Re: CVE request: Transmission can be made to crash remotely

[Jan Lieskovsky <jlieskov () redhat com>]

Hello Yves-Alexis,

  to follow up on this one. The source of the issue
seems to be underlying libutp code:
[1] https://trac.transmissionbt.com/ticket/5002#comment:22

more specifically the way how libutp (previously) handled
selective acknowledgements, which resulted in following two
(libutp) patches:
[2] https://github.com/bittorrent/libutp/issues/38
[3] https://github.com/bittorrent/libutp/issues/37

Transmission upstream corrected this issue in v2.74:
[4] https://trac.transmissionbt.com/query?milestone=2.74&group=component&order=severity

with the following patch:
[5] https://trac.transmissionbt.com/changeset/13646

Ad assigning CVE ids - I think one CVE id is enough.
The problem is in libutp code, and Transmission upstream
seems to commit their own change only due to libutp
(un)responsiveness:
[6] https://trac.transmissionbt.com/ticket/5002#comment:32

Thank you && Regards, Jan.
--
Jan iankko Lieskovsky / Red Hat Security Response Team

P.S.: All the links from above at one place are at:
      [7] https://bugzilla.redhat.com/show_bug.cgi?id=909934

----- Original Message -----
On dim., 2013-02-10 at 11:50 +0100, Josselin Mouette wrote:
> Package: transmission-daemon
> Version: 2.52-3
> Severity: grave
> Tags: security patch upstream
> Justification: user security hole
>
> The transmission-daemon package in wheezy crashes regularly. According 
> to upstream this is a remote security hole (at least a remote DoS, but 
> most probably there is a way to take control of the process).
>
> https://trac.transmissionbt.com/ticket/5044
> https://trac.transmissionbt.com/ticket/5002
>
> Apparently there is no CVE assigned. The bug is fixed upstream and I’m 
> attaching the patch. I’m currently testing a patched package, and will 
> report whether the fix is sufficient.
>
> Could a CVE be assigned for this?
>
> Thanks in advance,
> -- 
> Yves-Alexis