On Thu, 17 Aug 2017 20:41:10 +0000
Bart Van Assche <Bart.VanAssche@xxxxxxx> wrote:
> On Thu, 2017-08-17 at 16:30 -0400, Steven Rostedt wrote:
> > On Thu, 17 Aug 2017 12:24:39 -0400
> > Waiman Long <longman@xxxxxxxxxx> wrote:
> >
> > > On 08/17/2017 09:34 AM, Steven Rostedt wrote:
> > > > On Wed, 16 Aug 2017 16:40:40 -0400
> > > > Waiman Long <longman@xxxxxxxxxx> wrote:
> > > >
> > > > > The lockdep code had reported the following unsafe locking scenario:
> > > > >
> > > > > CPU0 CPU1
> > > > > ---- ----
> > > > > lock(s_active#228);
> > > > > lock(&bdev->bd_mutex/1);
> > > > > lock(s_active#228);
> > > > > lock(&bdev->bd_mutex);
> > > >
> > > > Can you show the exact locations of these locks. I have no idea where
> > > > this "s_active" is.
> > >
> > > The s_active isn't an actual lock. It is a reference count (kn->count)
> > > on the sysfs (kernfs) file. Removal of a sysfs file, however, require
> > > a wait until all the references are gone. The reference count is
> > > treated like a rwsem using lockdep instrumentation code.
> >
> > Which kernel is this? I don't see any lockdep annotation around
> > kn->count (doing a git grep, I find it referenced in fs/kernfs/dir.c)
>
> As far as I know the s_active lockdep annotations were introduced in 2007
> through commit 0ab66088c855 ("sysfs: implement sysfs_dirent active reference
> and immediate disconnect"). Today these annotations exist in kernfs:
> $ git grep -nHw dep_map fs/kernfs
> fs/kernfs/dir.c:421: rwsem_acquire_read(&kn->dep_map, 0, 1, _RET_IP_);
> fs/kernfs/dir.c:441: rwsem_release(&kn->dep_map, 1, _RET_IP_);
> fs/kernfs/dir.c:468: rwsem_acquire(&kn->dep_map, 0, 0, _RET_IP_);
> fs/kernfs/dir.c:470: lock_contended(&kn->dep_map, _RET_IP_);
> fs/kernfs/dir.c:478: lock_acquired(&kn->dep_map, _RET_IP_);
> fs/kernfs/dir.c:479: rwsem_release(&kn->dep_map, 1, _RET_IP_);
> fs/kernfs/dir.c:1385: rwsem_acquire(&kn->dep_map, 0, 1, _RET_IP_);
> fs/kernfs/file.c:1000: lockdep_init_map(&kn->dep_map, "s_active", key, 0);
>
Ah thanks. I was searching for the wrong thing. :-/
-- Steve
