Hi Ted, FYI, we noticed that simple dd write can trigger the below BUG in commit 208b9cc3a32f29b8a9bba682b5bfbb962bc9892d ("ext4: add a new spinlock i_raw_lock to protect the ext4's raw inode") [ 20.909408] EXT4-fs (sda1): mounted filesystem with ordered data mode. Opts: (null) [ 21.932207] BUG: sleeping function called from invalid context at fs/ext4/ext4_jbd2.c:252 [ 21.933168] in_atomic(): 1, irqs_disabled(): 0, pid: 3280, name: dd [ 21.933709] CPU: 1 PID: 3280 Comm: dd Not tainted 3.14.0-01910-gdeb73c7 #1 [ 21.934278] Hardware name: / , BIOS VVRBLI9J.86A.2891.2007.0511.1144 05/11/2007 [ 21.935205] 0000000000000000 ffff88007de99aa8 ffffffff819b8b88 ffff880038f10000 [ 21.936226] ffff88007de99ab8 ffffffff811030c2 ffff88007de99b18 ffffffff812bba58 [ 21.937231] 000000000000130c ffffffff812b5e7f ffff88007de99b00 ffffffff812d8046 [ 21.938231] Call Trace: [ 21.938627] [] dump_stack+0x4d/0x66 [ 21.939140] [] __might_sleep+0x10a/0x10c [ 21.939656] [] __ext4_handle_dirty_metadata+0x39/0x1b5 [ 21.940237] [] ? ext4_ext_tree_init+0x32/0x36 [ 21.940780] [] ? jbd2_journal_get_write_access+0x33/0x3c [ 21.941362] [] ext4_mark_iloc_dirty+0x577/0x630 [ 21.941915] [] ? ext4_ext_tree_init+0x32/0x36 [ 21.941916] [] ext4_mark_inode_dirty+0xa6/0x1f0 [ 21.941920] [] ext4_ext_tree_init+0x32/0x36 [ 21.941922] [] __ext4_new_inode+0xf29/0x10c9 [ 21.941925] [] ext4_create+0xac/0x156 [ 21.941930] [] vfs_create+0x68/0xa1 [ 21.941932] [] do_last.isra.52+0x518/0xa23 [ 21.941933] [] ? link_path_walk+0x69/0x71f [ 21.941935] [] path_openat+0x225/0x591 [ 21.941937] [] do_filp_open+0x3a/0x7f [ 21.941939] [] ? __alloc_fd+0x5e/0xef [ 21.941942] [] do_sys_open+0x147/0x1d6 [ 21.941943] [] ? __fdget+0x13/0x15 [ 21.941945] [] SyS_open+0x1e/0x20 [ 21.941947] [] system_call_fastpath+0x16/0x1b Thanks, Fengguang _______________________________________________ LKP mailing list LKP(a)linux.intel.com