netdev - Re: [PATCH 1/1] bridge: remove BR_GROUPFWD_RESTRICTED for arbitrary forwarding of reserved addresses

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]

Message-ID: <16029249.TkoLblyBmC@blindfold>
Date:   Mon, 01 Oct 2018 21:10:49 +0200
From:   Richard Weinberger <richard@....at>
To:     Ido Schimmel <idosch@...sch.org>
Cc:     Stephen Hemminger <stephen@...workplumber.org>,
        Florian Fainelli <f.fainelli@...il.com>,
        bernhard.thaler@...et.at, "David S. Miller" <davem@...emloft.net>,
        bridge@...ts.linux-foundation.org, netdev@...r.kernel.org,
        David Gstir <david@...ma-star.at>, nikolay@...ulusnetworks.com,
        roopa@...ulusnetworks.com
Subject: Re: [PATCH 1/1] bridge: remove BR_GROUPFWD_RESTRICTED for arbitrary forwarding of reserved addresses

Am Montag, 1. Oktober 2018, 21:04:33 CEST schrieb Ido Schimmel:
> On Mon, Oct 01, 2018 at 08:54:08PM +0200, Richard Weinberger wrote:
> > So the only option is having a bridge and transport STP via tc-mirred
> > or patching the bridge code (what we do right now).
> 
> And I vote for the first option. I understand it involves more typing,

hehe, it is not about typing. The setup is done by a script.
And we both know that getting a patch upstream is much more work than
typing a few lines of hacky bash scripts.
Since I want a decent solution and feedback I'm bringing this up here.

I could also just go with my in-house patch and don't tell anyone...

> but I see no reason to push more complexity into the kernel - and break
> standards - when you can relatively easily accomplish the same thing in
> other ways.

If having a bridge plus u32+mirred for STP bypass is the preferred solution,
I'm fine with it.
So far we use the kernel patch and didn't test the mirred-bypass a lot.
My fear was that mirred rules from eth0 to eth1 and back might cause a
loop or confuse the bridge...

> Adding Nik and Roopa who now maintain the bridge code and should
> eventually decide about this.

Thanks,
//richard