netdev - Re: [PATCH] net: split sk_reuse into sk_reuse and sk_force

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

Message-ID: <20180614200224.GA20731@outlook.office365.com>
Date:   Thu, 14 Jun 2018 13:02:29 -0700
From:   Andrei Vagin <avagin@...tuozzo.com>
To:     Eric Dumazet <edumazet@...gle.com>
Cc:     Andrey Vagin <avagin@...nvz.org>,
        David Miller <davem@...emloft.net>,
        netdev <netdev@...r.kernel.org>,
        Maciej Żenczykowski <maze@...gle.com>
Subject: Re: [PATCH] net: split sk_reuse into sk_reuse and sk_force_reuse

On Wed, Jun 13, 2018 at 06:17:41PM -0700, Eric Dumazet wrote:
> On Wed, Jun 13, 2018 at 5:56 PM Andrei Vagin <avagin@...nvz.org> wrote:
> 
> > The commit f396922d862a added a check to not allow changing
> > SO_REUSEADDR/SO_REUSEPORT on bound sockets. First, it doesn't
> > take into account that TCP_REPAIR changes SO_REUSEADDR. Second, now it
> > is impossible to restore a socket state and set SO_REUSEADDR,
> > because the kernel always sets SO_REUSEADDR into zero after disabling
> > the repair mode.
> >
> >
> Hi Andrey
> 
> This commit was reverted, do we still need  this patch ?

I have seen that this patch was reverted. Probably I had to check
net-next before sending it.

I'm agree with Maciej Żenczykowski that it makes this code better. I
have never understood why TCP_REPAIR drops SO_REUSEADDR. Now each time
when we use TCP_REPAIR, we have to save a value of SO_REUSEADDR and
restore it back after disabling TCP_REPAIR. With this patch, we will
able to enable/disable TCP_REPAIR and don't care about sk_reuse.

I will update the commit message and send the patch again.

Thanks,
Andrei