linux-kernel - Re: module loading with CAP_NET

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <1298565265.2613.16.camel@bwh-desktop>
Date:	Thu, 24 Feb 2011 16:34:25 +0000
From:	Ben Hutchings <bhutchings@...arflare.com>
To:	Vasiliy Kulikov <segoon@...nwall.com>
Cc:	netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
	Kees Cook <kees.cook@...onical.com>,
	Eugene Teo <eugene@...hat.com>,
	Dan Rosenberg <dan.j.rosenberg@...il.com>,
	"David S. Miller" <davem@...emloft.net>
Subject: Re: module loading with CAP_NET_ADMIN

On Thu, 2011-02-24 at 18:12 +0300, Vasiliy Kulikov wrote:
> Hi netdev folks,
> 
> I'd like to discuss the ability to load any modules from /lib/modules/
> by a process with CAP_NET_ADMIN.  Since Linux 2.6.32 [1] there is such
> possibility:
> 
> root@...atros:~# grep Cap /proc/$$/status
> CapInh: 0000000000000000
> CapPrm: fffffffc00001000
> CapEff: fffffffc00001000
> CapBnd: fffffffc00001000
> root@...atros:~# lsmod | grep xfs
> root@...atros:~# ifconfig xfs
> xfs: error fetching interface information: Device not found
> root@...atros:~# lsmod | grep xfs
> xfs                   767011  0 
> exportfs                4226  2 xfs,nfsd

Eek!

> Ability of CAP_NET_ADMIN to load the driver to work with a particular
> network device is rational;  however, one may load any module not even
> related to network this way.  Hopefully, this is not equal to
> CAP_SYS_MODULE since the module set is restricted to /lib/modules
> (additionally may be disabled with /proc/sys/kernel/modules_disabled),
> but the idea of non-netdev module loading is weird.
> 
> 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.

AFAIK these interface-name aliases are usually defined by distribution
configuration files rather than within the modules themselves.  And that
behaviour is pretty much obsolete now that we have hotplug and udev.

> This would allow to load only netdev
> modules via these ioctls.  I'm not sure what modules should be patches -
> at least real physical netdevices have names different from drivers'
> names, so they don't need patching.  I suppose the list is not big.

The only modules I can see that declare aliases like this are:

net/ipv4/ip_gre.c:MODULE_ALIAS("gre0");
net/ipv4/ipip.c:MODULE_ALIAS("tunl0");
net/ipv6/sit.c:MODULE_ALIAS("sit0");

Ben.

-- 
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/