Freeing vnodes.

[Jeff Roberson]

On Tue, 29 Mar 2005, Jeff Roberson wrote:

> On Tue, 29 Mar 2005, David Schultz wrote:
>
> > On Tue, Mar 29, 2005, Jeff Roberson wrote:
> > > On Tue, 29 Mar 2005, Jeff Roberson wrote:
> > >
> > > > On Mon, 28 Mar 2005, David Schultz wrote:
> > > >
> > > > > On Mon, Mar 28, 2005, Jeff Roberson wrote:
> > > > > > > > I am worried about the v_dd,v_ddid fields of a directory B that has the
> > > > > > > > to be released vnode A as parent. (Obviously in this case there is no
> > > > > > > > namecache entry with the vnode A as the directory (nc_dvp))
> > > > > > > >
> > > > > > > > Right now A is type stable - but if A is released, access to B->v_dd
> > > > > > > > may cause a page fault.
> > > > > > > >
> > > > > > > > Stephan
> > > > > > >
> > > > > > > Jeff,
> > > > > > >
> > > > > > > Do you plan to address the problem now that the code is checked in?
> > > > > >
> > > > > > Vnodes with children in the name cache are held with vhold() and not
> > > > > > recycled.
> > > > >
> > > > > Yes, but cache_purge() is called directly in a number of places
> > > > > where the vnode may have children, e.g. in mount.  So dangling
> > > > > references might still be possible unless cache_purge() fixes up
> > > > > the children's v_dd pointers appropriately.
> > > > >
> > > >
> > > > ah, indeed.  How does this look:
> > >
> > > Also, are the ids really necessary now that we don't reuse vnodes?
> > > Shouldn't the pointer be sufficient?
> >
> > I think so.  The patch I sent you a few days ago gets rid of v_id
> > except in vfs_cache_lookup(), where it is used to guarantee that
> > the vnode hasn't changed while sleeping in vn_lock().  With vnode
> > reclamation, that isn't safe anyway, so if you fix vfs_cache_lookup(),
> > we can kill v_id completely.
>
> You're right, cache_lookup() needs to be changed to return a referenced
> vnode.  There are only a few callers outside of vfs_cache.c that I'll have
> to change.  I'll put this on my todo list.  After that v_id can go away.
> I'll look at your patch again soon.

I fixed this.  You should wrap vn_fullpath() with Giant and then commit
your patch.  I'll do the VFS_LOCK_GIANT() bit.  Do you think we can get
rid of v_id and v_ddid after you commit your patch?

>
> >
> > Your patch looks okay at a glance, but shouldn't you be iterating over
> > v_cache_src instead of v_cache_dst?
> >
>
> Yes, I thought I double checked that.  I haven't done any name cache stuff
> in a while, and I always get confused whether src means sources for this
> vnode, or names this vnode is a source for.
> _______________________________________________
> freebsd-arch at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-arch
> To unsubscribe, send any mail to "freebsd-arch-unsubscribe at freebsd.org"
>