Re: [Lksctp-developers] miscomputation of sack chunk size

[Sridhar S. <sr...@us...>]

good catch.
will integrate along with your other patches.
-Sridhar


On Thu, 22 Jul 2004, Jorge Hernandez-Herrero wrote:

> Hi,
>
> There is a '+' missing in sctp_sm_pull_sack() when
> trying to account to for the variable length of the
> SACK chunk (due to the reported number of dup TSN and
> gap acks) which results in skb_pull() being
> called with the wrong length. I guess, it could cause
> some bad side effects when processing inbound SCTP
> packets that bundle SACK+DATA.
>
> --- lksctp-2.4.work/net/sctp/sm_statefuns.c     2004-06-07
> 18:19:45.000000000 -0500
> +++ PATCH_SACKSIZE/net/sctp/sm_statefuns.c      2004-07-22
> 08:04:29.000000000 -0500
> @@ -4709,7 +4709,7 @@ struct sctp_sackhdr *sctp_sm_pull_sack(s
>         num_blocks = ntohs(sack->num_gap_ack_blocks);
>         num_dup_tsns = ntohs(sack->num_dup_tsns);
>         len = sizeof(struct sctp_sackhdr);
> -       len = (num_blocks + num_dup_tsns) * sizeof(__u32);
> +       len += (num_blocks + num_dup_tsns) * sizeof(__u32);
>         if (len > chunk->skb->len)
>                 return NULL;
>
> I opened a new bug report and uploaded this patch in sourceforge
> for this problem.
>
> Please take a look at it.
>
> Thanks,
> Jorge
>
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by BEA Weblogic Workshop
> FREE Java Enterprise J2EE developer tools!
> Get your free copy of BEA WebLogic Workshop 8.1 today.
> http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click
> _______________________________________________
> Lksctp-developers mailing list
> Lks...@li...
> https://lists.sourceforge.net/lists/listinfo/lksctp-developers
>