On 10/31/2014, 11:12 AM, Florian Westphal wrote:
> Jiri Slaby <jslaby@xxxxxxx> wrote:
>>> I recommend to revert fe6cc55f3a9a ('net: ip, ipv6: handle gso skbs in
>>> forwarding path'), its 91e97e5763b4f0425ad90df39a23cdbd72671a7e in
>>> 3.12.y tree instead.
>>
>> But what about the ones below, which "Fixes fe6cc55" too? I am not sure
>> I want to revert all those. What are the consequences, if I skip only
>> this one (net: ip: push gso skb forwarding handling down the stack), please?
>
> I am not aware of any problem except the one in the commit log, i.e.
> udp UFO (udp offload) coming from a VM will be segmented in forward
> path which means the ip fragments will be pushed through the netfilter
> postrouting hooks individually (if the mtu is too small, else we won't
> segment).
>
> We did have a lot of other problems with fe6cc55f3a9a backport in the
> earlier kernels (3.2, 3.4), but afaics 3.12 is recent enough to not give
> more surprises.
>
> So perhaps the best solution is to just not put the pushdown into
> 3.12.y and otherwise leave it alone.
Ok, I dropped it. Thanks!
--
js
suse labs
--
To unsubscribe from this list: send the line "unsubscribe stable" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html