[Ns-developers] Ns-3 RC Daily Bug Status for Saturday June 21, 2008

[Raj Bhattacharjea]

On Sun, Jun 22, 2008 at 12:25 AM, Mathieu Lacage
<mathieu.lacage at sophia.inria.fr> wrote:
> On Sat, 2008-06-21 at 20:47 -0700, craigdo at ee.washington.edu wrote:
>
>> The "Evensky test program," tcp-2way.cc, is turning out to be one of
>> those
>> gifts that keeps on giving.  It has uncovered at least four separate
>> problems in the ns-3 code with perhaps two remaining problems pending:
>> a
>> lost socket address tag on a one-byte packet, and a relatively bizarre
>
> If you see a lost tag on a one-byte packet, it might have been caused by
> my previous fixes to the tag list: I was somewhat worried about such a
> potential problem but I was not able to come up with a testcase which
> exhibited the problem. Can you point me to a testcase ?
>
>> looking TCP header that Mathieu observed today.
>>
>> tcp-2way:HandleRead(Server(old)|, 0x8075128)
>> ns3::SocketAddressTag [0-500] address=02-06-0a:01:00:01:01:c0Packet
>> Dump: ns3::TcpHeader (17476 > 17476 [ RST ] Seq=1145324612
>> Ack=1145324612 Win=17476) Payload Fragment [46:546]
>>
>> Mathieu is looking into the remaining problems.  Raj, have you seen anything
>> like the TCP header in question?  It looks like a bunch of hexadecimal 4.
>
> That bug was a metadata problem and fixing it avoided the following
> crasher (I somewhat expected that) because the crasher was caused by
> believing the metadata while it was wrong.

I haven't seen anything like this in TCP, and Mathieu...is this
particular issue resolved then as a metadata error?


-- 
Raj Bhattacharjea
Georgia Institute of Technology
School of Electrical and Computer Engineering
Systems Analyst
404.894.2955