = pending = atomic_dec_return(&md->pending); + + disk_stat_add(dm_disk(md), ticks[rw], duration); + + return !pending; +} + /* * Add the bio to the list of deferred io. */ static int queue_io(struct mapped_device *md, struct bio *bio) { - down_write(&md->lock); + down_write(&md->io_lock); if (!test_bit(DMF_BLOCK_IO, &md->flags)) { - up_write(&md->lock); + up_write(&md->io_lock); return 1; } bio_list_add(&md->deferred, bio); - up_write(&md->lock); + up_write(&md->io_lock); return 0; /* deferred successfully */ } @@ -263,6 +415,41 @@ struct dm_table *dm_get_table(struct mapped_device *md) return t; } +/* + * Get the geometry associated with a dm device + */ +int dm_get_geometry(struct mapped_device *md, struct hd_geometry *geo) +{ + *geo = md->geometry; + + return 0; +} + +/* + * Set the geometry of a device. + */ +int dm_set_geometry(struct mapped_device *md, struct hd_geometry *geo) +{ + sector_t sz = (sector_t)geo->cylinders * geo->heads * geo->sectors; + + if (geo->start > sz) { + DMWARN("Start sector is beyond the geometry limits."); + return -EINVAL; + } + + md->geometry = *geo; + + return 0; +} + +/* + * Get the xid associated with a dm device + */ +xid_t dm_get_xid(struct mapped_device *md) +{ + return md->xid; +} + /*----------------------------------------------------------------- * CRUD START: * A more elegant soln is in the works that uses the queue @@ -272,21 +459,50 @@ struct dm_table *dm_get_table(struct mapped_device *md) * 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 inline void dec_pending(struct dm_io *io, int error) +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 (atomic_dec_and_test(&io->md->pending)) + 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); - bio_endio(io->bio, io->bio->bi_size, io->error); + 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); + } + free_io(io->md, io); } } @@ -295,7 +511,7 @@ static int clone_endio(struct bio *bio, unsigned int done, int error) { int r = 0; struct target_io *tio = bio->bi_private; - struct dm_io *io = tio->io; + struct mapped_device *md = tio->io->md; dm_endio_fn endio = tio->ti->type->end_io; if (bio->bi_size) @@ -306,17 +522,30 @@ 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(); + } } - free_tio(io->md, tio); - dec_pending(io, error); + dec_pending(tio->io, error); + + /* + * Store md for cleanup instead of tio which is about to get freed. + */ + bio->bi_private = md->bs; + bio_put(bio); + free_tio(md, tio); return r; } @@ -331,8 +560,8 @@ static sector_t max_io_len(struct mapped_device *md, */ if (ti->split_io) { sector_t boundary; - boundary = dm_round_up(offset + 1, ti->split_io) - offset; - + boundary = ((offset + ti->split_io) & ~(ti->split_io - 1)) + - offset; if (len > boundary) len = boundary; } @@ -344,6 +573,8 @@ static void __map_bio(struct dm_target *ti, struct bio *clone, struct target_io *tio) { int r; + sector_t sector; + struct mapped_device *md; /* * Sanity checks. @@ -359,17 +590,29 @@ static void __map_bio(struct dm_target *ti, struct bio *clone, * this io. */ 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 */ - generic_make_request(clone); - else if (r < 0) { - /* error the io and bail out */ - struct dm_io *io = tio->io; - free_tio(tio->io->md, tio); - dec_pending(io, -EIO); + blk_add_trace_remap(bdev_get_queue(clone->bi_bdev), clone, + tio->io->bio->bi_bdev->bd_dev, sector, + clone->bi_sector); + + generic_make_request(clone); + } 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); + /* + * Store bio_set for cleanup. + */ + clone->bi_private = md->bs; bio_put(clone); + free_tio(md, tio); + } else if (r) { + DMWARN("unimplemented target map return value: %d", r); + BUG(); } } @@ -383,18 +626,26 @@ struct clone_info { unsigned short idx; }; +static void dm_bio_destructor(struct bio *bio) +{ + struct bio_set *bs = bio->bi_private; + + bio_free(bio, bs); +} + /* * Creates a little bio that is just does part of a bvec. */ static struct bio *split_bvec(struct bio *bio, sector_t sector, unsigned short idx, unsigned int offset, - unsigned int len) + unsigned int len, struct bio_set *bs) { struct bio *clone; struct bio_vec *bv = bio->bi_io_vec + idx; - clone = bio_alloc(GFP_NOIO, 1); - memcpy(clone->bi_io_vec, bv, sizeof(*bv)); + clone = bio_alloc_bioset(GFP_NOIO, 1, bs); + clone->bi_destructor = dm_bio_destructor; + *clone->bi_io_vec = *bv; clone->bi_sector = sector; clone->bi_bdev = bio->bi_bdev; @@ -412,11 +663,13 @@ static struct bio *split_bvec(struct bio *bio, sector_t sector, */ static struct bio *clone_bio(struct bio *bio, sector_t sector, unsigned short idx, unsigned short bv_count, - unsigned int len) + unsigned int len, struct bio_set *bs) { struct bio *clone; - clone = bio_clone(bio, GFP_NOIO); + clone = bio_alloc_bioset(GFP_NOIO, bio->bi_max_vecs, bs); + __bio_clone(clone, bio); + clone->bi_destructor = dm_bio_destructor; clone->bi_sector = sector; clone->bi_idx = idx; clone->bi_vcnt = idx + bv_count; @@ -447,7 +700,8 @@ static void __clone_and_map(struct clone_info *ci) * the remaining io with a single clone. */ clone = clone_bio(bio, ci->sector, ci->idx, - bio->bi_vcnt - ci->idx, ci->sector_count); + bio->bi_vcnt - ci->idx, ci->sector_count, + ci->md->bs); __map_bio(ti, clone, tio); ci->sector_count = 0; @@ -470,7 +724,8 @@ static void __clone_and_map(struct clone_info *ci) len += bv_len; } - clone = clone_bio(bio, ci->sector, ci->idx, i - ci->idx, len); + clone = clone_bio(bio, ci->sector, ci->idx, i - ci->idx, len, + ci->md->bs); __map_bio(ti, clone, tio); ci->sector += len; @@ -479,30 +734,36 @@ static void __clone_and_map(struct clone_info *ci) } else { /* - * Create two copy bios to deal with io that has - * been split across a target. + * Handle a bvec that must be split between two or more targets. */ struct bio_vec *bv = bio->bi_io_vec + ci->idx; + sector_t remaining = to_sector(bv->bv_len); + unsigned int offset = 0; - clone = split_bvec(bio, ci->sector, ci->idx, - bv->bv_offset, max); - __map_bio(ti, clone, tio); + do { + if (offset) { + ti = dm_table_find_target(ci->map, ci->sector); + max = max_io_len(ci->md, ci->sector, ti); - ci->sector += max; - ci->sector_count -= max; - ti = dm_table_find_target(ci->map, ci->sector); - - len = to_sector(bv->bv_len) - max; - clone = split_bvec(bio, ci->sector, ci->idx, - bv->bv_offset + to_bytes(max), len); - tio = alloc_tio(ci->md); - tio->io = ci->io; - tio->ti = ti; - memset(&tio->info, 0, sizeof(tio->info)); - __map_bio(ti, clone, tio); + tio = alloc_tio(ci->md); + tio->io = ci->io; + tio->ti = ti; + memset(&tio->info, 0, sizeof(tio->info)); + } + + len = min(remaining, max); + + clone = split_bvec(bio, ci->sector, ci->idx, + bv->bv_offset + offset, len, + ci->md->bs); + + __map_bio(ti, clone, tio); + + ci->sector += len; + ci->sector_count -= len; + offset += to_bytes(len); + } while (remaining -= len); - ci->sector += len; - ci->sector_count -= len; ci->idx++; } } @@ -531,7 +792,7 @@ static void __split_bio(struct mapped_device *md, struct bio *bio) ci.sector_count = bio_sectors(bio); ci.idx = bio->bi_idx; - atomic_inc(&md->pending); + start_io_acct(ci.io); while (ci.sector_count) __clone_and_map(&ci); @@ -550,16 +811,20 @@ static void __split_bio(struct mapped_device *md, struct bio *bio) static int dm_request(request_queue_t *q, struct bio *bio) { int r; + int rw = bio_data_dir(bio); struct mapped_device *md = q->queuedata; - down_read(&md->lock); + down_read(&md->io_lock); + + disk_stat_inc(dm_disk(md), ios[rw]); + disk_stat_add(dm_disk(md), sectors[rw], bio_sectors(bio)); /* * If we're suspended we have to queue * this io for later. */ while (test_bit(DMF_BLOCK_IO, &md->flags)) { - up_read(&md->lock); + up_read(&md->io_lock); if (bio_rw(bio) == READA) { bio_io_error(bio, bio->bi_size); @@ -578,14 +843,29 @@ static int dm_request(request_queue_t *q, struct bio *bio) * We're in a while loop, because someone could suspend * before we get to the following read lock. */ - down_read(&md->lock); + down_read(&md->io_lock); } __split_bio(md, bio); - up_read(&md->lock); + up_read(&md->io_lock); return 0; } +static int dm_flush_all(request_queue_t *q, struct gendisk *disk, + sector_t *error_sector) +{ + struct mapped_device *md = q->queuedata; + struct dm_table *map = dm_get_table(md); + int ret = -ENXIO; + + if (map) { + ret = dm_table_flush_all(map); + dm_table_put(map); + } + + return ret; +} + static void dm_unplug_all(request_queue_t *q) { struct mapped_device *md = q->queuedata; @@ -613,53 +893,78 @@ static int dm_any_congested(void *congested_data, int bdi_bits) } /*----------------------------------------------------------------- - * A bitset is used to keep track of allocated minor numbers. + * An IDR is used to keep track of allocated minor numbers. *---------------------------------------------------------------*/ -static spinlock_t _minor_lock = SPIN_LOCK_UNLOCKED; -static unsigned long _minor_bits[MAX_DEVICES / BITS_PER_LONG]; +static DEFINE_IDR(_minor_idr); -static void free_minor(unsigned int minor) +static void free_minor(int minor) { spin_lock(&_minor_lock); - clear_bit(minor, _minor_bits); + idr_remove(&_minor_idr, minor); spin_unlock(&_minor_lock); } /* * See if the device with a specific minor # is free. */ -static int specific_minor(unsigned int minor) +static int specific_minor(struct mapped_device *md, int minor) { - int r = -EBUSY; + int r, m; - if (minor >= MAX_DEVICES) { - DMWARN("request for a mapped_device beyond MAX_DEVICES (%d)", - MAX_DEVICES); + if (minor >= (1 << MINORBITS)) return -EINVAL; - } + + r = idr_pre_get(&_minor_idr, GFP_KERNEL); + if (!r) + return -ENOMEM; spin_lock(&_minor_lock); - if (!test_and_set_bit(minor, _minor_bits)) - r = 0; - spin_unlock(&_minor_lock); + if (idr_find(&_minor_idr, minor)) { + r = -EBUSY; + goto out; + } + + r = idr_get_new_above(&_minor_idr, MINOR_ALLOCED, minor, &m); + if (r) + goto out; + + if (m != minor) { + idr_remove(&_minor_idr, m); + r = -EBUSY; + goto out; + } + +out: + spin_unlock(&_minor_lock); return r; } -static int next_free_minor(unsigned int *minor) +static int next_free_minor(struct mapped_device *md, int *minor) { - int r = -EBUSY; - unsigned int m; + int r, m; + + r = idr_pre_get(&_minor_idr, GFP_KERNEL); + if (!r) + return -ENOMEM; spin_lock(&_minor_lock); - m = find_first_zero_bit(_minor_bits, MAX_DEVICES); - if (m != MAX_DEVICES) { - set_bit(m, _minor_bits); - *minor = m; - r = 0; + + r = idr_get_new(&_minor_idr, MINOR_ALLOCED, &m); + if (r) { + goto out; } - spin_unlock(&_minor_lock); + if (m >= (1 << MINORBITS)) { + idr_remove(&_minor_idr, m); + r = -ENOSPC; + goto out; + } + + *minor = m; + +out: + spin_unlock(&_minor_lock); return r; } @@ -668,51 +973,70 @@ static struct block_device_operations dm_blk_dops; /* * Allocate and initialise a blank device with a given minor. */ -static struct mapped_device *alloc_dev(unsigned int minor, int persistent) +static struct mapped_device *alloc_dev(int minor) { int r; struct mapped_device *md = kmalloc(sizeof(*md), GFP_KERNEL); + void *old_md; if (!md) { DMWARN("unable to allocate device, out of memory."); return NULL; } + if (!try_module_get(THIS_MODULE)) + goto bad0; + /* get a minor number for the dev */ - r = persistent ? specific_minor(minor) : next_free_minor(&minor); + if (minor == DM_ANY_MINOR) + r = next_free_minor(md, &minor); + else + r = specific_minor(md, minor); if (r < 0) goto bad1; memset(md, 0, sizeof(*md)); - init_rwsem(&md->lock); + 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); atomic_set(&md->event_nr, 0); + md->xid = vx_current_xid(); md->queue = blk_alloc_queue(GFP_KERNEL); if (!md->queue) - goto bad1; + goto bad1_free_minor; md->queue->queuedata = md; md->queue->backing_dev_info.congested_fn = dm_any_congested; md->queue->backing_dev_info.congested_data = md; blk_queue_make_request(md->queue, dm_request); + blk_queue_bounce_limit(md->queue, BLK_BOUNCE_ANY); md->queue->unplug_fn = dm_unplug_all; + md->queue->issue_flush_fn = dm_flush_all; - md->io_pool = mempool_create(MIN_IOS, mempool_alloc_slab, - mempool_free_slab, _io_cache); - if (!md->io_pool) - goto bad2; + md->io_pool = mempool_create_slab_pool(MIN_IOS, _io_cache); + if (!md->io_pool) + goto bad2; - md->tio_pool = mempool_create(MIN_IOS, mempool_alloc_slab, - mempool_free_slab, _tio_cache); + md->tio_pool = mempool_create_slab_pool(MIN_IOS, _tio_cache); if (!md->tio_pool) goto bad3; + md->bs = bioset_create(16, 16, 4); + if (!md->bs) + goto bad_no_bioset; + md->disk = alloc_disk(1); if (!md->disk) goto bad4; + atomic_set(&md->pending, 0); + init_waitqueue_head(&md->wait); + init_waitqueue_head(&md->eventq); + md->disk->major = _major; md->disk->first_minor = minor; md->disk->fops = &dm_blk_dops; @@ -720,33 +1044,55 @@ static struct mapped_device *alloc_dev(unsigned int minor, int persistent) md->disk->private_data = md; sprintf(md->disk->disk_name, "dm-%d", minor); add_disk(md->disk); + format_dev_t(md->name, MKDEV(_major, minor)); - atomic_set(&md->pending, 0); - init_waitqueue_head(&md->wait); - init_waitqueue_head(&md->eventq); + /* Populate the mapping, nobody knows we exist yet */ + spin_lock(&_minor_lock); + old_md = idr_replace(&_minor_idr, md, minor); + spin_unlock(&_minor_lock); + + BUG_ON(old_md != MINOR_ALLOCED); return md; bad4: + bioset_free(md->bs); + bad_no_bioset: mempool_destroy(md->tio_pool); bad3: mempool_destroy(md->io_pool); bad2: - blk_put_queue(md->queue); + blk_cleanup_queue(md->queue); + bad1_free_minor: free_minor(minor); bad1: + module_put(THIS_MODULE); + bad0: kfree(md); return NULL; } static void free_dev(struct mapped_device *md) { - free_minor(md->disk->first_minor); + int minor = md->disk->first_minor; + + if (md->suspended_bdev) { + thaw_bdev(md->suspended_bdev, NULL); + bdput(md->suspended_bdev); + } mempool_destroy(md->tio_pool); mempool_destroy(md->io_pool); + bioset_free(md->bs); del_gendisk(md->disk); + free_minor(minor); + + spin_lock(&_minor_lock); + md->disk->private_data = NULL; + spin_unlock(&_minor_lock); + put_disk(md->disk); - blk_put_queue(md->queue); + blk_cleanup_queue(md->queue); + module_put(THIS_MODULE); kfree(md); } @@ -757,22 +1103,17 @@ static void event_callback(void *context) { struct mapped_device *md = (struct mapped_device *) context; - atomic_inc(&md->event_nr);; + atomic_inc(&md->event_nr); wake_up(&md->eventq); } -static void __set_size(struct gendisk *disk, sector_t size) +static void __set_size(struct mapped_device *md, sector_t size) { - struct block_device *bdev; + set_capacity(md->disk, size); - set_capacity(disk, size); - bdev = bdget_disk(disk, 0); - if (bdev) { - down(&bdev->bd_inode->i_sem); - i_size_write(bdev->bd_inode, (loff_t)size << SECTOR_SHIFT); - up(&bdev->bd_inode->i_sem); - bdput(bdev); - } + mutex_lock(&md->suspended_bdev->bd_inode->i_mutex); + i_size_write(md->suspended_bdev->bd_inode, (loff_t)size << SECTOR_SHIFT); + mutex_unlock(&md->suspended_bdev->bd_inode->i_mutex); } static int __bind(struct mapped_device *md, struct dm_table *t) @@ -781,17 +1122,26 @@ static int __bind(struct mapped_device *md, struct dm_table *t) sector_t size; size = dm_table_get_size(t); - __set_size(md->disk, size); + + /* + * Wipe any geometry if the size of the table changed. + */ + if (size != get_capacity(md->disk)) + memset(&md->geometry, 0, sizeof(md->geometry)); + + if (md->suspended_bdev) + __set_size(md, size); if (size == 0) return 0; + dm_table_get(t); + dm_table_event_callback(t, event_callback, md); + write_lock(&md->map_lock); md->map = t; + dm_table_set_restrictions(t, q); write_unlock(&md->map_lock); - dm_table_get(t); - dm_table_event_callback(md->map, event_callback, md); - dm_table_set_restrictions(t, q); return 0; } @@ -812,12 +1162,11 @@ static void __unbind(struct mapped_device *md) /* * Constructor for a new device. */ -static int create_aux(unsigned int minor, int persistent, - struct mapped_device **result) +int dm_create(int minor, struct mapped_device **result) { struct mapped_device *md; - md = alloc_dev(minor, persistent); + md = alloc_dev(minor); if (!md) return -ENXIO; @@ -825,14 +1174,48 @@ static int create_aux(unsigned int minor, int persistent, return 0; } -int dm_create(struct mapped_device **result) +static struct mapped_device *dm_find_md(dev_t dev) +{ + struct mapped_device *md; + unsigned minor = MINOR(dev); + + if (MAJOR(dev) != _major || minor >= (1 << MINORBITS)) + return NULL; + + spin_lock(&_minor_lock); + + md = idr_find(&_minor_idr, minor); + if (md && (md == MINOR_ALLOCED || + (dm_disk(md)->first_minor != minor) || + test_bit(DMF_FREEING, &md->flags))) { + md = NULL; + goto out; + } + +out: + spin_unlock(&_minor_lock); + + return md; +} + +struct mapped_device *dm_get_md(dev_t dev) +{ + struct mapped_device *md = dm_find_md(dev); + + if (md) + dm_get(md); + + return md; +} + +void *dm_get_mdptr(struct mapped_device *md) { - return create_aux(0, 0, result); + return md->interface_ptr; } -int dm_create_with_minor(unsigned int minor, struct mapped_device **result) +void dm_set_mdptr(struct mapped_device *md, void *ptr) { - return create_aux(minor, 1, result); + md->interface_ptr = ptr; } void dm_get(struct mapped_device *md) @@ -840,18 +1223,31 @@ void dm_get(struct mapped_device *md) atomic_inc(&md->holders); } +const char *dm_device_name(struct mapped_device *md) +{ + return md->name; +} +EXPORT_SYMBOL_GPL(dm_device_name); + void dm_put(struct mapped_device *md) { - struct dm_table *map = dm_get_table(md); + struct dm_table *map; - if (atomic_dec_and_test(&md->holders)) { - if (!test_bit(DMF_SUSPENDED, &md->flags) && map) - dm_table_suspend_targets(map); + BUG_ON(test_bit(DMF_FREEING, &md->flags)); + + if (atomic_dec_and_lock(&md->holders, &_minor_lock)) { + map = dm_get_table(md); + idr_replace(&_minor_idr, MINOR_ALLOCED, dm_disk(md)->first_minor); + set_bit(DMF_FREEING, &md->flags); + spin_unlock(&_minor_lock); + if (!dm_suspended(md)) { + dm_table_presuspend_targets(map); + dm_table_postsuspend_targets(map); + } __unbind(md); + dm_table_put(map); free_dev(md); } - - dm_table_put(map); } /* @@ -874,69 +1270,60 @@ static void __flush_deferred_io(struct mapped_device *md, struct bio *c) */ int dm_swap_table(struct mapped_device *md, struct dm_table *table) { - int r; + int r = -EINVAL; - down_write(&md->lock); + down(&md->suspend_lock); /* device must be suspended */ - if (!test_bit(DMF_SUSPENDED, &md->flags)) { - up_write(&md->lock); - return -EPERM; - } + if (!dm_suspended(md)) + goto out; + + /* without bdev, the device size cannot be changed */ + if (!md->suspended_bdev) + if (get_capacity(md->disk) != dm_table_get_size(table)) + goto out; __unbind(md); r = __bind(md, table); - if (r) - return r; - up_write(&md->lock); - return 0; +out: + up(&md->suspend_lock); + return r; } /* * Functions to lock and unlock any filesystem running on the * device. */ -static int __lock_fs(struct mapped_device *md) +static int lock_fs(struct mapped_device *md) { - struct block_device *bdev; + int r; - if (test_and_set_bit(DMF_FS_LOCKED, &md->flags)) - return 0; + WARN_ON(md->frozen_sb); - bdev = bdget_disk(md->disk, 0); - if (!bdev) { - DMWARN("bdget failed in __lock_fs"); - return -ENOMEM; + md->frozen_sb = freeze_bdev(md->suspended_bdev); + if (IS_ERR(md->frozen_sb)) { + r = PTR_ERR(md->frozen_sb); + md->frozen_sb = NULL; + return r; } - WARN_ON(md->frozen_sb); - md->frozen_sb = freeze_bdev(bdev); + set_bit(DMF_FROZEN, &md->flags); + /* don't bdput right now, we don't want the bdev - * to go away while it is locked. We'll bdput - * in __unlock_fs + * to go away while it is locked. */ return 0; } -static int __unlock_fs(struct mapped_device *md) +static void unlock_fs(struct mapped_device *md) { - struct block_device *bdev; - - if (!test_and_clear_bit(DMF_FS_LOCKED, &md->flags)) - return 0; - - bdev = bdget_disk(md->disk, 0); - if (!bdev) { - DMWARN("bdget failed in __unlock_fs"); - return -ENOMEM; - } + if (!test_bit(DMF_FROZEN, &md->flags)) + return; - thaw_bdev(bdev, md->frozen_sb); + thaw_bdev(md->suspended_bdev, md->frozen_sb); md->frozen_sb = NULL; - bdput(bdev); - bdput(bdev); - return 0; + clear_bit(DMF_FROZEN, &md->flags); } /* @@ -946,46 +1333,65 @@ static int __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 dm_suspend(struct mapped_device *md, unsigned suspend_flags) { - struct dm_table *map; + 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; - /* Flush I/O to the device. */ - down_read(&md->lock); - if (test_bit(DMF_BLOCK_IO, &md->flags)) { - up_read(&md->lock); - return -EINVAL; - } + down(&md->suspend_lock); - __lock_fs(md); - up_read(&md->lock); + if (dm_suspended(md)) + goto out_unlock; + + map = dm_get_table(md); /* - * First we set the BLOCK_IO flag so no more ios will be - * mapped. + * DMF_NOFLUSH_SUSPENDING must be set before presuspend. + * This flag is cleared before dm_suspend returns. */ - down_write(&md->lock); - if (test_bit(DMF_BLOCK_IO, &md->flags)) { - /* - * If we get here we know another thread is - * trying to suspend as well, so we leave the fs - * locked for this thread. - */ - up_write(&md->lock); - return -EINVAL; + 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); + + /* bdget() can stall if the pending I/Os are not flushed */ + if (!noflush) { + md->suspended_bdev = bdget_disk(md->disk, 0); + if (!md->suspended_bdev) { + DMWARN("bdget failed in dm_suspend"); + r = -ENOMEM; + goto flush_and_out; + } + } + + /* + * 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; } + /* + * First we set the BLOCK_IO flag so no more ios will be mapped. + */ + down_write(&md->io_lock); set_bit(DMF_BLOCK_IO, &md->flags); + add_wait_queue(&md->wait, &wait); - up_write(&md->lock); + up_write(&md->io_lock); /* unplug */ - map = dm_get_table(md); - if (map) { + if (map) dm_table_unplug_all(map); - dm_table_put(map); - } /* * Then we wait for the already mapped ios to @@ -1001,54 +1407,112 @@ int dm_suspend(struct mapped_device *md) } set_current_state(TASK_RUNNING); - down_write(&md->lock); + 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)) { - __unlock_fs(md); clear_bit(DMF_BLOCK_IO, &md->flags); - up_write(&md->lock); - return -EINTR; + def = bio_list_get(&md->deferred); + __flush_deferred_io(md, def); + up_write(&md->io_lock); + unlock_fs(md); + goto out; /* pushback list is already flushed, so skip flush */ } + up_write(&md->io_lock); + + dm_table_postsuspend_targets(map); set_bit(DMF_SUSPENDED, &md->flags); - map = dm_get_table(md); - if (map) - dm_table_suspend_targets(map); + 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); + md->suspended_bdev = NULL; + } + dm_table_put(map); - up_write(&md->lock); - return 0; +out_unlock: + up(&md->suspend_lock); + return r; } int dm_resume(struct mapped_device *md) { + int r = -EINVAL; struct bio *def; - struct dm_table *map = dm_get_table(md); + struct dm_table *map = NULL; - down_write(&md->lock); - if (!map || - !test_bit(DMF_SUSPENDED, &md->flags) || - !dm_table_get_size(map)) { - up_write(&md->lock); - dm_table_put(map); - return -EINVAL; - } + down(&md->suspend_lock); + if (!dm_suspended(md)) + goto out; - dm_table_resume_targets(map); - clear_bit(DMF_SUSPENDED, &md->flags); + map = dm_get_table(md); + if (!map || !dm_table_get_size(map)) + goto out; + + r = dm_table_resume_targets(map); + if (r) + goto out; + + down_write(&md->io_lock); clear_bit(DMF_BLOCK_IO, &md->flags); def = bio_list_get(&md->deferred); __flush_deferred_io(md, def); - up_write(&md->lock); - __unlock_fs(md); + up_write(&md->io_lock); + + unlock_fs(md); + + if (md->suspended_bdev) { + bdput(md->suspended_bdev); + md->suspended_bdev = NULL; + } + + clear_bit(DMF_SUSPENDED, &md->flags); + dm_table_unplug_all(map); + + kobject_uevent(&md->disk->kobj, KOBJ_CHANGE); + + r = 0; + +out: dm_table_put(map); + up(&md->suspend_lock); - return 0; + return r; } /*----------------------------------------------------------------- @@ -1079,12 +1543,27 @@ 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, + .ioctl = dm_blk_ioctl, + .getgeo = dm_blk_getgeo, .owner = THIS_MODULE }; +EXPORT_SYMBOL(dm_get_mapinfo); + /* * module hooks */ @@ -1094,5 +1573,5 @@ module_exit(dm_exit); module_param(major, uint, 0); MODULE_PARM_DESC(major, "The major number of the device mapper"); MODULE_DESCRIPTION(DM_NAME " driver"); -MODULE_AUTHOR("Joe Thornber "); +MODULE_AUTHOR("Joe Thornber "); MODULE_LICENSE("GPL");