syzbot

loop4: detected capacity change from 0 to 512
==================================================================
BUG: KCSAN: data-race in fsnotify_detach_mark / inotify_handle_inode_event

write to 0xffff8881049aea44 of 4 bytes by task 3291 on cpu 1:
 fsnotify_detach_mark+0xba/0x160 fs/notify/mark.c:554
 fsnotify_destroy_mark+0x70/0x150 fs/notify/mark.c:596
 __do_sys_inotify_rm_watch fs/notify/inotify/inotify_user.c:805 [inline]
 __se_sys_inotify_rm_watch+0xe8/0x170 fs/notify/inotify/inotify_user.c:786
 __x64_sys_inotify_rm_watch+0x31/0x40 fs/notify/inotify/inotify_user.c:786
 x64_sys_call+0x2e86/0x2ff0 arch/x86/include/generated/asm/syscalls_64.h:256
 do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline]
 do_syscall_64+0xd2/0x200 arch/x86/entry/syscall_64.c:94
 entry_SYSCALL_64_after_hwframe+0x77/0x7f

read to 0xffff8881049aea44 of 4 bytes by task 3587 on cpu 0:
 inotify_handle_inode_event+0x34e/0x3d0 fs/notify/inotify/inotify_fsnotify.c:132
 fsnotify_handle_inode_event+0x1bd/0x220 fs/notify/fsnotify.c:321
 fsnotify_handle_event fs/notify/fsnotify.c:376 [inline]
 send_to_group fs/notify/fsnotify.c:424 [inline]
 fsnotify+0x13c9/0x14b0 fs/notify/fsnotify.c:641
 __fsnotify_parent+0x2ed/0x330 fs/notify/fsnotify.c:287
 fsnotify_parent include/linux/fsnotify.h:96 [inline]
 fsnotify_path include/linux/fsnotify.h:113 [inline]
 fsnotify_file include/linux/fsnotify.h:127 [inline]
 fsnotify_close include/linux/fsnotify.h:457 [inline]
 __fput+0x1de/0x650 fs/file_table.c:454
 fput_close_sync+0x6e/0x120 fs/file_table.c:573
 __do_sys_close fs/open.c:1587 [inline]
 __se_sys_close fs/open.c:1572 [inline]
 __x64_sys_close+0x56/0xf0 fs/open.c:1572
 x64_sys_call+0x2738/0x2ff0 arch/x86/include/generated/asm/syscalls_64.h:4
 do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline]
 do_syscall_64+0xd2/0x200 arch/x86/entry/syscall_64.c:94
 entry_SYSCALL_64_after_hwframe+0x77/0x7f

value changed: 0x00000003 -> 0x00000000

Reported by Kernel Concurrency Sanitizer on:
CPU: 0 UID: 0 PID: 3587 Comm: syz.4.37 Not tainted syzkaller #0 PREEMPT(voluntary) 
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 08/18/2025
==================================================================
EXT4-fs: Ignoring removed mblk_io_submit option
EXT4-fs: Ignoring removed nomblk_io_submit option
EXT4-fs: Journaled quota options ignored when QUOTA feature is enabled
EXT4-fs (loop4): feature flags set on rev 0 fs, running e2fsck is recommended
EXT4-fs error (device loop4): ext4_mb_mark_diskspace_used:4183: comm syz.4.37: Allocating blocks 41-42 which overlap fs metadata
EXT4-fs error (device loop4): ext4_mb_mark_diskspace_used:4183: comm syz.4.37: Allocating blocks 41-42 which overlap fs metadata
EXT4-fs error (device loop4): ext4_acquire_dquot:6937: comm syz.4.37: Failed to acquire dquot type 1
EXT4-fs error (device loop4): mb_free_blocks:2017: group 0, inode 12: block 14:freeing already freed block (bit 14); block bitmap corrupt.
EXT4-fs error (device loop4): ext4_do_update_inode:5653: inode #12: comm syz.4.37: corrupted inode contents
EXT4-fs error (device loop4): ext4_dirty_inode:6538: inode #12: comm syz.4.37: mark_inode_dirty error
EXT4-fs error (device loop4): ext4_do_update_inode:5653: inode #12: comm syz.4.37: corrupted inode contents
EXT4-fs error (device loop4): __ext4_ext_dirty:206: inode #12: comm syz.4.37: mark_inode_dirty error
EXT4-fs error (device loop4): ext4_do_update_inode:5653: inode #12: comm syz.4.37: corrupted inode contents
EXT4-fs error (device loop4) in ext4_orphan_del:305: Corrupt filesystem
EXT4-fs error (device loop4): ext4_do_update_inode:5653: inode #12: comm syz.4.37: corrupted inode contents
EXT4-fs error (device loop4): ext4_truncate:4666: inode #12: comm syz.4.37: mark_inode_dirty error
EXT4-fs error (device loop4) in ext4_process_orphan:347: Corrupt filesystem
EXT4-fs (loop4): 1 truncate cleaned up
EXT4-fs (loop4): mounted filesystem 00000000-0000-0000-0000-000000000000 r/w without journal. Quota mode: writeback.
=======================================================
WARNING: The mand mount option has been deprecated and
         and is ignored by this kernel. Remove the mand
         option from the mount to silence this warning.
=======================================================
EXT4-fs: Journaled quota options ignored when QUOTA feature is enabled
syz.4.37 (3587) used greatest stack depth: 10032 bytes left