netdev - [Q] why do we ever try to call unix_dgram_peer_wake_disconnect() for SOCK

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

Date:   Sun, 24 Feb 2019 06:28:12 +0000
From:   Al Viro <viro@...iv.linux.org.uk>
To:     netdev@...r.kernel.org
Cc:     Rainer Weikusat <rweikusat@...ileactivedefense.com>
Subject: [Q] why do we ever try to call unix_dgram_peer_wake_disconnect() for
 SOCK_STREAM?

	What's the point of calling unix_dgram_peer_wake_disconnect() in
unix_release_sock() of SOCK_STREAM/SOCK_SEQPACKET sockets?  AFAICS,
for those we never call unix_dgram_peer_wake_connect(), i.e. ->peer_wake.private
of any non-SOCK_DGRAM socker has to remain NULL.  Which makes that call simply
	spin_lock(&unix_sk(skpair)->peer_wait.lock);
	spin_unlock(&unix_sk(skpair)->peer_wait.lock);
and that doesn't even serve as a barrier for anything else.  Should that
have been
                if (sk->sk_type == SOCK_STREAM || sk->sk_type == SOCK_SEQPACKET) {
                        unix_state_lock(skpair);
                        /* No more writes */
                        skpair->sk_shutdown = SHUTDOWN_MASK;
                        if (!skb_queue_empty(&sk->sk_receive_queue) || embrion)
                                skpair->sk_err = ECONNRESET;
                        unix_state_unlock(skpair);
                        skpair->sk_state_change(skpair);
                        sk_wake_async(skpair, SOCK_WAKE_WAITD, POLL_HUP);
                } else {
			unix_dgram_peer_wake_disconnect(sk, skpair);
		}
or am I missing something subtle here?