[lkml]   [2011]   [Feb]   [25]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
    SubjectRe: module loading with CAP_NET_ADMIN
    On Fri, Feb 25, 2011 at 18:29 +0300, Michael Tokarev wrote:
    > 25.02.2011 15:30, Vasiliy Kulikov wrote:
    > > On Thu, Feb 24, 2011 at 16:34 +0000, Ben Hutchings wrote:
    > >> On Thu, 2011-02-24 at 18:12 +0300, Vasiliy Kulikov wrote:
    > >>> My proposal is changing request_module("%s", name) to something like
    > >>> request_module("netdev-%s", name) inside of dev_load() and adding
    > >>> aliases to related drivers.
    > It is not the kernel patching which we should worry about, kernel
    > part is trivial.
    > What is not trivial is to patch all the systems out there who
    > autoloads network drivers based on /etc/modprobe.d/network-aliases.conf
    > (some local file), ie, numerous working setups which already
    > uses this mechanism since stone age. And patching these is
    > not trivial at all, unfortunately.
    > Somewhat weird setups (one can load the modules explicitly, and
    > nowadays this all is handled by udev anyway), but this change
    > will break some working systems.
    > Maybe the cost (some pain for some users) isn't large enough
    > but the outcome is good, and I think it _is_ good, but it needs
    > some wider discussion first, imho.
    > I can't think of a way to handle this without breaking stuff.

    Currently Linux slowly moves in the direction of rootless systems. This
    definitely need proper restrictions of CAP_* power. Network admin does
    nothing with general modules. It _has_ to break something one day
    because old assumptions about permission stuff don't conform CAP_*
    things: old assumptions are very closely connected with just everything.

    I'm not sure how this particular CAP_NET_ADMIN misuse should be fixed,
    maybe distributions should supply script to upgrade modprobe configs.
    Also note that change s/CAP_SYS_MODULE/CAP_NET_ADMIN/ was made in
    2.6.32, so there is a possibility that the set of affected distributions
    (that doesn't use udev stuff) is very small.

    Thanks for your input,

    Vasiliy Kulikov - bringing security into open computing environments

     \ /
      Last update: 2011-02-25 16:59    [W:0.022 / U:8.184 seconds]
    ©2003-2017 Jasper Spaans. hosted at Digital OceanAdvertise on this site