[Olsr-users] Received message to big to be forwarded

Henning Rogge (spam-protected)
Fri Apr 17 17:17:03 CEST 2015


No,

it will not... as far as I understand the source the "message
forwarding" function does not call this function at all, it just
pushes the message data, which in turn directly looks into the
remaining bytes.

So this patch should restrict message size, but not packet size for forwarding.

Henning

On Fri, Apr 17, 2015 at 5:15 PM, Ferry Huberts <(spam-protected)> wrote:
>
>
> On 17/04/15 16:41, Henning Rogge wrote:
>>
>> Jernej,
>>
>> can you test the following patch?
>>
>> diff --git a/src/net_olsr.c b/src/net_olsr.c
>> index afd24ec..4db54b1 100644
>> --- a/src/net_olsr.c
>> +++ b/src/net_olsr.c
>> @@ -258,7 +258,14 @@ net_outbuffer_push_reserved(struct interface_olsr
>> *ifp, const void *data, const
>>   int
>>   net_outbuffer_bytes_left(const struct interface_olsr *ifp)
>>   {
>> -  return ifp->netbuf.maxsize - ifp->netbuf.pending;
>> +  /* IPv6 minimum MTU - IPv6 header - UDP header - VLAN-Tag */
>> +  static int MAX_REMAINING = 1280 - 40 - 8 - 4;
>> +  int remaining = ifp->netbuf.maxsize - ifp->netbuf.pending;
>> +
>> +  if (remaining > MAX_REMAINING) {
>> +    return MAX_REMAINING;
>> +  }
>> +  return remaining;
>>   }
>>
>
> I don't fully agree with this.
> I see why you do it this way but it feels like a band-aid instead of a real
> fix.
>
> This will limit the packet size by about 15% causing more overhead and
> significant more media access for the same traffic volume.
>
>
>>
>>
>> Henning
>>
>> On Fri, Apr 17, 2015 at 4:32 PM, Henning Rogge <(spam-protected)> wrote:
>>>
>>> Sure,
>>>
>>> but that is not really a good idea for UDP Multicast... especially in
>>> networks with some amount of multicast packet loss.
>>>
>>> Henning
>>>
>>> On Fri, Apr 17, 2015 at 4:31 PM, Teco Boot <(spam-protected)> wrote:
>>>>
>>>> Clearing the DF bit enables IP stack fragmenting.
>>>>
>>>> Teco
>>>>
>>>>> Op 17 apr. 2015, om 16:08 heeft Henning Rogge <(spam-protected)> het
>>>>> volgende geschreven:
>>>>>
>>>>> That is what I mentioned earlier... limiting the maximum size of a
>>>>> message independent of the packet/MTU size.
>>>>>
>>>>> Henning
>>>>>
>>>>> On Fri, Apr 17, 2015 at 4:07 PM, Ferry Huberts <(spam-protected)>
>>>>> wrote:
>>>>>>
>>>>>>
>>>>>>
>>>>>> On 17/04/15 16:04, Henning Rogge wrote:
>>>>>>>
>>>>>>>
>>>>>>> Yes,
>>>>>>>
>>>>>>> but the fix does not handle mesh networks with different MTUs.
>>>>>>>
>>>>>>> the 160 neighbor nodes produces a TC as large as it can... according
>>>>>>> to the local MTU. when this message is forwarded (within UDP and olsr
>>>>>>> packets) it reaches a node with a link where the MTU is small enough
>>>>>>> that the message does not fit through.
>>>>>>>
>>>>>>
>>>>>> Which is what I meant earlier: olsrd should split the packet.
>>>>>>
>>>>>>
>>>>>>
>>>>>>> Henning
>>>>>>>
>>>>>>> On Fri, Apr 17, 2015 at 3:16 PM, Ferry Huberts <(spam-protected)>
>>>>>>> wrote:
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> On 17/04/15 15:11, Jernej Kos wrote:
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Hello!
>>>>>>>>>
>>>>>>>>> On 17. 04. 2015 15:08, Ferry Huberts wrote:
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> Which version of olsrd are you running?
>>>>>>>>>> The 160+ number is suspiciously like a problem that Henning fixed
>>>>>>>>>> a
>>>>>>>>>> while ago.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> We are running olsrd version 0.6.7.1.
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> That includes Henning's fix.
>>>>>>
>>>>>>
>>>>>>
>>>>>> --
>>>>>> Ferry Huberts
>>>>>
>>>>>
>>>>> --
>>>>> Olsr-users mailing list
>>>>> (spam-protected)
>>>>> https://lists.olsr.org/mailman/listinfo/olsr-users
>>>>
>>>>
>
> --
> Ferry Huberts




More information about the Olsr-users mailing list