[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20131224150543.66581c13@nehalam.linuxnetplumber.net>
Date: Tue, 24 Dec 2013 15:05:43 -0800
From: Stephen Hemminger <stephen@...workplumber.org>
To: Eric Dumazet <eric.dumazet@...il.com>
Cc: Sathya Perla <sathya.perla@...lex.com>, netdev@...r.kernel.org,
edumazet@...gle.com
Subject: Re: [PATCH net-next] vxlan: distribute vxlan tunneled traffic
across multiple TXQs
On Tue, 24 Dec 2013 10:39:06 -0800
Eric Dumazet <eric.dumazet@...il.com> wrote:
> On Mon, 2013-12-23 at 11:28 -0800, Stephen Hemminger wrote:
>
> > The idea is good, but without the destructor there is nothing to keep
> > the UDP socket from being destroyed while packet is being sent on another
> > CPU.
>
> I see no requirement of holding a reference on the vxlan UDP socket in
> transmit path.
>
> At the time vxlan_set_owner() is called, nothing requires access to the
> socket anymore. If you believe its needed, then its already too late.
>
> Sathya, your patch is a step in the right direction, but the skb_clone()
> thing should be done a bit differently.
>
> The trick would be to avoid the skb_clone() for the last
> vxlan_xmit_one() call.
>
That code was cloned from L2TP, I assumed that that it was required
to hold reference when calling UDP, since that is what socket layer does.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists