From mboxrd@z Thu Jan  1 00:00:00 1970
From: AMG Zollner Robert <robert@cloudmedia.eu>
Subject: [bug] cxgb4: vrf stopped working with cxgb4 card
Date: Mon, 4 Jun 2018 18:03:32 +0300
Message-ID: <b515a1c7-2cf8-1af8-372d-393420d298b8@cloudmedia.eu>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit
Cc: netdev@vger.kernel.org, dsa@cumulusnetworks.com
To: ganeshgr@chelsio.com
Return-path: <netdev-owner@vger.kernel.org>
Received: from web01.accessmedia.ro ([86.107.100.4]:45803 "EHLO
        web01.accessmedia.ro" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
        with ESMTP id S1753603AbeFDPR3 (ORCPT
        <rfc822;netdev@vger.kernel.org>); Mon, 4 Jun 2018 11:17:29 -0400
Content-Language: en-US
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>

I have noticed that vrf is not working with kernel v4.15.0 but was 
working with v4.13.0 when using cxgb4 Chelsio driver (T520-cr)

Setup:
Two metal servers with a T520-cr card each, directly connected without a 
switch in between.

        SVR1  only ipfwd                 SVR2     with vrf
.----------------------------. .----------------------------------.
|                            |         |             |
|    192.168.8.1 [  ens2f4]--|---------|--[ens1f4] 192.168.8.2   |
|    192.168.9.1 [ens2f4d1]--|---------|--<ens1f4d1> 192.168.9.2 VRF=10   |
`----------------------------' `----------------------------------'

When vrf is not working there are no error messages (dmesg or iproute 
commands), tcpdump on the interface (SVR2.ens1f4d1) enslaved in vrf 10 
shows packets(arp req/reply) coming in and going out, but outgoing 
packets(arp reply) do not reach the other server SVR1.ens2f4d1


Bisect:
Found this commit to be the problem after doing a git bisect between 
v4.13..v4.15:

commit ba581f77df23c8ee70b372966e69cf10bc5453d8
Author: Ganesh Goudar <ganeshgr@chelsio.com>
Date:   Sat Sep 23 16:07:28 2017 +0530

     cxgb4: do DCB state reset in couple of places

     reset the driver's DCB state in couple of places
     where it was missing.


A bisect step was considered good when:
- successful ping from SVR1 to SVR2.ens1f4d1 vrf interface
- successful ping from SVR2 global to SVR2 vrf interface trough SVR1(l3 
forwarding) (this check was redundant,both tests fail or pass simultaneous)

The problem is still present on recent kernels also, checked v4.16.0 and 
v4.17.rc7

Disabling DCB for the card support fixes the problem ( Compiling kernel 
with "CONFIG_CHELSIO_T4_DCB=n")



This is my first time reporting a bug to the linux kernel and hope I 
have included the right amount of information. Please let me know if I 
have missed something.



Thank you,
Zollner Robert


--------
Logs:

VRF configured using folowing commands:

#!/bin/sh

CHDEV=ens1f4
VRF=vrf-recv

sysctl -w net.ipv4.tcp_l3mdev_accept=1
sysctl -w net.ipv4.udp_l3mdev_accept=1
sysctl -w net.ipv4.conf.all.accept_local=1

ifconfig ${CHDEV}   192.168.8.2/24
ifconfig ${CHDEV}d1 192.168.9.2/24

ip link add ${VRF} type vrf table 10
ip link set dev ${VRF} up

ip rule add pref 32765 table local
ip rule del pref 0

ip route add table 10 unreachable default metric 4278198272

ip link set dev ${CHDEV}d1 master ${VRF}

ip route add table 10 default via 192.168.9.1
ip route add 192.168.9.0/24 via 192.168.8.1