Re: Resend: [PATCH] DCCP: Divide by zero fix

["Ian McDonald" <imcdnzl@gmail.com>]

On 3/1/06, Arnaldo Carvalho de Melo <acme@ghostprotocols.net> wrote:
> On 2/28/06, Ian McDonald <imcdnzl@gmail.com> wrote:
> > Arnaldo,
> >
> > Haven't seen this going in so resending...
>
> What a coincidence, I'm testing this patch right now, and I'm afraid it has
> some bug or makes it more likely that some other bug happens, the machine
> soft locks with it enabled and when using CCID3 as the default RX CCID in
> a TTCP test over loopback.
>
> Note that the problem happens more quickly if you test it on a machine with
> less than 64 MB (where the problem happens, but after more time).
>
OK. I am doing some testing on lockups at present from CCID3 which I
suspect are introduced recently.... they don't lock up on older code.
Having said that I'm running without that patch applied so I will
retest soon - but I don't believe it is the problem when I looked at
the scope of that code.

I have a suspicion that if I don't find anything with git-bisect for
CCID3 lockups that it might be when you use CCID3 and set
DCCP_INITIAL_SEND_ACK_VECTOR to 1....

Anyway another 3 or 4 reboots and I should give some more answers.

BTW my problems occur when using a netem box in the middle as per
setup described on DCCP testing on ODSL Wiki.

Ian
--
Ian McDonald
Web: http://wand.net.nz/~iam4
Blog: http://imcdnzl.blogspot.com
WAND Network Research Group
Department of Computer Science
University of Waikato
New Zealand