netdev - Re: [PATCH net-next 2/2] tcp/dccp: Remove unnecessary initialization of refcounted.

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Mon, 23 Mar 2020 11:47:17 -0700
From:   Eric Dumazet <edumazet@...gle.com>
To:     Kuniyuki Iwashima <kuniyu@...zon.co.jp>
Cc:     "David S . Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>,
        Gerrit Renker <gerrit@....abdn.ac.uk>,
        Alexey Kuznetsov <kuznet@....inr.ac.ru>,
        Hideaki YOSHIFUJI <yoshfuji@...ux-ipv6.org>,
        dccp@...r.kernel.org, netdev <netdev@...r.kernel.org>,
        Kuniyuki Iwashima <kuni1840@...il.com>,
        osa-contribution-log@...zon.com
Subject: Re: [PATCH net-next 2/2] tcp/dccp: Remove unnecessary initialization
 of refcounted.

On Mon, Mar 23, 2020 at 11:22 AM Kuniyuki Iwashima <kuniyu@...zon.co.jp> wrote:
>
> When we get a TCP_NEW_SYN_RECV/DCCP_NEW_SYN_RECV socket by
> __inet_lookup_skb(), refcounted is already set true, so it is not
> necessary to do it again.

This changelog is absolutely not accurate.

sk is a listener here.
(because sk was set to req->rsk_listener;)

Please do not add confusion by mixing different things.

I prefer not relying on the old value of 'refcounted', since we
switched sk value.

Note that we call reqsk_put(req); regardless of 'refcounted'

I would rather not change this code and make future backports more complicated.

Thanks.