On (25/05/21 12:18), Jan Kara wrote: > On Tue 20-05-25 21:35:12, Sergey Senozhatsky wrote: > > Once reply response is set for all outstanding requests > > wake_up_all() of the ->access_waitq waiters so that they > > can finish user-wait. Otherwise fsnotify_destroy_group() > > can wait forever for ->user_waits to reach 0 (which it > > never will.) > > > > Signed-off-by: Sergey Senozhatsky <senozhatsky@xxxxxxxxxxxx> > > We don't use exclusive waits with access_waitq so wake_up() and > wake_up_all() should do the same thing? Oh, non-exclusive waiters, I see. I totally missed that, thanks. So... the problem is somewhere else then. I'm currently looking at some crashes (across all LTS kernels) where group owner just gets stuck and then hung-task watchdog kicks in and panics the system. Basically just a single backtrace in the kernel logs: schedule+0x534/0x2540 fsnotify_destroy_group+0xa7/0x150 fanotify_release+0x147/0x160 ____fput+0xe4/0x2a0 task_work_run+0x71/0xb0 do_exit+0x1ea/0x800 do_group_exit+0x81/0x90 get_signal+0x32d/0x4e0 My assumption was that it's this wait: wait_event(group->notification_waitq, !atomic_read(&group->user_waits)); But I guess I was wrong.
