linux/fs/btrfs
Josef Bacik c495144bc6 btrfs: move the dio_sem higher up the callchain
We're getting a lockdep splat because we take the dio_sem under the
log_mutex.  What we really need is to protect fsync() from logging an
extent map for an extent we never waited on higher up, so just guard the
whole thing with dio_sem.

======================================================
WARNING: possible circular locking dependency detected
4.18.0-rc4-xfstests-00025-g5de5edbaf1d4 #411 Not tainted
------------------------------------------------------
aio-dio-invalid/30928 is trying to acquire lock:
0000000092621cfd (&mm->mmap_sem){++++}, at: get_user_pages_unlocked+0x5a/0x1e0

but task is already holding lock:
00000000cefe6b35 (&ei->dio_sem){++++}, at: btrfs_direct_IO+0x3be/0x400

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #5 (&ei->dio_sem){++++}:
       lock_acquire+0xbd/0x220
       down_write+0x51/0xb0
       btrfs_log_changed_extents+0x80/0xa40
       btrfs_log_inode+0xbaf/0x1000
       btrfs_log_inode_parent+0x26f/0xa80
       btrfs_log_dentry_safe+0x50/0x70
       btrfs_sync_file+0x357/0x540
       do_fsync+0x38/0x60
       __ia32_sys_fdatasync+0x12/0x20
       do_fast_syscall_32+0x9a/0x2f0
       entry_SYSENTER_compat+0x84/0x96

-> #4 (&ei->log_mutex){+.+.}:
       lock_acquire+0xbd/0x220
       __mutex_lock+0x86/0xa10
       btrfs_record_unlink_dir+0x2a/0xa0
       btrfs_unlink+0x5a/0xc0
       vfs_unlink+0xb1/0x1a0
       do_unlinkat+0x264/0x2b0
       do_fast_syscall_32+0x9a/0x2f0
       entry_SYSENTER_compat+0x84/0x96

-> #3 (sb_internal#2){.+.+}:
       lock_acquire+0xbd/0x220
       __sb_start_write+0x14d/0x230
       start_transaction+0x3e6/0x590
       btrfs_evict_inode+0x475/0x640
       evict+0xbf/0x1b0
       btrfs_run_delayed_iputs+0x6c/0x90
       cleaner_kthread+0x124/0x1a0
       kthread+0x106/0x140
       ret_from_fork+0x3a/0x50

-> #2 (&fs_info->cleaner_delayed_iput_mutex){+.+.}:
       lock_acquire+0xbd/0x220
       __mutex_lock+0x86/0xa10
       btrfs_alloc_data_chunk_ondemand+0x197/0x530
       btrfs_check_data_free_space+0x4c/0x90
       btrfs_delalloc_reserve_space+0x20/0x60
       btrfs_page_mkwrite+0x87/0x520
       do_page_mkwrite+0x31/0xa0
       __handle_mm_fault+0x799/0xb00
       handle_mm_fault+0x7c/0xe0
       __do_page_fault+0x1d3/0x4a0
       async_page_fault+0x1e/0x30

-> #1 (sb_pagefaults){.+.+}:
       lock_acquire+0xbd/0x220
       __sb_start_write+0x14d/0x230
       btrfs_page_mkwrite+0x6a/0x520
       do_page_mkwrite+0x31/0xa0
       __handle_mm_fault+0x799/0xb00
       handle_mm_fault+0x7c/0xe0
       __do_page_fault+0x1d3/0x4a0
       async_page_fault+0x1e/0x30

-> #0 (&mm->mmap_sem){++++}:
       __lock_acquire+0x42e/0x7a0
       lock_acquire+0xbd/0x220
       down_read+0x48/0xb0
       get_user_pages_unlocked+0x5a/0x1e0
       get_user_pages_fast+0xa4/0x150
       iov_iter_get_pages+0xc3/0x340
       do_direct_IO+0xf93/0x1d70
       __blockdev_direct_IO+0x32d/0x1c20
       btrfs_direct_IO+0x227/0x400
       generic_file_direct_write+0xcf/0x180
       btrfs_file_write_iter+0x308/0x58c
       aio_write+0xf8/0x1d0
       io_submit_one+0x3a9/0x620
       __ia32_compat_sys_io_submit+0xb2/0x270
       do_int80_syscall_32+0x5b/0x1a0
       entry_INT80_compat+0x88/0xa0

other info that might help us debug this:

Chain exists of:
  &mm->mmap_sem --> &ei->log_mutex --> &ei->dio_sem

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(&ei->dio_sem);
                               lock(&ei->log_mutex);
                               lock(&ei->dio_sem);
  lock(&mm->mmap_sem);

 *** DEADLOCK ***

1 lock held by aio-dio-invalid/30928:
 #0: 00000000cefe6b35 (&ei->dio_sem){++++}, at: btrfs_direct_IO+0x3be/0x400

stack backtrace:
CPU: 0 PID: 30928 Comm: aio-dio-invalid Not tainted 4.18.0-rc4-xfstests-00025-g5de5edbaf1d4 #411
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.11.0-2.el7 04/01/2014
Call Trace:
 dump_stack+0x7c/0xbb
 print_circular_bug.isra.37+0x297/0x2a4
 check_prev_add.constprop.45+0x781/0x7a0
 ? __lock_acquire+0x42e/0x7a0
 validate_chain.isra.41+0x7f0/0xb00
 __lock_acquire+0x42e/0x7a0
 lock_acquire+0xbd/0x220
 ? get_user_pages_unlocked+0x5a/0x1e0
 down_read+0x48/0xb0
 ? get_user_pages_unlocked+0x5a/0x1e0
 get_user_pages_unlocked+0x5a/0x1e0
 get_user_pages_fast+0xa4/0x150
 iov_iter_get_pages+0xc3/0x340
 do_direct_IO+0xf93/0x1d70
 ? __alloc_workqueue_key+0x358/0x490
 ? __blockdev_direct_IO+0x14b/0x1c20
 __blockdev_direct_IO+0x32d/0x1c20
 ? btrfs_run_delalloc_work+0x40/0x40
 ? can_nocow_extent+0x490/0x490
 ? kvm_clock_read+0x1f/0x30
 ? can_nocow_extent+0x490/0x490
 ? btrfs_run_delalloc_work+0x40/0x40
 btrfs_direct_IO+0x227/0x400
 ? btrfs_run_delalloc_work+0x40/0x40
 generic_file_direct_write+0xcf/0x180
 btrfs_file_write_iter+0x308/0x58c
 aio_write+0xf8/0x1d0
 ? kvm_clock_read+0x1f/0x30
 ? __might_fault+0x3e/0x90
 io_submit_one+0x3a9/0x620
 ? io_submit_one+0xe5/0x620
 __ia32_compat_sys_io_submit+0xb2/0x270
 do_int80_syscall_32+0x5b/0x1a0
 entry_INT80_compat+0x88/0xa0

CC: stable@vger.kernel.org # 4.14+
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
2018-10-19 12:20:04 +02:00
..
tests btrfs: tests: add separate stub for find_lock_delalloc_range 2018-10-15 17:23:34 +02:00
acl.c btrfs: remove unnecessary curly braces in btrfs_get_acl 2018-08-06 13:12:41 +02:00
async-thread.c btrfs: replace GPL boilerplate by SPDX -- sources 2018-04-12 16:29:51 +02:00
async-thread.h btrfs: replace GPL boilerplate by SPDX -- headers 2018-04-12 16:29:46 +02:00
backref.c Btrfs: preftree: use rb_first_cached 2018-10-15 17:23:33 +02:00
backref.h btrfs: replace GPL boilerplate by SPDX -- headers 2018-04-12 16:29:46 +02:00
btrfs_inode.h btrfs: Remove 'objectid' member from struct btrfs_root 2018-10-15 17:23:25 +02:00
check-integrity.c Btrfs: use args in the correct order for kcalloc in btrfsic_read_block 2018-10-15 17:23:30 +02:00
check-integrity.h btrfs: replace GPL boilerplate by SPDX -- headers 2018-04-12 16:29:46 +02:00
compression.c btrfs: remove unused pointer 'tree' in btrfs_submit_compressed_read 2018-10-15 17:23:28 +02:00
compression.h btrfs: compression: Add linux/sizes.h for compression.h 2018-05-29 18:13:00 +02:00
ctree.c Btrfs: fix deadlock when writing out free space caches 2018-10-17 17:46:24 +02:00
ctree.h btrfs: remove fs_info from btrfs_should_throttle_delayed_refs 2018-10-15 17:23:41 +02:00
dedupe.h btrfs: replace GPL boilerplate by SPDX -- headers 2018-04-12 16:29:46 +02:00
delayed-inode.c Btrfs: kill btrfs_clear_path_blocking 2018-10-15 17:23:38 +02:00
delayed-inode.h Btrfs: delayed-inode: use rb_first_cached for ins_root and del_root 2018-10-15 17:23:33 +02:00
delayed-ref.c btrfs: delayed-ref: extract find_first_ref_head from find_ref_head 2018-10-17 19:21:00 +02:00
delayed-ref.h btrfs: delayed-ref: pass delayed_refs directly to btrfs_delayed_ref_lock 2018-10-15 17:23:41 +02:00
dev-replace.c btrfs: dev-replace: remove pointless assert in write unlock 2018-10-15 17:23:38 +02:00
dev-replace.h btrfs: open code btrfs_after_dev_replace_commit 2018-10-15 17:23:37 +02:00
dir-item.c btrfs: Remove root parameter from btrfs_insert_dir_item 2018-10-15 17:23:25 +02:00
disk-io.c btrfs: assert on non-empty delayed iputs 2018-10-15 17:23:39 +02:00
disk-io.h btrfs: unify end_io callbacks of async_submit_bio 2018-08-06 13:12:55 +02:00
export.c btrfs: Remove 'objectid' member from struct btrfs_root 2018-10-15 17:23:25 +02:00
export.h btrfs: replace GPL boilerplate by SPDX -- headers 2018-04-12 16:29:46 +02:00
extent_io.c btrfs: tests: add separate stub for find_lock_delalloc_range 2018-10-15 17:23:34 +02:00
extent_io.h btrfs: tests: add separate stub for find_lock_delalloc_range 2018-10-15 17:23:34 +02:00
extent_map.c Btrfs: extent_map: use rb_first_cached 2018-10-15 17:23:33 +02:00
extent_map.h Btrfs: extent_map: use rb_first_cached 2018-10-15 17:23:33 +02:00
extent-tree.c btrfs: fix insert_reserved error handling 2018-10-19 12:20:03 +02:00
file-item.c btrfs: simplify pointer chasing of local fs_info variables 2018-08-06 13:12:43 +02:00
file.c btrfs: move the dio_sem higher up the callchain 2018-10-19 12:20:04 +02:00
free-space-cache.c btrfs: set max_extent_size properly 2018-10-19 12:20:03 +02:00
free-space-cache.h btrfs: replace GPL boilerplate by SPDX -- headers 2018-04-12 16:29:46 +02:00
free-space-tree.c btrfs: Remove fs_info from btrfs_del_root 2018-08-06 13:13:00 +02:00
free-space-tree.h btrfs: Remove fs_info argument from add_to_free_space_tree 2018-05-28 18:07:36 +02:00
inode-item.c btrfs: replace GPL boilerplate by SPDX -- sources 2018-04-12 16:29:51 +02:00
inode-map.c btrfs: prune unused includes 2018-08-06 13:12:43 +02:00
inode-map.h btrfs: replace GPL boilerplate by SPDX -- headers 2018-04-12 16:29:46 +02:00
inode.c btrfs: only free reserved extent if we didn't insert it 2018-10-19 12:20:03 +02:00
ioctl.c btrfs: Ensure btrfs_trim_fs can trim the whole filesystem 2018-10-15 17:23:32 +02:00
Kconfig btrfs: add SPDX header to Kconfig 2018-04-12 16:29:55 +02:00
locking.c btrfs: replace waitqueue_actvie with cond_wake_up 2018-05-28 18:23:09 +02:00
locking.h btrfs: replace GPL boilerplate by SPDX -- headers 2018-04-12 16:29:46 +02:00
lzo.c btrfs: lzo: Harden inline lzo compressed extent decompression 2018-05-30 16:46:43 +02:00
Makefile btrfs: Remove custom crc32c init code 2018-03-26 15:09:39 +02:00
math.h btrfs: replace GPL boilerplate by SPDX -- headers 2018-04-12 16:29:46 +02:00
ordered-data.c btrfs: prune unused includes 2018-08-06 13:12:43 +02:00
ordered-data.h btrfs: remove remaing full_sync logic from btrfs_sync_file 2018-08-06 13:12:31 +02:00
orphan.c btrfs: replace GPL boilerplate by SPDX -- sources 2018-04-12 16:29:51 +02:00
print-tree.c btrfs: annotate unlikely branches after V0 extent type removal 2018-08-06 13:12:41 +02:00
print-tree.h btrfs: print-tree: debugging output enhancement 2018-04-20 19:18:16 +02:00
props.c btrfs: property: Set incompat flag if lzo/zstd compression is set 2018-05-17 14:18:25 +02:00
props.h btrfs: replace GPL boilerplate by SPDX -- headers 2018-04-12 16:29:46 +02:00
qgroup.c btrfs: qgroup: move the qgroup->members check out from (!qgroup)'s else branch 2018-10-15 17:23:40 +02:00
qgroup.h btrfs: qgroup: Avoid calling qgroup functions if qgroup is not enabled 2018-10-15 17:23:40 +02:00
raid56.c btrfs: raid56: catch errors from full_stripe_write 2018-08-06 13:12:45 +02:00
raid56.h btrfs: replace GPL boilerplate by SPDX -- headers 2018-04-12 16:29:46 +02:00
rcu-string.h btrfs: replace GPL boilerplate by SPDX -- headers 2018-04-12 16:29:46 +02:00
reada.c btrfs: prune unused includes 2018-08-06 13:12:43 +02:00
ref-verify.c btrfs: Remove 'objectid' member from struct btrfs_root 2018-10-15 17:23:25 +02:00
ref-verify.h btrfs: replace GPL boilerplate by SPDX -- headers 2018-04-12 16:29:46 +02:00
relocation.c btrfs: relocation: Remove redundant tree level check 2018-10-15 17:23:40 +02:00
root-tree.c btrfs: Remove fs_info from btrfs_add_root_ref 2018-08-06 13:13:00 +02:00
scrub.c btrfs: open code btrfs_dev_replace_stats_inc 2018-10-15 17:23:37 +02:00
send.c Btrfs: unify error handling of btrfs_lookup_dir_item 2018-10-15 17:23:30 +02:00
send.h btrfs: replace GPL boilerplate by SPDX -- headers 2018-04-12 16:29:46 +02:00
struct-funcs.c btrfs: prune unused includes 2018-08-06 13:12:43 +02:00
super.c btrfs: Remove 'objectid' member from struct btrfs_root 2018-10-15 17:23:25 +02:00
sysfs.c btrfs: prune unused includes 2018-08-06 13:12:43 +02:00
sysfs.h btrfs: sysfs: Use enum/define value for feature array definitions 2018-05-28 18:23:39 +02:00
transaction.c btrfs: don't run delayed_iputs in commit 2018-10-19 12:20:03 +02:00
transaction.h btrfs: replace get_seconds with new 64bit time API 2018-08-06 13:12:29 +02:00
tree-checker.c btrfs: tree-checker: Check level for leaves and nodes 2018-10-15 17:23:37 +02:00
tree-checker.h btrfs: replace GPL boilerplate by SPDX -- headers 2018-04-12 16:29:46 +02:00
tree-defrag.c btrfs: replace GPL boilerplate by SPDX -- sources 2018-04-12 16:29:51 +02:00
tree-log.c btrfs: move the dio_sem higher up the callchain 2018-10-19 12:20:04 +02:00
tree-log.h btrfs: change btrfs_pin_log_trans to return void 2018-10-15 17:23:27 +02:00
ulist.c btrfs: replace GPL boilerplate by SPDX -- sources 2018-04-12 16:29:51 +02:00
ulist.h btrfs: replace GPL boilerplate by SPDX -- headers 2018-04-12 16:29:46 +02:00
uuid-tree.c btrfs: Remove fs_info argument from btrfs_uuid_tree_rem 2018-05-30 16:46:53 +02:00
volumes.c btrfs: open code btrfs_dev_replace_clear_lock_blocking 2018-10-15 17:23:37 +02:00
volumes.h btrfs: Make btrfs_find_device_by_devspec return btrfs_device directly 2018-10-15 17:23:30 +02:00
xattr.c btrfs: replace GPL boilerplate by SPDX -- sources 2018-04-12 16:29:51 +02:00
xattr.h btrfs: replace GPL boilerplate by SPDX -- headers 2018-04-12 16:29:46 +02:00
zlib.c btrfs: replace GPL boilerplate by SPDX -- sources 2018-04-12 16:29:51 +02:00
zstd.c btrfs: replace GPL boilerplate by SPDX -- sources 2018-04-12 16:29:51 +02:00