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-devel

RE: [Xen-devel] [Patch 2 of 2]: PV-domain SMP performance Linux-part

To: "James Harper" <james.harper@xxxxxxxxxxxxxxxx>, "Keir Fraser" <keir.fraser@xxxxxxxxxxxxx>, "Juergen Gross" <juergen.gross@xxxxxxxxxxxxxxxxxxx>, "Venefax" <venefax@xxxxxxxxx>
Subject: RE: [Xen-devel] [Patch 2 of 2]: PV-domain SMP performance Linux-part
From: "Steve Prochniak" <sprochniak@xxxxxxxxxxxxxxx>
Date: Fri, 16 Jan 2009 09:40:17 -0500
Cc: George Dunlap <George.Dunlap@xxxxxxxxxxxxx>, xen-devel@xxxxxxxxxxxxxxxxxxx
Delivery-date: Fri, 16 Jan 2009 06:37:35 -0800
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <AEC6C66638C05B468B556EA548C1A77D015503BE@trantor>
List-help: <mailto:xen-devel-request@lists.xensource.com?subject=help>
List-id: Xen developer discussion <xen-devel.lists.xensource.com>
List-post: <mailto:xen-devel@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
References: <AEC6C66638C05B468B556EA548C1A77D015503BC@trantor><C59614C1.21090%keir.fraser@xxxxxxxxxxxxx> <AEC6C66638C05B468B556EA548C1A77D015503BE@trantor>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
Thread-index: Acl3rwQU85YH/4sVSvS2s/NknOY4ggAEnvzwAAFbsOQAAK3twAAHnhlA
Thread-topic: [Xen-devel] [Patch 2 of 2]: PV-domain SMP performance Linux-part
'enlightened' windows OSs have SpinLock routines that make a hypercall
to yield CPU to another VM.  So if you comply with the windows
hypervisor spec, you can get a performance boost when virtualized.  This
doesn't help you out with Pre - vista versions though...  Look at the
disassembly for KfAcquireSpinLock.

-----Original Message-----
From: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
[mailto:xen-devel-bounces@xxxxxxxxxxxxxxxxxxx] On Behalf Of James Harper
Sent: Friday, January 16, 2009 6:02 AM
To: Keir Fraser; Juergen Gross; Venefax
Cc: George Dunlap; xen-devel@xxxxxxxxxxxxxxxxxxx
Subject: RE: [Xen-devel] [Patch 2 of 2]: PV-domain SMP performance
Linux-part

> On 16/01/2009 10:16, "James Harper" <james.harper@xxxxxxxxxxxxxxxx>
wrote:
> 
> > I had previously wondered about optimizing spinlocks, my idea was
> > basically for Xen to set a bit in a structure to indicate what vcpus
are
> > currently scheduled, and my modified spinlock acquire routine would
> > check if the current vcpu wants a spinlock that is held by a
currently
> > unscheduled vcpu, and if so yield to Xen to let the other vcpu
schedule.
> 
> That's a lot more like our existing Linux pv_ops spinlock handling
> (yield/block instead of spin) than Juergen's patch (don't deschedule
me
> while in a critical section). The difference from what you suggest is
that
> we heuristically detect unscheduled lock holders by spinning a short
> while.
> 
> You can pv up your Windows spinlocks in the block-instead-of-spin way
> already (and yield-instead-of-spin is obviously even easier).
> 

But only in spinlocks that I 'own' completely right? I'm more concerned
about spinlocks that I share with Windows (eg in NDIS).

James

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

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

<Prev in Thread] Current Thread [Next in Thread>