diff options
Diffstat (limited to 'drivers/md/dm.c')
| -rw-r--r-- | drivers/md/dm.c | 739 |
1 files changed, 521 insertions, 218 deletions
diff --git a/drivers/md/dm.c b/drivers/md/dm.c index f0f118ab20fa..6c83ab940af7 100644 --- a/drivers/md/dm.c +++ b/drivers/md/dm.c @@ -11,6 +11,7 @@ #include "dm-uevent.h" #include "dm-ima.h" +#include <linux/bio-integrity.h> #include <linux/init.h> #include <linux/module.h> #include <linux/mutex.h> @@ -402,15 +403,16 @@ static void do_deferred_remove(struct work_struct *w) dm_deferred_remove(); } -static int dm_blk_getgeo(struct block_device *bdev, struct hd_geometry *geo) +static int dm_blk_getgeo(struct gendisk *disk, struct hd_geometry *geo) { - struct mapped_device *md = bdev->bd_disk->private_data; + struct mapped_device *md = disk->private_data; return dm_get_geometry(md, geo); } static int dm_prepare_ioctl(struct mapped_device *md, int *srcu_idx, - struct block_device **bdev) + struct block_device **bdev, unsigned int cmd, + unsigned long arg, bool *forward) { struct dm_target *ti; struct dm_table *map; @@ -433,8 +435,8 @@ retry: if (dm_suspended_md(md)) return -EAGAIN; - r = ti->type->prepare_ioctl(ti, bdev); - if (r == -ENOTCONN && !fatal_signal_pending(current)) { + r = ti->type->prepare_ioctl(ti, bdev, cmd, arg, forward); + if (r == -ENOTCONN && *forward && !fatal_signal_pending(current)) { dm_put_live_table(md, *srcu_idx); fsleep(10000); goto retry; @@ -453,9 +455,10 @@ static int dm_blk_ioctl(struct block_device *bdev, blk_mode_t mode, { struct mapped_device *md = bdev->bd_disk->private_data; int r, srcu_idx; + bool forward = true; - r = dm_prepare_ioctl(md, &srcu_idx, &bdev); - if (r < 0) + r = dm_prepare_ioctl(md, &srcu_idx, &bdev, cmd, arg, &forward); + if (!forward || r < 0) goto out; if (r > 0) { @@ -487,18 +490,13 @@ u64 dm_start_time_ns_from_clone(struct bio *bio) } EXPORT_SYMBOL_GPL(dm_start_time_ns_from_clone); -static inline bool bio_is_flush_with_data(struct bio *bio) -{ - return ((bio->bi_opf & REQ_PREFLUSH) && bio->bi_iter.bi_size); -} - static inline unsigned int dm_io_sectors(struct dm_io *io, struct bio *bio) { /* * If REQ_PREFLUSH set, don't account payload, it will be * submitted (and accounted) after this flush completes. */ - if (bio_is_flush_with_data(bio)) + if (io->requeue_flush_with_data) return 0; if (unlikely(dm_io_flagged(io, DM_IO_WAS_SPLIT))) return io->sectors; @@ -570,13 +568,15 @@ static void dm_end_io_acct(struct dm_io *io) dm_io_acct(io, true); } -static struct dm_io *alloc_io(struct mapped_device *md, struct bio *bio) +static struct dm_io *alloc_io(struct mapped_device *md, struct bio *bio, gfp_t gfp_mask) { struct dm_io *io; struct dm_target_io *tio; struct bio *clone; - clone = bio_alloc_clone(NULL, bio, GFP_NOIO, &md->mempools->io_bs); + clone = bio_alloc_clone(NULL, bio, gfp_mask, &md->mempools->io_bs); + if (unlikely(!clone)) + return NULL; tio = clone_to_tio(clone); tio->flags = 0; dm_tio_set_flag(tio, DM_TIO_INSIDE_DM_IO); @@ -585,6 +585,7 @@ static struct dm_io *alloc_io(struct mapped_device *md, struct bio *bio) io = container_of(tio, struct dm_io, tio); io->magic = DM_IO_MAGIC; io->status = BLK_STS_OK; + io->requeue_flush_with_data = false; /* one ref is for submission, the other is for completion */ atomic_set(&io->io_count, 2); @@ -643,7 +644,7 @@ static struct bio *alloc_tio(struct clone_info *ci, struct dm_target *ti, /* Set default bdev, but target must bio_set_dev() before issuing IO */ clone->bi_bdev = md->disk->part0; - if (unlikely(ti->needs_bio_set_dev)) + if (likely(ti != NULL) && unlikely(ti->needs_bio_set_dev)) bio_set_dev(clone, md->disk->part0); if (len) { @@ -715,24 +716,6 @@ static void dm_put_live_table_fast(struct mapped_device *md) __releases(RCU) rcu_read_unlock(); } -static inline struct dm_table *dm_get_live_table_bio(struct mapped_device *md, - int *srcu_idx, blk_opf_t bio_opf) -{ - if (bio_opf & REQ_NOWAIT) - return dm_get_live_table_fast(md); - else - return dm_get_live_table(md, srcu_idx); -} - -static inline void dm_put_live_table_bio(struct mapped_device *md, int srcu_idx, - blk_opf_t bio_opf) -{ - if (bio_opf & REQ_NOWAIT) - dm_put_live_table_fast(md); - else - dm_put_live_table(md, srcu_idx); -} - static char *_dm_claim_ptr = "I belong to device-mapper"; /* @@ -742,6 +725,7 @@ static struct table_device *open_table_device(struct mapped_device *md, dev_t dev, blk_mode_t mode) { struct table_device *td; + struct file *bdev_file; struct block_device *bdev; u64 part_off; int r; @@ -751,12 +735,14 @@ static struct table_device *open_table_device(struct mapped_device *md, return ERR_PTR(-ENOMEM); refcount_set(&td->count, 1); - bdev = blkdev_get_by_dev(dev, mode, _dm_claim_ptr, NULL); - if (IS_ERR(bdev)) { - r = PTR_ERR(bdev); + bdev_file = bdev_file_open_by_dev(dev, mode, _dm_claim_ptr, NULL); + if (IS_ERR(bdev_file)) { + r = PTR_ERR(bdev_file); goto out_free_td; } + bdev = file_bdev(bdev_file); + /* * We can be called before the dm disk is added. In that case we can't * register the holder relation here. It will be done once add_disk was @@ -770,13 +756,15 @@ static struct table_device *open_table_device(struct mapped_device *md, td->dm_dev.mode = mode; td->dm_dev.bdev = bdev; - td->dm_dev.dax_dev = fs_dax_get_by_bdev(bdev, &part_off, NULL, NULL); + td->dm_dev.bdev_file = bdev_file; + td->dm_dev.dax_dev = fs_dax_get_by_bdev(bdev, &part_off, + NULL, NULL); format_dev_t(td->dm_dev.name, dev); list_add(&td->list, &md->table_devices); return td; out_blkdev_put: - blkdev_put(bdev, _dm_claim_ptr); + __fput_sync(bdev_file); out_free_td: kfree(td); return ERR_PTR(r); @@ -789,7 +777,13 @@ static void close_table_device(struct table_device *td, struct mapped_device *md { if (md->disk->slave_dir) bd_unlink_disk_holder(td->dm_dev.bdev, md->disk); - blkdev_put(td->dm_dev.bdev, _dm_claim_ptr); + + /* Leverage async fput() if DMF_DEFERRED_REMOVE set */ + if (unlikely(test_bit(DMF_DEFERRED_REMOVE, &md->flags))) + fput(td->dm_dev.bdev_file); + else + __fput_sync(td->dm_dev.bdev_file); + put_dax(td->dm_dev.dax_dev); list_del(&td->list); kfree(td); @@ -950,6 +944,7 @@ static void __dm_io_complete(struct dm_io *io, bool first_stage) struct mapped_device *md = io->md; blk_status_t io_error; bool requeued; + bool requeue_flush_with_data; requeued = dm_handle_requeue(io, first_stage); if (requeued && first_stage) @@ -966,6 +961,7 @@ static void __dm_io_complete(struct dm_io *io, bool first_stage) __dm_start_io_acct(io); dm_end_io_acct(io); } + requeue_flush_with_data = io->requeue_flush_with_data; free_io(io); smp_wmb(); this_cpu_dec(*md->pending_io); @@ -978,7 +974,7 @@ static void __dm_io_complete(struct dm_io *io, bool first_stage) if (requeued) return; - if (bio_is_flush_with_data(bio)) { + if (unlikely(requeue_flush_with_data)) { /* * Preflush done for flush with data, reissue * without REQ_PREFLUSH. @@ -1026,10 +1022,8 @@ static void dm_wq_requeue_work(struct work_struct *work) * * 2) io->orig_bio points to new cloned bio which matches the requeued dm_io. */ -static void dm_io_complete(struct dm_io *io) +static inline void dm_io_complete(struct dm_io *io) { - bool first_requeue; - /* * Only dm_io that has been split needs two stage requeue, otherwise * we may run into long bio clone chain during suspend and OOM could @@ -1038,12 +1032,7 @@ static void dm_io_complete(struct dm_io *io) * Also flush data dm_io won't be marked as DM_IO_WAS_SPLIT, so they * also aren't handled via the first stage requeue. */ - if (dm_io_flagged(io, DM_IO_WAS_SPLIT)) - first_requeue = true; - else - first_requeue = false; - - __dm_io_complete(io, first_requeue); + __dm_io_complete(io, dm_io_flagged(io, DM_IO_WAS_SPLIT)); } /* @@ -1086,22 +1075,6 @@ static inline struct queue_limits *dm_get_queue_limits(struct mapped_device *md) return &md->queue->limits; } -void disable_discard(struct mapped_device *md) -{ - struct queue_limits *limits = dm_get_queue_limits(md); - - /* device doesn't really support DISCARD, disable it */ - limits->max_discard_sectors = 0; -} - -void disable_write_zeroes(struct mapped_device *md) -{ - struct queue_limits *limits = dm_get_queue_limits(md); - - /* device doesn't really support WRITE ZEROES, disable it */ - limits->max_write_zeroes_sectors = 0; -} - static bool swap_bios_limit(struct dm_target *ti, struct bio *bio) { return unlikely((bio->bi_opf & REQ_SWAP) != 0) && unlikely(ti->limit_swap_bios); @@ -1112,17 +1085,17 @@ static void clone_endio(struct bio *bio) blk_status_t error = bio->bi_status; struct dm_target_io *tio = clone_to_tio(bio); struct dm_target *ti = tio->ti; - dm_endio_fn endio = ti->type->end_io; + dm_endio_fn endio = likely(ti != NULL) ? ti->type->end_io : NULL; struct dm_io *io = tio->io; struct mapped_device *md = io->md; if (unlikely(error == BLK_STS_TARGET)) { if (bio_op(bio) == REQ_OP_DISCARD && !bdev_max_discard_sectors(bio->bi_bdev)) - disable_discard(md); + blk_queue_disable_discard(md->queue); else if (bio_op(bio) == REQ_OP_WRITE_ZEROES && !bdev_write_zeroes_sectors(bio->bi_bdev)) - disable_write_zeroes(md); + blk_queue_disable_write_zeroes(md->queue); } if (static_branch_unlikely(&zoned_enabled) && @@ -1159,7 +1132,7 @@ static void clone_endio(struct bio *bio) } if (static_branch_unlikely(&swap_bios_enabled) && - unlikely(swap_bios_limit(ti, bio))) + likely(ti != NULL) && unlikely(swap_bios_limit(ti, bio))) up(&md->swap_bios_semaphore); free_tio(bio); @@ -1193,7 +1166,7 @@ static sector_t __max_io_len(struct dm_target *ti, sector_t sector, return len; return min_t(sector_t, len, min(max_sectors ? : queue_max_sectors(ti->table->md->queue), - blk_chunk_sectors_left(target_offset, max_granularity))); + blk_boundary_sectors_left(target_offset, max_granularity))); } static inline sector_t max_io_len(struct dm_target *ti, sector_t sector) @@ -1236,7 +1209,7 @@ static struct dm_target *dm_dax_get_live_target(struct mapped_device *md, static long dm_dax_direct_access(struct dax_device *dax_dev, pgoff_t pgoff, long nr_pages, enum dax_access_mode mode, void **kaddr, - pfn_t *pfn) + unsigned long *pfn) { struct mapped_device *md = dax_get_private(dax_dev); sector_t sector = pgoff * PAGE_SECTORS; @@ -1311,8 +1284,9 @@ out: /* * A target may call dm_accept_partial_bio only from the map routine. It is * allowed for all bio types except REQ_PREFLUSH, REQ_OP_ZONE_* zone management - * operations, REQ_OP_ZONE_APPEND (zone append writes) and any bio serviced by - * __send_duplicate_bios(). + * operations, zone append writes (native with REQ_OP_ZONE_APPEND or emulated + * with write BIOs flagged with BIO_EMULATES_ZONE_APPEND) and any bio serviced + * by __send_duplicate_bios(). * * dm_accept_partial_bio informs the dm that the target only wants to process * additional n_sectors sectors of the bio and the rest of the data should be @@ -1345,11 +1319,19 @@ void dm_accept_partial_bio(struct bio *bio, unsigned int n_sectors) unsigned int bio_sectors = bio_sectors(bio); BUG_ON(dm_tio_flagged(tio, DM_TIO_IS_DUPLICATE_BIO)); - BUG_ON(op_is_zone_mgmt(bio_op(bio))); - BUG_ON(bio_op(bio) == REQ_OP_ZONE_APPEND); BUG_ON(bio_sectors > *tio->len_ptr); BUG_ON(n_sectors > bio_sectors); + if (static_branch_unlikely(&zoned_enabled) && + unlikely(bdev_is_zoned(bio->bi_bdev))) { + enum req_op op = bio_op(bio); + + BUG_ON(op_is_zone_mgmt(op)); + BUG_ON(op == REQ_OP_WRITE); + BUG_ON(op == REQ_OP_WRITE_ZEROES); + BUG_ON(op == REQ_OP_ZONE_APPEND); + } + *tio->len_ptr -= bio_sectors - n_sectors; bio->bi_iter.bi_size = n_sectors << SECTOR_SHIFT; @@ -1433,17 +1415,11 @@ static void __map_bio(struct bio *clone) down(&md->swap_bios_semaphore); } - if (static_branch_unlikely(&zoned_enabled)) { - /* - * Check if the IO needs a special mapping due to zone append - * emulation on zoned target. In this case, dm_zone_map_bio() - * calls the target map operation. - */ - if (unlikely(dm_emulate_zone_append(md))) - r = dm_zone_map_bio(tio); - else - r = ti->type->map(ti, clone); - } else + if (likely(ti->type->map == linear_map)) + r = linear_map(ti, clone); + else if (ti->type->map == stripe_map) + r = stripe_map(ti, clone); + else r = ti->type->map(ti, clone); switch (r) { @@ -1497,7 +1473,7 @@ static void alloc_multiple_bios(struct bio_list *blist, struct clone_info *ci, for (try = 0; try < 2; try++) { int bio_nr; - if (try) + if (try && num_bios > 1) mutex_lock(&ci->io->md->table_devices_lock); for (bio_nr = 0; bio_nr < num_bios; bio_nr++) { bio = alloc_tio(ci, ti, bio_nr, len, @@ -1507,7 +1483,7 @@ static void alloc_multiple_bios(struct bio_list *blist, struct clone_info *ci, bio_list_add(blist, bio); } - if (try) + if (try && num_bios > 1) mutex_unlock(&ci->io->md->table_devices_lock); if (bio_nr == num_bios) return; @@ -1517,34 +1493,30 @@ static void alloc_multiple_bios(struct bio_list *blist, struct clone_info *ci, } } -static int __send_duplicate_bios(struct clone_info *ci, struct dm_target *ti, - unsigned int num_bios, unsigned int *len) +static unsigned int __send_duplicate_bios(struct clone_info *ci, struct dm_target *ti, + unsigned int num_bios, unsigned int *len) { struct bio_list blist = BIO_EMPTY_LIST; struct bio *clone; unsigned int ret = 0; - switch (num_bios) { - case 0: - break; - case 1: - if (len) - setup_split_accounting(ci, *len); - clone = alloc_tio(ci, ti, 0, len, GFP_NOIO); - __map_bio(clone); - ret = 1; - break; - default: - if (len) - setup_split_accounting(ci, *len); - /* dm_accept_partial_bio() is not supported with shared tio->len_ptr */ - alloc_multiple_bios(&blist, ci, ti, num_bios, len); - while ((clone = bio_list_pop(&blist))) { + if (WARN_ON_ONCE(num_bios == 0)) /* num_bios = 0 is a bug in caller */ + return 0; + + /* dm_accept_partial_bio() is not supported with shared tio->len_ptr */ + if (len) + setup_split_accounting(ci, *len); + + /* + * Using alloc_multiple_bios(), even if num_bios is 1, to consistently + * support allocating using GFP_NOWAIT with GFP_NOIO fallback. + */ + alloc_multiple_bios(&blist, ci, ti, num_bios, len); + while ((clone = bio_list_pop(&blist))) { + if (num_bios > 1) dm_tio_set_flag(clone_to_tio(clone), DM_TIO_IS_DUPLICATE_BIO); - __map_bio(clone); - ret += 1; - } - break; + __map_bio(clone); + ret += 1; } return ret; @@ -1554,26 +1526,60 @@ static void __send_empty_flush(struct clone_info *ci) { struct dm_table *t = ci->map; struct bio flush_bio; + blk_opf_t opf = REQ_OP_WRITE | REQ_PREFLUSH | REQ_SYNC; + + if ((ci->io->orig_bio->bi_opf & (REQ_IDLE | REQ_SYNC)) == + (REQ_IDLE | REQ_SYNC)) + opf |= REQ_IDLE; /* * Use an on-stack bio for this, it's safe since we don't * need to reference it after submit. It's just used as * the basis for the clone(s). */ - bio_init(&flush_bio, ci->io->md->disk->part0, NULL, 0, - REQ_OP_WRITE | REQ_PREFLUSH | REQ_SYNC); + bio_init(&flush_bio, ci->io->md->disk->part0, NULL, 0, opf); ci->bio = &flush_bio; ci->sector_count = 0; ci->io->tio.clone.bi_iter.bi_size = 0; - for (unsigned int i = 0; i < t->num_targets; i++) { - unsigned int bios; - struct dm_target *ti = dm_table_get_target(t, i); + if (!t->flush_bypasses_map) { + for (unsigned int i = 0; i < t->num_targets; i++) { + unsigned int bios; + struct dm_target *ti = dm_table_get_target(t, i); - atomic_add(ti->num_flush_bios, &ci->io->io_count); - bios = __send_duplicate_bios(ci, ti, ti->num_flush_bios, NULL); - atomic_sub(ti->num_flush_bios - bios, &ci->io->io_count); + if (unlikely(ti->num_flush_bios == 0)) + continue; + + atomic_add(ti->num_flush_bios, &ci->io->io_count); + bios = __send_duplicate_bios(ci, ti, ti->num_flush_bios, + NULL); + atomic_sub(ti->num_flush_bios - bios, &ci->io->io_count); + } + } else { + /* + * Note that there's no need to grab t->devices_lock here + * because the targets that support flush optimization don't + * modify the list of devices. + */ + struct list_head *devices = dm_table_get_devices(t); + unsigned int len = 0; + struct dm_dev_internal *dd; + list_for_each_entry(dd, devices, list) { + struct bio *clone; + /* + * Note that the structure dm_target_io is not + * associated with any target (because the device may be + * used by multiple targets), so we set tio->ti = NULL. + * We must check for NULL in the I/O processing path, to + * avoid NULL pointer dereference. + */ + clone = alloc_tio(ci, NULL, 0, &len, GFP_NOIO); + atomic_add(1, &ci->io->io_count); + bio_set_dev(clone, dd->dm_dev->bdev); + clone->bi_end_io = clone_endio; + dm_submit_bio_remap(clone, NULL); + } } /* @@ -1585,10 +1591,9 @@ static void __send_empty_flush(struct clone_info *ci) bio_uninit(ci->bio); } -static void __send_changing_extent_only(struct clone_info *ci, struct dm_target *ti, - unsigned int num_bios, - unsigned int max_granularity, - unsigned int max_sectors) +static void __send_abnormal_io(struct clone_info *ci, struct dm_target *ti, + unsigned int num_bios, unsigned int max_granularity, + unsigned int max_sectors) { unsigned int len, bios; @@ -1609,20 +1614,19 @@ static void __send_changing_extent_only(struct clone_info *ci, struct dm_target static bool is_abnormal_io(struct bio *bio) { - enum req_op op = bio_op(bio); - - if (op != REQ_OP_READ && op != REQ_OP_WRITE && op != REQ_OP_FLUSH) { - switch (op) { - case REQ_OP_DISCARD: - case REQ_OP_SECURE_ERASE: - case REQ_OP_WRITE_ZEROES: - return true; - default: - break; - } + switch (bio_op(bio)) { + case REQ_OP_READ: + case REQ_OP_WRITE: + case REQ_OP_FLUSH: + return false; + case REQ_OP_DISCARD: + case REQ_OP_SECURE_ERASE: + case REQ_OP_WRITE_ZEROES: + case REQ_OP_ZONE_RESET_ALL: + return true; + default: + return false; } - - return false; } static blk_status_t __process_abnormal_io(struct clone_info *ci, @@ -1643,14 +1647,10 @@ static blk_status_t __process_abnormal_io(struct clone_info *ci, case REQ_OP_SECURE_ERASE: num_bios = ti->num_secure_erase_bios; max_sectors = limits->max_secure_erase_sectors; - if (ti->max_secure_erase_granularity) - max_granularity = max_sectors; break; case REQ_OP_WRITE_ZEROES: num_bios = ti->num_write_zeroes_bios; max_sectors = limits->max_write_zeroes_sectors; - if (ti->max_write_zeroes_granularity) - max_granularity = max_sectors; break; default: break; @@ -1665,8 +1665,8 @@ static blk_status_t __process_abnormal_io(struct clone_info *ci, if (unlikely(!num_bios)) return BLK_STS_NOTSUPP; - __send_changing_extent_only(ci, ti, num_bios, - max_granularity, max_sectors); + __send_abnormal_io(ci, ti, num_bios, max_granularity, max_sectors); + return BLK_STS_OK; } @@ -1725,10 +1725,6 @@ static blk_status_t __split_and_process_bio(struct clone_info *ci) if (unlikely(!ti)) return BLK_STS_IOERR; - if (unlikely((ci->bio->bi_opf & REQ_NOWAIT) != 0) && - unlikely(!dm_target_supports_nowait(ti->type))) - return BLK_STS_NOTSUPP; - if (unlikely(ci->is_abnormal_io)) return __process_abnormal_io(ci, ti); @@ -1739,8 +1735,21 @@ static blk_status_t __split_and_process_bio(struct clone_info *ci) ci->submit_as_polled = !!(ci->bio->bi_opf & REQ_POLLED); len = min_t(sector_t, max_io_len(ti, ci->sector), ci->sector_count); + if (ci->bio->bi_opf & REQ_ATOMIC && len != ci->sector_count) + return BLK_STS_IOERR; + setup_split_accounting(ci, len); - clone = alloc_tio(ci, ti, 0, &len, GFP_NOIO); + + if (unlikely(ci->bio->bi_opf & REQ_NOWAIT)) { + if (unlikely(!dm_target_supports_nowait(ti->type))) + return BLK_STS_NOTSUPP; + + clone = alloc_tio(ci, ti, 0, &len, GFP_NOWAIT); + if (unlikely(!clone)) + return BLK_STS_AGAIN; + } else { + clone = alloc_tio(ci, ti, 0, &len, GFP_NOIO); + } __map_bio(clone); ci->sector += len; @@ -1749,11 +1758,11 @@ static blk_status_t __split_and_process_bio(struct clone_info *ci) return BLK_STS_OK; } -static void init_clone_info(struct clone_info *ci, struct mapped_device *md, +static void init_clone_info(struct clone_info *ci, struct dm_io *io, struct dm_table *map, struct bio *bio, bool is_abnormal) { ci->map = map; - ci->io = alloc_io(md, bio); + ci->io = io; ci->bio = bio; ci->is_abnormal_io = is_abnormal; ci->submit_as_polled = false; @@ -1766,6 +1775,165 @@ static void init_clone_info(struct clone_info *ci, struct mapped_device *md, ci->sector_count = 0; } +#ifdef CONFIG_BLK_DEV_ZONED +static inline bool dm_zone_bio_needs_split(struct bio *bio) +{ + /* + * Special case the zone operations that cannot or should not be split. + */ + switch (bio_op(bio)) { + case REQ_OP_ZONE_APPEND: + case REQ_OP_ZONE_FINISH: + case REQ_OP_ZONE_RESET: + case REQ_OP_ZONE_RESET_ALL: + return false; + default: + break; + } + + /* + * When mapped devices use the block layer zone write plugging, we must + * split any large BIO to the mapped device limits to not submit BIOs + * that span zone boundaries and to avoid potential deadlocks with + * queue freeze operations. + */ + return bio_needs_zone_write_plugging(bio) || bio_straddles_zones(bio); +} + +static inline bool dm_zone_plug_bio(struct mapped_device *md, struct bio *bio) +{ + if (!bio_needs_zone_write_plugging(bio)) + return false; + return blk_zone_plug_bio(bio, 0); +} + +static blk_status_t __send_zone_reset_all_emulated(struct clone_info *ci, + struct dm_target *ti) +{ + struct bio_list blist = BIO_EMPTY_LIST; + struct mapped_device *md = ci->io->md; + unsigned int zone_sectors = md->disk->queue->limits.chunk_sectors; + unsigned long *need_reset; + unsigned int i, nr_zones, nr_reset; + unsigned int num_bios = 0; + blk_status_t sts = BLK_STS_OK; + sector_t sector = ti->begin; + struct bio *clone; + int ret; + + nr_zones = ti->len >> ilog2(zone_sectors); + need_reset = bitmap_zalloc(nr_zones, GFP_NOIO); + if (!need_reset) + return BLK_STS_RESOURCE; + + ret = dm_zone_get_reset_bitmap(md, ci->map, ti->begin, + nr_zones, need_reset); + if (ret) { + sts = BLK_STS_IOERR; + goto free_bitmap; + } + + /* If we have no zone to reset, we are done. */ + nr_reset = bitmap_weight(need_reset, nr_zones); + if (!nr_reset) + goto free_bitmap; + + atomic_add(nr_zones, &ci->io->io_count); + + for (i = 0; i < nr_zones; i++) { + + if (!test_bit(i, need_reset)) { + sector += zone_sectors; + continue; + } + + if (bio_list_empty(&blist)) { + /* This may take a while, so be nice to others */ + if (num_bios) + cond_resched(); + + /* + * We may need to reset thousands of zones, so let's + * not go crazy with the clone allocation. + */ + alloc_multiple_bios(&blist, ci, ti, min(nr_reset, 32), + NULL); + } + + /* Get a clone and change it to a regular reset operation. */ + clone = bio_list_pop(&blist); + clone->bi_opf &= ~REQ_OP_MASK; + clone->bi_opf |= REQ_OP_ZONE_RESET | REQ_SYNC; + clone->bi_iter.bi_sector = sector; + clone->bi_iter.bi_size = 0; + __map_bio(clone); + + sector += zone_sectors; + num_bios++; + nr_reset--; + } + + WARN_ON_ONCE(!bio_list_empty(&blist)); + atomic_sub(nr_zones - num_bios, &ci->io->io_count); + ci->sector_count = 0; + +free_bitmap: + bitmap_free(need_reset); + + return sts; +} + +static void __send_zone_reset_all_native(struct clone_info *ci, + struct dm_target *ti) +{ + unsigned int bios; + + atomic_add(1, &ci->io->io_count); + bios = __send_duplicate_bios(ci, ti, 1, NULL); + atomic_sub(1 - bios, &ci->io->io_count); + + ci->sector_count = 0; +} + +static blk_status_t __send_zone_reset_all(struct clone_info *ci) +{ + struct dm_table *t = ci->map; + blk_status_t sts = BLK_STS_OK; + + for (unsigned int i = 0; i < t->num_targets; i++) { + struct dm_target *ti = dm_table_get_target(t, i); + + if (ti->zone_reset_all_supported) { + __send_zone_reset_all_native(ci, ti); + continue; + } + + sts = __send_zone_reset_all_emulated(ci, ti); + if (sts != BLK_STS_OK) + break; + } + + /* Release the reference that alloc_io() took for submission. */ + atomic_sub(1, &ci->io->io_count); + + return sts; +} + +#else +static inline bool dm_zone_bio_needs_split(struct bio *bio) +{ + return false; +} +static inline bool dm_zone_plug_bio(struct mapped_device *md, struct bio *bio) +{ + return false; +} +static blk_status_t __send_zone_reset_all(struct clone_info *ci) +{ + return BLK_STS_NOTSUPP; +} +#endif + /* * Entry point to split a bio into clones and submit them to the targets. */ @@ -1775,28 +1943,87 @@ static void dm_split_and_process_bio(struct mapped_device *md, struct clone_info ci; struct dm_io *io; blk_status_t error = BLK_STS_OK; - bool is_abnormal; + bool is_abnormal, need_split; is_abnormal = is_abnormal_io(bio); - if (unlikely(is_abnormal)) { + if (static_branch_unlikely(&zoned_enabled)) { + need_split = is_abnormal || dm_zone_bio_needs_split(bio); + } else { + need_split = is_abnormal; + } + + if (unlikely(need_split)) { /* * Use bio_split_to_limits() for abnormal IO (e.g. discard, etc) * otherwise associated queue_limits won't be imposed. + * Also split the BIO for mapped devices needing zone append + * emulation to ensure that the BIO does not cross zone + * boundaries. */ bio = bio_split_to_limits(bio); if (!bio) return; } - init_clone_info(&ci, md, map, bio, is_abnormal); - io = ci.io; + /* + * Use the block layer zone write plugging for mapped devices that + * need zone append emulation (e.g. dm-crypt). + */ + if (static_branch_unlikely(&zoned_enabled) && dm_zone_plug_bio(md, bio)) + return; + + /* Only support nowait for normal IO */ + if (unlikely(bio->bi_opf & REQ_NOWAIT) && !is_abnormal) { + /* + * Don't support NOWAIT for FLUSH because it may allocate + * multiple bios and there's no easy way how to undo the + * allocations. + */ + if (bio->bi_opf & REQ_PREFLUSH) { + bio_wouldblock_error(bio); + return; + } + io = alloc_io(md, bio, GFP_NOWAIT); + if (unlikely(!io)) { + /* Unable to do anything without dm_io. */ + bio_wouldblock_error(bio); + return; + } + } else { + io = alloc_io(md, bio, GFP_NOIO); + } + init_clone_info(&ci, io, map, bio, is_abnormal); - if (bio->bi_opf & REQ_PREFLUSH) { + if (unlikely((bio->bi_opf & REQ_PREFLUSH) != 0)) { + /* + * The "flush_bypasses_map" is set on targets where it is safe + * to skip the map function and submit bios directly to the + * underlying block devices - currently, it is set for dm-linear + * and dm-stripe. + * + * If we have just one underlying device (i.e. there is one + * linear target or multiple linear targets pointing to the same + * device), we can send the flush with data directly to it. + */ + if (bio->bi_iter.bi_size && map->flush_bypasses_map) { + struct list_head *devices = dm_table_get_devices(map); + if (devices->next == devices->prev) + goto send_preflush_with_data; + } + if (bio->bi_iter.bi_size) + io->requeue_flush_with_data = true; __send_empty_flush(&ci); /* dm_io_complete submits any data associated with flush */ goto out; } +send_preflush_with_data: + if (static_branch_unlikely(&zoned_enabled) && + (bio_op(bio) == REQ_OP_ZONE_RESET_ALL)) { + error = __send_zone_reset_all(&ci); + goto out; + } + error = __split_and_process_bio(&ci); if (error || !ci.sector_count) goto out; @@ -1833,13 +2060,17 @@ static void dm_submit_bio(struct bio *bio) struct mapped_device *md = bio->bi_bdev->bd_disk->private_data; int srcu_idx; struct dm_table *map; - blk_opf_t bio_opf = bio->bi_opf; - map = dm_get_live_table_bio(md, &srcu_idx, bio_opf); + map = dm_get_live_table(md, &srcu_idx); + if (unlikely(!map)) { + DMERR_LIMIT("%s: mapping table unavailable, erroring io", + dm_device_name(md)); + bio_io_error(bio); + goto out; + } - /* If suspended, or map not yet available, queue this IO for later */ - if (unlikely(test_bit(DMF_BLOCK_IO_FOR_SUSPEND, &md->flags)) || - unlikely(!map)) { + /* If suspended, queue this IO for later */ + if (unlikely(test_bit(DMF_BLOCK_IO_FOR_SUSPEND, &md->flags))) { if (bio->bi_opf & REQ_NOWAIT) bio_wouldblock_error(bio); else if (bio->bi_opf & REQ_RAHEAD) @@ -1851,7 +2082,7 @@ static void dm_submit_bio(struct bio *bio) dm_split_and_process_bio(md, map, bio); out: - dm_put_live_table_bio(md, srcu_idx, bio_opf); + dm_put_live_table(md, srcu_idx); } static bool dm_poll_dm_io(struct dm_io *io, struct io_comp_batch *iob, @@ -1999,7 +2230,6 @@ static void cleanup_mapped_device(struct mapped_device *md) md->dax_dev = NULL; } - dm_cleanup_zoned_dev(md); if (md->disk) { spin_lock(&_minor_lock); md->disk->private_data = NULL; @@ -2046,6 +2276,7 @@ static void cleanup_mapped_device(struct mapped_device *md) static struct mapped_device *alloc_dev(int minor) { int r, numa_node_id = dm_get_numa_node(); + struct dax_device *dax_dev; struct mapped_device *md; void *old_md; @@ -2090,9 +2321,11 @@ static struct mapped_device *alloc_dev(int minor) * established. If request-based table is loaded: blk-mq will * override accordingly. */ - md->disk = blk_alloc_disk(md->numa_node_id); - if (!md->disk) + md->disk = blk_alloc_disk(NULL, md->numa_node_id); + if (IS_ERR(md->disk)) { + md->disk = NULL; goto bad; + } md->queue = md->disk->queue; init_waitqueue_head(&md->wait); @@ -2114,15 +2347,15 @@ static struct mapped_device *alloc_dev(int minor) md->disk->private_data = md; sprintf(md->disk->disk_name, "dm-%d", minor); - if (IS_ENABLED(CONFIG_FS_DAX)) { - md->dax_dev = alloc_dax(md, &dm_dax_ops); - if (IS_ERR(md->dax_dev)) { - md->dax_dev = NULL; + dax_dev = alloc_dax(md, &dm_dax_ops); + if (IS_ERR(dax_dev)) { + if (PTR_ERR(dax_dev) != -EOPNOTSUPP) goto bad; - } - set_dax_nocache(md->dax_dev); - set_dax_nomc(md->dax_dev); - if (dax_add_host(md->dax_dev, md->disk)) + } else { + set_dax_nocache(dax_dev); + set_dax_nomc(dax_dev); + md->dax_dev = dax_dev; + if (dax_add_host(dax_dev, md->disk)) goto bad; } @@ -2205,21 +2438,35 @@ static struct dm_table *__bind(struct mapped_device *md, struct dm_table *t, struct queue_limits *limits) { struct dm_table *old_map; - sector_t size; + sector_t size, old_size; int ret; lockdep_assert_held(&md->suspend_lock); size = dm_table_get_size(t); + old_size = dm_get_size(md); + + if (!dm_table_supports_size_change(t, old_size, size)) { + old_map = ERR_PTR(-EINVAL); + goto out; + } + + set_capacity(md->disk, size); + + ret = dm_table_set_restrictions(t, md->queue, limits); + if (ret) { + set_capacity(md->disk, old_size); + old_map = ERR_PTR(ret); + goto out; + } + /* * Wipe any geometry if the size of the table changed. */ - if (size != dm_get_size(md)) + if (size != old_size) memset(&md->geometry, 0, sizeof(md->geometry)); - set_capacity(md->disk, size); - dm_table_event_callback(t, event_callback, md); if (dm_table_request_based(t)) { @@ -2237,10 +2484,10 @@ static struct dm_table *__bind(struct mapped_device *md, struct dm_table *t, * requests in the queue may refer to bio from the old bioset, * so you must walk through the queue to unprep. */ - if (!md->mempools) { + if (!md->mempools) md->mempools = t->mempools; - t->mempools = NULL; - } + else + dm_free_md_mempools(t->mempools); } else { /* * The md may already have mempools that need changing. @@ -2249,14 +2496,8 @@ static struct dm_table *__bind(struct mapped_device *md, struct dm_table *t, */ dm_free_md_mempools(md->mempools); md->mempools = t->mempools; - t->mempools = NULL; - } - - ret = dm_table_set_restrictions(t, md->queue, limits); - if (ret) { - old_map = ERR_PTR(ret); - goto out; } + t->mempools = NULL; old_map = rcu_dereference_protected(md->map, lockdep_is_held(&md->suspend_lock)); rcu_assign_pointer(md->map, (void *)t); @@ -2316,12 +2557,6 @@ void dm_unlock_md_type(struct mapped_device *md) mutex_unlock(&md->type_lock); } -void dm_set_md_type(struct mapped_device *md, enum dm_queue_mode type) -{ - BUG_ON(!mutex_is_locked(&md->type_lock)); - md->type = type; -} - enum dm_queue_mode dm_get_md_type(struct mapped_device *md) { return md->type; @@ -2342,22 +2577,15 @@ int dm_setup_md_queue(struct mapped_device *md, struct dm_table *t) struct table_device *td; int r; - switch (type) { - case DM_TYPE_REQUEST_BASED: + WARN_ON_ONCE(type == DM_TYPE_NONE); + + if (type == DM_TYPE_REQUEST_BASED) { md->disk->fops = &dm_rq_blk_dops; r = dm_mq_init_request_queue(md, t); if (r) { DMERR("Cannot initialize queue for request-based dm mapped device"); return r; } - break; - case DM_TYPE_BIO_BASED: - case DM_TYPE_DAX_BIO_BASED: - blk_queue_flag_set(QUEUE_FLAG_IO_STAT, md->queue); - break; - case DM_TYPE_NONE: - WARN_ON_ONCE(true); - break; } r = dm_calculate_queue_limits(t, &limits); @@ -2550,7 +2778,7 @@ static int dm_wait_for_bios_completion(struct mapped_device *md, unsigned int ta break; if (signal_pending_state(task_state, current)) { - r = -EINTR; + r = -ERESTARTSYS; break; } @@ -2575,7 +2803,7 @@ static int dm_wait_for_completion(struct mapped_device *md, unsigned int task_st break; if (signal_pending_state(task_state, current)) { - r = -EINTR; + r = -ERESTARTSYS; break; } @@ -2667,7 +2895,7 @@ static int lock_fs(struct mapped_device *md) WARN_ON(test_bit(DMF_FROZEN, &md->flags)); - r = freeze_bdev(md->disk->part0); + r = bdev_freeze(md->disk->part0); if (!r) set_bit(DMF_FROZEN, &md->flags); return r; @@ -2677,7 +2905,7 @@ static void unlock_fs(struct mapped_device *md) { if (!test_bit(DMF_FROZEN, &md->flags)) return; - thaw_bdev(md->disk->part0); + bdev_thaw(md->disk->part0); clear_bit(DMF_FROZEN, &md->flags); } @@ -2696,7 +2924,7 @@ static int __dm_suspend(struct mapped_device *md, struct dm_table *map, { bool do_lockfs = suspend_flags & DM_SUSPEND_LOCKFS_FLAG; bool noflush = suspend_flags & DM_SUSPEND_NOFLUSH_FLAG; - int r; + int r = 0; lockdep_assert_held(&md->suspend_lock); @@ -2748,8 +2976,10 @@ static int __dm_suspend(struct mapped_device *md, struct dm_table *map, * Stop md->queue before flushing md->wq in case request-based * dm defers requests to md->wq from md->queue. */ - if (dm_request_based(md)) + if (map && dm_request_based(md)) { dm_stop_queue(md->queue); + set_bit(DMF_QUEUE_STOPPED, &md->flags); + } flush_workqueue(md->wq); @@ -2758,7 +2988,8 @@ static int __dm_suspend(struct mapped_device *md, struct dm_table *map, * We call dm_wait_for_completion to wait for all existing requests * to finish. */ - r = dm_wait_for_completion(md, task_state); + if (map) + r = dm_wait_for_completion(md, task_state); if (!r) set_bit(dmf_suspended_flag, &md->flags); @@ -2771,7 +3002,7 @@ static int __dm_suspend(struct mapped_device *md, struct dm_table *map, if (r < 0) { dm_queue_flush(md); - if (dm_request_based(md)) + if (test_and_clear_bit(DMF_QUEUE_STOPPED, &md->flags)) dm_start_queue(md->queue); unlock_fs(md); @@ -2855,7 +3086,7 @@ static int __dm_resume(struct mapped_device *md, struct dm_table *map) * so that mapping of targets can work correctly. * Request-based dm is queueing the deferred I/Os in its request_queue. */ - if (dm_request_based(md)) + if (test_and_clear_bit(DMF_QUEUE_STOPPED, &md->flags)) dm_start_queue(md->queue); unlock_fs(md); @@ -2937,6 +3168,9 @@ static void __dm_internal_suspend(struct mapped_device *md, unsigned int suspend static void __dm_internal_resume(struct mapped_device *md) { + int r; + struct dm_table *map; + BUG_ON(!md->internal_suspend_count); if (--md->internal_suspend_count) @@ -2945,12 +3179,23 @@ static void __dm_internal_resume(struct mapped_device *md) if (dm_suspended_md(md)) goto done; /* resume from nested suspend */ - /* - * NOTE: existing callers don't need to call dm_table_resume_targets - * (which may fail -- so best to avoid it for now by passing NULL map) - */ - (void) __dm_resume(md, NULL); - + map = rcu_dereference_protected(md->map, lockdep_is_held(&md->suspend_lock)); + r = __dm_resume(md, map); + if (r) { + /* + * If a preresume method of some target failed, we are in a + * tricky situation. We can't return an error to the caller. We + * can't fake success because then the "resume" and + * "postsuspend" methods would not be paired correctly, and it + * would break various targets, for example it would cause list + * corruption in the "origin" target. + * + * So, we fake normal suspend here, to make sure that the + * "resume" and "postsuspend" methods will be paired correctly. + */ + DMERR("Preresume method failed: %d", r); + set_bit(DMF_SUSPENDED, &md->flags); + } done: clear_bit(DMF_SUSPENDED_INTERNALLY, &md->flags); smp_mb__after_atomic(); @@ -3141,6 +3386,59 @@ void dm_free_md_mempools(struct dm_md_mempools *pools) kfree(pools); } +struct dm_blkdev_id { + u8 *id; + enum blk_unique_id type; +}; + +static int __dm_get_unique_id(struct dm_target *ti, struct dm_dev *dev, + sector_t start, sector_t len, void *data) +{ + struct dm_blkdev_id *dm_id = data; + const struct block_device_operations *fops = dev->bdev->bd_disk->fops; + + if (!fops->get_unique_id) + return 0; + + return fops->get_unique_id(dev->bdev->bd_disk, dm_id->id, dm_id->type); +} + +/* + * Allow access to get_unique_id() for the first device returning a + * non-zero result. Reasonable use expects all devices to have the + * same unique id. + */ +static int dm_blk_get_unique_id(struct gendisk *disk, u8 *id, + enum blk_unique_id type) +{ + struct mapped_device *md = disk->private_data; + struct dm_table *table; + struct dm_target *ti; + int ret = 0, srcu_idx; + + struct dm_blkdev_id dm_id = { + .id = id, + .type = type, + }; + + table = dm_get_live_table(md, &srcu_idx); + if (!table || !dm_table_get_size(table)) + goto out; + + /* We only support devices that have a single target */ + if (table->num_targets != 1) + goto out; + ti = dm_table_get_target(table, 0); + + if (!ti->type->iterate_devices) + goto out; + + ret = ti->type->iterate_devices(ti, __dm_get_unique_id, &dm_id); +out: + dm_put_live_table(md, srcu_idx); + return ret; +} + struct dm_pr { u64 old_key; u64 new_key; @@ -3368,10 +3666,13 @@ static int dm_pr_clear(struct block_device *bdev, u64 key) struct mapped_device *md = bdev->bd_disk->private_data; const struct pr_ops *ops; int r, srcu_idx; + bool forward = true; - r = dm_prepare_ioctl(md, &srcu_idx, &bdev); + /* Not a real ioctl, but targets must not interpret non-DM ioctls */ + r = dm_prepare_ioctl(md, &srcu_idx, &bdev, 0, 0, &forward); if (r < 0) goto out; + WARN_ON_ONCE(!forward); ops = bdev->bd_disk->fops->pr_ops; if (ops && ops->pr_clear) @@ -3466,6 +3767,7 @@ static const struct block_device_operations dm_blk_dops = { .ioctl = dm_blk_ioctl, .getgeo = dm_blk_getgeo, .report_zones = dm_blk_report_zones, + .get_unique_id = dm_blk_get_unique_id, .pr_ops = &dm_pr_ops, .owner = THIS_MODULE }; @@ -3475,6 +3777,7 @@ static const struct block_device_operations dm_rq_blk_dops = { .release = dm_blk_close, .ioctl = dm_blk_ioctl, .getgeo = dm_blk_getgeo, + .get_unique_id = dm_blk_get_unique_id, .pr_ops = &dm_pr_ops, .owner = THIS_MODULE }; @@ -3504,5 +3807,5 @@ module_param(swap_bios, int, 0644); MODULE_PARM_DESC(swap_bios, "Maximum allowed inflight swap IOs"); MODULE_DESCRIPTION(DM_NAME " driver"); -MODULE_AUTHOR("Joe Thornber <dm-devel@redhat.com>"); +MODULE_AUTHOR("Joe Thornber <dm-devel@lists.linux.dev>"); MODULE_LICENSE("GPL"); |