netdev - Re: [PATCH 1/5 net-next] inet: replace ->bind_conflict with ->rcv_saddr

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:   Wed, 21 Dec 2016 13:28:22 -0500 (EST)
From:   David Miller <davem@...emloft.net>
To:     jbacik@...com
Cc:     hannes@...essinduktion.org, kraigatgoog@...il.com,
        eric.dumazet@...il.com, tom@...bertland.com,
        netdev@...r.kernel.org, kernel-team@...com
Subject: Re: [PATCH 1/5 net-next] inet: replace ->bind_conflict with
 ->rcv_saddr_equal

From: Josef Bacik <jbacik@...com>
Date: Tue, 20 Dec 2016 15:07:00 -0500

> The only difference between inet6_csk_bind_conflict and inet_csk_bind_conflict
> is how they check the rcv_saddr.  Since we want to be able to check the saddr in
> other places just drop the protocol specific ->bind_conflict and replace it with
> ->rcv_saddr_equal, then make inet_csk_bind_conflict the one true bind conflict
> function.
> 
> Signed-off-by: Josef Bacik <jbacik@...com>

This may be a nice cleanup and all, but realize that if we do actually
have to traverse a lot of sockets this code has become significantly
slower.

We now have to execute a hard to predict indirect call for every
socket we process on the list.

This is almost certainly why we had two seperate functions expanded
rather than having an AF-specific helper execute in the inner loop
of a generic function.