From 00251a527a6fae93ccd4322619b23db56ed82986 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 20 Mar 2018 15:35:50 +0100 Subject: btrfs: squeeze btrfs_dev_replace_continue_on_mount to its caller The function is called once and is fairly small, we can merge it with the caller. Reviewed-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/dev-replace.c | 16 +++------------- 1 file changed, 3 insertions(+), 13 deletions(-) (limited to 'fs/btrfs/dev-replace.c') diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c index f82be266ba4b..db4d08c65131 100644 --- a/fs/btrfs/dev-replace.c +++ b/fs/btrfs/dev-replace.c @@ -33,8 +33,6 @@ static void btrfs_dev_replace_update_device_in_mapping_tree( struct btrfs_device *srcdev, struct btrfs_device *tgtdev); static int btrfs_dev_replace_kthread(void *data); -static int btrfs_dev_replace_continue_on_mount(struct btrfs_fs_info *fs_info); - int btrfs_init_dev_replace(struct btrfs_fs_info *fs_info) { @@ -810,6 +808,7 @@ static int btrfs_dev_replace_kthread(void *data) struct btrfs_fs_info *fs_info = data; struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace; u64 progress; + int ret; progress = btrfs_dev_replace_progress(fs_info); progress = div_u64(progress, 10); @@ -820,23 +819,14 @@ static int btrfs_dev_replace_kthread(void *data) btrfs_dev_name(dev_replace->tgtdev), (unsigned int)progress); - btrfs_dev_replace_continue_on_mount(fs_info); - clear_bit(BTRFS_FS_EXCL_OP, &fs_info->flags); - - return 0; -} - -static int btrfs_dev_replace_continue_on_mount(struct btrfs_fs_info *fs_info) -{ - struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace; - int ret; - ret = btrfs_scrub_dev(fs_info, dev_replace->srcdev->devid, dev_replace->committed_cursor_left, btrfs_device_get_total_bytes(dev_replace->srcdev), &dev_replace->scrub_progress, 0, 1); ret = btrfs_dev_replace_finishing(fs_info, ret); WARN_ON(ret); + + clear_bit(BTRFS_FS_EXCL_OP, &fs_info->flags); return 0; } -- cgit From 6fc4749d25738e1a5e5b02d04a0a60bbae516652 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 20 Mar 2018 15:37:08 +0100 Subject: btrfs: make success path out of btrfs_init_dev_replace_tgtdev more clear This is a preparatory cleanup that will make clear that the only successful way out of btrfs_init_dev_replace_tgtdev will also set the device_out to a valid pointer. With this guarantee, the callers can be simplified. Reviewed-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/dev-replace.c | 1 - fs/btrfs/volumes.c | 8 +++++++- 2 files changed, 7 insertions(+), 2 deletions(-) (limited to 'fs/btrfs/dev-replace.c') diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c index db4d08c65131..e3ec0eb5789b 100644 --- a/fs/btrfs/dev-replace.c +++ b/fs/btrfs/dev-replace.c @@ -358,7 +358,6 @@ int btrfs_dev_replace_start(struct btrfs_fs_info *fs_info, dev_replace->cont_reading_from_srcdev_mode = read_src; WARN_ON(!src_device); dev_replace->srcdev = src_device; - WARN_ON(!tgt_device); dev_replace->tgtdev = tgt_device; btrfs_info_in_rcu(fs_info, diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 70a87d4fe5fe..a8f8a2e39da6 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -2612,6 +2612,12 @@ error: return ret; } +/* + * Initialize a new device for device replace target from a given source dev + * and path. + * + * Return 0 and new device in @device_out, otherwise return < 0 + */ int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info, const char *device_path, struct btrfs_device *srcdev, @@ -2698,7 +2704,7 @@ int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info, mutex_unlock(&fs_info->fs_devices->device_list_mutex); *device_out = device; - return ret; + return 0; error: blkdev_put(bdev, FMODE_EXCL); -- cgit From d48f39d5a529244f59454386208c6da92bb1c493 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 20 Mar 2018 16:09:48 +0100 Subject: btrfs: move btrfs_init_dev_replace_tgtdev to dev-replace.c and make static The function logically belongs there and there's only a single caller, no need to export it. No code changes. Reviewed-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/dev-replace.c | 99 ++++++++++++++++++++++++++++++++++++++++++++++++++ fs/btrfs/volumes.c | 99 -------------------------------------------------- fs/btrfs/volumes.h | 4 -- 3 files changed, 99 insertions(+), 103 deletions(-) (limited to 'fs/btrfs/dev-replace.c') diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c index e3ec0eb5789b..8531b5dae777 100644 --- a/fs/btrfs/dev-replace.c +++ b/fs/btrfs/dev-replace.c @@ -176,6 +176,105 @@ out: return ret; } +/* + * Initialize a new device for device replace target from a given source dev + * and path. + * + * Return 0 and new device in @device_out, otherwise return < 0 + */ +static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info, + const char *device_path, + struct btrfs_device *srcdev, + struct btrfs_device **device_out) +{ + struct btrfs_device *device; + struct block_device *bdev; + struct list_head *devices; + struct rcu_string *name; + u64 devid = BTRFS_DEV_REPLACE_DEVID; + int ret = 0; + + *device_out = NULL; + if (fs_info->fs_devices->seeding) { + btrfs_err(fs_info, "the filesystem is a seed filesystem!"); + return -EINVAL; + } + + bdev = blkdev_get_by_path(device_path, FMODE_WRITE | FMODE_EXCL, + fs_info->bdev_holder); + if (IS_ERR(bdev)) { + btrfs_err(fs_info, "target device %s is invalid!", device_path); + return PTR_ERR(bdev); + } + + filemap_write_and_wait(bdev->bd_inode->i_mapping); + + devices = &fs_info->fs_devices->devices; + list_for_each_entry(device, devices, dev_list) { + if (device->bdev == bdev) { + btrfs_err(fs_info, + "target device is in the filesystem!"); + ret = -EEXIST; + goto error; + } + } + + + if (i_size_read(bdev->bd_inode) < + btrfs_device_get_total_bytes(srcdev)) { + btrfs_err(fs_info, + "target device is smaller than source device!"); + ret = -EINVAL; + goto error; + } + + + device = btrfs_alloc_device(NULL, &devid, NULL); + if (IS_ERR(device)) { + ret = PTR_ERR(device); + goto error; + } + + name = rcu_string_strdup(device_path, GFP_KERNEL); + if (!name) { + btrfs_free_device(device); + ret = -ENOMEM; + goto error; + } + rcu_assign_pointer(device->name, name); + + mutex_lock(&fs_info->fs_devices->device_list_mutex); + set_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state); + device->generation = 0; + device->io_width = fs_info->sectorsize; + device->io_align = fs_info->sectorsize; + device->sector_size = fs_info->sectorsize; + device->total_bytes = btrfs_device_get_total_bytes(srcdev); + device->disk_total_bytes = btrfs_device_get_disk_total_bytes(srcdev); + device->bytes_used = btrfs_device_get_bytes_used(srcdev); + device->commit_total_bytes = srcdev->commit_total_bytes; + device->commit_bytes_used = device->bytes_used; + device->fs_info = fs_info; + device->bdev = bdev; + set_bit(BTRFS_DEV_STATE_IN_FS_METADATA, &device->dev_state); + set_bit(BTRFS_DEV_STATE_REPLACE_TGT, &device->dev_state); + device->mode = FMODE_EXCL; + device->dev_stats_valid = 1; + set_blocksize(device->bdev, BTRFS_BDEV_BLOCKSIZE); + device->fs_devices = fs_info->fs_devices; + list_add(&device->dev_list, &fs_info->fs_devices->devices); + fs_info->fs_devices->num_devices++; + fs_info->fs_devices->open_devices++; + mutex_unlock(&fs_info->fs_devices->device_list_mutex); + + *device_out = device; + return 0; + +error: + blkdev_put(bdev, FMODE_EXCL); + return ret; +} + /* * called from commit_transaction. Writes changed device replace state to * disk. diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 364f4e7206f4..488935e66779 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -2612,105 +2612,6 @@ error: return ret; } -/* - * Initialize a new device for device replace target from a given source dev - * and path. - * - * Return 0 and new device in @device_out, otherwise return < 0 - */ -int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info, - const char *device_path, - struct btrfs_device *srcdev, - struct btrfs_device **device_out) -{ - struct btrfs_device *device; - struct block_device *bdev; - struct list_head *devices; - struct rcu_string *name; - u64 devid = BTRFS_DEV_REPLACE_DEVID; - int ret = 0; - - *device_out = NULL; - if (fs_info->fs_devices->seeding) { - btrfs_err(fs_info, "the filesystem is a seed filesystem!"); - return -EINVAL; - } - - bdev = blkdev_get_by_path(device_path, FMODE_WRITE | FMODE_EXCL, - fs_info->bdev_holder); - if (IS_ERR(bdev)) { - btrfs_err(fs_info, "target device %s is invalid!", device_path); - return PTR_ERR(bdev); - } - - filemap_write_and_wait(bdev->bd_inode->i_mapping); - - devices = &fs_info->fs_devices->devices; - list_for_each_entry(device, devices, dev_list) { - if (device->bdev == bdev) { - btrfs_err(fs_info, - "target device is in the filesystem!"); - ret = -EEXIST; - goto error; - } - } - - - if (i_size_read(bdev->bd_inode) < - btrfs_device_get_total_bytes(srcdev)) { - btrfs_err(fs_info, - "target device is smaller than source device!"); - ret = -EINVAL; - goto error; - } - - - device = btrfs_alloc_device(NULL, &devid, NULL); - if (IS_ERR(device)) { - ret = PTR_ERR(device); - goto error; - } - - name = rcu_string_strdup(device_path, GFP_KERNEL); - if (!name) { - btrfs_free_device(device); - ret = -ENOMEM; - goto error; - } - rcu_assign_pointer(device->name, name); - - mutex_lock(&fs_info->fs_devices->device_list_mutex); - set_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state); - device->generation = 0; - device->io_width = fs_info->sectorsize; - device->io_align = fs_info->sectorsize; - device->sector_size = fs_info->sectorsize; - device->total_bytes = btrfs_device_get_total_bytes(srcdev); - device->disk_total_bytes = btrfs_device_get_disk_total_bytes(srcdev); - device->bytes_used = btrfs_device_get_bytes_used(srcdev); - device->commit_total_bytes = srcdev->commit_total_bytes; - device->commit_bytes_used = device->bytes_used; - device->fs_info = fs_info; - device->bdev = bdev; - set_bit(BTRFS_DEV_STATE_IN_FS_METADATA, &device->dev_state); - set_bit(BTRFS_DEV_STATE_REPLACE_TGT, &device->dev_state); - device->mode = FMODE_EXCL; - device->dev_stats_valid = 1; - set_blocksize(device->bdev, BTRFS_BDEV_BLOCKSIZE); - device->fs_devices = fs_info->fs_devices; - list_add(&device->dev_list, &fs_info->fs_devices->devices); - fs_info->fs_devices->num_devices++; - fs_info->fs_devices->open_devices++; - mutex_unlock(&fs_info->fs_devices->device_list_mutex); - - *device_out = device; - return 0; - -error: - blkdev_put(bdev, FMODE_EXCL); - return ret; -} - static noinline int btrfs_update_device(struct btrfs_trans_handle *trans, struct btrfs_device *device) { diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index 316cce159969..5737e6e68f8b 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -432,10 +432,6 @@ struct btrfs_device *btrfs_find_device(struct btrfs_fs_info *fs_info, u64 devid, u8 *uuid, u8 *fsid); int btrfs_shrink_device(struct btrfs_device *device, u64 new_size); int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *path); -int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info, - const char *device_path, - struct btrfs_device *srcdev, - struct btrfs_device **device_out); int btrfs_balance(struct btrfs_balance_control *bctl, struct btrfs_ioctl_balance_args *bargs); int btrfs_resume_balance_async(struct btrfs_fs_info *fs_info); -- cgit From 010a47bde94201d9abdab7ff04bedc17b6e8c357 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 20 Mar 2018 19:51:04 +0100 Subject: btrfs: add proper safety check before resuming dev-replace The device replace is paused by unmount or read only remount, and resumed on next mount or write remount. The exclusive status should be checked properly as it's a global invariant and we must not allow 2 operations run. In this case, the balance can be also paused and resumed under same conditions. It's always checked first so dev-replace could see the EXCL_OP already taken, BUT, the ioctl would never let start both at the same time. Replace the WARN_ON with message and return 0, indicating no error as this is purely theoretical and the user will be informed. Resolving that manually should be possible by waiting for the other operation to finish or cancel the paused state. Reviewed-by: Anand Jain Reviewed-by: Nikolay Borisov Signed-off-by: David Sterba --- fs/btrfs/dev-replace.c | 12 +++++++++++- 1 file changed, 11 insertions(+), 1 deletion(-) (limited to 'fs/btrfs/dev-replace.c') diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c index 8531b5dae777..9fe7be7fdbef 100644 --- a/fs/btrfs/dev-replace.c +++ b/fs/btrfs/dev-replace.c @@ -896,7 +896,17 @@ int btrfs_resume_dev_replace_async(struct btrfs_fs_info *fs_info) } btrfs_dev_replace_write_unlock(dev_replace); - WARN_ON(test_and_set_bit(BTRFS_FS_EXCL_OP, &fs_info->flags)); + /* + * This could collide with a paused balance, but the exclusive op logic + * should never allow both to start and pause. We don't want to allow + * dev-replace to start anyway. + */ + if (test_and_set_bit(BTRFS_FS_EXCL_OP, &fs_info->flags)) { + btrfs_info(fs_info, + "cannot resume dev-replace, other exclusive operation running"); + return 0; + } + task = kthread_run(btrfs_dev_replace_kthread, fs_info, "btrfs-devrepl"); return PTR_ERR_OR_ZERO(task); } -- cgit From a0fecc23718aa9ef020b8c86173a0b783ed37dcf Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 20 Mar 2018 23:44:50 +0100 Subject: btrfs: remove wrong use of volume_mutex from btrfs_dev_replace_start The volume mutex does not protect against anything in this case, the comment about scrub is right but not related to locking and looks confusing. The comment in btrfs_find_device_missing_or_by_path is wrong and confusing too. The device_list_mutex is not held here to protect device lookup, but in this case device replace cannot run in parallel with device removal (due to exclusive op protection), so we don't need further locking here. Reviewed-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/dev-replace.c | 7 +------ fs/btrfs/volumes.c | 4 ---- 2 files changed, 1 insertion(+), 10 deletions(-) (limited to 'fs/btrfs/dev-replace.c') diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c index 9fe7be7fdbef..d097701d494d 100644 --- a/fs/btrfs/dev-replace.c +++ b/fs/btrfs/dev-replace.c @@ -414,18 +414,13 @@ int btrfs_dev_replace_start(struct btrfs_fs_info *fs_info, struct btrfs_device *tgt_device = NULL; struct btrfs_device *src_device = NULL; - /* the disk copy procedure reuses the scrub code */ - mutex_lock(&fs_info->volume_mutex); ret = btrfs_find_device_by_devspec(fs_info, srcdevid, srcdev_name, &src_device); - if (ret) { - mutex_unlock(&fs_info->volume_mutex); + if (ret) return ret; - } ret = btrfs_init_dev_replace_tgtdev(fs_info, tgtdev_name, src_device, &tgt_device); - mutex_unlock(&fs_info->volume_mutex); if (ret) return ret; diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 07706c0a5781..9e5d27dd00b7 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -2218,10 +2218,6 @@ int btrfs_find_device_missing_or_by_path(struct btrfs_fs_info *fs_info, struct btrfs_device *tmp; devices = &fs_info->fs_devices->devices; - /* - * It is safe to read the devices since the volume_mutex - * is held by the caller. - */ list_for_each_entry(tmp, devices, dev_list) { if (test_bit(BTRFS_DEV_STATE_IN_FS_METADATA, &tmp->dev_state) && !tmp->bdev) { -- cgit From 82b3e53b8da19b25ef36b68316374df47f8fa268 Mon Sep 17 00:00:00 2001 From: Nikolay Borisov Date: Mon, 23 Apr 2018 10:54:13 +0300 Subject: btrfs: Remove delayed_iput parameter of btrfs_start_delalloc_roots This parameter was introduced alongside the function in eb73c1b7cea7 ("Btrfs: introduce per-subvolume delalloc inode list") to avoid deadlocks since this function was used in the transaction commit path. However, commit 8d875f95da43 ("btrfs: disable strict file flushes for renames and truncates") removed that usage, rendering the parameter obsolete. Signed-off-by: Nikolay Borisov Reviewed-by: Qu Wenruo Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 3 +-- fs/btrfs/dev-replace.c | 2 +- fs/btrfs/extent-tree.c | 4 ++-- fs/btrfs/inode.c | 5 ++--- fs/btrfs/ioctl.c | 2 +- 5 files changed, 7 insertions(+), 9 deletions(-) (limited to 'fs/btrfs/dev-replace.c') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 9cc90f407cae..0d15d98a964a 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -3205,8 +3205,7 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans, u32 min_type); int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput); -int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, int delay_iput, - int nr); +int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, int nr); int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end, unsigned int extra_bits, struct extent_state **cached_state, int dedupe); diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c index d097701d494d..12f703e127dd 100644 --- a/fs/btrfs/dev-replace.c +++ b/fs/btrfs/dev-replace.c @@ -594,7 +594,7 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info, * flush all outstanding I/O and inode extent mappings before the * copy operation is declared as being finished */ - ret = btrfs_start_delalloc_roots(fs_info, 0, -1); + ret = btrfs_start_delalloc_roots(fs_info, -1); if (ret) { mutex_unlock(&dev_replace->lock_finishing_cancel_unmount); return ret; diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 888a47894c30..38dd98bc50d7 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -4343,7 +4343,7 @@ commit_trans: need_commit--; if (need_commit > 0) { - btrfs_start_delalloc_roots(fs_info, 0, -1); + btrfs_start_delalloc_roots(fs_info, -1); btrfs_wait_ordered_roots(fs_info, U64_MAX, 0, (u64)-1); } @@ -4796,7 +4796,7 @@ static void btrfs_writeback_inodes_sb_nr(struct btrfs_fs_info *fs_info, * the filesystem is readonly(all dirty pages are written to * the disk). */ - btrfs_start_delalloc_roots(fs_info, 0, nr_items); + btrfs_start_delalloc_roots(fs_info, nr_items); if (!current->journal_info) btrfs_wait_ordered_roots(fs_info, nr_items, 0, (u64)-1); } diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index ce2f9288df3e..c2bb7c8e1fda 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -10269,8 +10269,7 @@ int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput) return ret; } -int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, int delay_iput, - int nr) +int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, int nr) { struct btrfs_root *root; struct list_head splice; @@ -10293,7 +10292,7 @@ int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, int delay_iput, &fs_info->delalloc_roots); spin_unlock(&fs_info->delalloc_root_lock); - ret = __start_delalloc_inodes(root, delay_iput, nr); + ret = __start_delalloc_inodes(root, 0, nr); btrfs_put_fs_root(root); if (ret < 0) goto out; diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 397f026b1e4c..67db5f6b0476 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -5307,7 +5307,7 @@ long btrfs_ioctl(struct file *file, unsigned int case BTRFS_IOC_SYNC: { int ret; - ret = btrfs_start_delalloc_roots(fs_info, 0, -1); + ret = btrfs_start_delalloc_roots(fs_info, -1); if (ret) return ret; ret = btrfs_sync_fs(inode->i_sb, 1); -- cgit From 093258e6ebaf178bb25da514f0d1f744968cc900 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Mon, 26 Feb 2018 16:15:17 +0100 Subject: btrfs: replace waitqueue_actvie with cond_wake_up Use the wrappers and reduce the amount of low-level details about the waitqueue management. Reviewed-by: Nikolay Borisov Signed-off-by: David Sterba --- fs/btrfs/compression.c | 7 +------ fs/btrfs/delayed-inode.c | 9 +++------ fs/btrfs/dev-replace.c | 10 ++++------ fs/btrfs/extent-tree.c | 7 +------ fs/btrfs/inode.c | 9 +++------ fs/btrfs/locking.c | 34 +++++++++++----------------------- fs/btrfs/ordered-data.c | 14 ++++---------- fs/btrfs/transaction.c | 7 +------ fs/btrfs/tree-log.c | 34 ++++++++++++---------------------- 9 files changed, 40 insertions(+), 91 deletions(-) (limited to 'fs/btrfs/dev-replace.c') diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index 1061575a7d25..d3e447b45bf7 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -990,12 +990,7 @@ static void __free_workspace(int type, struct list_head *workspace, btrfs_compress_op[idx]->free_workspace(workspace); atomic_dec(total_ws); wake: - /* - * Make sure counter is updated before we wake up waiters. - */ - smp_mb(); - if (waitqueue_active(ws_wait)) - wake_up(ws_wait); + cond_wake_up(ws_wait); } static void free_workspace(int type, struct list_head *ws) diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c index a8d492dbd3e7..fe6caa7e698b 100644 --- a/fs/btrfs/delayed-inode.c +++ b/fs/btrfs/delayed-inode.c @@ -460,13 +460,10 @@ static void finish_one_item(struct btrfs_delayed_root *delayed_root) { int seq = atomic_inc_return(&delayed_root->items_seq); - /* - * atomic_dec_return implies a barrier for waitqueue_active - */ + /* atomic_dec_return implies a barrier */ if ((atomic_dec_return(&delayed_root->items) < - BTRFS_DELAYED_BACKGROUND || seq % BTRFS_DELAYED_BATCH == 0) && - waitqueue_active(&delayed_root->wait)) - wake_up(&delayed_root->wait); + BTRFS_DELAYED_BACKGROUND || seq % BTRFS_DELAYED_BATCH == 0)) + cond_wake_up_nomb(&delayed_root->wait); } static void __btrfs_remove_delayed_item(struct btrfs_delayed_item *delayed_item) diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c index 12f703e127dd..89946285203d 100644 --- a/fs/btrfs/dev-replace.c +++ b/fs/btrfs/dev-replace.c @@ -1009,9 +1009,9 @@ void btrfs_dev_replace_clear_lock_blocking( ASSERT(atomic_read(&dev_replace->read_locks) > 0); ASSERT(atomic_read(&dev_replace->blocking_readers) > 0); read_lock(&dev_replace->lock); - if (atomic_dec_and_test(&dev_replace->blocking_readers) && - waitqueue_active(&dev_replace->read_lock_wq)) - wake_up(&dev_replace->read_lock_wq); + /* Barrier implied by atomic_dec_and_test */ + if (atomic_dec_and_test(&dev_replace->blocking_readers)) + cond_wake_up_nomb(&dev_replace->read_lock_wq); } void btrfs_bio_counter_inc_noblocked(struct btrfs_fs_info *fs_info) @@ -1022,9 +1022,7 @@ void btrfs_bio_counter_inc_noblocked(struct btrfs_fs_info *fs_info) void btrfs_bio_counter_sub(struct btrfs_fs_info *fs_info, s64 amount) { percpu_counter_sub(&fs_info->bio_counter, amount); - - if (waitqueue_active(&fs_info->replace_wait)) - wake_up(&fs_info->replace_wait); + cond_wake_up_nomb(&fs_info->replace_wait); } void btrfs_bio_counter_inc_blocked(struct btrfs_fs_info *fs_info) diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 5be54cedb56f..fa2ed14532c1 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -11081,12 +11081,7 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range) void btrfs_end_write_no_snapshotting(struct btrfs_root *root) { percpu_counter_dec(&root->subv_writers->counter); - /* - * Make sure counter is updated before we wake up waiters. - */ - smp_mb(); - if (waitqueue_active(&root->subv_writers->wait)) - wake_up(&root->subv_writers->wait); + cond_wake_up(&root->subv_writers->wait); } int btrfs_start_write_no_snapshotting(struct btrfs_root *root) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 6757fe136177..563e63fa2fce 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -1156,13 +1156,10 @@ static noinline void async_cow_submit(struct btrfs_work *work) nr_pages = (async_cow->end - async_cow->start + PAGE_SIZE) >> PAGE_SHIFT; - /* - * atomic_sub_return implies a barrier for waitqueue_active - */ + /* atomic_sub_return implies a barrier */ if (atomic_sub_return(nr_pages, &fs_info->async_delalloc_pages) < - 5 * SZ_1M && - waitqueue_active(&fs_info->async_submit_wait)) - wake_up(&fs_info->async_submit_wait); + 5 * SZ_1M) + cond_wake_up_nomb(&fs_info->async_submit_wait); if (async_cow->inode) submit_compressed_extents(async_cow->inode, async_cow); diff --git a/fs/btrfs/locking.c b/fs/btrfs/locking.c index e4faefac9d16..1da768e5ef75 100644 --- a/fs/btrfs/locking.c +++ b/fs/btrfs/locking.c @@ -66,22 +66,16 @@ void btrfs_clear_lock_blocking_rw(struct extent_buffer *eb, int rw) write_lock(&eb->lock); WARN_ON(atomic_read(&eb->spinning_writers)); atomic_inc(&eb->spinning_writers); - /* - * atomic_dec_and_test implies a barrier for waitqueue_active - */ - if (atomic_dec_and_test(&eb->blocking_writers) && - waitqueue_active(&eb->write_lock_wq)) - wake_up(&eb->write_lock_wq); + /* atomic_dec_and_test implies a barrier */ + if (atomic_dec_and_test(&eb->blocking_writers)) + cond_wake_up_nomb(&eb->write_lock_wq); } else if (rw == BTRFS_READ_LOCK_BLOCKING) { BUG_ON(atomic_read(&eb->blocking_readers) == 0); read_lock(&eb->lock); atomic_inc(&eb->spinning_readers); - /* - * atomic_dec_and_test implies a barrier for waitqueue_active - */ - if (atomic_dec_and_test(&eb->blocking_readers) && - waitqueue_active(&eb->read_lock_wq)) - wake_up(&eb->read_lock_wq); + /* atomic_dec_and_test implies a barrier */ + if (atomic_dec_and_test(&eb->blocking_readers)) + cond_wake_up_nomb(&eb->read_lock_wq); } } @@ -221,12 +215,9 @@ void btrfs_tree_read_unlock_blocking(struct extent_buffer *eb) } btrfs_assert_tree_read_locked(eb); WARN_ON(atomic_read(&eb->blocking_readers) == 0); - /* - * atomic_dec_and_test implies a barrier for waitqueue_active - */ - if (atomic_dec_and_test(&eb->blocking_readers) && - waitqueue_active(&eb->read_lock_wq)) - wake_up(&eb->read_lock_wq); + /* atomic_dec_and_test implies a barrier */ + if (atomic_dec_and_test(&eb->blocking_readers)) + cond_wake_up_nomb(&eb->read_lock_wq); atomic_dec(&eb->read_locks); } @@ -275,12 +266,9 @@ void btrfs_tree_unlock(struct extent_buffer *eb) if (blockers) { WARN_ON(atomic_read(&eb->spinning_writers)); atomic_dec(&eb->blocking_writers); - /* - * Make sure counter is updated before we wake up waiters. - */ + /* Use the lighter barrier after atomic */ smp_mb__after_atomic(); - if (waitqueue_active(&eb->write_lock_wq)) - wake_up(&eb->write_lock_wq); + cond_wake_up_nomb(&eb->write_lock_wq); } else { WARN_ON(atomic_read(&eb->spinning_writers) != 1); atomic_dec(&eb->spinning_writers); diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c index 6db8bb2f2c28..2e1a1694a33d 100644 --- a/fs/btrfs/ordered-data.c +++ b/fs/btrfs/ordered-data.c @@ -343,11 +343,8 @@ int btrfs_dec_test_first_ordered_pending(struct inode *inode, if (entry->bytes_left == 0) { ret = test_and_set_bit(BTRFS_ORDERED_IO_DONE, &entry->flags); - /* - * Implicit memory barrier after test_and_set_bit - */ - if (waitqueue_active(&entry->wait)) - wake_up(&entry->wait); + /* test_and_set_bit implies a barrier */ + cond_wake_up_nomb(&entry->wait); } else { ret = 1; } @@ -410,11 +407,8 @@ have_entry: if (entry->bytes_left == 0) { ret = test_and_set_bit(BTRFS_ORDERED_IO_DONE, &entry->flags); - /* - * Implicit memory barrier after test_and_set_bit - */ - if (waitqueue_active(&entry->wait)) - wake_up(&entry->wait); + /* test_and_set_bit implies a barrier */ + cond_wake_up_nomb(&entry->wait); } else { ret = 1; } diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index c944b4769e3c..ff841abb756e 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -877,12 +877,7 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans, atomic_dec(&cur_trans->num_writers); extwriter_counter_dec(cur_trans, trans->type); - /* - * Make sure counter is updated before we wake up waiters. - */ - smp_mb(); - if (waitqueue_active(&cur_trans->writer_wait)) - wake_up(&cur_trans->writer_wait); + cond_wake_up(&cur_trans->writer_wait); btrfs_put_transaction(cur_trans); if (current->journal_info == trans) diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index 2009cea65d89..f8220ec02036 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -222,11 +222,8 @@ int btrfs_pin_log_trans(struct btrfs_root *root) void btrfs_end_log_trans(struct btrfs_root *root) { if (atomic_dec_and_test(&root->log_writers)) { - /* - * Implicit memory barrier after atomic_dec_and_test - */ - if (waitqueue_active(&root->log_writer_wait)) - wake_up(&root->log_writer_wait); + /* atomic_dec_and_test implies a barrier */ + cond_wake_up_nomb(&root->log_writer_wait); } } @@ -2988,11 +2985,8 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans, mutex_lock(&log_root_tree->log_mutex); if (atomic_dec_and_test(&log_root_tree->log_writers)) { - /* - * Implicit memory barrier after atomic_dec_and_test - */ - if (waitqueue_active(&log_root_tree->log_writer_wait)) - wake_up(&log_root_tree->log_writer_wait); + /* atomic_dec_and_test implies a barrier */ + cond_wake_up_nomb(&log_root_tree->log_writer_wait); } if (ret) { @@ -3116,13 +3110,11 @@ out_wake_log_root: mutex_unlock(&log_root_tree->log_mutex); /* - * The barrier before waitqueue_active is needed so all the updates - * above are seen by the woken threads. It might not be necessary, but - * proving that seems to be hard. + * The barrier before waitqueue_active (in cond_wake_up) is needed so + * all the updates above are seen by the woken threads. It might not be + * necessary, but proving that seems to be hard. */ - smp_mb(); - if (waitqueue_active(&log_root_tree->log_commit_wait[index2])) - wake_up(&log_root_tree->log_commit_wait[index2]); + cond_wake_up(&log_root_tree->log_commit_wait[index2]); out: mutex_lock(&root->log_mutex); btrfs_remove_all_log_ctxs(root, index1, ret); @@ -3131,13 +3123,11 @@ out: mutex_unlock(&root->log_mutex); /* - * The barrier before waitqueue_active is needed so all the updates - * above are seen by the woken threads. It might not be necessary, but - * proving that seems to be hard. + * The barrier before waitqueue_active (in cond_wake_up) is needed so + * all the updates above are seen by the woken threads. It might not be + * necessary, but proving that seems to be hard. */ - smp_mb(); - if (waitqueue_active(&root->log_commit_wait[index1])) - wake_up(&root->log_commit_wait[index1]); + cond_wake_up(&root->log_commit_wait[index1]); return ret; } -- cgit From b25e59e2b2ee394a2a7e981b99cbd92ef6c8be85 Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Thu, 12 Apr 2018 10:29:36 +0800 Subject: btrfs: drop uuid_mutex in btrfs_dev_replace_finishing btrfs_dev_replace_finishing updates devices (soruce and target) which are within the btrfs_fs_devices::devices or withint the cloned seed devices (btrfs_fs_devices::seed::devices), so we don't need the global uuid_mutex. The device replace context is also locked by its own locks. Signed-off-by: Anand Jain Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/dev-replace.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'fs/btrfs/dev-replace.c') diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c index 89946285203d..e2ba0419297a 100644 --- a/fs/btrfs/dev-replace.c +++ b/fs/btrfs/dev-replace.c @@ -609,7 +609,6 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info, ret = btrfs_commit_transaction(trans); WARN_ON(ret); - mutex_lock(&uuid_mutex); /* keep away write_all_supers() during the finishing procedure */ mutex_lock(&fs_info->fs_devices->device_list_mutex); mutex_lock(&fs_info->chunk_mutex); @@ -636,7 +635,6 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info, btrfs_dev_replace_write_unlock(dev_replace); mutex_unlock(&fs_info->chunk_mutex); mutex_unlock(&fs_info->fs_devices->device_list_mutex); - mutex_unlock(&uuid_mutex); btrfs_rm_dev_replace_blocked(fs_info); if (tgt_device) btrfs_destroy_dev_replace_tgtdev(fs_info, tgt_device); @@ -687,7 +685,6 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info, */ mutex_unlock(&fs_info->chunk_mutex); mutex_unlock(&fs_info->fs_devices->device_list_mutex); - mutex_unlock(&uuid_mutex); /* replace the sysfs entry */ btrfs_sysfs_rm_device_link(fs_info->fs_devices, src_device); -- cgit