netdev - Re: [PATCH net-next] sfc: remove udp_tnl_has

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: <20200701114336.62b57cc4@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
Date:   Wed, 1 Jul 2020 11:43:36 -0700
From:   Jakub Kicinski <kuba@...nel.org>
To:     Edward Cree <ecree@...arflare.com>
Cc:     <davem@...emloft.net>, <netdev@...r.kernel.org>,
        <mhabets@...arflare.com>, <linux-net-drivers@...arflare.com>
Subject: Re: [PATCH net-next] sfc: remove udp_tnl_has_port

On Wed, 1 Jul 2020 15:32:03 +0100 Edward Cree wrote:
> On 30/06/2020 23:50, Jakub Kicinski wrote:
> > Nothing seems to have ever been calling this.
> >
> > Signed-off-by: Jakub Kicinski <kuba@...nel.org>  
> That was intended to be used by encap offloads (TX csum and TSO), which
>  we only recently realised we hadn't upstreamed the rest of; the
>  udp_tnl_has_port method would be called from our ndo_features_check().
> I'll try to get to upstreaming that support after ef100 is in, hopefully
>  within this cycle, but if you don't want this dead code lying around in
>  the meantime then have an
> Acked-by: Edward Cree <ecree@...arflare.com>
>  and I can revert it when I add the code that calls it.
> (And don't worry, ef100 doesn't use ugly port-based offloads; it does
>  proper CHECKSUM_PARTIAL and GSO_PARTIAL, so it won't have this stuff.)

There's a number of drivers which try to match the UDP ports. That
seems fragile to me. Is it actually required for HW to operate
correctly? 

Aren't the ports per ns in the kernel? There's no guarantee that some
other netns won't send a TSO skb and whatever other UDP encap.