On Wed, Oct 24, 2018 at 10:13:03AM +0100, fdman...@kernel.org wrote: > From: Filipe Manana <fdman...@suse.com> > > When we are writing out a free space cache, during the transaction commit > phase, we can end up in a deadlock which results in a stack trace like the > following: > > schedule+0x28/0x80 > btrfs_tree_read_lock+0x8e/0x120 [btrfs] > ? finish_wait+0x80/0x80 > btrfs_read_lock_root_node+0x2f/0x40 [btrfs] > btrfs_search_slot+0xf6/0x9f0 [btrfs] > ? evict_refill_and_join+0xd0/0xd0 [btrfs] > ? inode_insert5+0x119/0x190 > btrfs_lookup_inode+0x3a/0xc0 [btrfs] > ? kmem_cache_alloc+0x166/0x1d0 > btrfs_iget+0x113/0x690 [btrfs] > __lookup_free_space_inode+0xd8/0x150 [btrfs] > lookup_free_space_inode+0x5b/0xb0 [btrfs] > load_free_space_cache+0x7c/0x170 [btrfs] > ? cache_block_group+0x72/0x3b0 [btrfs] > cache_block_group+0x1b3/0x3b0 [btrfs] > ? finish_wait+0x80/0x80 > find_free_extent+0x799/0x1010 [btrfs] > btrfs_reserve_extent+0x9b/0x180 [btrfs] > btrfs_alloc_tree_block+0x1b3/0x4f0 [btrfs] > __btrfs_cow_block+0x11d/0x500 [btrfs] > btrfs_cow_block+0xdc/0x180 [btrfs] > btrfs_search_slot+0x3bd/0x9f0 [btrfs] > btrfs_lookup_inode+0x3a/0xc0 [btrfs] > ? kmem_cache_alloc+0x166/0x1d0 > btrfs_update_inode_item+0x46/0x100 [btrfs] > cache_save_setup+0xe4/0x3a0 [btrfs] > btrfs_start_dirty_block_groups+0x1be/0x480 [btrfs] > btrfs_commit_transaction+0xcb/0x8b0 [btrfs] > > At cache_save_setup() we need to update the inode item of a block group's > cache which is located in the tree root (fs_info->tree_root), which means > that it may result in COWing a leaf from that tree. If that happens we > need to find a free metadata extent and while looking for one, if we find > a block group which was not cached yet we attempt to load its cache by > calling cache_block_group(). However this function will try to load the > inode of the free space cache, which requires finding the matching inode > item in the tree root - if that inode item is located in the same leaf as > the inode item of the space cache we are updating at cache_save_setup(), > we end up in a deadlock, since we try to obtain a read lock on the same > extent buffer that we previously write locked. > > So fix this by using the tree root's commit root when searching for a > block group's free space cache inode item when we are attempting to load > a free space cache. This is safe since block groups once loaded stay in > memory forever, as well as their caches, so after they are first loaded > we will never need to read their inode items again. For new block groups, > once they are created they get their ->cached field set to > BTRFS_CACHE_FINISHED meaning we will not need to read their inode item. > > Reported-by: Andrew Nelson <andrew.s.nel...@gmail.com> > Link: > https://lore.kernel.org/linux-btrfs/captelenq9x5kowuq+fa7h1r3nsjg8vyith8+ifjurc_duhh...@mail.gmail.com/ > Fixes: 9d66e233c704 ("Btrfs: load free space cache if it exists") > Tested-by: Andrew Nelson <andrew.s.nel...@gmail.com> > Signed-off-by: Filipe Manana <fdman...@suse.com> > --- >
Now my goal is to see how many times I can get you to redo this thing. Why not instead just do if (btrfs_is_free_space_inode(inode)) path->search_commit_root = 1; in read_locked_inode? That would be cleaner. If we don't want to do that for the inode cache (I'm not sure if it's ok or not) we could just do if (root == fs_info->tree_root) instead. Thanks, Josef
