From 87d8fe1ee6b8d2f95076142d58c440dba4e7bdc2 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Sat, 3 Jan 2009 09:47:09 -0500 Subject: add releasepage hooks to block devices which can be used by file systems Implement blkdev_releasepage() to release the buffer_heads and pages after we release private data belonging to a mounted filesystem. Cc: Toshiyuki Okajima Cc: linux-fsdevel@vger.kernel.org Signed-off-by: "Theodore Ts'o" --- fs/block_dev.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+) (limited to 'fs/block_dev.c') diff --git a/fs/block_dev.c b/fs/block_dev.c index 349a26c10001..1dd07e66e98a 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -1220,6 +1220,20 @@ static long block_ioctl(struct file *file, unsigned cmd, unsigned long arg) return blkdev_ioctl(bdev, mode, cmd, arg); } +/* + * Try to release a page associated with block device when the system + * is under memory pressure. + */ +static int blkdev_releasepage(struct page *page, gfp_t wait) +{ + struct super_block *super = BDEV_I(page->mapping->host)->bdev.bd_super; + + if (super && super->s_op->bdev_try_to_free_page) + return super->s_op->bdev_try_to_free_page(super, page, wait); + + return try_to_free_buffers(page); +} + static const struct address_space_operations def_blk_aops = { .readpage = blkdev_readpage, .writepage = blkdev_writepage, @@ -1227,6 +1241,7 @@ static const struct address_space_operations def_blk_aops = { .write_begin = blkdev_write_begin, .write_end = blkdev_write_end, .writepages = generic_writepages, + .releasepage = blkdev_releasepage, .direct_IO = blkdev_direct_IO, }; -- cgit v1.2.3 From d3374825ce57ba2214d375023979f6197ccc1385 Mon Sep 17 00:00:00 2001 From: NeilBrown Date: Fri, 9 Jan 2009 08:31:10 +1100 Subject: md: make devices disappear when they are no longer needed. Currently md devices, once created, never disappear until the module is unloaded. This is essentially because the gendisk holds a reference to the mddev, and the mddev holds a reference to the gendisk, this a circular reference. If we drop the reference from mddev to gendisk, then we need to ensure that the mddev is destroyed when the gendisk is destroyed. However it is not possible to hook into the gendisk destruction process to enable this. So we drop the reference from the gendisk to the mddev and destroy the gendisk when the mddev gets destroyed. However this has a complication. Between the call __blkdev_get->get_gendisk->kobj_lookup->md_probe and the call __blkdev_get->md_open there is no obvious way to hold a reference on the mddev any more, so unless something is done, it will disappear and gendisk will be destroyed prematurely. Also, once we decide to destroy the mddev, there will be an unlockable moment before the gendisk is unlinked (blk_unregister_region) during which a new reference to the gendisk can be created. We need to ensure that this reference can not be used. i.e. the ->open must fail. So: 1/ in md_probe we set a flag in the mddev (hold_active) which indicates that the array should be treated as active, even though there are no references, and no appearance of activity. This is cleared by md_release when the device is closed if it is no longer needed. This ensures that the gendisk will survive between md_probe and md_open. 2/ In md_open we check if the mddev we expect to open matches the gendisk that we did open. If there is a mismatch we return -ERESTARTSYS and modify __blkdev_get to retry from the top in that case. In the -ERESTARTSYS sys case we make sure to wait until the old gendisk (that we succeeded in opening) is really gone so we loop at most once. Some udev configurations will always open an md device when it first appears. If we allow an md device that was just created by an open to disappear on an immediate close, then this can race with such udev configurations and result in an infinite loop the device being opened and closed, then re-open due to the 'ADD' even from the first open, and then close and so on. So we make sure an md device, once created by an open, remains active at least until some md 'ioctl' has been made on it. This means that all normal usage of md devices will allow them to disappear promptly when not needed, but the worst that an incorrect usage will do it cause an inactive md device to be left in existence (it can easily be removed). As an array can be stopped by writing to a sysfs attribute echo clear > /sys/block/mdXXX/md/array_state we need to use scheduled work for deleting the gendisk and other kobjects. This allows us to wait for any pending gendisk deletion to complete by simply calling flush_scheduled_work(). Signed-off-by: NeilBrown --- drivers/md/md.c | 61 +++++++++++++++++++++++++++++++++++++---------- fs/block_dev.c | 14 +++++++++++ include/linux/raid/md_k.h | 4 ++++ 3 files changed, 67 insertions(+), 12 deletions(-) (limited to 'fs/block_dev.c') diff --git a/drivers/md/md.c b/drivers/md/md.c index 970a8c42ba92..38697283aaf4 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -214,16 +214,33 @@ static inline mddev_t *mddev_get(mddev_t *mddev) return mddev; } +static void mddev_delayed_delete(struct work_struct *ws) +{ + mddev_t *mddev = container_of(ws, mddev_t, del_work); + kobject_del(&mddev->kobj); + kobject_put(&mddev->kobj); +} + static void mddev_put(mddev_t *mddev) { if (!atomic_dec_and_lock(&mddev->active, &all_mddevs_lock)) return; - if (!mddev->raid_disks && list_empty(&mddev->disks)) { + if (!mddev->raid_disks && list_empty(&mddev->disks) && + !mddev->hold_active) { list_del(&mddev->all_mddevs); - spin_unlock(&all_mddevs_lock); - kobject_put(&mddev->kobj); - } else - spin_unlock(&all_mddevs_lock); + if (mddev->gendisk) { + /* we did a probe so need to clean up. + * Call schedule_work inside the spinlock + * so that flush_scheduled_work() after + * mddev_find will succeed in waiting for the + * work to be done. + */ + INIT_WORK(&mddev->del_work, mddev_delayed_delete); + schedule_work(&mddev->del_work); + } else + kfree(mddev); + } + spin_unlock(&all_mddevs_lock); } static mddev_t * mddev_find(dev_t unit) @@ -242,6 +259,7 @@ static mddev_t * mddev_find(dev_t unit) if (new) { list_add(&new->all_mddevs, &all_mddevs); + mddev->hold_active = UNTIL_IOCTL; spin_unlock(&all_mddevs_lock); return new; } @@ -3435,6 +3453,8 @@ md_attr_store(struct kobject *kobj, struct attribute *attr, if (!capable(CAP_SYS_ADMIN)) return -EACCES; rv = mddev_lock(mddev); + if (mddev->hold_active == UNTIL_IOCTL) + mddev->hold_active = 0; if (!rv) { rv = entry->store(mddev, page, length); mddev_unlock(mddev); @@ -3484,6 +3504,11 @@ static struct kobject *md_probe(dev_t dev, int *part, void *data) if (!mddev) return NULL; + /* wait for any previous instance if this device + * to be completed removed (mddev_delayed_delete). + */ + flush_scheduled_work(); + mutex_lock(&disks_mutex); if (mddev->gendisk) { mutex_unlock(&disks_mutex); @@ -3520,7 +3545,7 @@ static struct kobject *md_probe(dev_t dev, int *part, void *data) disk->private_data = mddev; disk->queue = mddev->queue; /* Allow extended partitions. This makes the - * 'mdp' device redundant, but we can really + * 'mdp' device redundant, but we can't really * remove it now. */ disk->flags |= GENHD_FL_EXT_DEVT; @@ -3536,6 +3561,7 @@ static struct kobject *md_probe(dev_t dev, int *part, void *data) kobject_uevent(&mddev->kobj, KOBJ_ADD); mddev->sysfs_state = sysfs_get_dirent(mddev->kobj.sd, "array_state"); } + mddev_put(mddev); return NULL; } @@ -5054,6 +5080,9 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode, done_unlock: abort_unlock: + if (mddev->hold_active == UNTIL_IOCTL && + err != -EINVAL) + mddev->hold_active = 0; mddev_unlock(mddev); return err; @@ -5070,14 +5099,25 @@ static int md_open(struct block_device *bdev, fmode_t mode) * Succeed if we can lock the mddev, which confirms that * it isn't being stopped right now. */ - mddev_t *mddev = bdev->bd_disk->private_data; + mddev_t *mddev = mddev_find(bdev->bd_dev); int err; + if (mddev->gendisk != bdev->bd_disk) { + /* we are racing with mddev_put which is discarding this + * bd_disk. + */ + mddev_put(mddev); + /* Wait until bdev->bd_disk is definitely gone */ + flush_scheduled_work(); + /* Then retry the open from the top */ + return -ERESTARTSYS; + } + BUG_ON(mddev != bdev->bd_disk->private_data); + if ((err = mutex_lock_interruptible_nested(&mddev->reconfig_mutex, 1))) goto out; err = 0; - mddev_get(mddev); atomic_inc(&mddev->openers); mddev_unlock(mddev); @@ -6436,11 +6476,8 @@ static __exit void md_exit(void) unregister_sysctl_table(raid_table_header); remove_proc_entry("mdstat", NULL); for_each_mddev(mddev, tmp) { - struct gendisk *disk = mddev->gendisk; - if (!disk) - continue; export_array(mddev); - mddev_put(mddev); + mddev->hold_active = 0; } } diff --git a/fs/block_dev.c b/fs/block_dev.c index b957717e25ab..8ebbfdf708c2 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -1005,6 +1005,7 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part) } lock_kernel(); + restart: ret = -ENXIO; disk = get_gendisk(bdev->bd_dev, &partno); @@ -1025,6 +1026,19 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part) if (disk->fops->open) { ret = disk->fops->open(bdev, mode); + if (ret == -ERESTARTSYS) { + /* Lost a race with 'disk' being + * deleted, try again. + * See md.c + */ + disk_put_part(bdev->bd_part); + bdev->bd_part = NULL; + module_put(disk->fops->owner); + put_disk(disk); + bdev->bd_disk = NULL; + mutex_unlock(&bdev->bd_mutex); + goto restart; + } if (ret) goto out_clear; } diff --git a/include/linux/raid/md_k.h b/include/linux/raid/md_k.h index 8f9a54c1fb0e..e3d17c7f954e 100644 --- a/include/linux/raid/md_k.h +++ b/include/linux/raid/md_k.h @@ -137,6 +137,8 @@ struct mddev_s struct gendisk *gendisk; struct kobject kobj; + int hold_active; +#define UNTIL_IOCTL 1 /* Superblock information */ int major_version, @@ -246,6 +248,8 @@ struct mddev_s */ struct sysfs_dirent *sysfs_action; /* handle for 'sync_action' */ + struct work_struct del_work; /* used for delayed sysfs removal */ + spinlock_t write_lock; wait_queue_head_t sb_wait; /* for waiting on superblock updates */ atomic_t pending_writes; /* number of active superblock writes */ -- cgit v1.2.3 From fcccf502540e3d752d33b2d8e976034dee81f9f7 Mon Sep 17 00:00:00 2001 From: Takashi Sato Date: Fri, 9 Jan 2009 16:40:59 -0800 Subject: filesystem freeze: implement generic freeze feature The ioctls for the generic freeze feature are below. o Freeze the filesystem int ioctl(int fd, int FIFREEZE, arg) fd: The file descriptor of the mountpoint FIFREEZE: request code for the freeze arg: Ignored Return value: 0 if the operation succeeds. Otherwise, -1 o Unfreeze the filesystem int ioctl(int fd, int FITHAW, arg) fd: The file descriptor of the mountpoint FITHAW: request code for unfreeze arg: Ignored Return value: 0 if the operation succeeds. Otherwise, -1 Error number: If the filesystem has already been unfrozen, errno is set to EINVAL. [akpm@linux-foundation.org: fix CONFIG_BLOCK=n] Signed-off-by: Takashi Sato Signed-off-by: Masayuki Hamaguchi Cc: Cc: Cc: Christoph Hellwig Cc: Dave Kleikamp Cc: Dave Chinner Cc: Alasdair G Kergon Cc: Al Viro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/block_dev.c | 2 ++ fs/buffer.c | 74 +++++++++++++++++++++++++++++++++++++++------ fs/ioctl.c | 46 ++++++++++++++++++++++++++++ include/linux/buffer_head.h | 11 ++++++- include/linux/fs.h | 7 +++++ 5 files changed, 130 insertions(+), 10 deletions(-) (limited to 'fs/block_dev.c') diff --git a/fs/block_dev.c b/fs/block_dev.c index ac7031f12ea5..b3c1efff5e1d 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -285,6 +285,8 @@ static void init_once(void *foo) INIT_LIST_HEAD(&bdev->bd_holder_list); #endif inode_init_once(&ei->vfs_inode); + /* Initialize mutex for freeze. */ + mutex_init(&bdev->bd_fsfreeze_mutex); } static inline void __bd_forget(struct inode *inode) diff --git a/fs/buffer.c b/fs/buffer.c index 87f9e537b8c3..b6e8b8632e2f 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -203,10 +203,25 @@ int fsync_bdev(struct block_device *bdev) * happen on bdev until thaw_bdev() is called. * 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 + * actually. */ struct super_block *freeze_bdev(struct block_device *bdev) { struct super_block *sb; + int error = 0; + + mutex_lock(&bdev->bd_fsfreeze_mutex); + if (bdev->bd_fsfreeze_count > 0) { + bdev->bd_fsfreeze_count++; + sb = get_super(bdev); + mutex_unlock(&bdev->bd_fsfreeze_mutex); + return sb; + } + bdev->bd_fsfreeze_count++; down(&bdev->bd_mount_sem); sb = get_super(bdev); @@ -221,11 +236,24 @@ struct super_block *freeze_bdev(struct block_device *bdev) sync_blockdev(sb->s_bdev); - if (sb->s_op->freeze_fs) - sb->s_op->freeze_fs(sb); + if (sb->s_op->freeze_fs) { + error = sb->s_op->freeze_fs(sb); + if (error) { + printk(KERN_ERR + "VFS:Filesystem freeze failed\n"); + sb->s_frozen = SB_UNFROZEN; + drop_super(sb); + up(&bdev->bd_mount_sem); + bdev->bd_fsfreeze_count--; + mutex_unlock(&bdev->bd_fsfreeze_mutex); + return ERR_PTR(error); + } + } } sync_blockdev(bdev); + mutex_unlock(&bdev->bd_fsfreeze_mutex); + return sb; /* thaw_bdev releases s->s_umount and bd_mount_sem */ } EXPORT_SYMBOL(freeze_bdev); @@ -237,20 +265,48 @@ EXPORT_SYMBOL(freeze_bdev); * * Unlocks the filesystem and marks it writeable again after freeze_bdev(). */ -void thaw_bdev(struct block_device *bdev, struct super_block *sb) +int thaw_bdev(struct block_device *bdev, struct super_block *sb) { + int error = 0; + + mutex_lock(&bdev->bd_fsfreeze_mutex); + if (!bdev->bd_fsfreeze_count) { + mutex_unlock(&bdev->bd_fsfreeze_mutex); + return -EINVAL; + } + + bdev->bd_fsfreeze_count--; + if (bdev->bd_fsfreeze_count > 0) { + if (sb) + drop_super(sb); + mutex_unlock(&bdev->bd_fsfreeze_mutex); + return 0; + } + if (sb) { BUG_ON(sb->s_bdev != bdev); - - if (sb->s_op->unfreeze_fs) - sb->s_op->unfreeze_fs(sb); - sb->s_frozen = SB_UNFROZEN; - smp_wmb(); - wake_up(&sb->s_wait_unfrozen); + if (!(sb->s_flags & MS_RDONLY)) { + if (sb->s_op->unfreeze_fs) { + error = sb->s_op->unfreeze_fs(sb); + if (error) { + printk(KERN_ERR + "VFS:Filesystem thaw failed\n"); + sb->s_frozen = SB_FREEZE_TRANS; + bdev->bd_fsfreeze_count++; + mutex_unlock(&bdev->bd_fsfreeze_mutex); + return error; + } + } + sb->s_frozen = SB_UNFROZEN; + smp_wmb(); + wake_up(&sb->s_wait_unfrozen); + } drop_super(sb); } up(&bdev->bd_mount_sem); + mutex_unlock(&bdev->bd_fsfreeze_mutex); + return 0; } EXPORT_SYMBOL(thaw_bdev); diff --git a/fs/ioctl.c b/fs/ioctl.c index cc3f1aa1cf7b..20b0a8a24c6b 100644 --- a/fs/ioctl.c +++ b/fs/ioctl.c @@ -439,6 +439,43 @@ static int ioctl_fioasync(unsigned int fd, struct file *filp, return error; } +static int ioctl_fsfreeze(struct file *filp) +{ + struct super_block *sb = filp->f_path.dentry->d_inode->i_sb; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + /* If filesystem doesn't support freeze feature, return. */ + if (sb->s_op->freeze_fs == NULL) + return -EOPNOTSUPP; + + /* If a blockdevice-backed filesystem isn't specified, return. */ + if (sb->s_bdev == NULL) + return -EINVAL; + + /* Freeze */ + sb = freeze_bdev(sb->s_bdev); + if (IS_ERR(sb)) + return PTR_ERR(sb); + return 0; +} + +static int ioctl_fsthaw(struct file *filp) +{ + struct super_block *sb = filp->f_path.dentry->d_inode->i_sb; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + /* If a blockdevice-backed filesystem isn't specified, return EINVAL. */ + if (sb->s_bdev == NULL) + return -EINVAL; + + /* Thaw */ + return thaw_bdev(sb->s_bdev, sb); +} + /* * When you add any new common ioctls to the switches above and below * please update compat_sys_ioctl() too. @@ -486,6 +523,15 @@ int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd, } else error = -ENOTTY; break; + + case FIFREEZE: + error = ioctl_fsfreeze(filp); + break; + + case FITHAW: + error = ioctl_fsthaw(filp); + break; + default: if (S_ISREG(filp->f_path.dentry->d_inode->i_mode)) error = file_ioctl(filp, cmd, arg); diff --git a/include/linux/buffer_head.h b/include/linux/buffer_head.h index 8605f8a74df9..bd7ac793be19 100644 --- a/include/linux/buffer_head.h +++ b/include/linux/buffer_head.h @@ -171,7 +171,7 @@ void __wait_on_buffer(struct buffer_head *); wait_queue_head_t *bh_waitq_head(struct buffer_head *bh); int fsync_bdev(struct block_device *); struct super_block *freeze_bdev(struct block_device *); -void thaw_bdev(struct block_device *, struct super_block *); +int thaw_bdev(struct block_device *, struct super_block *); int fsync_super(struct super_block *); int fsync_no_super(struct block_device *); struct buffer_head *__find_get_block(struct block_device *bdev, sector_t block, @@ -346,6 +346,15 @@ static inline int remove_inode_buffers(struct inode *inode) { return 1; } static inline int sync_mapping_buffers(struct address_space *mapping) { return 0; } static inline void invalidate_bdev(struct block_device *bdev) {} +static inline struct super_block *freeze_bdev(struct block_device *sb) +{ + return NULL; +} + +static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb) +{ + return 0; +} #endif /* CONFIG_BLOCK */ #endif /* _LINUX_BUFFER_HEAD_H */ diff --git a/include/linux/fs.h b/include/linux/fs.h index 3e59182de9df..6022f44043f2 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -234,6 +234,8 @@ struct inodes_stat_t { #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */ #define FIBMAP _IO(0x00,1) /* bmap access */ #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */ +#define FIFREEZE _IOWR('X', 119, int) /* Freeze */ +#define FITHAW _IOWR('X', 120, int) /* Thaw */ #define FS_IOC_GETFLAGS _IOR('f', 1, long) #define FS_IOC_SETFLAGS _IOW('f', 2, long) @@ -591,6 +593,11 @@ struct block_device { * care to not mess up bd_private for that case. */ unsigned long bd_private; + + /* The counter of freeze processes */ + int bd_fsfreeze_count; + /* Mutex for freeze */ + struct mutex bd_fsfreeze_mutex; }; /* -- cgit v1.2.3