netdev - Re: [PATCH net-next 03/10] mptcp: add SIOCINQ, OUTQ and OUTQNSD ioctls

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]

Date:   Tue, 07 Dec 2021 11:30:42 +0100
From:   Paolo Abeni <pabeni@...hat.com>
To:     Jakub Kicinski <kuba@...nel.org>,
        Mat Martineau <mathew.j.martineau@...ux.intel.com>
Cc:     netdev@...r.kernel.org, Florian Westphal <fw@...len.de>,
        davem@...emloft.net, matthieu.baerts@...sares.net,
        mptcp@...ts.linux.dev
Subject: Re: [PATCH net-next 03/10] mptcp: add SIOCINQ, OUTQ and OUTQNSD
 ioctls

Hello,

On Mon, 2021-12-06 at 17:16 -0800, Jakub Kicinski wrote:
> On Fri,  3 Dec 2021 14:35:34 -0800 Mat Martineau wrote:
> > +		if (sk->sk_state == TCP_LISTEN)
> > +			return -EINVAL;
> > +
> > +		lock_sock(sk);
> > +		__mptcp_move_skbs(msk);
> > +		answ = mptcp_inq_hint(sk);
> > +		release_sock(sk);
> 
> The raciness is not harmful here?

Thank you for the careful review!

AFAICS the race between the socket state test and the receive queue
lenght estimate is not harmful: if a socket is concurrently
disconnected and moved to a listen status, mptcp_inq_hint() will return
0, as plain TCP would do in the same scenario - all the fields accessed
by __mptcp_move_skbs() and mptcp_inq_hint() will be in a consistant
status (modulo unknown bugs).

Cheers,

Paolo