From mboxrd@z Thu Jan 1 00:00:00 1970 From: Konrad Rzeszutek Wilk Subject: Re: [PATCH 3/3] xen/block: add multi-page ring support Date: Tue, 9 Jun 2015 10:21:27 -0400 Message-ID: <20150609142126.GJ15200__36591.5526789196$1433859810$gmane$org@x230> References: <1433310003-13089-1-git-send-email-bob.liu@oracle.com> <1433310003-13089-3-git-send-email-bob.liu@oracle.com> <5576A8C0.8000804@oracle.com> <9AAE0902D5BC7E449B7C8E4E778ABCD0259410F3@AMSPEX01CL01.citrite.net> <20150609133938.GA15200@x230> <5576F325.5050304@citrix.com> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Return-path: Content-Disposition: inline In-Reply-To: <5576F325.5050304@citrix.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: Roger Pau =?iso-8859-1?Q?Monn=E9?= Cc: Rafal Mielniczuk , "linux-kernel@vger.kernel.org" , Marcus Granado , "xen-devel@lists.xen.org" , Julien Grall , "justing@spectralogic.com" , Paul Durrant , David Vrabel List-Id: xen-devel@lists.xenproject.org On Tue, Jun 09, 2015 at 04:07:33PM +0200, Roger Pau Monn=E9 wrote: > El 09/06/15 a les 15.39, Konrad Rzeszutek Wilk ha escrit: > > On Tue, Jun 09, 2015 at 08:52:53AM +0000, Paul Durrant wrote: > >>> -----Original Message----- > >>> From: Bob Liu [mailto:bob.liu@oracle.com] > >>> Sent: 09 June 2015 09:50 > >>> To: Bob Liu > >>> Cc: xen-devel@lists.xen.org; David Vrabel; justing@spectralogic.com; > >>> konrad.wilk@oracle.com; Roger Pau Monne; Paul Durrant; Julien Grall; = linux- > >>> kernel@vger.kernel.org > >>> Subject: Re: [PATCH 3/3] xen/block: add multi-page ring support > >>> > >>> > >>> On 06/03/2015 01:40 PM, Bob Liu wrote: > >>>> Extend xen/block to support multi-page ring, so that more requests c= an be > >>>> issued by using more than one pages as the request ring between blkf= ront > >>>> and backend. > >>>> As a result, the performance can get improved significantly. > >>>> > >>>> We got some impressive improvements on our highend iscsi storage clu= ster > >>>> backend. If using 64 pages as the ring, the IOPS increased about 15 = times > >>>> for the throughput testing and above doubled for the latency testing. > >>>> > >>>> The reason was the limit on outstanding requests is 32 if use only o= ne-page > >>>> ring, but in our case the iscsi lun was spread across about 100 phys= ical > >>>> drives, 32 was really not enough to keep them busy. > >>>> > >>>> Changes in v2: > >>>> - Rebased to 4.0-rc6. > >>>> - Document on how multi-page ring feature working to linux io/blkif= .h. > >>>> > >>>> Changes in v3: > >>>> - Remove changes to linux io/blkif.h and follow the protocol defined > >>>> in io/blkif.h of XEN tree. > >>>> - Rebased to 4.1-rc3 > >>>> > >>>> Changes in v4: > >>>> - Turn to use 'ring-page-order' and 'max-ring-page-order'. > >>>> - A few comments from Roger. > >>>> > >>>> Changes in v5: > >>>> - Clarify with 4k granularity to comment > >>>> - Address more comments from Roger > >>>> > >>>> Signed-off-by: Bob Liu > >>> > >>> Also tested the windows PV driver which also works fine when multi-pa= ge > >>> ring feature > >>> was enabled in Linux backend. > >>> http://www.xenproject.org/downloads/windows-pv-drivers.html > >>> > >> > >> Great! Thanks for verifying that :-) > > = > > Woot! Bob, could you repost the blkif.h patch for the Xen tree > > pleas e and also mention the testing part in it please? I think this > > was the only big 'what if?!' question holding this up. > > = > > = > > Roger, I put them (patches) on devel/for-jens-4.2 on > > = > > git://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen.git > > = > > I think these two patches: > > drivers: xen-blkback: delay pending_req allocation to connect_ring > > xen/block: add multi-page ring support > > = > > are the only ones that haven't been Acked by you (or maybe they > > have and I missed the Ack?) > = > Hello, > = > I was waiting to Ack those because the XenServer storage performance > folks found out that these patches cause a performance regression on > some of their tests. I'm adding them to the conversation so they can This is with multi-page enabled or with the patches but multi-page disabled (baseline)? > provide more details about the issues they found, and whether we should > hold pushing this patches or not. Or surely fix whatever is causing this. > = > Roger. > =