Re: CVE-NONE kernel: PHONET signedness issue

[Michael Gilbert <michael.s.gilbert () gmail com>]

On Thu, 6 Jan 2011 13:08:59 -0500, Dan Rosenberg wrote:
> This is a slippery slope.  I'm in favor of not having a CVE assigned
> for this issue.
>
> Otherwise, wouldn't we need a CVE for every vector that allows
> transitioning from various capabilities to root?  The capability
> system may be poorly designed to allow such transitions, but I don't
> think they represent unexpected behavior.

What's the point of a capabilities system if its equivalent to root
in the majority of cases anyway?  For file access/operations, there is
always sudo and the /etc/sudoers file for making it easy to access to
stuff thats accessed often without a password.  For port binding, the
capabilities system makes sense; and according to Brad Spengler's list,
those caps don't appear to be root equivalent so that could stay.
Otherwise, I don't see the point.

I'm not sure if there is a written security model for the capabilities
system, but this looks to me like it would be a violation of it.

Best wishes,
Mike