Bug 405058

Removing and inserting network card module should work too (modprobe -r module_name; modprobe module_name) and it can be added as workaround into suspend, resume scripts. This is probably bug in network driver.

Please attach output from "hwinfo --network" to see which network card you have.

Cyril, output attached. My LAN is integrated on the Gigabyte GA-M57SLI-S4  motherboard, Marvell 88E1116 chipset. 
Best wishes.


28: None 00.0: 10701 Ethernet
  [Created at net.124]
  Unique ID: usDW.ndpeucax6V1
  Parent ID: rBUF.nENVDTNIEjC
  SysFS ID: /class/net/eth0
  SysFS Device Link: /devices/pci0000:00/0000:00:08.0
  Hardware Class: network interface
  Model: "Ethernet network interface"
  Driver: "forcedeth"
  Driver Modules: "forcedeth"
  Device File: eth0
  HW Address: 00:1d:7d:a6:aa:57
  Link detected: yes
  Config Status: cfg=no, avail=yes, need=no, active=unknown
  Attached to: #11 (Ethernet controller)

29: None 00.0: 10700 Loopback
  [Created at net.124]
  Unique ID: ZsBS.GQNx7L4uPNA
  SysFS ID: /class/net/lo
  Hardware Class: network interface
  Model: "Loopback network interface"
  Device File: lo
  Link detected: yes
  Config Status: cfg=no, avail=yes, need=no, active=unknown

Please next time remove NEEDINFO status by selecting "This comment/attachment 
provides ..." checkbox after supplying needed information. Reassigning to kernel maintainers.

Does rmmoding/insmoding the forcedeth module help?

Pavel - thanks for the clue. I did not know the name of the module before. My knowledge of text commands is not good, but I tried "modprobe -r forcedeth", followed by "modprobe forcedeth", and am happy to report that I was reconnected to my router. All is now well, but how can I make this automatic?

Ok, so problem is in the forcedeth module, good.

You should be able to add it to the blacklist in powersave, but preffered solution is to fix forcedeth module not to do that. .... I took a look and forcedeth module seems to have some suspend/resume support.

Stefan, do we have machine needing forcedeth in our test pool? Does suspend/resume work there?

#5: what was exact sentence of events? suspend, modprobe -r, modprobe?

It was suspend, resume (followed by a check to make sure router was not available), then modprobe -r, modprobe. Router was then immediately available.

Peter: Can you verify this is still a problem on recent vanilla kernel?

Stefan: can we add forcedeth to list of modules for unloading / can I get machine with forcedeth network card?

Pavel - yes the problem is still there.

I suspect 2.6.22.12 kernel doesn't affect by this bug, due to bug #342049, comment #14.

This bug is reported in Kernel Bug Tracker too:
http://bugzilla.kernel.org/show_bug.cgi?id=10487

I don't have forcedeth nearby and Rafael is already solving similar problem in upstream => reassign...

I commited patches for SLES10SP1/2, but I have no way to test them. It seems like redhat has a testcase...

Anyway, patch is pretty straightforward, so...

Oops, sorry, wrong bug. Ignore comment #13.

I can confirm that forcedeth is fine vs suspend/hibernation on my test box with 2.6.27-rc4 .

Cyril, would it be possible to test 2.6.27-rc4 from kernel.org?

2.6.27-rc4 works for me too.

So for the next product, with a new kernel, this is fixed.

For 11.0, just create a pm-utils hook /etc/pm/sleep.d/05forcedeth_reload containing:
-----8<-----
#!/bin/bash

case $1 in
    thaw|resume)
        echo "*** reloading forcedeth ***"
        modprobe -r forcedeth
        modprobe forcedeth
        ;;
    *)  ;;
esac
-----8<-----
and do a "chmod +x /etc/pm/sleep.d/05forcedeth_reload"

This will unload and reload forcedeth after resume.

I run into the same issue with a MSI MS-7226 main board.  This as has two MCP55 Ethernet (rev a2) on board.

Using a newer kernel like kernel-default-2.6.27-rc6.17.1 as available from Kernel:/HEAD/openSUSE_Factory of the Build Service solved this issue in my case.

Unfortunately I have to use the workaround as provided with comment #18 as in the case of the kernel update I can't longer use X.org.

http://linux.derkeiler.com/Mailing-Lists/Kernel/2008-05/msg08220.html might be the same issue.

(In reply to comment #18 from Stefan Seyfried)
> So for the next product, with a new kernel, this is fixed.
> 
> For 11.0, just create a pm-utils hook /etc/pm/sleep.d/05forcedeth_reload
> containing:
> -----8<-----
> #!/bin/bash
> 
> case $1 in
>     thaw|resume)
>         echo "*** reloading forcedeth ***"
>         modprobe -r forcedeth
>         modprobe forcedeth
>         ;;
>     *)  ;;
> esac
> -----8<-----
> and do a "chmod +x /etc/pm/sleep.d/05forcedeth_reload"
> 
> This will unload and reload forcedeth after resume.
> 

I tried this fix, but my PC locked up completely after returning from suspend.

Had to reboot.

Linux 75-92-212-161.sea.clearwire-dns.net 2.6.25.16-0.1-default #1 SMP 2008-08-21 00:34:25 +0200 x86_64 x86_64 x86_64 GNU/Linux

(In reply to comment #20 from John Bailo)
> (In reply to comment #18 from Stefan Seyfried)

> > This will unload and reload forcedeth after resume.
> > 
> 
> I tried this fix, but my PC locked up completely after returning from suspend.

But it works if you unload and reload the module manually after resume?