e1000e broken after resume on x230 [was: Uhhuh. NMI received for unknown reason 2c on CPU 0.]

On 02/15/2013 09:54 AM, Konstantin Khlebnikov wrote:

Borislav Petkov wrote:

quoted

On Thu, Feb 14, 2013 at 10:17:46AM -0700, Bjorn Helgaas wrote:

quoted

It's pretty late for v3.8, but let me know if you think they're
critical.

Ok, I meant those:

http://marc.info/?l=linux-kernel&m=135984592927219

They fix the link detection issue on my x230. So let's see. The first
one is:

* https://patchwork.kernel.org/patch/2001211/ (e1000e: fix pci device
enable
counter balance)

please use this instead:

Hi, I am a bit confused. Is this fixed in -next yet? And if so, is it
known which commit IDs are needed to fix the issue in 3.7 (see below)?

[PATCH v2 1/7] e1000e: fix pci-device enable-counter balance
https://lkml.org/lkml/2013/2/4/190

from v2 patchset: https://lkml.org/lkml/2013/2/4/185

So this is now in -next as:
commit e34f7147d93afe5efc574734bbff6584c0cc4a02
Author: Konstantin Khlebnikov [off-list ref]
Date:   Mon Feb 25 09:19:04 2013 +0400

    e1000e: fix pci-device enable-counter balance

quoted

I don't see that one in your tree.

* pci-pm-clear-state_saved-during-suspend.patch

This is:
commit 82fee4d67ab86d6fe5eb0f9a9e988ca9d654d765
Author: Rafael J. Wysocki [off-list ref]
Date:   Mon Feb 4 15:56:05 2013 +0400

    PCI/PM: Clear state_saved during suspend

quoted

Rafael said this one is a real bugfix. Looks like e1000e maintainers are
picking that one?

* pci-pm-fix-e1000e-runtime-suspend.patch

Is this one replaced by a different fix in the end? Which one? I don't
think it is in -next yet, right?

confused,
-- 
js
suse labs