3.9.0 dmesg reports that my NIC is hanging

3.9.0 dmesg reports that my NIC is hanging

[John]

After upgrading to the official Arch Linux 3.9-2 kernel package, dmesg reports that my NIC is hanging:

[    5.955720] e1000e 0000:00:19.0 eno1: changing MTU from 1500 to 4000
[    8.464507] e1000e 0000:00:19.0 eno1: Detected Hardware Unit Hang:
  TDH                  <0>
  TDT                  <2>
  next_to_use          <2>
  next_to_clean        <0>
buffer_info[next_to_clean]:
  time_stamp           <fffea787>
  next_to_watch        <0>
  jiffies              <fffeaa30>
  next_to_watch.status <0>
MAC Status             <40080080>
PHY Status             <7949>
PHY 1000BASE-T Status  <0>
PHY Extended Status    <3000>
PCI Status             <10>

Not too sure what else to post.  I am not subscribed to lkml so please cc my email in your reply.


Link to complete dmesg: http://pastebin.com/zRBajGrY
Seems similar to: bugzilla.redhat.com/show_bug.cgi?id=785806

Re: 3.9.0 dmesg reports that my NIC is hanging

[Bjorn Helgaas]

[+cc Jeff, e1000-devel (from MAINTAINERS)]

On Sat, May 4, 2013 at 1:56 PM, John <da_audiophile@yahoo.com> wrote:
> After upgrading to the official Arch Linux 3.9-2 kernel package, dmesg reports that my NIC is hanging:
>
> [    5.955720] e1000e 0000:00:19.0 eno1: changing MTU from 1500 to 4000
> [    8.464507] e1000e 0000:00:19.0 eno1: Detected Hardware Unit Hang:
>   TDH                  <0>
>   TDT                  <2>
>   next_to_use          <2>
>   next_to_clean        <0>
> buffer_info[next_to_clean]:
>   time_stamp           <fffea787>
>   next_to_watch        <0>
>   jiffies              <fffeaa30>
>   next_to_watch.status <0>
> MAC Status             <40080080>
> PHY Status             <7949>
> PHY 1000BASE-T Status  <0>
> PHY Extended Status    <3000>
> PCI Status             <10>
>
> Not too sure what else to post.  I am not subscribed to lkml so please cc my email in your reply.
>
>
> Link to complete dmesg: http://pastebin.com/zRBajGrY
> Seems similar to: bugzilla.redhat.com/show_bug.cgi?id=785806

It sounds like this is a regression, so it might be useful to know
what the newest working kernel was, and maybe a dmesg log from it as
well, though I don't see any obvious clues in the 3.9.0-2-ARCH dmesg
you collected.

Bjorn

Re: 3.9.0 dmesg reports that my NIC is hanging

[John]

----- Original Message -----
> From: Bjorn Helgaas <bhelgaas@google.com>
> To: John <da_audiophile@yahoo.com>
> Cc: lkml <linux-kernel@vger.kernel.org>; Jeff Kirsher <jeffrey.t.kirsher@intel.com>; "e1000-devel@lists.sourceforge.net" <e1000-devel@lists.sourceforge.net>
> Sent: Monday, May 6, 2013 1:29 PM
> Subject: Re: 3.9.0 dmesg reports that my NIC is hanging
> 
> [+cc Jeff, e1000-devel (from MAINTAINERS)]
> 
> On Sat, May 4, 2013 at 1:56 PM, John <da_audiophile@yahoo.com> wrote:
>>  After upgrading to the official Arch Linux 3.9-2 kernel package, dmesg 
> reports that my NIC is hanging:
>> 
>>  [    5.955720] e1000e 0000:00:19.0 eno1: changing MTU from 1500 to 4000
>>  [    8.464507] e1000e 0000:00:19.0 eno1: Detected Hardware Unit Hang:
>>    TDH                  <0>
>>    TDT                  <2>
>>    next_to_use          <2>
>>    next_to_clean        <0>
>>  buffer_info[next_to_clean]:
>>    time_stamp           <fffea787>
>>    next_to_watch        <0>
>>    jiffies              <fffeaa30>
>>    next_to_watch.status <0>
>>  MAC Status             <40080080>
>>  PHY Status             <7949>
>>  PHY 1000BASE-T Status  <0>
>>  PHY Extended Status    <3000>
>>  PCI Status             <10>
>> 
>>  Not too sure what else to post.  I am not subscribed to lkml so please cc 
> my email in your reply.
>> 
>> 
>>  Link to complete dmesg: http://pastebin.com/zRBajGrY
>>  Seems similar to: bugzilla.redhat.com/show_bug.cgi?id=785806
> 
> It sounds like this is a regression, so it might be useful to know
> what the newest working kernel was, and maybe a dmesg log from it as
> well, though I don't see any obvious clues in the 3.9.0-2-ARCH dmesg
> you collected.
> 
> Bjorn


Thank you for the reply, Bjorn.  3.8.11-1-ARCH works just fine for me.  Here is the dmesg from 3.8.11-1-ARCH per your request: http://pastebin.com/cUHwrQfq

Re: 3.9.0 dmesg reports that my NIC is hanging

[Bjorn Helgaas]

On Tue, May 7, 2013 at 3:07 PM, John <da_audiophile@yahoo.com> wrote:
>
>
>
>
> ----- Original Message -----
>> From: Bjorn Helgaas <bhelgaas@google.com>
>> To: John <da_audiophile@yahoo.com>
>> Cc: lkml <linux-kernel@vger.kernel.org>; Jeff Kirsher <jeffrey.t.kirsher@intel.com>; "e1000-devel@lists.sourceforge.net" <e1000-devel@lists.sourceforge.net>
>> Sent: Monday, May 6, 2013 1:29 PM
>> Subject: Re: 3.9.0 dmesg reports that my NIC is hanging
>>
>> [+cc Jeff, e1000-devel (from MAINTAINERS)]
>>
>> On Sat, May 4, 2013 at 1:56 PM, John <da_audiophile@yahoo.com> wrote:
>>>  After upgrading to the official Arch Linux 3.9-2 kernel package, dmesg
>> reports that my NIC is hanging:
>>>
>>>  [    5.955720] e1000e 0000:00:19.0 eno1: changing MTU from 1500 to 4000
>>>  [    8.464507] e1000e 0000:00:19.0 eno1: Detected Hardware Unit Hang:
>>>    TDH                  <0>
>>>    TDT                  <2>
>>>    next_to_use          <2>
>>>    next_to_clean        <0>
>>>  buffer_info[next_to_clean]:
>>>    time_stamp           <fffea787>
>>>    next_to_watch        <0>
>>>    jiffies              <fffeaa30>
>>>    next_to_watch.status <0>
>>>  MAC Status             <40080080>
>>>  PHY Status             <7949>
>>>  PHY 1000BASE-T Status  <0>
>>>  PHY Extended Status    <3000>
>>>  PCI Status             <10>
>>>
>>>  Not too sure what else to post.  I am not subscribed to lkml so please cc
>> my email in your reply.
>>>
>>>
>>>  Link to complete dmesg: http://pastebin.com/zRBajGrY
>>>  Seems similar to: bugzilla.redhat.com/show_bug.cgi?id=785806
>>
>> It sounds like this is a regression, so it might be useful to know
>> what the newest working kernel was, and maybe a dmesg log from it as
>> well, though I don't see any obvious clues in the 3.9.0-2-ARCH dmesg
>> you collected.
>>
>> Bjorn
>
>
> Thank you for the reply, Bjorn.  3.8.11-1-ARCH works just fine for me.  Here is the dmesg from 3.8.11-1-ARCH per your request: http://pastebin.com/cUHwrQfq

Sorry this thread died.  Did this ever get resolved?

If not, can you collect "lspci -vv" output for the whole system on
both the working kernel and the failing one?

There are reports of similar symptoms at [1] and [2].  I can't tell
yet if you're seeing the same problem, but for [1], booting with
"pci=pcie_bus_peer2peer" was a workaround.

Bjorn

[1] http://lkml.kernel.org/r/509B5038.8090304@oracle.com [2012-11-08]
[2] http://lkml.kernel.org/r/4FFA9B96.6040901@oracle.com [2012-07-09]