WARNING - OLD ARCHIVES

This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
   
 
 
Xen 
 
Home Products Support Community News
 
   
 

xen-ia64-devel

RE: [Xen-ia64-devel] SMP guest and itc

To: "Dong, Eddie" <eddie.dong@xxxxxxxxx>, "Williamson, Alex (Linux Kernel Dev)" <alex.williamson@xxxxxx>, "Tristan Gingold" <Tristan.Gingold@xxxxxxxx>
Subject: RE: [Xen-ia64-devel] SMP guest and itc
From: "Magenheimer, Dan (HP Labs Fort Collins)" <dan.magenheimer@xxxxxx>
Date: Fri, 10 Feb 2006 10:15:47 -0800
Cc: xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
Delivery-date: Fri, 10 Feb 2006 18:27:36 +0000
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
List-help: <mailto:xen-ia64-devel-request@lists.xensource.com?subject=help>
List-id: Discussion of the ia64 port of Xen <xen-ia64-devel.lists.xensource.com>
List-post: <mailto:xen-ia64-devel@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-ia64-devel>, <mailto:xen-ia64-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-ia64-devel>, <mailto:xen-ia64-devel-request@lists.xensource.com?subject=unsubscribe>
Sender: xen-ia64-devel-bounces@xxxxxxxxxxxxxxxxxxx
Thread-index: AcYtmn6i7qz6HFxIT1CqTf846bPMiAAZCGiAABtqfXA=
Thread-topic: [Xen-ia64-devel] SMP guest and itc
Just to briefly reiterate my opinion on the general
timer tick issue:  The vast majority of timer ticks
get immediately reflected to the currently executing
domain.   Since ticks are very frequent (though recently
reduced in Linux/ia64 from 1024/sec to 250/sec), there
needs to be a "fast path" that can easily determine
if this is a "immediately reflect" tick or not.  It
shouldn't be necessary to go through all the save/restore
overhead to call C and then march down linked
lists only to determine that this is indeed one of the
99% of ticks that should be immediately reflected.

That said, I agree that Xen needs to pre-synchronize ITC
for host SMP to work properly.  I don't think guest SMP
for Linux will force ITC to be paravirtualized, but
I guess Tristan will let us know if he discovers otherwise.

Dan

> -----Original Message-----
> From: xen-ia64-devel-bounces@xxxxxxxxxxxxxxxxxxx 
> [mailto:xen-ia64-devel-bounces@xxxxxxxxxxxxxxxxxxx] On Behalf 
> Of Dong, Eddie
> Sent: Thursday, February 09, 2006 10:04 PM
> To: Williamson, Alex (Linux Kernel Dev); Tristan Gingold
> Cc: xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
> Subject: RE: [Xen-ia64-devel] SMP guest and itc
> 
> Tristan:
>       This is a known issue and I ever disucssed time virtualization
> with Dan for a while. 
>       I prefer #1. That let domU see its ITC start from 0 at boot
> time? And further more, we need to support Windows eventually.
>       In the meantime, I think both #1 and #3 should co-exist. Yes,
> isn't the host ITC pre-synchronized in Xen?
> Eddie.
> 
> Alex Williamson wrote:
> > On Thu, 2006-02-09 at 08:32 +0100, Tristan Gingold wrote:
> >> Hi,
> >> 
> >> during boot, linux kernel tries to synchronize itc of all
> >> processors.  However currently itc is not virtualized:
> >> (XEN) vcpu_set_itc: Setting ar.itc is currently disabled
> >> 
> >> What should we do ?
> >> a) virtualize ITC
> >> b) para-virtualize ITC (ie, modifying linux kernel)
> >> c) pre-synchronize ITC
> >> 
> >> I really prefer c).
> > 
> >    I'm not sure I understand c).  Aren't the ITCs already
> > pre-synchronized by xen in smp_callin()?   Thanks,
> > 
> >     Alex
> 
> 
> _______________________________________________
> Xen-ia64-devel mailing list
> Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-ia64-devel
> 

_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ia64-devel