Bug#456273: Bug #456273

[Paul Gear <paul@gear.dyndns.org>]

Moritz Muehlenhoff wrote:
> On Tue, Jan 22, 2008 at 11:07:58PM +0100, maximilian attems wrote:
>> On Wed, Jan 23, 2008 at 07:51:43AM +1000, Paul Gear wrote:
>>> I assume by kerneloops you mean the ksymoops package?  Is there anything
>>> special i need to do to enable serial output?
>> no i mean the kerneloops package.
>>
>> yes a serial cable and the corresponding console boot arg.
>>  
>>>> sorry for loosing track but you are already running 2.6.24-rc8?
>>> No, but i've only tried up to 2.6.24-rc7.  Nothing since 2.6.18-4 has
>>> worked on my system, except for Xen kernels.  I'll install 2.6.24-rc8
>>> and test now.
>> ok.
>>  
>>> Is there an easy way i can get a diff of the source differences between
>>> 2.6.18-4 and 2.6.18-5 so i can better report the issues to upstream?
>> yes if you checkout the subversion etch tree the series file nr 5 would
>> have the patches that had been added and you could do a binary search on
>> them enabling only half and so on until the faulty patch shows up.
> 
> Does this error still occur with more recent kernel versions?

The following kernels work:

linux-image-2.6.18-3-amd64               2.6.18-7
linux-image-2.6.18-3-vserver-amd64       2.6.18-7
linux-image-2.6.18-4-amd64               2.6.18.dfsg.1-12etch2
linux-image-2.6.18-4-vserver-amd64       2.6.18.dfsg.1-12etch2
linux-image-2.6.18-4-xen-vserver-amd64   2.6.18.dfsg.1-12etch2
linux-image-2.6.18-6-xen-vserver-amd64   2.6.18.dfsg.1-23etch1
linux-image-2.6.26-bpo.1-xen-amd64       2.6.26-12~bpo40+1

The following kernels don't:

linux-image-2.6.18-6-amd64               2.6.18.dfsg.1-23etch1
linux-image-2.6.18-6-vserver-amd64       2.6.18.dfsg.1-23etch1
linux-image-2.6.22-3-amd64               2.6.22-6~bpo40+1
linux-image-2.6.22-3-vserver-amd64       2.6.22-6~bpo40+1
linux-image-2.6.23-1-amd64               2.6.23-2
linux-image-2.6.24-etchnhalf.1-amd64     2.6.24-6~etchnhalf.7
linux-image-2.6.24-rc6-amd64
2.6.24~rc6-1~experimental.1~snapshot.10030
linux-image-2.6.24-rc7-amd64
2.6.24~rc7-1~experimental.1~snapshot.10109
linux-image-2.6.24-rc8-amd64
2.6.24~rc8-1~experimental.2~snapshot.10176
linux-image-2.6.26-bpo.1-amd64           2.6.26-12~bpo40+1
linux-image-2.6.26-bpo.1-vserver-amd64   2.6.26-12~bpo40+1

I'm no expert, but i can't help but think that there is a bug in this
model of the Intel 965G/ICH8 chipset that the normal kernel is
exercising (due to a change that happened at 2.6.18-4) but Xen is not,
due to different behaviour of the AGP or MM subsystem.

Paul