Review Request 119011: KInit: call setgroups(0, 0) before calling setgid()

[Thiago Macieira]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/119011/#review88231
-----------------------------------------------------------


This broke the ability of users to have more than one group (usermod), for groups like vboxusers and systemd-journal. Now, start_kdeinit unconditionally drops all groups and that's wrong.

It should call getgrouplist(3) and set those groups on the user.

Besides, I'm not convinced the rpmlint warning was correct.

- Thiago Macieira


On Julho 1, 2014, 10:21 a.m., Daniel Vrátil wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://git.reviewboard.kde.org/r/119011/
> -----------------------------------------------------------
> 
> (Updated Julho 1, 2014, 10:21 a.m.)
> 
> 
> Review request for KDE Frameworks.
> 
> 
> Repository: kinit
> 
> 
> Description
> -------
> 
> While packaging kinit, we got a warning from rpmlint that start_kdeinit calls setgid() without calling setgroups() first. From rpmlint:
> 
>    This executable is calling setuid and setgid without setgroups or initgroups.
>    There is a high probability this mean it didn't relinquish all groups, and
>    this would be a potential security issue to be fixed. Seek POS36-C on the web
>    for details about the problem.
> 
> The reasoning is that when you drop privileges from root to regular user, there might be some extra groups left that, if not cleared, might grant the process privileges to do superuser things.
> 
> The code does not check for return value, as the call will fail if we are not a superuser.
> 
> This oneliner makes rpmlint happy and maybe prevents a security issue.
> 
> 
> Diffs
> -----
> 
>   src/start_kdeinit/start_kdeinit.c 07a28d3 
> 
> Diff: https://git.reviewboard.kde.org/r/119011/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Daniel Vrátil
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-frameworks-devel/attachments/20151111/b9580232/attachment.html>