From 3b6571c180da85e43550c608e954ab7b2a31d954 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 27 May 2016 13:03:04 -0400 Subject: Btrfs: don't BUG_ON() in btrfs_orphan_add This is just a screwup for developers, so change it to an ASSERT() so developers notice when things go wrong and deal with the error appropriately if ASSERT() isn't enabled. Thanks, Signed-off-by: Josef Bacik Reviewed-by: Mark Fasheh Signed-off-by: David Sterba --- fs/btrfs/inode.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index e5558d9e396e..bb62418b8023 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -3271,7 +3271,16 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode) /* grab metadata reservation from transaction handle */ if (reserve) { ret = btrfs_orphan_reserve_metadata(trans, inode); - BUG_ON(ret); /* -ENOSPC in reservation; Logic error? JDM */ + ASSERT(!ret); + if (ret) { + atomic_dec(&root->orphan_inodes); + clear_bit(BTRFS_INODE_ORPHAN_META_RESERVED, + &BTRFS_I(inode)->runtime_flags); + if (insert) + clear_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, + &BTRFS_I(inode)->runtime_flags); + return ret; + } } /* insert an orphan item to track this unlinked/truncated file */ -- cgit v1.2.3-70-g09d2 From 31b9655f439a26856edca0f3f8daa368a61f16d5 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Mon, 11 Apr 2016 17:37:40 -0400 Subject: Btrfs: track transid for delayed ref flushing Using the offwakecputime bpf script I noticed most of our time was spent waiting on the delayed ref throttling. This is what is supposed to happen, but sometimes the transaction can commit and then we're waiting for throttling that doesn't matter anymore. So change this stuff to be a little smarter by tracking the transid we were in when we initiated the throttling. If the transaction we get is different then we can just bail out. This resulted in a 50% speedup in my fs_mark test, and reduced the amount of time spent throttling by 60 seconds over the entire run (which is about 30 minutes). Thanks, Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/ctree.h | 2 +- fs/btrfs/extent-tree.c | 15 ++++++++++++--- fs/btrfs/inode.c | 1 + fs/btrfs/transaction.c | 3 ++- 4 files changed, 16 insertions(+), 5 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 101c3cfd3f7c..4274a7bfdaed 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -2518,7 +2518,7 @@ void btrfs_put_block_group(struct btrfs_block_group_cache *cache); int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans, struct btrfs_root *root, unsigned long count); int btrfs_async_run_delayed_refs(struct btrfs_root *root, - unsigned long count, int wait); + unsigned long count, u64 transid, int wait); int btrfs_lookup_data_extent(struct btrfs_root *root, u64 start, u64 len); int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 bytenr, diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 29e5d000bbee..ecfa52002363 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -2835,6 +2835,7 @@ int btrfs_should_throttle_delayed_refs(struct btrfs_trans_handle *trans, struct async_delayed_refs { struct btrfs_root *root; + u64 transid; int count; int error; int sync; @@ -2850,9 +2851,16 @@ static void delayed_ref_async_start(struct btrfs_work *work) async = container_of(work, struct async_delayed_refs, work); - trans = btrfs_join_transaction(async->root); + trans = btrfs_attach_transaction(async->root); if (IS_ERR(trans)) { - async->error = PTR_ERR(trans); + if (PTR_ERR(trans) != -ENOENT) + async->error = PTR_ERR(trans); + goto done; + } + + /* Don't bother flushing if we got into a different transaction */ + if (trans->transid != async->transid) { + btrfs_end_transaction(trans, async->root); goto done; } @@ -2876,7 +2884,7 @@ done: } int btrfs_async_run_delayed_refs(struct btrfs_root *root, - unsigned long count, int wait) + unsigned long count, u64 transid, int wait) { struct async_delayed_refs *async; int ret; @@ -2888,6 +2896,7 @@ int btrfs_async_run_delayed_refs(struct btrfs_root *root, async->root = root->fs_info->tree_root; async->count = count; async->error = 0; + async->transid = transid; if (wait) async->sync = 1; else diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index bb62418b8023..78582e339f33 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -4558,6 +4558,7 @@ delete: BUG_ON(ret); if (btrfs_should_throttle_delayed_refs(trans, root)) btrfs_async_run_delayed_refs(root, + trans->transid, trans->delayed_ref_updates * 2, 0); if (be_nice) { if (truncate_space_check(trans, root, diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index 765845742fde..948aa186b353 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -818,6 +818,7 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans, { struct btrfs_transaction *cur_trans = trans->transaction; struct btrfs_fs_info *info = root->fs_info; + u64 transid = trans->transid; unsigned long cur = trans->delayed_ref_updates; int lock = (trans->type != TRANS_JOIN_NOLOCK); int err = 0; @@ -905,7 +906,7 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans, kmem_cache_free(btrfs_trans_handle_cachep, trans); if (must_run_delayed_refs) { - btrfs_async_run_delayed_refs(root, cur, + btrfs_async_run_delayed_refs(root, cur, transid, must_run_delayed_refs == 1); } return err; -- cgit v1.2.3-70-g09d2 From 02dbfc99b424dde3cf0a492ed3bec4f222441754 Mon Sep 17 00:00:00 2001 From: Omar Sandoval Date: Fri, 20 May 2016 13:50:33 -0700 Subject: Btrfs: fix ->iterate_shared() by upgrading i_rwsem for delayed nodes Commit fe742fd4f90f ("Revert "btrfs: switch to ->iterate_shared()"") backed out the conversion to ->iterate_shared() for Btrfs because the delayed inode handling in btrfs_real_readdir() is racy. However, we can still do readdir in parallel if there are no delayed nodes. This is a temporary fix which upgrades the shared inode lock to an exclusive lock only when we have delayed items until we come up with a more complete solution. While we're here, rename the btrfs_{get,put}_delayed_items functions to make it very clear that they're just for readdir. Tested with xfstests and by doing a parallel kernel build: while make tinyconfig && make -j4 && git clean dqfx; do : done along with a bunch of parallel finds in another shell: while true; do for ((i=0; i<4; i++)); do find . >/dev/null & done wait done Signed-off-by: Omar Sandoval Signed-off-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/delayed-inode.c | 27 ++++++++++++++++++++++----- fs/btrfs/delayed-inode.h | 10 ++++++---- fs/btrfs/inode.c | 10 ++++++---- 3 files changed, 34 insertions(+), 13 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c index 61561c2a3f96..d3aaabbfada0 100644 --- a/fs/btrfs/delayed-inode.c +++ b/fs/btrfs/delayed-inode.c @@ -1606,15 +1606,23 @@ int btrfs_inode_delayed_dir_index_count(struct inode *inode) return 0; } -void btrfs_get_delayed_items(struct inode *inode, struct list_head *ins_list, - struct list_head *del_list) +bool btrfs_readdir_get_delayed_items(struct inode *inode, + struct list_head *ins_list, + struct list_head *del_list) { struct btrfs_delayed_node *delayed_node; struct btrfs_delayed_item *item; delayed_node = btrfs_get_delayed_node(inode); if (!delayed_node) - return; + return false; + + /* + * We can only do one readdir with delayed items at a time because of + * item->readdir_list. + */ + inode_unlock_shared(inode); + inode_lock(inode); mutex_lock(&delayed_node->mutex); item = __btrfs_first_delayed_insertion_item(delayed_node); @@ -1641,10 +1649,13 @@ void btrfs_get_delayed_items(struct inode *inode, struct list_head *ins_list, * requeue or dequeue this delayed node. */ atomic_dec(&delayed_node->refs); + + return true; } -void btrfs_put_delayed_items(struct list_head *ins_list, - struct list_head *del_list) +void btrfs_readdir_put_delayed_items(struct inode *inode, + struct list_head *ins_list, + struct list_head *del_list) { struct btrfs_delayed_item *curr, *next; @@ -1659,6 +1670,12 @@ void btrfs_put_delayed_items(struct list_head *ins_list, if (atomic_dec_and_test(&curr->refs)) kfree(curr); } + + /* + * The VFS is going to do up_read(), so we need to downgrade back to a + * read lock. + */ + downgrade_write(&inode->i_rwsem); } int btrfs_should_delete_dir_index(struct list_head *del_list, diff --git a/fs/btrfs/delayed-inode.h b/fs/btrfs/delayed-inode.h index 0167853c84ae..2495b3d4075f 100644 --- a/fs/btrfs/delayed-inode.h +++ b/fs/btrfs/delayed-inode.h @@ -137,10 +137,12 @@ void btrfs_kill_all_delayed_nodes(struct btrfs_root *root); void btrfs_destroy_delayed_inodes(struct btrfs_root *root); /* Used for readdir() */ -void btrfs_get_delayed_items(struct inode *inode, struct list_head *ins_list, - struct list_head *del_list); -void btrfs_put_delayed_items(struct list_head *ins_list, - struct list_head *del_list); +bool btrfs_readdir_get_delayed_items(struct inode *inode, + struct list_head *ins_list, + struct list_head *del_list); +void btrfs_readdir_put_delayed_items(struct inode *inode, + struct list_head *ins_list, + struct list_head *del_list); int btrfs_should_delete_dir_index(struct list_head *del_list, u64 index); int btrfs_readdir_delayed_dir_index(struct dir_context *ctx, diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index d2be95cfb6d1..969a25c5abcb 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -5757,6 +5757,7 @@ static int btrfs_real_readdir(struct file *file, struct dir_context *ctx) int name_len; int is_curr = 0; /* ctx->pos points to the current index? */ bool emitted; + bool put = false; /* FIXME, use a real flag for deciding about the key type */ if (root->fs_info->tree_root == root) @@ -5774,7 +5775,8 @@ static int btrfs_real_readdir(struct file *file, struct dir_context *ctx) if (key_type == BTRFS_DIR_INDEX_KEY) { INIT_LIST_HEAD(&ins_list); INIT_LIST_HEAD(&del_list); - btrfs_get_delayed_items(inode, &ins_list, &del_list); + put = btrfs_readdir_get_delayed_items(inode, &ins_list, + &del_list); } key.type = key_type; @@ -5921,8 +5923,8 @@ next: nopos: ret = 0; err: - if (key_type == BTRFS_DIR_INDEX_KEY) - btrfs_put_delayed_items(&ins_list, &del_list); + if (put) + btrfs_readdir_put_delayed_items(inode, &ins_list, &del_list); btrfs_free_path(path); return ret; } @@ -10534,7 +10536,7 @@ static const struct inode_operations btrfs_dir_ro_inode_operations = { static const struct file_operations btrfs_dir_file_operations = { .llseek = generic_file_llseek, .read = generic_read_dir, - .iterate = btrfs_real_readdir, + .iterate_shared = btrfs_real_readdir, .unlocked_ioctl = btrfs_ioctl, #ifdef CONFIG_COMPAT .compat_ioctl = btrfs_compat_ioctl, -- cgit v1.2.3-70-g09d2