Mailing List Archive

i40e kernel panic
Hello,

I am able to reproduce a kernel panic using the i40e-1.5.18-zc drivers.
When I have a PF_RING application running using an interface (in this case
eth2)
If I run tcpdump on this interface at the same time, then exit tcpdump it
will consistently cause a kernel panic. The logs indicate this happens when
tcpdump tries to leave promiscuous mode.
>From what I can see, the device resets when promiscuous mode is toggled
which causes PF_RING to raise an error since the device is in use.

This is the kernel log of when it had happened:
http://hastebin.com/hibonigira.md

After this happens, the system hard-locks and requires physically rebooted.

Thanks
Re: i40e kernel panic [ In reply to ]
Hi Josiah
thank you for reporting,
it seems to be due to a card reset (on this card model, setting the promisc leads to a card reset),
it should be handled by avoiding handling promisc state changes when the card is in use, I will
check it. Please open an issue on github to keep track of it.

Alfredo

> On 27 Jul 2016, at 11:42, Josiah White <josiah@protrafsolutions.com> wrote:
>
> Hello,
>
> I am able to reproduce a kernel panic using the i40e-1.5.18-zc drivers.
> When I have a PF_RING application running using an interface (in this case eth2)
> If I run tcpdump on this interface at the same time, then exit tcpdump it will consistently cause a kernel panic. The logs indicate this happens when tcpdump tries to leave promiscuous mode.
> From what I can see, the device resets when promiscuous mode is toggled which causes PF_RING to raise an error since the device is in use.
>
> This is the kernel log of when it had happened:
> http://hastebin.com/hibonigira.md <http://hastebin.com/hibonigira.md>
>
> After this happens, the system hard-locks and requires physically rebooted.
>
> Thanks
> _______________________________________________
> Ntop-misc mailing list
> Ntop-misc@listgateway.unipi.it
> http://listgateway.unipi.it/mailman/listinfo/ntop-misc
Re: i40e kernel panic [ In reply to ]
Hello,

I have opened the issue:
https://github.com/ntop/PF_RING/issues/118

Thanks again

On Wed, Jul 27, 2016 at 6:11 AM, Alfredo Cardigliano <cardigliano@ntop.org>
wrote:

> Hi Josiah
> thank you for reporting,
> it seems to be due to a card reset (on this card model, setting the
> promisc leads to a card reset),
> it should be handled by avoiding handling promisc state changes when the
> card is in use, I will
> check it. Please open an issue on github to keep track of it.
>
> Alfredo
>
> On 27 Jul 2016, at 11:42, Josiah White <josiah@protrafsolutions.com>
> wrote:
>
> Hello,
>
> I am able to reproduce a kernel panic using the i40e-1.5.18-zc drivers.
> When I have a PF_RING application running using an interface (in this case
> eth2)
> If I run tcpdump on this interface at the same time, then exit tcpdump it
> will consistently cause a kernel panic. The logs indicate this happens when
> tcpdump tries to leave promiscuous mode.
> From what I can see, the device resets when promiscuous mode is toggled
> which causes PF_RING to raise an error since the device is in use.
>
> This is the kernel log of when it had happened:
> http://hastebin.com/hibonigira.md
>
> After this happens, the system hard-locks and requires physically rebooted.
>
> Thanks
> _______________________________________________
> Ntop-misc mailing list
> Ntop-misc@listgateway.unipi.it
> http://listgateway.unipi.it/mailman/listinfo/ntop-misc
>
>
>
> _______________________________________________
> Ntop-misc mailing list
> Ntop-misc@listgateway.unipi.it
> http://listgateway.unipi.it/mailman/listinfo/ntop-misc
>
Re: i40e kernel panic [ In reply to ]
Thank you

Alfredo

> On 27 Jul 2016, at 12:30, Josiah White <josiah@protrafsolutions.com> wrote:
>
> Hello,
>
> I have opened the issue:
> https://github.com/ntop/PF_RING/issues/118 <https://github.com/ntop/PF_RING/issues/118>
>
> Thanks again
>
> On Wed, Jul 27, 2016 at 6:11 AM, Alfredo Cardigliano <cardigliano@ntop.org <mailto:cardigliano@ntop.org>> wrote:
> Hi Josiah
> thank you for reporting,
> it seems to be due to a card reset (on this card model, setting the promisc leads to a card reset),
> it should be handled by avoiding handling promisc state changes when the card is in use, I will
> check it. Please open an issue on github to keep track of it.
>
> Alfredo
>
>> On 27 Jul 2016, at 11:42, Josiah White <josiah@protrafsolutions.com <mailto:josiah@protrafsolutions.com>> wrote:
>>
>> Hello,
>>
>> I am able to reproduce a kernel panic using the i40e-1.5.18-zc drivers.
>> When I have a PF_RING application running using an interface (in this case eth2)
>> If I run tcpdump on this interface at the same time, then exit tcpdump it will consistently cause a kernel panic. The logs indicate this happens when tcpdump tries to leave promiscuous mode.
>> From what I can see, the device resets when promiscuous mode is toggled which causes PF_RING to raise an error since the device is in use.
>>
>> This is the kernel log of when it had happened:
>> http://hastebin.com/hibonigira.md <http://hastebin.com/hibonigira.md>
>>
>> After this happens, the system hard-locks and requires physically rebooted.
>>
>> Thanks
>> _______________________________________________
>> Ntop-misc mailing list
>> Ntop-misc@listgateway.unipi.it <mailto:Ntop-misc@listgateway.unipi.it>
>> http://listgateway.unipi.it/mailman/listinfo/ntop-misc <http://listgateway.unipi.it/mailman/listinfo/ntop-misc>
>
> _______________________________________________
> Ntop-misc mailing list
> Ntop-misc@listgateway.unipi.it <mailto:Ntop-misc@listgateway.unipi.it>
> http://listgateway.unipi.it/mailman/listinfo/ntop-misc <http://listgateway.unipi.it/mailman/listinfo/ntop-misc>
>
> _______________________________________________
> Ntop-misc mailing list
> Ntop-misc@listgateway.unipi.it
> http://listgateway.unipi.it/mailman/listinfo/ntop-misc