diff options
Diffstat (limited to 'block')
| -rw-r--r-- | block/Kconfig | 20 | ||||
| -rw-r--r-- | block/badblocks.c | 6 | ||||
| -rw-r--r-- | block/bdev.c | 258 | ||||
| -rw-r--r-- | block/blk-core.c | 14 | ||||
| -rw-r--r-- | block/blk-mq.c | 14 | ||||
| -rw-r--r-- | block/blk-sysfs.c | 2 | ||||
| -rw-r--r-- | block/fops.c | 23 |
7 files changed, 220 insertions, 117 deletions
diff --git a/block/Kconfig b/block/Kconfig index 55ae2286a4ded..1de4682d48ccb 100644 --- a/block/Kconfig +++ b/block/Kconfig @@ -78,6 +78,26 @@ config BLK_DEV_INTEGRITY_T10 select CRC_T10DIF select CRC64_ROCKSOFT +config BLK_DEV_WRITE_MOUNTED + bool "Allow writing to mounted block devices" + default y + help + When a block device is mounted, writing to its buffer cache is very + likely going to cause filesystem corruption. It is also rather easy to + crash the kernel in this way since the filesystem has no practical way + of detecting these writes to buffer cache and verifying its metadata + integrity. However there are some setups that need this capability + like running fsck on read-only mounted root device, modifying some + features on mounted ext4 filesystem, and similar. If you say N, the + kernel will prevent processes from writing to block devices that are + mounted by filesystems which provides some more protection from runaway + privileged processes and generally makes it much harder to crash + filesystem drivers. Note however that this does not prevent + underlying device(s) from being modified by other means, e.g. by + directly submitting SCSI commands or through access to lower layers of + storage stack. If in doubt, say Y. The configuration can be overridden + with the bdev_allow_write_mounted boot option. + config BLK_DEV_ZONED bool "Zoned block device support" select MQ_IOSCHED_DEADLINE diff --git a/block/badblocks.c b/block/badblocks.c index fc92d4e18aa3c..db4ec8b9b2a8c 100644 --- a/block/badblocks.c +++ b/block/badblocks.c @@ -1312,12 +1312,14 @@ re_check: prev = prev_badblocks(bb, &bad, hint); /* start after all badblocks */ - if ((prev + 1) >= bb->count && !overlap_front(bb, prev, &bad)) { + if ((prev >= 0) && + ((prev + 1) >= bb->count) && !overlap_front(bb, prev, &bad)) { len = sectors; goto update_sectors; } - if (overlap_front(bb, prev, &bad)) { + /* Overlapped with front badblocks record */ + if ((prev >= 0) && overlap_front(bb, prev, &bad)) { if (BB_ACK(p[prev])) acked_badblocks++; else diff --git a/block/bdev.c b/block/bdev.c index 750aec178b6ab..e9f1b12bd75c7 100644 --- a/block/bdev.c +++ b/block/bdev.c @@ -30,6 +30,9 @@ #include "../fs/internal.h" #include "blk.h" +/* Should we allow writing to mounted block devices? */ +static bool bdev_allow_write_mounted = IS_ENABLED(CONFIG_BLK_DEV_WRITE_MOUNTED); + struct bdev_inode { struct block_device bdev; struct inode vfs_inode; @@ -207,85 +210,88 @@ int sync_blockdev_range(struct block_device *bdev, loff_t lstart, loff_t lend) EXPORT_SYMBOL(sync_blockdev_range); /** - * freeze_bdev - lock a filesystem and force it into a consistent state + * bdev_freeze - lock a filesystem and force it into a consistent state * @bdev: blockdevice to lock * * If a superblock is found on this device, we take the s_umount semaphore * on it to make sure nobody unmounts until the snapshot creation is done. * The reference counter (bd_fsfreeze_count) guarantees that only the last * unfreeze process can unfreeze the frozen filesystem actually when multiple - * freeze requests arrive simultaneously. It counts up in freeze_bdev() and - * count down in thaw_bdev(). When it becomes 0, thaw_bdev() will unfreeze + * freeze requests arrive simultaneously. It counts up in bdev_freeze() and + * count down in bdev_thaw(). When it becomes 0, thaw_bdev() will unfreeze * actually. + * + * Return: On success zero is returned, negative error code on failure. */ -int freeze_bdev(struct block_device *bdev) +int bdev_freeze(struct block_device *bdev) { - struct super_block *sb; int error = 0; mutex_lock(&bdev->bd_fsfreeze_mutex); - if (++bdev->bd_fsfreeze_count > 1) - goto done; - - sb = get_active_super(bdev); - if (!sb) - goto sync; - if (sb->s_op->freeze_super) - error = sb->s_op->freeze_super(sb, FREEZE_HOLDER_USERSPACE); - else - error = freeze_super(sb, FREEZE_HOLDER_USERSPACE); - deactivate_super(sb); - if (error) { - bdev->bd_fsfreeze_count--; - goto done; + if (atomic_inc_return(&bdev->bd_fsfreeze_count) > 1) { + mutex_unlock(&bdev->bd_fsfreeze_mutex); + return 0; } - bdev->bd_fsfreeze_sb = sb; -sync: - sync_blockdev(bdev); -done: + mutex_lock(&bdev->bd_holder_lock); + if (bdev->bd_holder_ops && bdev->bd_holder_ops->freeze) { + error = bdev->bd_holder_ops->freeze(bdev); + lockdep_assert_not_held(&bdev->bd_holder_lock); + } else { + mutex_unlock(&bdev->bd_holder_lock); + error = sync_blockdev(bdev); + } + + if (error) + atomic_dec(&bdev->bd_fsfreeze_count); + mutex_unlock(&bdev->bd_fsfreeze_mutex); return error; } -EXPORT_SYMBOL(freeze_bdev); +EXPORT_SYMBOL(bdev_freeze); /** - * thaw_bdev - unlock filesystem + * bdev_thaw - unlock filesystem * @bdev: blockdevice to unlock * - * Unlocks the filesystem and marks it writeable again after freeze_bdev(). + * Unlocks the filesystem and marks it writeable again after bdev_freeze(). + * + * Return: On success zero is returned, negative error code on failure. */ -int thaw_bdev(struct block_device *bdev) +int bdev_thaw(struct block_device *bdev) { - struct super_block *sb; - int error = -EINVAL; + int error = -EINVAL, nr_freeze; mutex_lock(&bdev->bd_fsfreeze_mutex); - if (!bdev->bd_fsfreeze_count) + + /* + * If this returns < 0 it means that @bd_fsfreeze_count was + * already 0 and no decrement was performed. + */ + nr_freeze = atomic_dec_if_positive(&bdev->bd_fsfreeze_count); + if (nr_freeze < 0) goto out; error = 0; - if (--bdev->bd_fsfreeze_count > 0) + if (nr_freeze > 0) goto out; - sb = bdev->bd_fsfreeze_sb; - if (!sb) - goto out; + mutex_lock(&bdev->bd_holder_lock); + if (bdev->bd_holder_ops && bdev->bd_holder_ops->thaw) { + error = bdev->bd_holder_ops->thaw(bdev); + lockdep_assert_not_held(&bdev->bd_holder_lock); + } else { + mutex_unlock(&bdev->bd_holder_lock); + } - if (sb->s_op->thaw_super) - error = sb->s_op->thaw_super(sb, FREEZE_HOLDER_USERSPACE); - else - error = thaw_super(sb, FREEZE_HOLDER_USERSPACE); if (error) - bdev->bd_fsfreeze_count++; - else - bdev->bd_fsfreeze_sb = NULL; + atomic_inc(&bdev->bd_fsfreeze_count); out: mutex_unlock(&bdev->bd_fsfreeze_mutex); return error; } -EXPORT_SYMBOL(thaw_bdev); +EXPORT_SYMBOL(bdev_thaw); /* * pseudo-fs @@ -729,9 +735,60 @@ void blkdev_put_no_open(struct block_device *bdev) { put_device(&bdev->bd_device); } - + +static bool bdev_writes_blocked(struct block_device *bdev) +{ + return bdev->bd_writers == -1; +} + +static void bdev_block_writes(struct block_device *bdev) +{ + bdev->bd_writers = -1; +} + +static void bdev_unblock_writes(struct block_device *bdev) +{ + bdev->bd_writers = 0; +} + +static bool bdev_may_open(struct block_device *bdev, blk_mode_t mode) +{ + if (bdev_allow_write_mounted) + return true; + /* Writes blocked? */ + if (mode & BLK_OPEN_WRITE && bdev_writes_blocked(bdev)) + return false; + if (mode & BLK_OPEN_RESTRICT_WRITES && bdev->bd_writers > 0) + return false; + return true; +} + +static void bdev_claim_write_access(struct block_device *bdev, blk_mode_t mode) +{ + if (bdev_allow_write_mounted) + return; + + /* Claim exclusive or shared write access. */ + if (mode & BLK_OPEN_RESTRICT_WRITES) + bdev_block_writes(bdev); + else if (mode & BLK_OPEN_WRITE) + bdev->bd_writers++; +} + +static void bdev_yield_write_access(struct block_device *bdev, blk_mode_t mode) +{ + if (bdev_allow_write_mounted) + return; + + /* Yield exclusive or shared write access. */ + if (mode & BLK_OPEN_RESTRICT_WRITES) + bdev_unblock_writes(bdev); + else if (mode & BLK_OPEN_WRITE) + bdev->bd_writers--; +} + /** - * blkdev_get_by_dev - open a block device by device number + * bdev_open_by_dev - open a block device by device number * @dev: device number of block device to open * @mode: open mode (BLK_OPEN_*) * @holder: exclusive holder identifier @@ -743,32 +800,46 @@ void blkdev_put_no_open(struct block_device *bdev) * * Use this interface ONLY if you really do not have anything better - i.e. when * you are behind a truly sucky interface and all you are given is a device - * number. Everything else should use blkdev_get_by_path(). + * number. Everything else should use bdev_open_by_path(). * * CONTEXT: * Might sleep. * * RETURNS: - * Reference to the block_device on success, ERR_PTR(-errno) on failure. + * Handle with a reference to the block_device on success, ERR_PTR(-errno) on + * failure. */ -struct block_device *blkdev_get_by_dev(dev_t dev, blk_mode_t mode, void *holder, - const struct blk_holder_ops *hops) +struct bdev_handle *bdev_open_by_dev(dev_t dev, blk_mode_t mode, void *holder, + const struct blk_holder_ops *hops) { - bool unblock_events = true; + struct bdev_handle *handle = kmalloc(sizeof(struct bdev_handle), + GFP_KERNEL); struct block_device *bdev; + bool unblock_events = true; struct gendisk *disk; int ret; + if (!handle) + return ERR_PTR(-ENOMEM); + ret = devcgroup_check_permission(DEVCG_DEV_BLOCK, MAJOR(dev), MINOR(dev), ((mode & BLK_OPEN_READ) ? DEVCG_ACC_READ : 0) | ((mode & BLK_OPEN_WRITE) ? DEVCG_ACC_WRITE : 0)); if (ret) - return ERR_PTR(ret); + goto free_handle; + + /* Blocking writes requires exclusive opener */ + if (mode & BLK_OPEN_RESTRICT_WRITES && !holder) { + ret = -EINVAL; + goto free_handle; + } bdev = blkdev_get_no_open(dev); - if (!bdev) - return ERR_PTR(-ENXIO); + if (!bdev) { + ret = -ENXIO; + goto free_handle; + } disk = bdev->bd_disk; if (holder) { @@ -791,12 +862,16 @@ struct block_device *blkdev_get_by_dev(dev_t dev, blk_mode_t mode, void *holder, goto abort_claiming; if (!try_module_get(disk->fops->owner)) goto abort_claiming; + ret = -EBUSY; + if (!bdev_may_open(bdev, mode)) + goto abort_claiming; if (bdev_is_partition(bdev)) ret = blkdev_get_part(bdev, mode); else ret = blkdev_get_whole(bdev, mode); if (ret) goto put_module; + bdev_claim_write_access(bdev, mode); if (holder) { bd_finish_claiming(bdev, holder, hops); @@ -817,7 +892,10 @@ struct block_device *blkdev_get_by_dev(dev_t dev, blk_mode_t mode, void *holder, if (unblock_events) disk_unblock_events(disk); - return bdev; + handle->bdev = bdev; + handle->holder = holder; + handle->mode = mode; + return handle; put_module: module_put(disk->fops->owner); abort_claiming: @@ -827,34 +905,14 @@ abort_claiming: disk_unblock_events(disk); put_blkdev: blkdev_put_no_open(bdev); +free_handle: + kfree(handle); return ERR_PTR(ret); } -EXPORT_SYMBOL(blkdev_get_by_dev); - -struct bdev_handle *bdev_open_by_dev(dev_t dev, blk_mode_t mode, void *holder, - const struct blk_holder_ops *hops) -{ - struct bdev_handle *handle = kmalloc(sizeof(*handle), GFP_KERNEL); - struct block_device *bdev; - - if (!handle) - return ERR_PTR(-ENOMEM); - bdev = blkdev_get_by_dev(dev, mode, holder, hops); - if (IS_ERR(bdev)) { - kfree(handle); - return ERR_CAST(bdev); - } - handle->bdev = bdev; - handle->holder = holder; - if (holder) - mode |= BLK_OPEN_EXCL; - handle->mode = mode; - return handle; -} EXPORT_SYMBOL(bdev_open_by_dev); /** - * blkdev_get_by_path - open a block device by name + * bdev_open_by_path - open a block device by name * @path: path to the block device to open * @mode: open mode (BLK_OPEN_*) * @holder: exclusive holder identifier @@ -868,29 +926,9 @@ EXPORT_SYMBOL(bdev_open_by_dev); * Might sleep. * * RETURNS: - * Reference to the block_device on success, ERR_PTR(-errno) on failure. + * Handle with a reference to the block_device on success, ERR_PTR(-errno) on + * failure. */ -struct block_device *blkdev_get_by_path(const char *path, blk_mode_t mode, - void *holder, const struct blk_holder_ops *hops) -{ - struct block_device *bdev; - dev_t dev; - int error; - - error = lookup_bdev(path, &dev); - if (error) - return ERR_PTR(error); - - bdev = blkdev_get_by_dev(dev, mode, holder, hops); - if (!IS_ERR(bdev) && (mode & BLK_OPEN_WRITE) && bdev_read_only(bdev)) { - blkdev_put(bdev, holder); - return ERR_PTR(-EACCES); - } - - return bdev; -} -EXPORT_SYMBOL(blkdev_get_by_path); - struct bdev_handle *bdev_open_by_path(const char *path, blk_mode_t mode, void *holder, const struct blk_holder_ops *hops) { @@ -913,8 +951,9 @@ struct bdev_handle *bdev_open_by_path(const char *path, blk_mode_t mode, } EXPORT_SYMBOL(bdev_open_by_path); -void blkdev_put(struct block_device *bdev, void *holder) +void bdev_release(struct bdev_handle *handle) { + struct block_device *bdev = handle->bdev; struct gendisk *disk = bdev->bd_disk; /* @@ -928,8 +967,10 @@ void blkdev_put(struct block_device *bdev, void *holder) sync_blockdev(bdev); mutex_lock(&disk->open_mutex); - if (holder) - bd_end_claim(bdev, holder); + bdev_yield_write_access(bdev, handle->mode); + + if (handle->holder) + bd_end_claim(bdev, handle->holder); /* * Trigger event checking and tell drivers to flush MEDIA_CHANGE @@ -946,12 +987,6 @@ void blkdev_put(struct block_device *bdev, void *holder) module_put(disk->fops->owner); blkdev_put_no_open(bdev); -} -EXPORT_SYMBOL(blkdev_put); - -void bdev_release(struct bdev_handle *handle) -{ - blkdev_put(handle->bdev, handle->holder); kfree(handle); } EXPORT_SYMBOL(bdev_release); @@ -1102,3 +1137,12 @@ void bdev_statx_dioalign(struct inode *inode, struct kstat *stat) blkdev_put_no_open(bdev); } + +static int __init setup_bdev_allow_write_mounted(char *str) +{ + if (kstrtobool(str, &bdev_allow_write_mounted)) + pr_warn("Invalid option string for bdev_allow_write_mounted:" + " '%s'\n", str); + return 1; +} +__setup("bdev_allow_write_mounted=", setup_bdev_allow_write_mounted); diff --git a/block/blk-core.c b/block/blk-core.c index 9520ccab30500..11342af420d0c 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -501,9 +501,17 @@ static inline void bio_check_ro(struct bio *bio) if (op_is_write(bio_op(bio)) && bdev_read_only(bio->bi_bdev)) { if (op_is_flush(bio->bi_opf) && !bio_sectors(bio)) return; - pr_warn_ratelimited("Trying to write to read-only block-device %pg\n", - bio->bi_bdev); - /* Older lvm-tools actually trigger this */ + + if (bio->bi_bdev->bd_ro_warned) + return; + + bio->bi_bdev->bd_ro_warned = true; + /* + * Use ioctl to set underlying disk of raid/dm to read-only + * will trigger this. + */ + pr_warn("Trying to write to read-only block-device %pg\n", + bio->bi_bdev); } } diff --git a/block/blk-mq.c b/block/blk-mq.c index fb29ff5cc281d..c11c97afa0bc1 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -1513,14 +1513,26 @@ void blk_mq_delay_kick_requeue_list(struct request_queue *q, } EXPORT_SYMBOL(blk_mq_delay_kick_requeue_list); +static bool blk_is_flush_data_rq(struct request *rq) +{ + return (rq->rq_flags & RQF_FLUSH_SEQ) && !is_flush_rq(rq); +} + static bool blk_mq_rq_inflight(struct request *rq, void *priv) { /* * If we find a request that isn't idle we know the queue is busy * as it's checked in the iter. * Return false to stop the iteration. + * + * In case of queue quiesce, if one flush data request is completed, + * don't count it as inflight given the flush sequence is suspended, + * and the original flush data request is invisible to driver, just + * like other pending requests because of quiesce */ - if (blk_mq_request_started(rq)) { + if (blk_mq_request_started(rq) && !(blk_queue_quiesced(rq->q) && + blk_is_flush_data_rq(rq) && + blk_mq_request_completed(rq))) { bool *busy = priv; *busy = true; diff --git a/block/blk-sysfs.c b/block/blk-sysfs.c index 40bab5975c561..6b2429cad81af 100644 --- a/block/blk-sysfs.c +++ b/block/blk-sysfs.c @@ -610,6 +610,7 @@ static ssize_t queue_wb_lat_store(struct request_queue *q, const char *page, QUEUE_RW_ENTRY(queue_wb_lat, "wbt_lat_usec"); #endif +/* Common attributes for bio-based and request-based queues. */ static struct attribute *queue_attrs[] = { &queue_ra_entry.attr, &queue_max_hw_sectors_entry.attr, @@ -654,6 +655,7 @@ static struct attribute *queue_attrs[] = { NULL, }; +/* Request-based queue attributes that are not relevant for bio-based queues. */ static struct attribute *blk_mq_queue_attrs[] = { &queue_requests_entry.attr, &elv_iosched_entry.attr, diff --git a/block/fops.c b/block/fops.c index 0abaac705dafb..0cf8cf72cdfa1 100644 --- a/block/fops.c +++ b/block/fops.c @@ -410,9 +410,24 @@ static int blkdev_get_block(struct inode *inode, sector_t iblock, return 0; } -static int blkdev_writepage(struct page *page, struct writeback_control *wbc) +/* + * We cannot call mpage_writepages() as it does not take the buffer lock. + * We must use block_write_full_folio() directly which holds the buffer + * lock. The buffer lock provides the synchronisation with writeback + * that filesystems rely on when they use the blockdev's mapping. + */ +static int blkdev_writepages(struct address_space *mapping, + struct writeback_control *wbc) { - return block_write_full_page(page, blkdev_get_block, wbc); + struct blk_plug plug; + int err; + + blk_start_plug(&plug); + err = write_cache_pages(mapping, wbc, block_write_full_folio, + blkdev_get_block); + blk_finish_plug(&plug); + + return err; } static int blkdev_read_folio(struct file *file, struct folio *folio) @@ -449,7 +464,7 @@ const struct address_space_operations def_blk_aops = { .invalidate_folio = block_invalidate_folio, .read_folio = blkdev_read_folio, .readahead = blkdev_readahead, - .writepage = blkdev_writepage, + .writepages = blkdev_writepages, .write_begin = blkdev_write_begin, .write_end = blkdev_write_end, .migrate_folio = buffer_migrate_folio_norefs, @@ -500,7 +515,7 @@ const struct address_space_operations def_blk_aops = { .readahead = blkdev_readahead, .writepages = blkdev_writepages, .is_partially_uptodate = iomap_is_partially_uptodate, - .error_remove_page = generic_error_remove_page, + .error_remove_folio = generic_error_remove_folio, .migrate_folio = filemap_migrate_folio, }; #endif /* CONFIG_BUFFER_HEAD */ |