Sorry, I don't understand... On 04/30, Peter Zijlstra wrote: > > Thaw then does the reverse, frobs lockdep Yes, in particular it does lockdep_sb_freeze_acquire() percpu_rwsem_acquire() sem->rw_sem.owner = current; > and then does: percpu_up_write(). > > percpu_up_write() on up_write(). And note how __up_write() has: > > DEBUG_RWSEMS_WARN_ON(sem->owner != current, sem); and everything looks correct, sem->owner == current by the time thaw_super_locked() does percpu_up_write/up_write. Oleg.