I have non-hardened related problems on my laptop since 3.8.4.
The integrated e1000e NIC drops carrier unexpectedly. In the logs it seems
the NIC gets turned of for a moment, but won't come back despite another
message in the log. On that moment the wireless card is enabled
temporarily. It seemed to be kinda power-management issue from the
beginning. It is very annoying. The NIC cen be brought back by
unloading-reloading the driver. Since I'm using a hardened kernel with
modules locked, I can't do that. And I also don't want to play that game.
It doesn't happen on my server having integrated e1000e NICs on the mobo.
Recently I had time to dig into that, and have found some relevant links:
https://bbs.archlinux.org/viewtopic.php?id=159454 https://lkml.org/lkml/2013/1/18/147 https://bugzilla.kernel.org/show_bug.cgi?id=52021 https://lkml.org/lkml/2013/1/5/100 It's a known regression and there's a workaround (sysctl or udev). There
is a 5 element patch set also available by Konstantin Khlebnikov. That
solves the problem for me. But it's still not committed in the kernel. I
don't know if it will be committed or not, or what is the reason it's not
already included in the current kernel source.
I just wanted to let those know, who may also struggle with such a problem.
Regards:
Dw.
--
dr Tóth Attila, Radiológus, 06-20-825-8057
Attila Toth MD, Radiologist, +36-20-825-8057
2013.Ãprilis 19.(P) 14:19 idÅ‘pontban Anthony G. Basile ezt Ãrta:
> On 04/17/2013 10:24 AM, Alexander Tsoy wrote:
>> Ð’ Wed, 17 Apr 2013 16:07:39 +0300
>> ungifted01@gmail.com пишет:
>>
>>> Hi!
>>>
>>>
>>> Can't build hardened 3.8.3 (x86)
>>>
>>>
>>> CC kernel/kallsyms.o
>>> CC kernel/acct.o
>>> CC kernel/user_namespace.o
>>> kernel/user_namespace.c: In function 'userns_install':
>>> kernel/user_namespace.c:808:25: error: invalid operands to binary !=
>>> (have 'atomic_t' and 'int')
>>> make[1]: *** [kernel/user_namespace.o] Error 1
>>> make: *** [kernel] Error 2
>>>
>>>
>>
>> Hello
>>
>> This is a known issue
>> http://forums.grsecurity.net/viewtopic.php?t=3358&p=12707
>>
>
> Thanks for bringing that to my attention. Looks like 3.8.3 was not a
> good choice. Let me see if I can back port the fix or stabilize the
> next bump.
>
> --
> Anthony G. Basile, Ph. D.
> Chair of Information Technology
> D'Youville College
> Buffalo, NY 14201
> (716) 829-8197
>
>