netdev - Re: [PATCH] unix: properly account for FDs passed over unix sockets

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: <20151230131447.453c7d53@lxorguk.ukuu.org.uk>
Date:	Wed, 30 Dec 2015 13:14:47 +0000
From:	One Thousand Gnomes <gnomes@...rguk.ukuu.org.uk>
To:	Willy Tarreau <w@....eu>
Cc:	Hannes Frederic Sowa <hannes@...essinduktion.org>,
	"David S. Miller" <davem@...emloft.net>, netdev@...r.kernel.org,
	linux-kernel@...r.kernel.org,
	Linus Torvalds <torvalds@...ux-foundation.org>,
	Eric Dumazet <edumazet@...gle.com>, socketpair@...il.com
Subject: Re: [PATCH] unix: properly account for FDs passed over unix sockets

> > Another idea would be to add the amount of memory used to manage the fds 
> > to sock_rmem/wmem but I don't see any advantages or disadvantages.
> 
> Compared to the impact of the pending data in pipes themselves in flight,
> this would remain fairly minimal.

The true size of the memory pinned by a file handle cannot be determined
by the socket layer anyway. If the handle is for a device it may be
pinning lots of memory in the device driver. It's up to the rest of the
system to manage and contain those resource allocations correctly which a
fair bit of the code (except graphics) does properly.

The resource counting does create the reverse problem that user A running
a program written by hostile user B may find it takes all their resources,
stuffs them down an AF_UNIX socket and passes the last reference to the
socket via a setuid app somewhere where the user can't get it back.

For an end user case that's pretty irrelevant - you run hostile user B's
app then you already lost. For a service it's an interesting attack vector
but not one I can see any obvious ways to exploit because (with the
possible exception of dbus) services simply don't pass file handles down
sockets to untrusted targets.

Alan
--
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