[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20180530182442.wbgke26kaoau2du2@salvia>
Date: Wed, 30 May 2018 20:24:42 +0200
From: Pablo Neira Ayuso <pablo@...filter.org>
To: "Jason A. Donenfeld" <Jason@...c4.com>
Cc: Felix Fietkau <nbd@....name>, netfilter-devel@...r.kernel.org,
Netdev <netdev@...r.kernel.org>, openwrt-devel@...ts.openwrt.org,
Jaap Buurman <jaapbuurman@...il.com>,
WireGuard mailing list <wireguard@...ts.zx2c4.com>
Subject: Re: Missing skb->dst with flow offloading
On Wed, May 30, 2018 at 08:14:42PM +0200, Jason A. Donenfeld wrote:
> Hey Pablo,
>
> On Wed, May 30, 2018 at 8:05 PM, Pablo Neira Ayuso <pablo@...filter.org> wrote:
> > If there a more drivers in-tree that need this, we may add
> > skb_dst_set_noref() calls to _hook function in the flowtable codebase.
>
> Can I, then, take that as an implicit acknowledgement that this
> observed behavior on OpenWRT is to be expected with the current state
> of events, and that I should patch my driver accordingly?
>
> As one example of this in tree, take a look at vxlan -- it's using it
> for the mtu/pmtu exactly as WireGuard does.
May it crash the kernel because it's assuming is set? If so, then
I'd appreciate if you send us a patch to
netfilter-devel@...r.kernel.org.
Please, use the nf-next.git tree to patch nf_flow_offload_ip_hook()
and nf_flow_offload_ip6_hook(), it's rather late, we'll request a
-stable submission for this if needed.
Thanks.
Powered by blists - more mailing lists