From d2c609b834d62f1e91f1635a27dca29f7806d3d6 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Mon, 15 Aug 2016 12:10:33 -0400 Subject: btrfs: properly track when rescan worker is running The qgroup_flags field is overloaded such that it reflects the on-disk status of qgroups and the runtime state. The BTRFS_QGROUP_STATUS_FLAG_RESCAN flag is used to indicate that a rescan operation is in progress, but if the file system is unmounted while a rescan is running, the rescan operation is paused. If the file system is then mounted read-only, the flag will still be present but the rescan operation will not have been resumed. When we go to umount, btrfs_qgroup_wait_for_completion will see the flag and interpret it to mean that the rescan worker is still running and will wait for a completion that will never come. This patch uses a separate flag to indicate when the worker is running. The locking and state surrounding the qgroup rescan worker needs a lot of attention beyond this patch but this is enough to avoid a hung umount. Cc: # v4.4+ Signed-off-by; Jeff Mahoney Reviewed-by: Qu Wenruo Signed-off-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/qgroup.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) (limited to 'fs/btrfs/qgroup.c') diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c index 93ee1c18ef9d..3fe295e2c84a 100644 --- a/fs/btrfs/qgroup.c +++ b/fs/btrfs/qgroup.c @@ -2303,6 +2303,10 @@ static void btrfs_qgroup_rescan_worker(struct btrfs_work *work) int err = -ENOMEM; int ret = 0; + mutex_lock(&fs_info->qgroup_rescan_lock); + fs_info->qgroup_rescan_running = true; + mutex_unlock(&fs_info->qgroup_rescan_lock); + path = btrfs_alloc_path(); if (!path) goto out; @@ -2369,6 +2373,9 @@ out: } done: + mutex_lock(&fs_info->qgroup_rescan_lock); + fs_info->qgroup_rescan_running = false; + mutex_unlock(&fs_info->qgroup_rescan_lock); complete_all(&fs_info->qgroup_rescan_completion); } @@ -2494,7 +2501,7 @@ int btrfs_qgroup_wait_for_completion(struct btrfs_fs_info *fs_info) mutex_lock(&fs_info->qgroup_rescan_lock); spin_lock(&fs_info->qgroup_lock); - running = fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN; + running = fs_info->qgroup_rescan_running; spin_unlock(&fs_info->qgroup_lock); mutex_unlock(&fs_info->qgroup_rescan_lock); -- cgit v1.2.3-70-g09d2 From d06f23d6a947c9abae41dc46be69a56baf36f436 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Mon, 8 Aug 2016 22:08:06 -0400 Subject: btrfs: waiting on qgroup rescan should not always be interruptible We wait on qgroup rescan completion in three places: file system shutdown, the quota disable ioctl, and the rescan wait ioctl. If the user sends a signal while we're waiting, we continue happily along. This is expected behavior for the rescan wait ioctl. It's racy in the shutdown path but mostly works due to other unrelated synchronization points. In the quota disable path, it Oopses the kernel pretty much immediately. Cc: # v4.4+ Signed-off-by: Jeff Mahoney Reviewed-by: David Sterba Signed-off-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 2 +- fs/btrfs/ioctl.c | 2 +- fs/btrfs/qgroup.c | 12 +++++++++--- fs/btrfs/qgroup.h | 3 ++- 4 files changed, 13 insertions(+), 6 deletions(-) (limited to 'fs/btrfs/qgroup.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index f5b2a7fb4575..7857f64e1cae 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -3864,7 +3864,7 @@ void close_ctree(struct btrfs_root *root) smp_mb(); /* wait for the qgroup rescan worker to stop */ - btrfs_qgroup_wait_for_completion(fs_info); + btrfs_qgroup_wait_for_completion(fs_info, false); /* wait for the uuid_scan task to finish */ down(&fs_info->uuid_tree_rescan_sem); diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 14ed1e9e6bc8..b2a2da5893af 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -5084,7 +5084,7 @@ static long btrfs_ioctl_quota_rescan_wait(struct file *file, void __user *arg) if (!capable(CAP_SYS_ADMIN)) return -EPERM; - return btrfs_qgroup_wait_for_completion(root->fs_info); + return btrfs_qgroup_wait_for_completion(root->fs_info, true); } static long _btrfs_ioctl_set_received_subvol(struct file *file, diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c index 3fe295e2c84a..13eb6a7a4db1 100644 --- a/fs/btrfs/qgroup.c +++ b/fs/btrfs/qgroup.c @@ -995,7 +995,7 @@ int btrfs_quota_disable(struct btrfs_trans_handle *trans, goto out; fs_info->quota_enabled = 0; fs_info->pending_quota_state = 0; - btrfs_qgroup_wait_for_completion(fs_info); + btrfs_qgroup_wait_for_completion(fs_info, false); spin_lock(&fs_info->qgroup_lock); quota_root = fs_info->quota_root; fs_info->quota_root = NULL; @@ -2494,7 +2494,8 @@ btrfs_qgroup_rescan(struct btrfs_fs_info *fs_info) return 0; } -int btrfs_qgroup_wait_for_completion(struct btrfs_fs_info *fs_info) +int btrfs_qgroup_wait_for_completion(struct btrfs_fs_info *fs_info, + bool interruptible) { int running; int ret = 0; @@ -2505,9 +2506,14 @@ int btrfs_qgroup_wait_for_completion(struct btrfs_fs_info *fs_info) spin_unlock(&fs_info->qgroup_lock); mutex_unlock(&fs_info->qgroup_rescan_lock); - if (running) + if (!running) + return 0; + + if (interruptible) ret = wait_for_completion_interruptible( &fs_info->qgroup_rescan_completion); + else + wait_for_completion(&fs_info->qgroup_rescan_completion); return ret; } diff --git a/fs/btrfs/qgroup.h b/fs/btrfs/qgroup.h index 710887c06aaf..af3e5578cad7 100644 --- a/fs/btrfs/qgroup.h +++ b/fs/btrfs/qgroup.h @@ -46,7 +46,8 @@ int btrfs_quota_disable(struct btrfs_trans_handle *trans, struct btrfs_fs_info *fs_info); int btrfs_qgroup_rescan(struct btrfs_fs_info *fs_info); void btrfs_qgroup_rescan_resume(struct btrfs_fs_info *fs_info); -int btrfs_qgroup_wait_for_completion(struct btrfs_fs_info *fs_info); +int btrfs_qgroup_wait_for_completion(struct btrfs_fs_info *fs_info, + bool interruptible); int btrfs_add_qgroup_relation(struct btrfs_trans_handle *trans, struct btrfs_fs_info *fs_info, u64 src, u64 dst); int btrfs_del_qgroup_relation(struct btrfs_trans_handle *trans, -- cgit v1.2.3-70-g09d2 From cb93b52cc005ba0e470845b519c662e661d5113c Mon Sep 17 00:00:00 2001 From: Qu Wenruo Date: Mon, 15 Aug 2016 10:36:50 +0800 Subject: btrfs: qgroup: Refactor btrfs_qgroup_insert_dirty_extent() Refactor btrfs_qgroup_insert_dirty_extent() function, to two functions: 1. btrfs_qgroup_insert_dirty_extent_nolock() Almost the same with original code. For delayed_ref usage, which has delayed refs locked. Change the return value type to int, since caller never needs the pointer, but only needs to know if they need to free the allocated memory. 2. btrfs_qgroup_insert_dirty_extent() The more encapsulated version. Will do the delayed_refs lock, memory allocation, quota enabled check and other things. The original design is to keep exported functions to minimal, but since more btrfs hacks exposed, like replacing path in balance, we need to record dirty extents manually, so we have to add such functions. Also, add comment for both functions, to info developers how to keep qgroup correct when doing hacks. Cc: Mark Fasheh Signed-off-by: Qu Wenruo Reviewed-and-Tested-by: Goldwyn Rodrigues Signed-off-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/delayed-ref.c | 7 ++----- fs/btrfs/extent-tree.c | 37 +++++-------------------------------- fs/btrfs/qgroup.c | 41 +++++++++++++++++++++++++++++++++++------ fs/btrfs/qgroup.h | 33 +++++++++++++++++++++++++++++---- 4 files changed, 71 insertions(+), 47 deletions(-) (limited to 'fs/btrfs/qgroup.c') diff --git a/fs/btrfs/delayed-ref.c b/fs/btrfs/delayed-ref.c index d9ddcfc18c91..ac02e041464b 100644 --- a/fs/btrfs/delayed-ref.c +++ b/fs/btrfs/delayed-ref.c @@ -541,7 +541,6 @@ add_delayed_ref_head(struct btrfs_fs_info *fs_info, struct btrfs_delayed_ref_head *existing; struct btrfs_delayed_ref_head *head_ref = NULL; struct btrfs_delayed_ref_root *delayed_refs; - struct btrfs_qgroup_extent_record *qexisting; int count_mod = 1; int must_insert_reserved = 0; @@ -606,10 +605,8 @@ add_delayed_ref_head(struct btrfs_fs_info *fs_info, qrecord->num_bytes = num_bytes; qrecord->old_roots = NULL; - qexisting = btrfs_qgroup_insert_dirty_extent(fs_info, - delayed_refs, - qrecord); - if (qexisting) + if(btrfs_qgroup_insert_dirty_extent_nolock(fs_info, + delayed_refs, qrecord)) kfree(qrecord); } diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index d2cc16ff3e21..d10872748c43 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -8521,35 +8521,6 @@ reada: wc->reada_slot = slot; } -/* - * These may not be seen by the usual inc/dec ref code so we have to - * add them here. - */ -static int record_one_subtree_extent(struct btrfs_trans_handle *trans, - struct btrfs_root *root, u64 bytenr, - u64 num_bytes) -{ - struct btrfs_qgroup_extent_record *qrecord; - struct btrfs_delayed_ref_root *delayed_refs; - - qrecord = kmalloc(sizeof(*qrecord), GFP_NOFS); - if (!qrecord) - return -ENOMEM; - - qrecord->bytenr = bytenr; - qrecord->num_bytes = num_bytes; - qrecord->old_roots = NULL; - - delayed_refs = &trans->transaction->delayed_refs; - spin_lock(&delayed_refs->lock); - if (btrfs_qgroup_insert_dirty_extent(trans->fs_info, - delayed_refs, qrecord)) - kfree(qrecord); - spin_unlock(&delayed_refs->lock); - - return 0; -} - static int account_leaf_items(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct extent_buffer *eb) @@ -8583,7 +8554,8 @@ static int account_leaf_items(struct btrfs_trans_handle *trans, num_bytes = btrfs_file_extent_disk_num_bytes(eb, fi); - ret = record_one_subtree_extent(trans, root, bytenr, num_bytes); + ret = btrfs_qgroup_insert_dirty_extent(trans, root->fs_info, + bytenr, num_bytes, GFP_NOFS); if (ret) return ret; } @@ -8732,8 +8704,9 @@ walk_down: btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK); path->locks[level] = BTRFS_READ_LOCK_BLOCKING; - ret = record_one_subtree_extent(trans, root, child_bytenr, - root->nodesize); + ret = btrfs_qgroup_insert_dirty_extent(trans, + root->fs_info, child_bytenr, + root->nodesize, GFP_NOFS); if (ret) goto out; } diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c index 13eb6a7a4db1..8db2e29fdcf4 100644 --- a/fs/btrfs/qgroup.c +++ b/fs/btrfs/qgroup.c @@ -1453,10 +1453,9 @@ int btrfs_qgroup_prepare_account_extents(struct btrfs_trans_handle *trans, return ret; } -struct btrfs_qgroup_extent_record * -btrfs_qgroup_insert_dirty_extent(struct btrfs_fs_info *fs_info, - struct btrfs_delayed_ref_root *delayed_refs, - struct btrfs_qgroup_extent_record *record) +int btrfs_qgroup_insert_dirty_extent_nolock(struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_root *delayed_refs, + struct btrfs_qgroup_extent_record *record) { struct rb_node **p = &delayed_refs->dirty_extent_root.rb_node; struct rb_node *parent_node = NULL; @@ -1475,12 +1474,42 @@ btrfs_qgroup_insert_dirty_extent(struct btrfs_fs_info *fs_info, else if (bytenr > entry->bytenr) p = &(*p)->rb_right; else - return entry; + return 1; } rb_link_node(&record->node, parent_node, p); rb_insert_color(&record->node, &delayed_refs->dirty_extent_root); - return NULL; + return 0; +} + +int btrfs_qgroup_insert_dirty_extent(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, u64 bytenr, u64 num_bytes, + gfp_t gfp_flag) +{ + struct btrfs_qgroup_extent_record *record; + struct btrfs_delayed_ref_root *delayed_refs; + int ret; + + if (!fs_info->quota_enabled || bytenr == 0 || num_bytes == 0) + return 0; + if (WARN_ON(trans == NULL)) + return -EINVAL; + record = kmalloc(sizeof(*record), gfp_flag); + if (!record) + return -ENOMEM; + + delayed_refs = &trans->transaction->delayed_refs; + record->bytenr = bytenr; + record->num_bytes = num_bytes; + record->old_roots = NULL; + + spin_lock(&delayed_refs->lock); + ret = btrfs_qgroup_insert_dirty_extent_nolock(fs_info, delayed_refs, + record); + spin_unlock(&delayed_refs->lock); + if (ret > 0) + kfree(record); + return 0; } #define UPDATE_NEW 0 diff --git a/fs/btrfs/qgroup.h b/fs/btrfs/qgroup.h index af3e5578cad7..1bc64c864b62 100644 --- a/fs/btrfs/qgroup.h +++ b/fs/btrfs/qgroup.h @@ -64,10 +64,35 @@ void btrfs_free_qgroup_config(struct btrfs_fs_info *fs_info); struct btrfs_delayed_extent_op; int btrfs_qgroup_prepare_account_extents(struct btrfs_trans_handle *trans, struct btrfs_fs_info *fs_info); -struct btrfs_qgroup_extent_record * -btrfs_qgroup_insert_dirty_extent(struct btrfs_fs_info *fs_info, - struct btrfs_delayed_ref_root *delayed_refs, - struct btrfs_qgroup_extent_record *record); +/* + * Insert one dirty extent record into @delayed_refs, informing qgroup to + * account that extent at commit trans time. + * + * No lock version, caller must acquire delayed ref lock and allocate memory. + * + * Return 0 for success insert + * Return >0 for existing record, caller can free @record safely. + * Error is not possible + */ +int btrfs_qgroup_insert_dirty_extent_nolock( + struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_root *delayed_refs, + struct btrfs_qgroup_extent_record *record); + +/* + * Insert one dirty extent record into @delayed_refs, informing qgroup to + * account that extent at commit trans time. + * + * Better encapsulated version. + * + * Return 0 if the operation is done. + * Return <0 for error, like memory allocation failure or invalid parameter + * (NULL trans) + */ +int btrfs_qgroup_insert_dirty_extent(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, u64 bytenr, u64 num_bytes, + gfp_t gfp_flag); + int btrfs_qgroup_account_extent(struct btrfs_trans_handle *trans, struct btrfs_fs_info *fs_info, -- cgit v1.2.3-70-g09d2