linux-kernel - Re: [PATCH 2/4] fs: define a firmware security filesystem named fwsecurityfs

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: <94fe007e8eab8bc7ae3f56b88ad94646b4673657.camel@HansenPartnership.com>
Date:   Mon, 21 Nov 2022 12:33:55 -0500
From:   James Bottomley <James.Bottomley@...senPartnership.com>
To:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Cc:     Nayna <nayna@...ux.vnet.ibm.com>, Nayna Jain <nayna@...ux.ibm.com>,
        linuxppc-dev@...ts.ozlabs.org, linux-fsdevel@...r.kernel.org,
        linux-efi@...r.kernel.org,
        linux-security-module <linux-security-module@...r.kernel.org>,
        linux-kernel@...r.kernel.org,
        Michael Ellerman <mpe@...erman.id.au>, npiggin@...il.com,
        christophe.leroy@...roup.eu, Dov Murik <dovmurik@...ux.ibm.com>,
        George Wilson <gcwilson@...ux.ibm.com>,
        Matthew Garrett <mjg59@...f.ucam.org>,
        Dave Hansen <dave.hansen@...el.com>,
        Benjamin Herrenschmidt <benh@...nel.crashing.org>,
        Paul Mackerras <paulus@...ba.org>,
        Russell Currey <ruscur@...sell.cc>,
        Andrew Donnellan <ajd@...ux.ibm.com>,
        Stefan Berger <stefanb@...ux.ibm.com>
Subject: Re: [PATCH 2/4] fs: define a firmware security filesystem named
 fwsecurityfs

On Mon, 2022-11-21 at 16:05 +0100, Greg Kroah-Hartman wrote:
> On Mon, Nov 21, 2022 at 09:03:18AM -0500, James Bottomley wrote:
> > On Mon, 2022-11-21 at 12:05 +0100, Greg Kroah-Hartman wrote:
> > > On Sun, Nov 20, 2022 at 10:14:26PM -0500, James Bottomley wrote:
[...]
> > > > I already explained in the email that sysfs contains APIs like
> > > > simple_pin_... which are completely inimical to namespacing.
> > > 
> > > Then how does the networking code handle the namespace stuff in
> > > sysfs? That seems to work today, or am I missing something?
> > 
> > have you actually tried?
> > 
> > jejb@...grow:~> sudo unshare --net bash
> > lingrow:/home/jejb # ls /sys/class/net/
> > lo  tun0  tun10  wlan0
> > lingrow:/home/jejb # ip link show
> > 1: lo: <LOOPBACK> mtu 65536 qdisc noop state DOWN mode DEFAULT
> > group
> > default qlen 1000
> >     link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
> > 
> > So, as you see, I've entered a network namespace and ip link shows
> > me the only interface I can see in that namespace (a down loopback)
> > but sysfs shows me every interface on the system outside the
> > namespace.
> 
> Then all of the code in include/kobject_ns.h is not being used?  We
> have a whole kobject namespace set up for networking, I just assumed
> they were using it.  If not, I'm all for ripping it out.

Hm, looking at the implementation, it seems to trigger off the
superblock (meaning you have to remount inside a mount namespace) and
it only works to control visibility in label based namespaces, so this
does actually work

jejb@...grow:~/git/linux> sudo unshare  --net --mount bash 
lingrow:/home/jejb # mount -t sysfs none /sys
lingrow:/home/jejb # ls /sys/class/net/
lo

The label based approach means that any given file can be shown in one
and only one namespace, which works for net, but not much else
(although it probably could be adapted).

James