[Vivid/Utopic/Trusty][SRU][PATCH] x86: pvclock: Really remove the sched notifier for cross-cpu migrations

Chris J Arges chris.j.arges at canonical.com
Fri Sep 11 14:14:03 UTC 2015


On 09/11/2015 08:55 AM, Seth Forshee wrote:
> On Fri, Sep 11, 2015 at 08:06:37AM -0500, Chris J Arges wrote:
>> FYI. I still think this patch makes sense for our distro kernel based on
>> the commit description alone. However the original reporter is unsure of
>> this fixing issues as reflected in the bug.
> 
> I'm not convinced.
> 
> Based on the description this patch seems to be of the "remove some
> unneeded cruft" variety rather than of the "fix a bug" variety. We don't
> have any testing to indicate that it does fix any bugs afaict, so on
> that count it fails to adhere to our SRU policy.
> 
> According to LP #1450584 we do have testing which says that
> 0a4e6be9ca17c54817cf814b4b5aa60478c6df27 and
> 80f7fdb1c7f0f9266421f823964fd1962681f6ce fix a bug. If I apply this
> patch to vivid I find that those commits have been completely reverted,
> so wouldn't this reintroduce that bug?
> 
> Thanks,
> Seth
> 

Agree, I think more testing should be done, I'll see if I repro the
original issue with/without this patch. Please hold off on accepting
this patch.
Thanks,
--chris




More information about the kernel-team mailing list