From: Peter Zijlstra Subject: Re: [PATCH 00/20] x86: ticket lock rewrite and paravirtualization Date: Mon, 15 Nov 2010 21:14:48 +0100 Message-ID: <1289852088.2109.553.camel@laptop> References: <4CDDBBD3.5050903@zytor.com> <4CDDBCE4.80906@goop.org> <4CDDBDB5.8000800@zytor.com> <4CE1915F.60507@goop.org> <4CE1920F.5000509@zytor.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Return-path: In-Reply-To: <4CE1920F.5000509@zytor.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: "H. Peter Anvin" Cc: Jeremy Fitzhardinge , Xen-devel , Mathieu Desnoyers , Srivatsa, Vaddagiri , Linux Kernel Mailing List , Jan Beulich , Virtualization , Kivity Linux, Avi List-Id: xen-devel@lists.xenproject.org On Mon, 2010-11-15 at 12:03 -0800, H. Peter Anvin wrote: > On 11/15/2010 12:00 PM, Jeremy Fitzhardinge wrote: > >=20 > > Another approach I discussed with PeterZ and Mathieu is to steal the LS= B > > of the ticket counters (halving the max CPU count) to use as a "there i= s > > someone in slowpath waiting on this lock". But I haven't spent the tim= e > > to work out an algorithm to maintain that flag (or flags, since there > > are bits available) in a correct and efficient way. > >=20 >=20 > Definitely worth pondering. Right, so the idea was to make the ticket increment 2, which would leave the LSB of both the head and tail available. I think that if one were to set both (using a cmpxchg), the ticket fast-path wouldn't need any changes since head=3D=3Dtail is still the correct condition for acquisition= . Then the unlock needs an added conditional: if (tail & 1)=20 unlock_slowpath()