linux/fs
Filipe Manana 519b7e13b5 btrfs: lock the inode in shared mode before starting fiemap
Currently fiemap does not take the inode's lock (VFS lock), it only locks
a file range in the inode's io tree. This however can lead to a deadlock
if we have a concurrent fsync on the file and fiemap code triggers a fault
when accessing the user space buffer with fiemap_fill_next_extent(). The
deadlock happens on the inode's i_mmap_lock semaphore, which is taken both
by fsync and btrfs_page_mkwrite(). This deadlock was recently reported by
syzbot and triggers a trace like the following:

   task:syz-executor361 state:D stack:20264 pid:5668  ppid:5119   flags:0x00004004
   Call Trace:
    <TASK>
    context_switch kernel/sched/core.c:5293 [inline]
    __schedule+0x995/0xe20 kernel/sched/core.c:6606
    schedule+0xcb/0x190 kernel/sched/core.c:6682
    wait_on_state fs/btrfs/extent-io-tree.c:707 [inline]
    wait_extent_bit+0x577/0x6f0 fs/btrfs/extent-io-tree.c:751
    lock_extent+0x1c2/0x280 fs/btrfs/extent-io-tree.c:1742
    find_lock_delalloc_range+0x4e6/0x9c0 fs/btrfs/extent_io.c:488
    writepage_delalloc+0x1ef/0x540 fs/btrfs/extent_io.c:1863
    __extent_writepage+0x736/0x14e0 fs/btrfs/extent_io.c:2174
    extent_write_cache_pages+0x983/0x1220 fs/btrfs/extent_io.c:3091
    extent_writepages+0x219/0x540 fs/btrfs/extent_io.c:3211
    do_writepages+0x3c3/0x680 mm/page-writeback.c:2581
    filemap_fdatawrite_wbc+0x11e/0x170 mm/filemap.c:388
    __filemap_fdatawrite_range mm/filemap.c:421 [inline]
    filemap_fdatawrite_range+0x175/0x200 mm/filemap.c:439
    btrfs_fdatawrite_range fs/btrfs/file.c:3850 [inline]
    start_ordered_ops fs/btrfs/file.c:1737 [inline]
    btrfs_sync_file+0x4ff/0x1190 fs/btrfs/file.c:1839
    generic_write_sync include/linux/fs.h:2885 [inline]
    btrfs_do_write_iter+0xcd3/0x1280 fs/btrfs/file.c:1684
    call_write_iter include/linux/fs.h:2189 [inline]
    new_sync_write fs/read_write.c:491 [inline]
    vfs_write+0x7dc/0xc50 fs/read_write.c:584
    ksys_write+0x177/0x2a0 fs/read_write.c:637
    do_syscall_x64 arch/x86/entry/common.c:50 [inline]
    do_syscall_64+0x3d/0xb0 arch/x86/entry/common.c:80
    entry_SYSCALL_64_after_hwframe+0x63/0xcd
   RIP: 0033:0x7f7d4054e9b9
   RSP: 002b:00007f7d404fa2f8 EFLAGS: 00000246 ORIG_RAX: 0000000000000001
   RAX: ffffffffffffffda RBX: 00007f7d405d87a0 RCX: 00007f7d4054e9b9
   RDX: 0000000000000090 RSI: 0000000020000000 RDI: 0000000000000006
   RBP: 00007f7d405a51d0 R08: 0000000000000000 R09: 0000000000000000
   R10: 0000000000000000 R11: 0000000000000246 R12: 61635f65646f6e69
   R13: 65646f7475616f6e R14: 7261637369646f6e R15: 00007f7d405d87a8
    </TASK>
   INFO: task syz-executor361:5697 blocked for more than 145 seconds.
         Not tainted 6.2.0-rc3-syzkaller-00376-g7c6984405241 #0
   "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
   task:syz-executor361 state:D stack:21216 pid:5697  ppid:5119   flags:0x00004004
   Call Trace:
    <TASK>
    context_switch kernel/sched/core.c:5293 [inline]
    __schedule+0x995/0xe20 kernel/sched/core.c:6606
    schedule+0xcb/0x190 kernel/sched/core.c:6682
    rwsem_down_read_slowpath+0x5f9/0x930 kernel/locking/rwsem.c:1095
    __down_read_common+0x54/0x2a0 kernel/locking/rwsem.c:1260
    btrfs_page_mkwrite+0x417/0xc80 fs/btrfs/inode.c:8526
    do_page_mkwrite+0x19e/0x5e0 mm/memory.c:2947
    wp_page_shared+0x15e/0x380 mm/memory.c:3295
    handle_pte_fault mm/memory.c:4949 [inline]
    __handle_mm_fault mm/memory.c:5073 [inline]
    handle_mm_fault+0x1b79/0x26b0 mm/memory.c:5219
    do_user_addr_fault+0x69b/0xcb0 arch/x86/mm/fault.c:1428
    handle_page_fault arch/x86/mm/fault.c:1519 [inline]
    exc_page_fault+0x7a/0x110 arch/x86/mm/fault.c:1575
    asm_exc_page_fault+0x22/0x30 arch/x86/include/asm/idtentry.h:570
   RIP: 0010:copy_user_short_string+0xd/0x40 arch/x86/lib/copy_user_64.S:233
   Code: 74 0a 89 (...)
   RSP: 0018:ffffc9000570f330 EFLAGS: 00050202
   RAX: ffffffff843e6601 RBX: 00007fffffffefc8 RCX: 0000000000000007
   RDX: 0000000000000000 RSI: ffffc9000570f3e0 RDI: 0000000020000120
   RBP: ffffc9000570f490 R08: 0000000000000000 R09: fffff52000ae1e83
   R10: fffff52000ae1e83 R11: 1ffff92000ae1e7c R12: 0000000000000038
   R13: ffffc9000570f3e0 R14: 0000000020000120 R15: ffffc9000570f3e0
    copy_user_generic arch/x86/include/asm/uaccess_64.h:37 [inline]
    raw_copy_to_user arch/x86/include/asm/uaccess_64.h:58 [inline]
    _copy_to_user+0xe9/0x130 lib/usercopy.c:34
    copy_to_user include/linux/uaccess.h:169 [inline]
    fiemap_fill_next_extent+0x22e/0x410 fs/ioctl.c:144
    emit_fiemap_extent+0x22d/0x3c0 fs/btrfs/extent_io.c:3458
    fiemap_process_hole+0xa00/0xad0 fs/btrfs/extent_io.c:3716
    extent_fiemap+0xe27/0x2100 fs/btrfs/extent_io.c:3922
    btrfs_fiemap+0x172/0x1e0 fs/btrfs/inode.c:8209
    ioctl_fiemap fs/ioctl.c:219 [inline]
    do_vfs_ioctl+0x185b/0x2980 fs/ioctl.c:810
    __do_sys_ioctl fs/ioctl.c:868 [inline]
    __se_sys_ioctl+0x83/0x170 fs/ioctl.c:856
    do_syscall_x64 arch/x86/entry/common.c:50 [inline]
    do_syscall_64+0x3d/0xb0 arch/x86/entry/common.c:80
    entry_SYSCALL_64_after_hwframe+0x63/0xcd
   RIP: 0033:0x7f7d4054e9b9
   RSP: 002b:00007f7d390d92f8 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
   RAX: ffffffffffffffda RBX: 00007f7d405d87b0 RCX: 00007f7d4054e9b9
   RDX: 0000000020000100 RSI: 00000000c020660b RDI: 0000000000000005
   RBP: 00007f7d405a51d0 R08: 00007f7d390d9700 R09: 0000000000000000
   R10: 00007f7d390d9700 R11: 0000000000000246 R12: 61635f65646f6e69
   R13: 65646f7475616f6e R14: 7261637369646f6e R15: 00007f7d405d87b8
    </TASK>

What happens is the following:

1) Task A is doing an fsync, enters btrfs_sync_file() and flushes delalloc
   before locking the inode and the i_mmap_lock semaphore, that is, before
   calling btrfs_inode_lock();

2) After task A flushes delalloc and before it calls btrfs_inode_lock(),
   another task dirties a page;

3) Task B starts a fiemap without FIEMAP_FLAG_SYNC, so the page dirtied
   at step 2 remains dirty and unflushed. Then when it enters
   extent_fiemap() and it locks a file range that includes the range of
   the page dirtied in step 2;

4) Task A calls btrfs_inode_lock() and locks the inode (VFS lock) and the
   inode's i_mmap_lock semaphore in write mode. Then it tries to flush
   delalloc by calling start_ordered_ops(), which will block, at
   find_lock_delalloc_range(), when trying to lock the range of the page
   dirtied at step 2, since this range was locked by the fiemap task (at
   step 3);

5) Task B generates a page fault when accessing the user space fiemap
   buffer with a call to fiemap_fill_next_extent().

   The fault handler needs to call btrfs_page_mkwrite() for some other
   page of our inode, and there we deadlock when trying to lock the
   inode's i_mmap_lock semaphore in read mode, since the fsync task locked
   it in write mode (step 4) and the fsync task can not progress because
   it's waiting to lock a file range that is currently locked by us (the
   fiemap task, step 3).

Fix this by taking the inode's lock (VFS lock) in shared mode when
entering fiemap. This effectively serializes fiemap with fsync (except the
most expensive part of fsync, the log sync), preventing this deadlock.

Reported-by: syzbot+cc35f55c41e34c30dcb5@syzkaller.appspotmail.com
Link: https://lore.kernel.org/linux-btrfs/00000000000032dc7305f2a66f46@google.com/
CC: stable@vger.kernel.org # 6.1+
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
2023-02-09 17:02:27 +01:00
..
9p 9p: Fix some kernel-doc comments 2022-07-02 18:52:21 +09:00
adfs fs: Convert block_read_full_page() to block_read_full_folio() 2022-05-09 16:21:44 -04:00
affs affs: move from strlcpy with unused retval to strscpy 2022-08-19 13:03:10 +02:00
afs afs: Fix server->active leak in afs_put_server 2022-11-30 10:02:37 -08:00
autofs autofs: remove unused ino field inode 2022-07-17 17:31:42 -07:00
befs befs: Convert befs_symlink_read_folio() to use a folio 2022-08-02 12:34:03 -04:00
bfs fs: Convert block_read_full_page() to block_read_full_folio() 2022-05-09 16:21:44 -04:00
btrfs btrfs: lock the inode in shared mode before starting fiemap 2023-02-09 17:02:27 +01:00
cachefiles cachefiles: use vfs_tmpfile_open() helper 2022-09-24 07:00:00 +02:00
ceph ceph: fix NULL pointer dereference for req->r_session 2022-11-14 10:29:05 +01:00
cifs cifs: fix missing unlock in cifs_file_copychunk_range() 2022-11-21 10:27:03 -06:00
coda coda: Convert coda_symlink_filler() to use a folio 2022-08-02 12:34:03 -04:00
configfs
cramfs cramfs: read_mapping_page() is synchronous 2022-08-02 12:34:02 -04:00
crypto fscrypt: fix keyring memory leak on mount failure 2022-10-19 20:54:43 -07:00
debugfs debugfs: use DEFINE_SHOW_ATTRIBUTE to define debugfs_regset32_fops 2022-09-24 15:00:48 +02:00
devpts
dlm Networking changes for 6.1. 2022-10-04 13:38:03 -07:00
ecryptfs whack-a-mole: constifying struct path * 2022-10-06 17:31:02 -07:00
efivarfs efi: efivars: Fix variable writes without query_variable_store() 2022-10-21 11:09:40 +02:00
efs efs: Convert efs symlinks to read_folio 2022-05-09 16:21:45 -04:00
erofs Changes since last update: 2022-11-15 10:30:34 -08:00
exfat treewide: use get_random_u32() when possible 2022-10-11 17:42:58 -06:00
exportfs Change calling conventions for filldir_t 2022-08-17 17:25:04 -04:00
ext2 treewide: use prandom_u32_max() when possible, part 2 2022-10-11 17:42:58 -06:00
ext4 ext4: fix use-after-free in ext4_ext_shift_extents 2022-11-07 12:53:43 -05:00
f2fs Random number generator fixes for Linux 6.1-rc1. 2022-10-16 15:27:07 -07:00
fat treewide: use get_random_u32() when possible 2022-10-11 17:42:58 -06:00
freevxfs freevxfs: Convert vxfs_immed_read_folio() to use a folio 2022-08-02 12:34:03 -04:00
fscache fscache: fix OOB Read in __fscache_acquire_volume 2022-11-23 10:31:13 -08:00
fuse fuse: lock inode unconditionally in fuse_fallocate() 2022-11-23 09:10:42 +01:00
gfs2 gfs2 debugfs improvements 2022-10-10 20:13:22 -07:00
hfs hfs: replace kmap() with kmap_local_page() in btree.c 2022-09-11 21:55:09 -07:00
hfsplus hfsplus: convert kmap() to kmap_local_page() in btree.c 2022-09-11 21:55:05 -07:00
hostfs hostfs: move from strlcpy with unused retval to strscpy 2022-09-19 22:46:25 +02:00
hpfs hpfs: Convert symlinks to read_folio 2022-05-09 16:21:45 -04:00
hugetlbfs hugetlbfs: don't delete error page from pagecache 2022-11-08 15:57:22 -08:00
iomap iomap: add a tracepoint for mappings returned by map_blocks 2022-10-02 11:42:19 -07:00
isofs - hfs and hfsplus kmap API modernization from Fabio Francesco 2022-10-12 11:00:22 -07:00
jbd2 - Yu Zhao's Multi-Gen LRU patches are here. They've been under test in 2022-10-10 17:53:04 -07:00
jffs2 mtd: always initialize 'stats' in struct mtd_oob_ops 2022-09-21 10:38:07 +02:00
jfs Folio changes for 6.0 2022-08-03 10:35:43 -07:00
kernfs kernfs: Fix spurious lockdep warning in kernfs_find_and_get_node_by_id() 2022-11-10 19:03:42 +01:00
ksmbd vfs: fix copy_file_range() averts filesystem freeze protection 2022-11-25 00:52:28 -05:00
lockd SUNRPC: Parametrize how much of argsize should be zeroed 2022-09-26 14:02:42 -04:00
minix vfs: open inside ->tmpfile() 2022-09-24 07:00:00 +02:00
netfs netfs: Fix dodgy maths 2022-11-15 16:56:07 +00:00
nfs nfs4: Fix kmemleak when allocate slot failed 2022-10-27 15:52:11 -04:00
nfs_common
nfsd Amir's copy_file_range() fix 2022-11-27 12:40:06 -08:00
nilfs2 nilfs2: fix NULL pointer dereference in nilfs_palloc_commit_free_entry() 2022-11-30 14:49:40 -08:00
nls
notify Merge tag 'fsnotify-for_v6.1-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs 2022-10-07 08:28:50 -07:00
ntfs - hfs and hfsplus kmap API modernization from Fabio Francesco 2022-10-12 11:00:22 -07:00
ntfs3 treewide: use get_random_u32() when possible 2022-10-11 17:42:58 -06:00
ocfs2 ocfs2: clear dinode links count in case of error 2022-10-20 21:27:22 -07:00
omfs fs: Convert block_read_full_page() to block_read_full_folio() 2022-05-09 16:21:44 -04:00
openpromfs fs: allocate inode by using alloc_inode_sb() 2022-03-22 15:57:03 -07:00
orangefs Orangefs: change iterate to iterate_shared 2022-10-13 09:56:14 -07:00
overlayfs tmpfile API change 2022-10-10 19:45:17 -07:00
proc proc/meminfo: fix spacing in SecPageTables 2022-11-22 18:50:44 -08:00
pstore Revert "pstore: migrate to crypto acomp interface" 2022-09-30 08:16:06 -07:00
qnx4 fs: Convert block_read_full_page() to block_read_full_folio() 2022-05-09 16:21:44 -04:00
qnx6 fs/qnx6: delete unnecessary checks before brelse() 2022-09-11 21:55:07 -07:00
quota quota: Add more checking after reading from quota file 2022-09-29 15:37:30 +02:00
ramfs tmpfile API change 2022-10-10 19:45:17 -07:00
reiserfs - hfs and hfsplus kmap API modernization from Fabio Francesco 2022-10-12 11:00:22 -07:00
romfs romfs: Convert romfs to read_folio 2022-05-09 16:21:46 -04:00
smbfs_common smb3: define missing create contexts 2022-10-05 01:55:27 -05:00
squashfs squashfs: fix buffer release race condition in readahead code 2022-10-28 13:37:21 -07:00
sysfs kobject: kobj_type: remove default_attrs 2022-04-05 15:39:19 +02:00
sysv Not a lot of material this cycle. Many singleton patches against various 2022-05-27 11:22:03 -07:00
tracefs tracefs: Only clobber mode/uid/gid on remount if asked 2022-09-08 17:10:54 -04:00
ubifs Random number generator fixes for Linux 6.1-rc1. 2022-10-16 15:27:07 -07:00
udf udf: Fix a slab-out-of-bounds write bug in udf_find_entry() 2022-11-09 12:24:42 +01:00
ufs ufs: replace ll_rw_block() 2022-09-11 20:26:07 -07:00
unicode
vboxsf vboxsf: Convert vboxsf to read_folio 2022-05-09 16:21:46 -04:00
verity for-6.1-tag 2022-10-06 17:36:48 -07:00
xfs xfs: rename XFS_REFC_COW_START to _COWFLAG 2022-10-31 08:58:22 -07:00
zonefs zonefs: Fix active zone accounting 2022-11-25 17:01:22 +09:00
aio.c aio: use atomic_try_cmpxchg in __get_reqs_available 2022-09-11 21:55:08 -07:00
anon_inodes.c dynamic_dname(): drop unused dentry argument 2022-08-20 11:34:04 -04:00
attr.c vfs: Check the truncate maximum size in inode_newsize_ok() 2022-08-08 10:39:29 -07:00
bad_inode.c vfs: open inside ->tmpfile() 2022-09-24 07:00:00 +02:00
binfmt_elf_fdpic.c
binfmt_elf_test.c
binfmt_elf.c fs/binfmt_elf: Fix memory leak in load_elf_binary() 2022-10-25 15:11:21 -07:00
binfmt_flat.c binfmt_flat: Remove shared library support 2022-04-22 10:57:18 -07:00
binfmt_misc.c
binfmt_script.c
buffer.c - hfs and hfsplus kmap API modernization from Fabio Francesco 2022-10-12 11:00:22 -07:00
char_dev.c
compat_binfmt_elf.c
coredump.c - Yu Zhao's Multi-Gen LRU patches are here. They've been under test in 2022-10-10 17:53:04 -07:00
d_path.c d_path.c: typo fix... 2022-08-20 11:34:33 -04:00
dax.c Merge branch 'for-6.0/dax' into libnvdimm-fixes 2022-09-24 18:14:12 -07:00
dcache.c tmpfile API change 2022-10-10 19:45:17 -07:00
direct-io.c block: remove PSI accounting from the bio layer 2022-09-20 08:24:38 -06:00
drop_caches.c
eventfd.c eventfd: guard wake_up in eventfd fs calls as well 2022-09-21 10:30:42 -06:00
eventpoll.c epoll: use try_cmpxchg in list_add_tail_lockless 2022-09-11 21:55:07 -07:00
exec.c 23 hotfixes. 2022-10-29 17:49:33 -07:00
fcntl.c keep iocb_flags() result cached in struct file 2022-06-10 16:10:23 -04:00
fhandle.c do_sys_name_to_handle(): constify path 2022-09-01 17:36:39 -04:00
file_table.c locks: fix TOCTOU race when granting write lease 2022-08-16 10:59:54 -04:00
file.c fs: use acquire ordering in __fget_light() 2022-10-31 15:30:11 -04:00
filesystems.c
fs_context.c
fs_parser.c
fs_pin.c
fs_struct.c
fs_types.c
fs-writeback.c fs: do not update freeing inode i_io_list 2022-11-22 17:00:00 -05:00
fsopen.c uninline may_mount() and don't opencode it in fspick(2)/fsopen(2) 2022-05-19 23:25:10 -04:00
init.c
inode.c saner inode_init_always() 2022-10-06 16:49:00 -07:00
internal.h whack-a-mole: constifying struct path * 2022-10-06 17:31:02 -07:00
ioctl.c Fixes for 5.18-rc1: 2022-04-01 19:35:56 -07:00
Kconfig hugetlb: make hugetlb depends on SYSFS or SYSCTL 2022-09-11 20:26:10 -07:00
Kconfig.binfmt Xtensa updates for v6.1 2022-10-10 14:21:11 -07:00
kernel_read_file.c fs/kernel_read_file: allow to read files up-to ssize_t 2022-06-16 19:58:21 -07:00
libfs.c fs: uninline inode_maybe_inc_iversion() 2022-10-03 14:21:43 -07:00
locks.c locks: Fix dropped call to ->fl_release_private() 2022-08-17 15:08:58 -04:00
Makefile a.out: Remove the a.out implementation 2022-09-27 07:11:02 -07:00
mbcache.c mbcache: Avoid nesting of cache->c_list_lock under bit locks 2022-09-30 23:46:52 -04:00
mount.h switch try_to_unlazy_next() to __legitimize_mnt() 2022-07-05 16:18:21 -04:00
mpage.c Folio changes for 6.0 2022-08-03 10:35:43 -07:00
namei.c vfs: vfs_tmpfile: ensure O_EXCL flag is enforced 2022-11-19 02:22:11 -05:00
namespace.c fs: require CAP_SYS_ADMIN in target namespace for idmapped mounts 2022-08-17 11:27:11 +02:00
no-block.c
nsfs.c dynamic_dname(): drop unused dentry argument 2022-08-20 11:34:04 -04:00
open.c struct file-related stuff 2022-10-06 17:13:18 -07:00
pipe.c dynamic_dname(): drop unused dentry argument 2022-08-20 11:34:04 -04:00
pnode.c
pnode.h
posix_acl.c - Yu Zhao's Multi-Gen LRU patches are here. They've been under test in 2022-10-10 17:53:04 -07:00
proc_namespace.c vfs: escape hash as well 2022-06-28 13:58:05 -04:00
read_write.c vfs: fix copy_file_range() averts filesystem freeze protection 2022-11-25 00:52:28 -05:00
readdir.c Change calling conventions for filldir_t 2022-08-17 17:25:04 -04:00
remap_range.c - The usual batches of cleanups from Baoquan He, Muchun Song, Miaohe 2022-08-05 16:32:45 -07:00
select.c
seq_file.c rxrpc: Fix locking issue 2022-05-22 21:03:01 +01:00
signalfd.c
splice.c iter_to_pipe(): switch to advancing variant of iov_iter_get_pages() 2022-08-08 22:37:23 -04:00
stack.c
stat.c vfs: support STATX_DIOALIGN on block devices 2022-09-11 19:47:12 -05:00
statfs.c
super.c fscrypt: fix keyring memory leak on mount failure 2022-10-19 20:54:43 -07:00
sync.c riscv: compat: syscall: Add compat_sys_call_table implementation 2022-04-26 13:36:25 -07:00
sysctls.c
timerfd.c
userfaultfd.c fs/userfaultfd: Fix maple tree iterator in userfaultfd_unregister() 2022-11-07 12:58:26 -08:00
utimes.c
xattr.c xattr: always us is_posix_acl_xattr() helper 2022-09-21 12:01:29 +02:00