[ofa-general] Re: [PATCH v3 08/13] QLogic VNIC: sysfs interface implementation for the driver

[David Dillow]

On Thu, 2008-06-05 at 09:38 -0700, Roland Dreier wrote:
> > That said, given that SRP's been using sysfs since it went in, is there
>  > a reason to move to configfs other than it's the new preferred way to do
>  > it? Given the desire to not break ABI's -- and IIRC sysfs was declared to
>  > be under that unbrella -- wouldn't we have to at least carry both
>  > interfaces for a while, assuming we can even get rid of the sysfs one?
> 
> Yes, we'd definitely be carrying both interfaces for at least a year.
> 
> Looking further into this, I'm not sure it makes much sense either.
> Another problem with configfs is that the lifetime of the object is
> controlled by userspace.  So if we lose a connection to a target,
> the object will persist in configfs until userspace notices.

Yep. Though I'd like to be able to have a persistent connection and
retry capability in the kernel, so that I don't need to have srp_daemon,
and the persistence would be fine in that case. I've not fully convinced
myself that doing a persistent connection in the kernel is a good idea,
but a deep background task is trying to do the OFED patches to the
initiator in a way that is acceptable for inclusion in mainline, and
they would easily fall out of that work.
-- 
Dave Dillow
National Center for Computational Science
Oak Ridge National Laboratory
(865) 241-6602 office