netdev - Re: vmw_vsock sk_ack_backlog double decrement bug

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] [day] [month] [year] [list]

Message-ID: <BY2PR0501MB2056DD2972A874DEC66728ECDA2E0@BY2PR0501MB2056.namprd05.prod.outlook.com>
Date:	Fri, 24 Jun 2016 13:09:18 +0000
From:	"Jorgen S. Hansen" <jhansen@...are.com>
To:	Stefan Hajnoczi <stefanha@...hat.com>
CC:	"netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: Re: vmw_vsock sk_ack_backlog double decrement bug

Hi Stefan,

Good catch. Thanks for pointing this out. I'll take care of fixing and testing this.

 Thanks,
 Jørgen

________________________________________
From: Stefan Hajnoczi <stefanha@...hat.com>
Sent: Thursday, June 23, 2016 5:40 PM
To: Jorgen S. Hansen
Cc: netdev@...r.kernel.org
Subject: vmw_vsock sk_ack_backlog double decrement bug

Hi Jorgen,
virtio-vsock doesn't use vsock_pending_work() but I may have spotted a
problem that affects the VMCI transport.  I'm not sending a patch
because I can't test it.

1. During vsock_accept() listener->sk_ack_backlog is decremented.
2. vsock_pending_work() will decrement listener->sk_ack_backlog again if
   vsk->rejected.

The result is that sk_ack_backlog can be invalid.  It only happens in
the case where the listener socket has an error.  Maybe in practice it's
not a problem because the server application will close the listener
socket if there is an error...

Stefan