Re: sky2 checksum errors

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Thu, Feb 18, 2010 at 11:38 PM, Stephen Hemminger
<shemminger@xxxxxxxxxx> wrote:
> On Thu, 18 Feb 2010 16:45:59 +0600
> Roman Tsisyk <roman@xxxxxxxxxx> wrote:
>
>> We have motherboard with Marvell 88E8056 ethernet controllers.
>> ------
>> 02:00.0 Ethernet controller: Marvell Technology Group Ltd. 88E8056
>> PCI-E Gigabit Ethernet Controller (rev 14)
>> 03:00.0 Ethernet controller: Marvell Technology Group Ltd. 88E8056
>> PCI-E Gigabit Ethernet Controller (rev 14)
>> ------
>>
>> It is used for routing network packets between two interfaces. This
>> system has been tested for several days using pktgen with different
>> packet size and bandwidth settings and found problem similar to
>> http://marc.info/?l=linux-net&m=121967539203137&w=2
>> Firmware was updated according to that thread instructions, thereafter
>> rev. changed from 12 to 14. But interfaces periodically starting to
>> loose a large number of packets and netdev_rx_csum_fault threw.
>>
>
> It would be good to know the size of packets and the state of the
> receive status ring when the receive checksum error occurred.
> This would require additional instrumentation.
>
> What is your kernel configuration? What kernel version?
> There were several recent stability fixes for sky2 driver relating
> to DMA mapping.
>
>
>
> --
>

You speak about http://patchwork.kernel.org/patch/76165/ ? Is it
related to this problem?

Kernel version was 2.6.30.8 (.9-10 released later) from Debian sources
(no network related patches has been applied) and default config
(http://debian.pastebin.com/m3d905d6). I didn't try a new version
because systems with similar hardware and completely identical kernel
but tg3 NICs are working fine. Probably problem in sky2 or Marvell
chip.

Flow was mixed, with various packet sizes and approximately 100-200 kpps
throughput. After these messages network was working but throughput
decreased to 4-5 kpps.

How to check state of receive status ring at the moment? On test
configuration random packets with sizes from 100 to 1500 and
throughput 500-600 kpps cannot crash system, but with real traffic it
happens randomly. Situation is really difficult to reproduce. Even so,
I can try to add verbose debug outputs that you need and to check it.

-- 
WBR, Tsisyk Roman
--
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[Index of Archives]     [Netdev]     [Ethernet Bridging]     [Linux 802.1Q VLAN]     [Linux Wireless]     [Kernel Newbies]     [Security]     [Linux for Hams]     [Netfilter]     [Git]     [Bugtraq]     [Yosemite News and Information]     [MIPS Linux]     [ARM Linux]     [Linux RAID]     [Linux PCI]     [Linux Admin]     [Samba]

  Powered by Linux