svn commit: r194763 - in head/sys: conf dev/gem modules/gem

[Marius Strobl]

On Tue, Jun 23, 2009 at 05:03:20PM -0400, John Baldwin wrote:
> On Tuesday 23 June 2009 4:36:59 pm Marius Strobl wrote:
> > Author: marius
> > Date: Tue Jun 23 20:36:59 2009
> > New Revision: 194763
> > URL: http://svn.freebsd.org/changeset/base/194763
> > 
> > Log:
> >   - Initialize the ifnet structure, especially if_dname, before probing
> >     the PHYs as some PHY drivers use it (but probably shouldn't). How
> >     gem(4) has worked with brgphy(4) on powerpc without this so far is
> >     unclear to me.
> >   - Introduce a dying flag which is set during detach and checked in
> >     gem_ioctl() in order to prevent active BPF listeners to clear
> >     promiscuous mode which may lead to the tick callout being restarted
> >     which will trigger a panic once it's actually gone.
> 
> This should not be needed assuming you follow a model of:
> 
> gem_detach()
> {
> 
> 	ether_ifdetach(ifp);	/* calls bpfdetach() */
> 	GEM_LOCK(sc);
> 	gem_stop(sc);
> 	GEM_UNLOCK(sc);
> 	...
> }
> 
> If you are invoking gem_stop() prior to ether_ifdetach() then that is your 
> real bug. :)
> 

Okay, I'll let yongari@ comment on this as he was the one
who claimed that either clearing IFF_UP in the driver before
calling ether_ifdetach(9) (which I think is a layering
violation) or the committed approach is necessary in order
to solve the problem :)

Marius