btrfs: fix lockdep splat in add_missing_dev
commit fccc0007b8dc952c6bc0805cdf842eb8ea06a639 upstream. Nikolay reported a lockdep splat in generic/476 that I could reproduce with btrfs/187. ====================================================== WARNING: possible circular locking dependency detected 5.9.0-rc2+ #1 Tainted: G W ------------------------------------------------------ kswapd0/100 is trying to acquire lock: ffff9e8ef38b6268 (&delayed_node->mutex){+.+.}-{3:3}, at: __btrfs_release_delayed_node.part.0+0x3f/0x330 but task is already holding lock: ffffffffa9d74700 (fs_reclaim){+.+.}-{0:0}, at: __fs_reclaim_acquire+0x5/0x30 which lock already depends on the new lock. the existing dependency chain (in reverse order) is: -> #2 (fs_reclaim){+.+.}-{0:0}: fs_reclaim_acquire+0x65/0x80 slab_pre_alloc_hook.constprop.0+0x20/0x200 kmem_cache_alloc_trace+0x3a/0x1a0 btrfs_alloc_device+0x43/0x210 add_missing_dev+0x20/0x90 read_one_chunk+0x301/0x430 btrfs_read_sys_array+0x17b/0x1b0 open_ctree+0xa62/0x1896 btrfs_mount_root.cold+0x12/0xea legacy_get_tree+0x30/0x50 vfs_get_tree+0x28/0xc0 vfs_kern_mount.part.0+0x71/0xb0 btrfs_mount+0x10d/0x379 legacy_get_tree+0x30/0x50 vfs_get_tree+0x28/0xc0 path_mount+0x434/0xc00 __x64_sys_mount+0xe3/0x120 do_syscall_64+0x33/0x40 entry_SYSCALL_64_after_hwframe+0x44/0xa9 -> #1 (&fs_info->chunk_mutex){+.+.}-{3:3}: __mutex_lock+0x7e/0x7e0 btrfs_chunk_alloc+0x125/0x3a0 find_free_extent+0xdf6/0x1210 btrfs_reserve_extent+0xb3/0x1b0 btrfs_alloc_tree_block+0xb0/0x310 alloc_tree_block_no_bg_flush+0x4a/0x60 __btrfs_cow_block+0x11a/0x530 btrfs_cow_block+0x104/0x220 btrfs_search_slot+0x52e/0x9d0 btrfs_lookup_inode+0x2a/0x8f __btrfs_update_delayed_inode+0x80/0x240 btrfs_commit_inode_delayed_inode+0x119/0x120 btrfs_evict_inode+0x357/0x500 evict+0xcf/0x1f0 vfs_rmdir.part.0+0x149/0x160 do_rmdir+0x136/0x1a0 do_syscall_64+0x33/0x40 entry_SYSCALL_64_after_hwframe+0x44/0xa9 -> #0 (&delayed_node->mutex){+.+.}-{3:3}: __lock_acquire+0x1184/0x1fa0 lock_acquire+0xa4/0x3d0 __mutex_lock+0x7e/0x7e0 __btrfs_release_delayed_node.part.0+0x3f/0x330 btrfs_evict_inode+0x24c/0x500 evict+0xcf/0x1f0 dispose_list+0x48/0x70 prune_icache_sb+0x44/0x50 super_cache_scan+0x161/0x1e0 do_shrink_slab+0x178/0x3c0 shrink_slab+0x17c/0x290 shrink_node+0x2b2/0x6d0 balance_pgdat+0x30a/0x670 kswapd+0x213/0x4c0 kthread+0x138/0x160 ret_from_fork+0x1f/0x30 other info that might help us debug this: Chain exists of: &delayed_node->mutex --> &fs_info->chunk_mutex --> fs_reclaim Possible unsafe locking scenario: CPU0 CPU1 ---- ---- lock(fs_reclaim); lock(&fs_info->chunk_mutex); lock(fs_reclaim); lock(&delayed_node->mutex); *** DEADLOCK *** 3 locks held by kswapd0/100: #0: ffffffffa9d74700 (fs_reclaim){+.+.}-{0:0}, at: __fs_reclaim_acquire+0x5/0x30 #1: ffffffffa9d65c50 (shrinker_rwsem){++++}-{3:3}, at: shrink_slab+0x115/0x290 #2: ffff9e8e9da260e0 (&type->s_umount_key#48){++++}-{3:3}, at: super_cache_scan+0x38/0x1e0 stack backtrace: CPU: 1 PID: 100 Comm: kswapd0 Tainted: G W 5.9.0-rc2+ #1 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-2.fc32 04/01/2014 Call Trace: dump_stack+0x92/0xc8 check_noncircular+0x12d/0x150 __lock_acquire+0x1184/0x1fa0 lock_acquire+0xa4/0x3d0 ? __btrfs_release_delayed_node.part.0+0x3f/0x330 __mutex_lock+0x7e/0x7e0 ? __btrfs_release_delayed_node.part.0+0x3f/0x330 ? __btrfs_release_delayed_node.part.0+0x3f/0x330 ? lock_acquire+0xa4/0x3d0 ? btrfs_evict_inode+0x11e/0x500 ? find_held_lock+0x2b/0x80 __btrfs_release_delayed_node.part.0+0x3f/0x330 btrfs_evict_inode+0x24c/0x500 evict+0xcf/0x1f0 dispose_list+0x48/0x70 prune_icache_sb+0x44/0x50 super_cache_scan+0x161/0x1e0 do_shrink_slab+0x178/0x3c0 shrink_slab+0x17c/0x290 shrink_node+0x2b2/0x6d0 balance_pgdat+0x30a/0x670 kswapd+0x213/0x4c0 ? _raw_spin_unlock_irqrestore+0x46/0x60 ? add_wait_queue_exclusive+0x70/0x70 ? balance_pgdat+0x670/0x670 kthread+0x138/0x160 ? kthread_create_worker_on_cpu+0x40/0x40 ret_from_fork+0x1f/0x30 This is because we are holding the chunk_mutex when we call btrfs_alloc_device, which does a GFP_KERNEL allocation. We don't want to switch that to a GFP_NOFS lock because this is the only place where it matters. So instead use memalloc_nofs_save() around the allocation in order to avoid the lockdep splat. Reported-by: Nikolay Borisov <nborisov@suse.com> CC: stable@vger.kernel.org # 4.4+ Reviewed-by: Anand Jain <anand.jain@oracle.com> Signed-off-by: Josef Bacik <josef@toxicpanda.com> Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This commit is contained in:
Josef Bacik
Greg Kroah-Hartman
parent
a6fdfac3fc
commit
b33e13e4b9
1 changed files with 10 additions and 0 deletions
|
|
@ -4,6 +4,7 @@
|
|||
*/
|
||||
|
||||
#include <linux/sched.h>
|
||||
#include <linux/sched/mm.h>
|
||||
#include <linux/bio.h>
|
||||
#include <linux/slab.h>
|
||||
#include <linux/buffer_head.h>
|
||||
|
|
@ -6292,8 +6293,17 @@ static struct btrfs_device *add_missing_dev(struct btrfs_fs_devices *fs_devices,
|
|||
u64 devid, u8 *dev_uuid)
|
||||
{
|
||||
struct btrfs_device *device;
|
||||
unsigned int nofs_flag;
|
||||
|
||||
/*
|
||||
* We call this under the chunk_mutex, so we want to use NOFS for this
|
||||
* allocation, however we don't want to change btrfs_alloc_device() to
|
||||
* always do NOFS because we use it in a lot of other GFP_KERNEL safe
|
||||
* places.
|
||||
*/
|
||||
nofs_flag = memalloc_nofs_save();
|
||||
device = btrfs_alloc_device(NULL, &devid, dev_uuid);
|
||||
memalloc_nofs_restore(nofs_flag);
|
||||
if (IS_ERR(device))
|
||||
return device;
|
||||
|
||||
|
|
|
|||
Loading…
Reference in a new issue