netdev - Re: [PATCH] dl2k: Tighten ioctl permissions

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 for Android: free password hash cracker in your pocket

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

Message-Id: <20120425.182954.804210541527871356.davem@davemloft.net>
Date:	Wed, 25 Apr 2012 18:29:54 -0400 (EDT)
From:	David Miller <davem@...emloft.net>
To:	bhutchings@...arflare.com
Cc:	jeffm@...e.com, netdev@...r.kernel.org
Subject: Re: [PATCH] dl2k: Tighten ioctl permissions

From: Ben Hutchings <bhutchings@...arflare.com>
Date: Wed, 25 Apr 2012 23:26:42 +0100

> On Wed, 2012-04-25 at 15:33 -0400, Jeff Mahoney wrote:
>> dl2k's rio_ioctl function defines several ioctls that involve
>>  operations that should be denied to regular users.
>> 
>>  SIOCDEVPRIVATE + 2 is a renumbered SIOCSMIIREG.
> 
> There was an early convention that SIOCDEVPRIVATE + {0,1,2} were MDIO
> operations.  (This was a bad idea, because you can't safely send them to
> an arbitrary driver... not that that stopped people doing it.  Now it's
> neither safe to send them from userland, nor to implement any other
> semantics for these ioctl numbers in a driver.)
> 
> Please fix the numbering instead; it will make standard MII/MDIO
> utilities work and the capability check (in dev_ioctl()) comes for free.
> 
>>  SIOCDEVPRIVATE + 5 calls netif_stop_queue.
>>  SIOCDEVPRIVATE + 6 calls netif_wake_queue.
> [...]
> 
> And SIOCDEVPRIVATE + {7,8} spam the kernel log, so they should perhaps
> be considered privileged too.

And I would also say that the netif_{stop,wake}_queue ones should just
be deleted.  There is no sane way, even as a debugging facility, we
can let useland trigger these conditions.

And if we could, it belongs in a generic facility no private ioctls
which are heavily discouraged anyways.

I seriously would suggest ditching all of these private ioctls from
the dl2k driver, I bet no binary in the world out there even exists
which handles these strange renumbered MII operations.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html