[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <17596.53679.608190.420651@hound.rchland.ibm.com>
Date: Tue, 18 Jul 2006 07:18:55 -0500
From: boutcher@...umn.edu (Dave Boutcher)
To: Arjan van de Ven <arjan@...radead.org>
Cc: Keir Fraser <Keir.Fraser@...cam.ac.uk>,
Ian Pratt <ian.pratt@...source.com>,
Jeremy Fitzhardinge <jeremy@...p.org>,
xen-devel@...ts.xensource.com, linux-kernel@...r.kernel.org,
virtualization@...ts.osdl.org, Andrew Morton <akpm@...l.org>,
netdev@...r.kernel.org, Chris Wright <chrisw@...s-sol.org>
Subject: Re: [RFC PATCH 32/33] Add the Xen virtual network device driver.
On Tue, 18 Jul 2006 12:42:28 +0200, Arjan van de Ven <arjan@...radead.org> said:
>
> On Tue, 2006-07-18 at 11:35 +0100, Keir Fraser wrote:
>> On 18 Jul 2006, at 11:27, Arjan van de Ven wrote:
>>
>> > Hmmm maybe it's me, but something bugs me if a NIC driver is going to
>> > send IP level ARP packets... that just feels very very wrong and is a
>> > blatant layering violation.... shouldn't the ifup/ifconfig scripts just
>> > be fixed instead if this is critical behavior?
>>
>> Maybe we should be faking this out from our hotplug scripts in the
>> control VM, although triggering this from user space is probably a bit
>> of a pain. Regardless, the function can be removed from the driver if
>> it's too distasteful: it's only a performance 'hack' to get network
>> traffic more quickly redirected when migrating a VM between physical
>> hosts. Things won't break horribly if it's removed.
>
>
> that sounds like a perfect job for a udev-kind of notification to
> userspace, eg "hey I migrated", and then udev/hotplug can just do this
> sending of arps.. maybe they also want to do other things just after a
> migration (which sounds entirely reasonable to me) such as renewing the
> dhcp lease, resetting the periodic userspace hardware polling timers in
> "hal", network manager etc etc...
I think we had this discussion the last time these patches came
around. Other virtual environments (VMWare, POWER) have the same
issue, and there are probably other scenarios (bonding?) that could
use the same function. This doesn't belong in the driver.
Dave B
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists