On Tue 2018-04-03 11:13:33, Peter Zijlstra wrote: > On Tue, Apr 03, 2018 at 10:50:03AM +0200, Petr Mladek wrote: > > raw_spin_lock() succeeded here. Therefore lockdep was still working > > at this stage. > > What does the success of raw_spin_lock() have to do with lockdep ?
It means that also lockdep succeeded there. Therefore the general protection fault in __lock_acquire() was specific to the spinlock taken by try_to_wake_up(). I though that it might had been an useful information. Anyway, the other replay from Dmitry suggested that it was a known bug, see https://lkml.kernel.org/r/CACT4Y+ZJ2QD7MPy4hB-M=mz2LuVu3bRbLg1QdY=z=+su1qw...@mail.gmail.com Best Regards, Petr
