From: "Jan Beulich" Subject: RE: [Patch 2 of 2]: PV-domain SMP performance Linux-part Date: Fri, 16 Jan 2009 11:18:08 +0000 Message-ID: <49707B00.76E4.0078.0@novell.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: quoted-printable Return-path: In-Reply-To: Content-Disposition: inline List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: James Harper Cc: George Dunlap , xen-devel@lists.xensource.com, Juergen Gross , Keir Fraser , Venefax List-Id: xen-devel@lists.xenproject.org >>> "James Harper" 16.01.09 12:01 >>> >> You can pv up your Windows spinlocks in the block-instead-of-spin way >> already (and yield-instead-of-spin is obviously even easier). >>=20 > >But only in spinlocks that I 'own' completely right? I'm more concerned >about spinlocks that I share with Windows (eg in NDIS). As long as you can hook the respective OS interface (and you know it is always used), you could do this on all spinlocks, since all that's needed = is logic in the acquire/release code (unless you want to add state to each lock [like the spinning CPUs bitmap Keir mentioned], but that shouldn't be necessary to achieve the intended effect). Jan