CVE-2025-38347

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> f2fs: fix to do sanity check on ino and xnid<br /> <br /> syzbot reported a f2fs bug as below:<br /> <br /> INFO: task syz-executor140:5308 blocked for more than 143 seconds.<br /> Not tainted 6.14.0-rc7-syzkaller-00069-g81e4f8d68c66 #0<br /> "echo 0 &gt; /proc/sys/kernel/hung_task_timeout_secs" disables this message.<br /> task:syz-executor140 state:D stack:24016 pid:5308 tgid:5308 ppid:5306 task_flags:0x400140 flags:0x00000006<br /> Call Trace:<br /> <br /> context_switch kernel/sched/core.c:5378 [inline]<br /> __schedule+0x190e/0x4c90 kernel/sched/core.c:6765<br /> __schedule_loop kernel/sched/core.c:6842 [inline]<br /> schedule+0x14b/0x320 kernel/sched/core.c:6857<br /> io_schedule+0x8d/0x110 kernel/sched/core.c:7690<br /> folio_wait_bit_common+0x839/0xee0 mm/filemap.c:1317<br /> __folio_lock mm/filemap.c:1664 [inline]<br /> folio_lock include/linux/pagemap.h:1163 [inline]<br /> __filemap_get_folio+0x147/0xb40 mm/filemap.c:1917<br /> pagecache_get_page+0x2c/0x130 mm/folio-compat.c:87<br /> find_get_page_flags include/linux/pagemap.h:842 [inline]<br /> f2fs_grab_cache_page+0x2b/0x320 fs/f2fs/f2fs.h:2776<br /> __get_node_page+0x131/0x11b0 fs/f2fs/node.c:1463<br /> read_xattr_block+0xfb/0x190 fs/f2fs/xattr.c:306<br /> lookup_all_xattrs fs/f2fs/xattr.c:355 [inline]<br /> f2fs_getxattr+0x676/0xf70 fs/f2fs/xattr.c:533<br /> __f2fs_get_acl+0x52/0x870 fs/f2fs/acl.c:179<br /> f2fs_acl_create fs/f2fs/acl.c:375 [inline]<br /> f2fs_init_acl+0xd7/0x9b0 fs/f2fs/acl.c:418<br /> f2fs_init_inode_metadata+0xa0f/0x1050 fs/f2fs/dir.c:539<br /> f2fs_add_inline_entry+0x448/0x860 fs/f2fs/inline.c:666<br /> f2fs_add_dentry+0xba/0x1e0 fs/f2fs/dir.c:765<br /> f2fs_do_add_link+0x28c/0x3a0 fs/f2fs/dir.c:808<br /> f2fs_add_link fs/f2fs/f2fs.h:3616 [inline]<br /> f2fs_mknod+0x2e8/0x5b0 fs/f2fs/namei.c:766<br /> vfs_mknod+0x36d/0x3b0 fs/namei.c:4191<br /> unix_bind_bsd net/unix/af_unix.c:1286 [inline]<br /> unix_bind+0x563/0xe30 net/unix/af_unix.c:1379<br /> __sys_bind_socket net/socket.c:1817 [inline]<br /> __sys_bind+0x1e4/0x290 net/socket.c:1848<br /> __do_sys_bind net/socket.c:1853 [inline]<br /> __se_sys_bind net/socket.c:1851 [inline]<br /> __x64_sys_bind+0x7a/0x90 net/socket.c:1851<br /> do_syscall_x64 arch/x86/entry/common.c:52 [inline]<br /> do_syscall_64+0xf3/0x230 arch/x86/entry/common.c:83<br /> entry_SYSCALL_64_after_hwframe+0x77/0x7f<br /> <br /> Let&amp;#39;s dump and check metadata of corrupted inode, it shows its xattr_nid<br /> is the same to its i_ino.<br /> <br /> dump.f2fs -i 3 chaseyu.img.raw<br /> i_xattr_nid [0x 3 : 3]<br /> <br /> So that, during mknod in the corrupted directory, it tries to get and<br /> lock inode page twice, result in deadlock.<br /> <br /> - f2fs_mknod<br /> - f2fs_add_inline_entry<br /> - f2fs_get_inode_page --- lock dir&amp;#39;s inode page<br /> - f2fs_init_acl<br /> - f2fs_acl_create(dir,..)<br /> - __f2fs_get_acl<br /> - f2fs_getxattr<br /> - lookup_all_xattrs<br /> - __get_node_page --- try to lock dir&amp;#39;s inode page<br /> <br /> In order to fix this, let&amp;#39;s add sanity check on ino and xnid.