On Wed, Jul 05, 2006 at 11:55:32AM +0900, YAMAMOTO Takashi wrote:
> > > > @@ -616,8 +617,9 @@ dounmount(struct mount *mp, int flags, s
> > > >  	if (LIST_FIRST(&mp->mnt_vnodelist) != NULL)
> > > >  		panic("unmount: dangling vnode");
> > > >  	mp->mnt_iflag |= IMNT_GONE;
> > > >  	lockmgr(&mp->mnt_lock, LK_RELEASE | LK_INTERLOCK, &mountlist_slock);
> > > > +	lockmgr(&mp->mnt_trans_lock, LK_DRAIN, NULL);
> > > 
> > > is this LK_DRAIN necessary?
> > 
> > The lockmgr man page states this as the right way to free memory
> > containing a lock.
> 
> can there be lwps waiting for or holding this lock at this point?
> iirc, LK_DRAIN is merely another kind of exclusive lock.

Depends on the gates we set.  I suppose we may have lwps holding the
lock during a forced unmount.  Draining before free will wait until
they have finished their transaction and error out new transactions
trying to get the lock.

> > > > +int
> > > > +vn_trans_lock(struct mount *mp, int flags)
> > > > +{
> > > > +	int error, pflags;
> > > > +	struct vn_trans_entry *vte, *new_vte;
> > > > +
> > > > +	KASSERT((flags & ~(LK_SHARED|LK_EXCLUSIVE|LK_NOWAIT)) == 0);
> > > > +
> > > > +	if (mp == NULL)
> > > > +		return 0;
> > > 
> > > when mp can be NULL?
> > 
> > At least on startup (cpu_initialize).
> 
> i couldn't find "cpu_initialize" other then sparc64 locore one.
> is it what you meant?

Yes. I meant the startup where we have vnodes with "vp->v_mount == NULL".

-- 
Juergen Hannken-Illjes - hannken@eis.cs.tu-bs.de - TU Braunschweig (Germany)