[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] A clocksource question



On 03/11/2010 05:55 AM, Jeremy Fitzhardinge wrote:
> On 03/10/2010 05:19 PM, Joanna Rutkowska wrote:
>> On 11/03/2010 02:06, Jeremy Fitzhardinge wrote:
>>   
>>> On 03/10/2010 04:52 PM, Joanna Rutkowska wrote:
>>>     
>>>> BTW, how does the clocksource=jiffies work on a pvops kernel in Dom0?
>>>>
>>>>        
>>> Not very well.  clocksource=jiffies just sets up timer interrupts at
>>> approx 100ms intervals and assumes that's 100ms.  You get very low res
>>> timers and timekeeping.
>>>
>>>      
>> So, how to explain that there is no wallclock drift *at all*, even in a
>> long run -- uptime of a few days, in Dom0, when it uses the jiffies
>> source?
>>    
> 
> You said earlier that you were seeing clock drift with
> clocksource=jiffies in dom0.
> 
Correct, sorry. With jiffies I do get a wallclock drift, yes.

>> Anyway, I assume that the "xen" clock source is much more fine grained
>> (1ms?)
> 
> The xen clocksource has nanosecond resolution.  But clocksources are
> different from event sources, and so the ns resolution of time
> measurement doesn't have much relationship to the timer precision (which
> is always going to use the xen event source, which is also ns
> resolution, but it will tend to fold together timer events which are
> closer than 50us).
> 
>>   and so, maybe my kbd hiccups are caused by some code executed by
>> the timer interrupt too frequently (maybe too much code executes per
>> each timer interrupt, because of some other bug)? Just a though...
>>    
> 
> Possibly, but doesn't seem too likely.  Can you tell if your hiccups are
> accompanied by bursts of timer interrupts in /proc/interrupts?
> 
> There is another possibility, which is that the scheduler is getting
> confused by xen's scheduler clock.  Rather than just scheduling based on
> real time, we try to take into account time stolen from a vcpu so that
> it isn't credited against the process (which may have had all its time
> stolen by another domain).  But that could just be confusing things.
> 
> Does this help?
> 
> diff --git a/arch/x86/xen/time.c b/arch/x86/xen/time.c

Will build a new kernel with this and let you know.

j.

Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.