qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "G Portokalidis" <georgios.portokalidis@gmail.com>
To: qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] qemu and kernel 2.6.18
Date: Tue, 17 Oct 2006 14:29:28 +0200	[thread overview]
Message-ID: <ef735050610170529u72166656ved0f15e41818c1a6@mail.gmail.com> (raw)
In-Reply-To: <20061016103648.8f2dfe23.chris@friedhoff.org>

[-- Attachment #1: Type: text/plain, Size: 2832 bytes --]

I thought the best way to overcome the restriction imposed in tun/tap
interfaces is to set qemu as suid, and revoke privileges as soon as
the network interfaces are configured, and before any virtual block
devices are opened.

I wrote this little patch, which hopefully does just that.

Cheers,
George


On 16/10/06, chris friedhoff <chris@friedhoff.org> wrote:
> Hi,
>
> checking the Changelog for 2.6.18 (and diffing) one can see, that the CAP_NET_ADMIN requirement was added for the tun/tap inerface in tun.c. The question is, is it acceptable for a user to add a tun/tap interface in a running system. It was before 2.6.18. A different approach is, to grant the user or the process the CAP_NET_ADMIN capabillity.
> If the user of the system running qemu is in control of the system, he might start qemu as root. The tun/tap interface is created (due to root-rights), but i wouldn't like to see windows running in a root started qemu.
> Running qemu with kqemu, you need to be able to modprobe kqemu in a root context. If a user has the right to modprobe a module, he can do anything. What might be critical is that a user-context started app might create also a tun/tap interface (for evil reason) without the knowledge of the user. But than one has to ask, what kind of software is he running.
> But nevertheless, if you patch the kernel, you have to know what you do ...
>
> The changelog: http://lwn.net/Articles/190305/ (search for tuntap)
>
> Chris
>
> ########################################
>
> On Sun, 15 Oct 2006 15:31:11 +0800
> Tace <tacetan@gmail.com> wrote:
>
> > Hi,
> >     That might be some security issues with removal of that capability
> > check. I think it is not a good idea to remove it.
> >
> > 2006/10/14, chris friedhoff <chris@friedhoff.org>:
> > > Hello,
> > >
> > > bringing up the tun/tap interface depends now on the capability CAP_NET_ADMIN, which usually only root has.
> > > This patch just removes this dependency, so normal user rights suffices again to bring up the tun/tap interface.
> > >
> > > diff -ruN linux-2.6.18-orig/drivers/net/tun.c linux-2.6.18/drivers/net/tun.c
> > > --- linux-2.6.18-orig/drivers/net/tun.c 2006-09-20 05:42:06.000000000 +0200
> > > +++ linux-2.6.18/drivers/net/tun.c      2006-10-02 09:21:52.000000000 +0200
> > > @@ -489,9 +489,6 @@
> > >
> > >                 err = -EINVAL;
> > >
> > > -               if (!capable(CAP_NET_ADMIN))
> > > -                       return -EPERM;
> > > -
> > >                 /* Set dev type */
> > >                 if (ifr->ifr_flags & IFF_TUN) {
> > >                         /* TUN device */
> > >
>
> --------------------
> Chris Friedhoff
> chris@friedhoff.org
>
>
> _______________________________________________
> Qemu-devel mailing list
> Qemu-devel@nongnu.org
> http://lists.nongnu.org/mailman/listinfo/qemu-devel
>

[-- Attachment #2: qemu-0.8.2-linux-2.6.18.patch --]
[-- Type: application/octet-stream, Size: 875 bytes --]

diff -Nuar qemu-0.8.2/Makefile.target qemu-0.8.2-linux-2.6.18/Makefile.target
--- qemu-0.8.2/Makefile.target	2006-07-22 19:23:34.000000000 +0200
+++ qemu-0.8.2-linux-2.6.18/Makefile.target	2006-10-17 14:20:41.000000000 +0200
@@ -530,7 +530,7 @@
 
 install: all 
 ifneq ($(PROGS),)
-	$(INSTALL) -m 755 -s $(PROGS) "$(DESTDIR)$(bindir)"
+	$(INSTALL) -m 4755 -s $(PROGS) "$(DESTDIR)$(bindir)"
 endif
 
 ifneq ($(wildcard .depend),)
diff -Nuar qemu-0.8.2/vl.c qemu-0.8.2-linux-2.6.18/vl.c
--- qemu-0.8.2/vl.c	2006-07-22 19:23:34.000000000 +0200
+++ qemu-0.8.2-linux-2.6.18/vl.c	2006-10-17 14:14:49.000000000 +0200
@@ -6069,6 +6069,10 @@
             exit(1);
     }
 
+    /* Revoke privileges here. Real user still needs to have permissions on
+     * block devs */
+    setuid(getuid());
+
     /* init the memory */
     phys_ram_size = ram_size + vga_ram_size + bios_size;
 

  reply	other threads:[~2006-10-17 12:29 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-10-13 15:31 [Qemu-devel] qemu and kernel 2.6.18 G Portokalidis
2006-10-13 17:00 ` WaxDragon
2006-10-14  7:47   ` chris friedhoff
2006-10-15  7:31     ` Tace
2006-10-16  8:36       ` chris friedhoff
2006-10-17 12:29         ` G Portokalidis [this message]
2006-11-06 16:20           ` Chris Friedhoff

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=ef735050610170529u72166656ved0f15e41818c1a6@mail.gmail.com \
    --to=georgios.portokalidis@gmail.com \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).