Re: 2.6.23.1-rt4 and kvm 48

[Carsten Emde <Carsten.Emde-Q945KHDl0DbYtjvyW6yDsg@public.gmane.org>]

Dor,

>>>> I tried to recreate your problem using 2.6.23-1 and latest rt patch 
>>>> (rt5).
>>>> The problem is that the kernel is not stable at all, I can't even
>>>> compile the code over vnc -
>>>> my connection is constantly lost. So it might not be kvm problem?
>>>> Can you try is with -no-kvm and see if it's working - then its just a
>>>> regular userspace process.
>>>> Anyway if all other things are stable on your end, can you send us
>>>> dmesg/strace outputs?
>>> Thanks, Dor, for offering help.
>>> The most recent RT-Preempt patch (2.6.23.1-rt5) still has an SMP issue
>>> which need to be sorted out first.
>>> When a working RT-Preempt patch is available, I will be glad to send
>>> dmesg outputs. We may then share our activities to create the next
>>> "Latest stable" RT-Preempt kernel that includes kvm support.
>> Wow, that smp issue may be a problem I've got a Intel Core 2 cpu with
>> just 2 cores.
>> but I've not seen a dump like this in a while (lustre?).
We see this all the time when we try to run unmodified kvm in an 
rt-kernel, nothing special.

>> I've updated to 2.6.23-rt5 and kvm-50 and I got this dump....
>> [..]
>> BUG: scheduling while atomic: 
>> [..]
>> This loops forever basically...
>> I'm really confused about this one. I've got the full minicom.cap file
>> from the session that produced this its attached zipped up.
>> If you'd like any other information I'd be glad to post it.
No, please not. Our log files are full of it. The "BUG: scheduling while 
atomic" problem is not related to the current SMP issue. It is the 
result of running rt-unaware code in an rt-kernel.

> If the rt patch is not smp stable you can boot with single cpu and run 
> kvm there.
> Maybe just pinning the guest to specific cpu will help although I had 
> troubles with X too.
This does not fix the problem. Save your energy. We live with this 
situation since kvm-28. Please wait until 2.6.23.<x>-rt<y> is stable. 
Thereafter, let's create together a kvm version that works in an 
rt-kernel and let's make sure that from this moment onwards subsequent 
kvm upgrade changes are tested in an rt-kernel before being released.

	--cbe

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/