Re: if_igb crashes system

From: matt <sendtomatt_at_gmail.com>
Date: Sat, 03 Mar 2012 15:54:26 -0800
On 03/02/12 12:07, matt wrote:
> On 03/01/12 21:24, matt wrote:
>> bringing up igb0 is currently causing my 10-CURRENT box to become very
>> non-responsive (as though downclocked to some 100 KHZ...very, very, very
>> slow).
>>
>> cmdwatch 'vmstat -i' shows the interrupts get assigned for igb0, and
>> about 1 second later the machine is essentially non-responsive.
>>
>> anyone else seeing if_igb issues suddenly?
>> the machine is running off a cvsup from about 75 minutes ago.
>>
>> I was only able to access the system after removing interface configs
>> from rc.conf, which made me suspect polling initially, however even
>> issuing "ifconfig igb0 up" with no media causes the issue to appear.
>>
>> Let me know what additional diagnostic data may help?
>>
>> Matt
>>
> Last time this machine was updated was last week.
> I'm going to have to install something else to test the hardware, but it
> had no problems until first reboot with new kernel.
> I have blown away /usr/src and re fetched it...no change.
> I've disabled pf and tried to bring it up...no change
> I've tried to bring up the other interface in the pair...no change (igb1).
> It doesn't look like the msix commit from jhb has anything to do with
> it...didn't exactly revert but changed return(0); back to
> return(msgs);...no change
>
> I'm down to either the hardware conveniently failed during upgrade, or
> some earlier commit.
> Luckily I have an em-based port as well.
> I'm testing hardware next and then going further back into the week on
> the commit log?
> Anyone else with igb out there and working with latest HEAD?
>
> Matt
>
>
igb is working now, I rebuilt world as of this afternoon and it appears
to be stable.
Sorry for the noise & thanks,

Matt
Received on Sat Mar 03 2012 - 22:55:47 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:40:24 UTC