netdev - Re: [PATCH net] selftests: forwarding: vxlan_bridge

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: <87r1ytpkdu.fsf@mellanox.com>
Date:   Mon, 17 Feb 2020 11:40:13 +0100
From:   Petr Machata <petrm@...lanox.com>
To:     Hangbin Liu <liuhangbin@...il.com>
Cc:     Petr Machata <pmachata@...il.com>, netdev@...r.kernel.org,
        Ido Schimmel <idosch@...lanox.com>,
        "David S . Miller" <davem@...emloft.net>,
        Peter Dawson <petedaws@...il.com>,
        Daniel Borkmann <daniel@...earbox.net>
Subject: Re: [PATCH net] selftests: forwarding: vxlan_bridge_1d: fix tos value


Hangbin Liu <liuhangbin@...il.com> writes:

> On Fri, Feb 14, 2020 at 11:54:09AM +0100, Petr Machata wrote:
>> >> > After commit 71130f29979c ("vxlan: fix tos value before xmit") we start
>> >> > strict vxlan xmit tos value by RT_TOS(), which limits the tos value less
>> >>
>> >> I don't understand how it is OK to slice the TOS field like this. It
>> >> could contain a DSCP value, which will be mangled.
>> >
>> > Thanks for this remind. I re-checked the tos definition and found a summary
>> > from Peter Dawson[1].
>> >
>> > IPv4/6 Header:0 |0 1 2 3 |0 1 2 3 |0 1 2 3 |0 1 2 3 |
>> > RFC2460(IPv6)   |Version | Traffic Class   |        |
>> > RFC2474(IPv6)   |Version | DSCP        |ECN|        |
>> > RFC2474(IPv4)   |Version |  IHL   |    DSCP     |ECN|
>> > RFC1349(IPv4)   |Version |  IHL   | PREC |  TOS   |X|
>> > RFC791 (IPv4)   |Version |  IHL   |      TOS        |
>> >
>> > According to this I think our current IPTOS_TOS_MASK should be updated to 0xFC
>> > based on RFC2474. But I'm not sure if there will have compatibility issue.
>> > What do you think?
>>
>> Looking at the various uses of RT_TOS, it looks like they tend to be
>> used in tunneling and routing code. I think that in both cases it makes
>> sense to convert to 0xfc. But I'm not ready to vouch for this :)
>
> Yes, I also could not... Maybe David or Daniel could help give some comments?
>
>>
>> What is the problem that commit 71130f29979c aims to solve? It's not
>> clear to me from the commit message. What issues arise if the TOS is
>> copied as is?
>
> As the commit said, we should not use config tos directly. We should remove
> the precedence field based on RFC1349 or ENC field based on RFC2474.

Well, RFC1349 didn't know about DSCP. I do not think it is possible to
conform to both RFC1349 and RFC2474 at the same time.

RFC2474 states that "DS field [...] is intended to supersede the
existing definitions of the IPv4 TOS octet [RFC791] and the IPv6 Traffic
Class octet [IPv6]". So the field should be assumed to contain DSCP from
that point on. In my opinion, that makes commit 71130f29979c incorrect.

(And other similar uses of RT_TOS in other tunneling devices likewise.)