Re: Attempting to checksum a non-TCP/UDP packet errors rolling across screen

Re: Attempting to checksum a non-TCP/UDP packet errors rolling across screen

[Pasi Kärkkäinen]

On Wed, Apr 21, 2010 at 11:41:35AM -0400, Gerry Reno wrote:
> Pasi Kärkkäinen wrote:
>> On Sat, Apr 10, 2010 at 02:19:39PM +0300, Pasi Kärkkäinen wrote:
>>   
>>> On Fri, Apr 09, 2010 at 06:40:04PM -0700, Brandon Lamb wrote:
>>>     
>>>> Hello,
>>>>
>>>> Just did a fresh install of xen-4.0 and using the 2.6.31.13 dom0 kernel.
>>>>
>>>> I have 6 linux debian guests running, and on my host console i am
>>>> getting these errors rolling pretty constantly.
>>>>
>>>> "Attempting to checksum a non-TCP/UDP packet, dropping a protocol 1 packet"
>>>>
>>>> I tried doing the ethtool -K eth0 tx off and same thing on peth0,
>>>> using bridged mode for network config.
>>>>
>>>> Any tips? Google is failing me, im not sure what to even search for
>>>> other than "xen Attempting to checksum a non-TCP/UDP packet".
>>>>
>>>> Im wondering if this has something to do with why my guests were
>>>> seeming to have network timeout/freezing problems, I thought maybe
>>>> upgrading from 3.4 to 4.0 would magically solve it.
>>>>
>>>>       
>>> All the networking stuff is done on the dom0 Linux kernel,
>>> so Xen upgrade doesn't affect that.
>>>
>>> Although most probably you also upgraded dom0 kernel..
>>>
>>> Can you please post output of:
>>>
>>> - ethtool peth0
>>> - ethtool -i peth0
>>> - brctl show
>>> - ifconfig -a
>>>
>>>     
>>
>> And also:
>> - ethtool -k peth0
>>
>> -- Pasi
>>   
>
> Hi Pasi,
>  I saw this post and I'm experiencing this checksumming problem as well.  
> The setup is eucalyptus and whenever we try to ping an 'instance'  
> machine from outside the cloud controller we see these errors.  Also  
> when we try to ping from inside the instance to the public network.  We  
> are running in NOVLAN mode across a bridge.
>
> Here are some details:  stats from the 'node controller' (which launches  
> the instances)
>
> root@grp-01-23-02:~# ethtool eth0
> Settings for eth0:
>        Supported ports: [ TP MII ]
>        Supported link modes:   10baseT/Half 10baseT/Full
>                                100baseT/Half 100baseT/Full
>                                1000baseT/Half 1000baseT/Full
>        Supports auto-negotiation: Yes
>        Advertised link modes:  10baseT/Half 10baseT/Full
>                                100baseT/Half 100baseT/Full
>                                1000baseT/Half 1000baseT/Full
>        Advertised auto-negotiation: Yes
>        Speed: 1000Mb/s
>        Duplex: Full
>        Port: MII
>        PHYAD: 0
>        Transceiver: internal
>        Auto-negotiation: on
>        Supports Wake-on: pumbg
>        Wake-on: g
>        Current message level: 0x00000033 (51)
>        Link detected: yes
> root@grp-01-23-02:~#
> root@grp-01-23-02:~#
> root@grp-01-23-02:~# ethtool -i eth0
> driver: r8169
> version: 2.3LK-NAPI
> firmware-version:
> bus-info: 0000:02:00.0
>

Ok, so Realtek NIC.


> root@grp-01-23-02:~#
> root@grp-01-23-02:~#
> root@grp-01-23-02:~# which brctl
> /usr/sbin/brctl
> You have new mail in /var/mail/root
> root@grp-01-23-02:~#
> root@grp-01-23-02:~# brctl show
> bridge name     bridge id               STP enabled     interfaces
> br0             8000.00241d1905df       no              eth0
>                                                        vif3.0
>                                                        vif4.0
> tmpbridge               8000.000000000000       no
> virbr0          8000.000000000000       yes
> root@grp-01-23-02:~#

<snip>

> root@grp-01-23-02:~# ethtool -k eth0
> Offload parameters for eth0:
> Cannot get device flags: Operation not supported
> rx-checksumming: on
>

Did you try disabling rx-checksumming? 

> tx-checksumming: off
> scatter-gather: off
> tcp-segmentation-offload: off
> udp-fragmentation-offload: off
> generic-segmentation-offload: off
> generic-receive-offload: off
> large-receive-offload: off
>
> We are running ubuntu enterprise cloud 1.6 with a pv_ops dom0 kernel  
> (2.6.31.6) compiled from Jeremy's git tree back about December.
>

Ok. Did you try upgrading the dom0 kernel? 

-- Pasi

Re: [Xen-devel] Re: Attempting to checksum a non-TCP/UDP packet errors rolling across screen

[Pasi Kärkkäinen]

On Wed, Apr 21, 2010 at 07:07:14PM +0300, Pasi Kärkkäinen wrote:
> On Wed, Apr 21, 2010 at 11:41:35AM -0400, Gerry Reno wrote:
> > Pasi Kärkkäinen wrote:
> >> On Sat, Apr 10, 2010 at 02:19:39PM +0300, Pasi Kärkkäinen wrote:
> >>   
> >>> On Fri, Apr 09, 2010 at 06:40:04PM -0700, Brandon Lamb wrote:
> >>>     
> >>>> Hello,
> >>>>
> >>>> Just did a fresh install of xen-4.0 and using the 2.6.31.13 dom0 kernel.
> >>>>
> >>>> I have 6 linux debian guests running, and on my host console i am
> >>>> getting these errors rolling pretty constantly.
> >>>>
> >>>> "Attempting to checksum a non-TCP/UDP packet, dropping a protocol 1 packet"
> >>>>
> >>>> I tried doing the ethtool -K eth0 tx off and same thing on peth0,
> >>>> using bridged mode for network config.
> >>>>
> >>>> Any tips? Google is failing me, im not sure what to even search for
> >>>> other than "xen Attempting to checksum a non-TCP/UDP packet".
> >>>>
> >>>> Im wondering if this has something to do with why my guests were
> >>>> seeming to have network timeout/freezing problems, I thought maybe
> >>>> upgrading from 3.4 to 4.0 would magically solve it.
> >>>>
> >>>>       
> >>> All the networking stuff is done on the dom0 Linux kernel,
> >>> so Xen upgrade doesn't affect that.
> >>>
> >>> Although most probably you also upgraded dom0 kernel..
> >>>
> >>> Can you please post output of:
> >>>
> >>> - ethtool peth0
> >>> - ethtool -i peth0
> >>> - brctl show
> >>> - ifconfig -a
> >>>
> >>>     
> >>
> >> And also:
> >> - ethtool -k peth0
> >>
> >> -- Pasi
> >>   
> >
> > Hi Pasi,
> >  I saw this post and I'm experiencing this checksumming problem as well.  
> > The setup is eucalyptus and whenever we try to ping an 'instance'  
> > machine from outside the cloud controller we see these errors.  Also  
> > when we try to ping from inside the instance to the public network.  We  
> > are running in NOVLAN mode across a bridge.
> >
> > Here are some details:  stats from the 'node controller' (which launches  
> > the instances)
> >
> > root@grp-01-23-02:~# ethtool eth0
> > Settings for eth0:
> >        Supported ports: [ TP MII ]
> >        Supported link modes:   10baseT/Half 10baseT/Full
> >                                100baseT/Half 100baseT/Full
> >                                1000baseT/Half 1000baseT/Full
> >        Supports auto-negotiation: Yes
> >        Advertised link modes:  10baseT/Half 10baseT/Full
> >                                100baseT/Half 100baseT/Full
> >                                1000baseT/Half 1000baseT/Full
> >        Advertised auto-negotiation: Yes
> >        Speed: 1000Mb/s
> >        Duplex: Full
> >        Port: MII
> >        PHYAD: 0
> >        Transceiver: internal
> >        Auto-negotiation: on
> >        Supports Wake-on: pumbg
> >        Wake-on: g
> >        Current message level: 0x00000033 (51)
> >        Link detected: yes
> > root@grp-01-23-02:~#
> > root@grp-01-23-02:~#
> > root@grp-01-23-02:~# ethtool -i eth0
> > driver: r8169
> > version: 2.3LK-NAPI
> > firmware-version:
> > bus-info: 0000:02:00.0
> >
> 
> Ok, so Realtek NIC.
> 
> 
> > root@grp-01-23-02:~#
> > root@grp-01-23-02:~#
> > root@grp-01-23-02:~# which brctl
> > /usr/sbin/brctl
> > You have new mail in /var/mail/root
> > root@grp-01-23-02:~#
> > root@grp-01-23-02:~# brctl show
> > bridge name     bridge id               STP enabled     interfaces
> > br0             8000.00241d1905df       no              eth0
> >                                                        vif3.0
> >                                                        vif4.0
> > tmpbridge               8000.000000000000       no
> > virbr0          8000.000000000000       yes
> > root@grp-01-23-02:~#
> 
> <snip>
> 
> > root@grp-01-23-02:~# ethtool -k eth0
> > Offload parameters for eth0:
> > Cannot get device flags: Operation not supported
> > rx-checksumming: on
> >
> 
> Did you try disabling rx-checksumming? 
> 
> > tx-checksumming: off
> > scatter-gather: off
> > tcp-segmentation-offload: off
> > udp-fragmentation-offload: off
> > generic-segmentation-offload: off
> > generic-receive-offload: off
> > large-receive-offload: off
> >
> > We are running ubuntu enterprise cloud 1.6 with a pv_ops dom0 kernel  
> > (2.6.31.6) compiled from Jeremy's git tree back about December.
> >
> 
> Ok. Did you try upgrading the dom0 kernel? 
> 

Replying here aswell.. this was a bug in the netback driver.

It should be fixed in the latest xen/stable-2.6.32.x tree (2.6.32.13).

-- Pasi