Date:	Wed, 16 Apr 2014 15:58:42 +0200
From:	Yegor Yefremov <yegorslists@...glemail.com>
To:	Mugunthan V N <mugunthanvnm@...com>
Cc:	David Miller <davem@...emloft.net>, netdev <netdev@...r.kernel.org>
Subject: Re: [net PATCH 1/1] drivers: net: cpsw: Add default vlan for dual
 emac case also

On Wed, Apr 16, 2014 at 7:57 AM, Mugunthan V N <mugunthanvnm@...com> wrote:
> On Tuesday 15 April 2014 07:16 PM, Yegor Yefremov wrote:
>> On Wed, Apr 9, 2014 at 8:56 PM, David Miller <davem@...emloft.net> wrote:
>>> From: Mugunthan V N <mugunthanvnm@...com>
>>> Date: Wed, 9 Apr 2014 11:34:40 +0530
>>>
>>>> Dual EMAC works with VLAN segregation of the ports, so default vlan needs
>>>> to be added in dual EMAC case else default vlan will be tagged for all
>>>> egress packets and vlan unaware switches/servers will drop packets
>>>> from the EVM.
>> Hi Mugunthan,
>>
>> though this patch fixes drop packets issue, it introduces another
>> issue. Just connect both interfaces to one switch and configure
>> different subnets. As soon as you get a broadcast you have an endless
>> loop and network communication is not possible at all. Before this
>> patch you could operate as if you had really two independent NICs.
>>
>>
>
> This patch doesn't add any multicast or unicast entry in the ALE table,
> so there is no possible way that a broadcast packet is switched within
> the two external ports, and this patch removes the encapsulation of the
> default vlan id which is used for internal packet switching inside CPSW
> switch from host and any packet received from external port will be
> tagged with port vlan so internal switching will happen with that port
> vlan-id which will be only that physical port and host port. so the
> scenario which you mentioned cannot happen theoretically untill you
> enable bridge between the dual ethernet interface and connect both to
> the same interface.

Have you tried the described setup? My system is Buildroot on
am335x-evmsk with almost empty /etc/network/interfaces

# Configure Loopback
auto lo
iface lo inet loopback

No bridges:

# brctl show
bridge name     bridge id               STP enabled     interfaces

And following address setup made manually:

1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN
group default
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast
state UP group default qlen 1000
    link/ether 00:18:31:93:f6:49 brd ff:ff:ff:ff:ff:ff
    inet 192.168.1.234/24 brd 192.168.1.255 scope global eth0
       valid_lft forever preferred_lft forever
3: eth1: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast
state DOWN group default qlen 1000
    link/ether 56:90:76:5b:b5:f3 brd ff:ff:ff:ff:ff:ff
    inet 192.168.4.234/24 brd 192.168.4.255 scope global eth1
       valid_lft forever preferred_lft forever
4: wlan0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN group
default qlen 1000
    link/ether de:ad:be:ef:00:00 brd ff:ff:ff:ff:ff:ff

Routing table:

# ip route show
192.168.1.0/24 dev eth0  proto kernel  scope link  src 192.168.1.234
192.168.4.0/24 dev eth1  proto kernel  scope link  src 192.168.4.234

Same behavior on 3.14.0-rc4-12740-gcc09039-dirty and 3.15.0-rc1-00002-ge0f8779

As soon as insert the cable into eth1, my switch shows very high
activity and the whole LAN in the company is dead till I remove the
cable. I have no problems with the version before the fix. I don't
really know, what happens inside the driver, but could make the same
setup (best of all in a LAN with many Windows PCs, that send lots of
broadcast) and let me know if you see the same behavior.

Regards,
Yegor
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Powered by blists - more mailing lists