From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 7C16FC38A02 for ; Fri, 28 Oct 2022 13:34:11 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S230458AbiJ1NeK (ORCPT ); Fri, 28 Oct 2022 09:34:10 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:59838 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S230490AbiJ1NeG (ORCPT ); Fri, 28 Oct 2022 09:34:06 -0400 Received: from mail-ed1-x52c.google.com (mail-ed1-x52c.google.com [IPv6:2a00:1450:4864:20::52c]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 4176F1D067E for ; Fri, 28 Oct 2022 06:34:02 -0700 (PDT) Received: by mail-ed1-x52c.google.com with SMTP id i21so7879839edj.10 for ; Fri, 28 Oct 2022 06:34:02 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=bytedance-com.20210112.gappssmtp.com; s=20210112; h=content-transfer-encoding:cc:to:subject:message-id:date:from :in-reply-to:references:mime-version:from:to:cc:subject:date :message-id:reply-to; bh=e9wwq0N7Q/p0HNN6IR/TsA18qAO4m6l6eUZVnQdWFgM=; b=wRePGTkMXdMf/zQdaoSKJKStCoIkXi9M1Neybau8ShPum5JeloJLVMZrlMaOCZbKCv Rny4Gl9ZBK3I5P8POGInr0HCcCb3GfJ+hQVHqo8qOiaViUYsXfTN3csC7mka79HJXW7L o91OMhZ2+iXyNDj1KBsGwmPXH1AyIPhzM1JorCSP1RjjWLzy99JmYsJNaVcl5rY/LLtZ qYgERP3JEdxwiB7bTCLi0YXDFjQNykLdGI7bzhfnepfRSJ4USP6IsKQJg/Q2rOkcEYZk F+K9yXPmaDQ45j3hoHfJssoK2eGzvd1kJt0DKDKHQBZ6/a3gpDcSzEORB0KiI26niXLg v4sQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=content-transfer-encoding:cc:to:subject:message-id:date:from :in-reply-to:references:mime-version:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=e9wwq0N7Q/p0HNN6IR/TsA18qAO4m6l6eUZVnQdWFgM=; b=eqEyG6koT6ghfVCpezEG17v9AOD4LY9oF0OH+7TaHNSVelW67CQxzRfSiB7mkvkyG+ P7fG9l1j9Zy/QfXDqQGqZDEBbnI0Aola3QOTBiymcOsCLcjpsH+plMDX7gottC7cKQv2 ho1jj29k4vVYhFfgDvIzCgYjQIxd3jnaW8xdmM0HaO39hwVBOp4eOyV2E76hNrQLIHsW UR2QHRnFw6TjuFhailx4o1GDij06WWhTeiboH5pM6RQreg1pBNka6fjtQM84uqwZAJgv +/7HsWv8nKwypRVC7PNZR09YbI3gFX4QIFGs83DD/TV6rwoFpeHnpfCGGuNqq3SHshIM ZqGw== X-Gm-Message-State: ACrzQf30L+UEOeTQJT8aG65KZ/HeVkLx1P4lzpb9Ip1IoHNTf/TcJLHe thF8aovORnqA7EW35Qlg0OnukSl/w/GYRRCpRazc+fxOlw== X-Google-Smtp-Source: AMsMyM7lmgzWJ8wLxRQFR8+2QFn3steG7yk2UO2LGPDChWLE9Q8r4zSVg2Y4kmqtLa1VaUiABetZX6t2CDrNBjqWfso= X-Received: by 2002:a50:c387:0:b0:461:3ae6:8cdf with SMTP id h7-20020a50c387000000b004613ae68cdfmr37368962edf.131.1666964040549; Fri, 28 Oct 2022 06:34:00 -0700 (PDT) MIME-Version: 1.0 References: <50827796-af93-4af5-4121-dc13c31a67fc@linux.alibaba.com> <1d051d63-ce34-1bb3-2256-4ced4be6d690@redhat.com> In-Reply-To: From: Yongji Xie Date: Fri, 28 Oct 2022 21:33:49 +0800 Message-ID: Subject: Re: ublk-qcow2: ublk-qcow2 is available To: Stefan Hajnoczi Cc: Jason Wang , "Michael S. Tsirkin" , Ming Lei , Ziyang Zhang , Stefan Hajnoczi , io-uring@vger.kernel.org, linux-block@vger.kernel.org, linux-kernel , "Denis V. Lunev" , Xiaoguang Wang Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Precedence: bulk List-ID: X-Mailing-List: io-uring@vger.kernel.org On Tue, Oct 25, 2022 at 8:02 PM Stefan Hajnoczi wrote: > > On Tue, 25 Oct 2022 at 04:17, Yongji Xie wrote: > > > > On Fri, Oct 21, 2022 at 2:30 PM Jason Wang wrote: > > > > > > > > > =E5=9C=A8 2022/10/21 13:33, Yongji Xie =E5=86=99=E9=81=93: > > > > On Tue, Oct 18, 2022 at 10:54 PM Stefan Hajnoczi wrote: > > > >> On Tue, 18 Oct 2022 at 09:17, Yongji Xie = wrote: > > > >>> On Tue, Oct 18, 2022 at 2:59 PM Ming Lei = wrote: > > > >>>> On Mon, Oct 17, 2022 at 07:11:59PM +0800, Yongji Xie wrote: > > > >>>>> On Fri, Oct 14, 2022 at 8:57 PM Ming Lei wrote: > > > >>>>>> On Thu, Oct 13, 2022 at 02:48:04PM +0800, Yongji Xie wrote: > > > >>>>>>> On Wed, Oct 12, 2022 at 10:22 PM Stefan Hajnoczi wrote: > > > >>>>>>>> On Sat, 8 Oct 2022 at 04:43, Ziyang Zhang wrote: > > > >>>>>>>>> On 2022/10/5 12:18, Ming Lei wrote: > > > >>>>>>>>>> On Tue, Oct 04, 2022 at 09:53:32AM -0400, Stefan Hajnoczi = wrote: > > > >>>>>>>>>>> On Tue, 4 Oct 2022 at 05:44, Ming Lei wrote: > > > >>>>>>>>>>>> On Mon, Oct 03, 2022 at 03:53:41PM -0400, Stefan Hajnocz= i wrote: > > > >>>>>>>>>>>>> On Fri, Sep 30, 2022 at 05:24:11PM +0800, Ming Lei wrot= e: > > > >>>>>>>>>>>>>> ublk-qcow2 is available now. > > > >>>>>>>>>>>>> Cool, thanks for sharing! > > > >>>>>>>>>>>>> > > > >>>>>>>>>>>>>> So far it provides basic read/write function, and comp= ression and snapshot > > > >>>>>>>>>>>>>> aren't supported yet. The target/backend implementatio= n is completely > > > >>>>>>>>>>>>>> based on io_uring, and share the same io_uring with ub= lk IO command > > > >>>>>>>>>>>>>> handler, just like what ublk-loop does. > > > >>>>>>>>>>>>>> > > > >>>>>>>>>>>>>> Follows the main motivations of ublk-qcow2: > > > >>>>>>>>>>>>>> > > > >>>>>>>>>>>>>> - building one complicated target from scratch helps l= ibublksrv APIs/functions > > > >>>>>>>>>>>>>> become mature/stable more quickly, since qcow2 is c= omplicated and needs more > > > >>>>>>>>>>>>>> requirement from libublksrv compared with other sim= ple ones(loop, null) > > > >>>>>>>>>>>>>> > > > >>>>>>>>>>>>>> - there are several attempts of implementing qcow2 dri= ver in kernel, such as > > > >>>>>>>>>>>>>> ``qloop`` [2], ``dm-qcow2`` [3] and ``in kernel qco= w2(ro)`` [4], so ublk-qcow2 > > > >>>>>>>>>>>>>> might useful be for covering requirement in this fi= eld > > > >>>>>>>>>>>>>> > > > >>>>>>>>>>>>>> - performance comparison with qemu-nbd, and it was my = 1st thought to evaluate > > > >>>>>>>>>>>>>> performance of ublk/io_uring backend by writing one= ublk-qcow2 since ublksrv > > > >>>>>>>>>>>>>> is started > > > >>>>>>>>>>>>>> > > > >>>>>>>>>>>>>> - help to abstract common building block or design pat= tern for writing new ublk > > > >>>>>>>>>>>>>> target/backend > > > >>>>>>>>>>>>>> > > > >>>>>>>>>>>>>> So far it basically passes xfstest(XFS) test by using = ublk-qcow2 block > > > >>>>>>>>>>>>>> device as TEST_DEV, and kernel building workload is ve= rified too. Also > > > >>>>>>>>>>>>>> soft update approach is applied in meta flushing, and = meta data > > > >>>>>>>>>>>>>> integrity is guaranteed, 'make test T=3Dqcow2/040' cov= ers this kind of > > > >>>>>>>>>>>>>> test, and only cluster leak is reported during this te= st. > > > >>>>>>>>>>>>>> > > > >>>>>>>>>>>>>> The performance data looks much better compared with q= emu-nbd, see > > > >>>>>>>>>>>>>> details in commit log[1], README[5] and STATUS[6]. And= the test covers both > > > >>>>>>>>>>>>>> empty image and pre-allocated image, for example of pr= e-allocated qcow2 > > > >>>>>>>>>>>>>> image(8GB): > > > >>>>>>>>>>>>>> > > > >>>>>>>>>>>>>> - qemu-nbd (make test T=3Dqcow2/002) > > > >>>>>>>>>>>>> Single queue? > > > >>>>>>>>>>>> Yeah. > > > >>>>>>>>>>>> > > > >>>>>>>>>>>>>> randwrite(4k): jobs 1, iops 24605 > > > >>>>>>>>>>>>>> randread(4k): jobs 1, iops 30938 > > > >>>>>>>>>>>>>> randrw(4k): jobs 1, iops read 13981 write 14001 > > > >>>>>>>>>>>>>> rw(512k): jobs 1, iops read 724 write 728 > > > >>>>>>>>>>>>> Please try qemu-storage-daemon's VDUSE export type as w= ell. The > > > >>>>>>>>>>>>> command-line should be similar to this: > > > >>>>>>>>>>>>> > > > >>>>>>>>>>>>> # modprobe virtio_vdpa # attaches vDPA devices to ho= st kernel > > > >>>>>>>>>>>> Not found virtio_vdpa module even though I enabled all t= he following > > > >>>>>>>>>>>> options: > > > >>>>>>>>>>>> > > > >>>>>>>>>>>> --- vDPA drivers > > > >>>>>>>>>>>> vDPA device simulator core > > > >>>>>>>>>>>> vDPA simulator for networking device > > > >>>>>>>>>>>> vDPA simulator for block device > > > >>>>>>>>>>>> VDUSE (vDPA Device in Userspace) suppor= t > > > >>>>>>>>>>>> Intel IFC VF vDPA driver > > > >>>>>>>>>>>> Virtio PCI bridge vDPA driver > > > >>>>>>>>>>>> vDPA driver for Alibaba ENI > > > >>>>>>>>>>>> > > > >>>>>>>>>>>> BTW, my test environment is VM and the shared data is do= ne in VM too, and > > > >>>>>>>>>>>> can virtio_vdpa be used inside VM? > > > >>>>>>>>>>> I hope Xie Yongji can help explain how to benchmark VDUSE= . > > > >>>>>>>>>>> > > > >>>>>>>>>>> virtio_vdpa is available inside guests too. Please check = that > > > >>>>>>>>>>> VIRTIO_VDPA ("vDPA driver for virtio devices") is enabled= in "Virtio > > > >>>>>>>>>>> drivers" menu. > > > >>>>>>>>>>> > > > >>>>>>>>>>>>> # modprobe vduse > > > >>>>>>>>>>>>> # qemu-storage-daemon \ > > > >>>>>>>>>>>>> --blockdev file,filename=3Dtest.qcow2,cache.dire= ct=3Dof|off,aio=3Dnative,node-name=3Dfile \ > > > >>>>>>>>>>>>> --blockdev qcow2,file=3Dfile,node-name=3Dqcow2 \ > > > >>>>>>>>>>>>> --object iothread,id=3Diothread0 \ > > > >>>>>>>>>>>>> --export vduse-blk,id=3Dvduse0,name=3Dvduse0,num= -queues=3D$(nproc),node-name=3Dqcow2,writable=3Don,iothread=3Diothread0 > > > >>>>>>>>>>>>> # vdpa dev add name vduse0 mgmtdev vduse > > > >>>>>>>>>>>>> > > > >>>>>>>>>>>>> A virtio-blk device should appear and xfstests can be r= un on it > > > >>>>>>>>>>>>> (typically /dev/vda unless you already have other virti= o-blk devices). > > > >>>>>>>>>>>>> > > > >>>>>>>>>>>>> Afterwards you can destroy the device using: > > > >>>>>>>>>>>>> > > > >>>>>>>>>>>>> # vdpa dev del vduse0 > > > >>>>>>>>>>>>> > > > >>>>>>>>>>>>>> - ublk-qcow2 (make test T=3Dqcow2/022) > > > >>>>>>>>>>>>> There are a lot of other factors not directly related t= o NBD vs ublk. In > > > >>>>>>>>>>>>> order to get an apples-to-apples comparison with qemu-*= a ublk export > > > >>>>>>>>>>>>> type is needed in qemu-storage-daemon. That way only th= e difference is > > > >>>>>>>>>>>>> the ublk interface and the rest of the code path is ide= ntical, making it > > > >>>>>>>>>>>>> possible to compare NBD, VDUSE, ublk, etc more precisel= y. > > > >>>>>>>>>>>> Maybe not true. > > > >>>>>>>>>>>> > > > >>>>>>>>>>>> ublk-qcow2 uses io_uring to handle all backend IO(includ= e meta IO) completely, > > > >>>>>>>>>>>> and so far single io_uring/pthread is for handling all q= cow2 IOs and IO > > > >>>>>>>>>>>> command. > > > >>>>>>>>>>> qemu-nbd doesn't use io_uring to handle the backend IO, s= o we don't > > > >>>>>>>>>> I tried to use it via --aio=3Dio_uring for setting up qemu= -nbd, but not succeed. > > > >>>>>>>>>> > > > >>>>>>>>>>> know whether the benchmark demonstrates that ublk is fast= er than NBD, > > > >>>>>>>>>>> that the ublk-qcow2 implementation is faster than qemu-nb= d's qcow2, > > > >>>>>>>>>>> whether there are miscellaneous implementation difference= s between > > > >>>>>>>>>>> ublk-qcow2 and qemu-nbd (like using the same io_uring con= text for both > > > >>>>>>>>>>> ublk and backend IO), or something else. > > > >>>>>>>>>> The theory shouldn't be too complicated: > > > >>>>>>>>>> > > > >>>>>>>>>> 1) io uring passthough(pt) communication is fast than sock= et, and io command > > > >>>>>>>>>> is carried over io_uring pt commands, and should be fast t= han virio > > > >>>>>>>>>> communication too. > > > >>>>>>>>>> > > > >>>>>>>>>> 2) io uring io handling is fast than libaio which is taken= in the > > > >>>>>>>>>> test on qemu-nbd, and all qcow2 backend io(include meta io= ) is handled > > > >>>>>>>>>> by io_uring. > > > >>>>>>>>>> > > > >>>>>>>>>> https://github.com/ming1/ubdsrv/blob/master/tests/common/q= cow2_common > > > >>>>>>>>>> > > > >>>>>>>>>> 3) ublk uses one single io_uring to handle all io commands= and qcow2 > > > >>>>>>>>>> backend IOs, so batching handling is common, and it is eas= y to see > > > >>>>>>>>>> dozens of IOs/io commands handled in single syscall, or ev= en more. > > > >>>>>>>>>> > > > >>>>>>>>>>> I'm suggesting measuring changes to just 1 variable at a = time. > > > >>>>>>>>>>> Otherwise it's hard to reach a conclusion about the root = cause of the > > > >>>>>>>>>>> performance difference. Let's learn why ublk-qcow2 perfor= ms well. > > > >>>>>>>>>> Turns out the latest Fedora 37-beta doesn't support vdpa y= et, so I built > > > >>>>>>>>>> qemu from the latest github tree, and finally it starts to= work. And test kernel > > > >>>>>>>>>> is v6.0 release. > > > >>>>>>>>>> > > > >>>>>>>>>> Follows the test result, and all three devices are setup a= s single > > > >>>>>>>>>> queue, and all tests are run in single job, still done in = one VM, and > > > >>>>>>>>>> the test images are stored on XFS/virito-scsi backed SSD. > > > >>>>>>>>>> > > > >>>>>>>>>> The 1st group tests all three block device which is backed= by empty > > > >>>>>>>>>> qcow2 image. > > > >>>>>>>>>> > > > >>>>>>>>>> The 2nd group tests all the three block devices backed by = pre-allocated > > > >>>>>>>>>> qcow2 image. > > > >>>>>>>>>> > > > >>>>>>>>>> Except for big sequential IO(512K), there is still not sma= ll gap between > > > >>>>>>>>>> vdpa-virtio-blk and ublk. > > > >>>>>>>>>> > > > >>>>>>>>>> 1. run fio on block device over empty qcow2 image > > > >>>>>>>>>> 1) qemu-nbd > > > >>>>>>>>>> running qcow2/001 > > > >>>>>>>>>> run perf test on empty qcow2 image via nbd > > > >>>>>>>>>> fio (nbd(/mnt/data/ublk_null_8G_nYbgF.qcow2), libai= o, bs 4k, dio, hw queues:1)... > > > >>>>>>>>>> randwrite: jobs 1, iops 8549 > > > >>>>>>>>>> randread: jobs 1, iops 34829 > > > >>>>>>>>>> randrw: jobs 1, iops read 11363 write 11333 > > > >>>>>>>>>> rw(512k): jobs 1, iops read 590 write 597 > > > >>>>>>>>>> > > > >>>>>>>>>> > > > >>>>>>>>>> 2) ublk-qcow2 > > > >>>>>>>>>> running qcow2/021 > > > >>>>>>>>>> run perf test on empty qcow2 image via ublk > > > >>>>>>>>>> fio (ublk/qcow2( -f /mnt/data/ublk_null_8G_s761j.qc= ow2), libaio, bs 4k, dio, hw queues:1, uring_comp: 0, get_data: 0). > > > >>>>>>>>>> randwrite: jobs 1, iops 16086 > > > >>>>>>>>>> randread: jobs 1, iops 172720 > > > >>>>>>>>>> randrw: jobs 1, iops read 35760 write 35702 > > > >>>>>>>>>> rw(512k): jobs 1, iops read 1140 write 1149 > > > >>>>>>>>>> > > > >>>>>>>>>> 3) vdpa-virtio-blk > > > >>>>>>>>>> running debug/test_dev > > > >>>>>>>>>> run io test on specified device > > > >>>>>>>>>> fio (vdpa(/dev/vdc), libaio, bs 4k, dio, hw queues:= 1)... > > > >>>>>>>>>> randwrite: jobs 1, iops 8626 > > > >>>>>>>>>> randread: jobs 1, iops 126118 > > > >>>>>>>>>> randrw: jobs 1, iops read 17698 write 17665 > > > >>>>>>>>>> rw(512k): jobs 1, iops read 1023 write 1031 > > > >>>>>>>>>> > > > >>>>>>>>>> > > > >>>>>>>>>> 2. run fio on block device over pre-allocated qcow2 image > > > >>>>>>>>>> 1) qemu-nbd > > > >>>>>>>>>> running qcow2/002 > > > >>>>>>>>>> run perf test on pre-allocated qcow2 image via nbd > > > >>>>>>>>>> fio (nbd(/mnt/data/ublk_data_8G_sc0SB.qcow2), libai= o, bs 4k, dio, hw queues:1)... > > > >>>>>>>>>> randwrite: jobs 1, iops 21439 > > > >>>>>>>>>> randread: jobs 1, iops 30336 > > > >>>>>>>>>> randrw: jobs 1, iops read 11476 write 11449 > > > >>>>>>>>>> rw(512k): jobs 1, iops read 718 write 722 > > > >>>>>>>>>> > > > >>>>>>>>>> 2) ublk-qcow2 > > > >>>>>>>>>> running qcow2/022 > > > >>>>>>>>>> run perf test on pre-allocated qcow2 image via ublk > > > >>>>>>>>>> fio (ublk/qcow2( -f /mnt/data/ublk_data_8G_yZiaJ.qc= ow2), libaio, bs 4k, dio, hw queues:1, uring_comp: 0, get_data: 0). > > > >>>>>>>>>> randwrite: jobs 1, iops 98757 > > > >>>>>>>>>> randread: jobs 1, iops 110246 > > > >>>>>>>>>> randrw: jobs 1, iops read 47229 write 47161 > > > >>>>>>>>>> rw(512k): jobs 1, iops read 1416 write 1427 > > > >>>>>>>>>> > > > >>>>>>>>>> 3) vdpa-virtio-blk > > > >>>>>>>>>> running debug/test_dev > > > >>>>>>>>>> run io test on specified device > > > >>>>>>>>>> fio (vdpa(/dev/vdc), libaio, bs 4k, dio, hw queues:= 1)... > > > >>>>>>>>>> randwrite: jobs 1, iops 47317 > > > >>>>>>>>>> randread: jobs 1, iops 74092 > > > >>>>>>>>>> randrw: jobs 1, iops read 27196 write 27234 > > > >>>>>>>>>> rw(512k): jobs 1, iops read 1447 write 1458 > > > >>>>>>>>>> > > > >>>>>>>>>> > > > >>>>>>>>> Hi All, > > > >>>>>>>>> > > > >>>>>>>>> We are interested in VDUSE vs UBLK, too. And I have tested = them with nullblk backend. > > > >>>>>>>>> Let me share some results here. > > > >>>>>>>>> > > > >>>>>>>>> I setup UBLK with: > > > >>>>>>>>> ublk add -t loop -f /dev/nullb0 -d QUEUE_DEPTH -q NR_QUE= UE > > > >>>>>>>>> > > > >>>>>>>>> I setup VDUSE with: > > > >>>>>>>>> qemu-storage-daemon \ > > > >>>>>>>>> --chardev socket,id=3Dcharmonitor,path=3D/tmp/qmp.s= ock,server=3Don,wait=3Doff \ > > > >>>>>>>>> --monitor chardev=3Dcharmonitor \ > > > >>>>>>>>> --blockdev driver=3Dhost_device,cache.direct=3Don,f= ilename=3D/dev/nullb0,node-name=3Ddisk0 \ > > > >>>>>>>>> --export vduse-blk,id=3Dtest,node-name=3Ddisk0,name= =3Dvduse_test,writable=3Don,num-queues=3DNR_QUEUE,queue-size=3DQUEUE_DEPTH > > > >>>>>>>>> > > > >>>>>>>>> Here QUEUE_DEPTH is 1, 32 or 128 and NR_QUEUE is 1 or 4. > > > >>>>>>>>> > > > >>>>>>>>> Note: > > > >>>>>>>>> (1) VDUSE requires QUEUE_DEPTH >=3D 2. I cannot setup QUEUE= _DEPTH to 1. > > > >>>>>>>>> (2) I use qemu 7.1.0-rc3. It supports vduse-blk. > > > >>>>>>>>> (3) I do not use ublk null target so that the test is fair. > > > >>>>>>>>> (4) I setup fio with direct=3D1, bs=3D4k. > > > >>>>>>>>> > > > >>>>>>>>> ------------------------------ > > > >>>>>>>>> 1 job 1 iodepth, lat=EF=BC=88usec) > > > >>>>>>>>> vduse ublk > > > >>>>>>>>> seq-read 22.55 11.15 > > > >>>>>>>>> rand-read 22.49 11.17 > > > >>>>>>>>> seq-write 25.67 10.25 > > > >>>>>>>>> rand-write 24.13 10.16 > > > >>>>>>>> Thanks for sharing. Any idea what the bottlenecks are for vd= use and ublk? > > > >>>>>>>> > > > >>>>>>> I think one reason for the latency gap of sync I/O is that vd= use uses > > > >>>>>>> workqueue in the I/O completion path but ublk doesn't. > > > >>>>>>> > > > >>>>>>> And one bottleneck for the async I/O in vduse is that vduse w= ill do > > > >>>>>>> memcpy inside the critical section of virtqueue's spinlock in= the > > > >>>>>>> virtio-blk driver. That will hurt the performance heavily whe= n > > > >>>>>>> virtio_queue_rq() and virtblk_done() run concurrently. And it= can be > > > >>>>>>> mitigated by the advance DMA mapping feature [1] or irq bindi= ng > > > >>>>>>> support [2]. > > > >>>>>> Hi Yongji, > > > >>>>>> > > > >>>>>> Yeah, that is the cost you paid for virtio. Wrt. userspace blo= ck device > > > >>>>>> or other sort of userspace devices, cmd completion is driven b= y > > > >>>>>> userspace, not sure if one such 'irq' is needed. > > > >>>>> I'm not sure, it can be an optional feature in the future if ne= eded. > > > >>>>> > > > >>>>>> Even not sure if virtio > > > >>>>>> ring is one good choice for such use case, given io_uring has = been proved > > > >>>>>> as very efficient(should be better than virtio ring, IMO). > > > >>>>>> > > > >>>>> Since vduse is aimed at creating a generic userspace device fra= mework, > > > >>>>> virtio should be the right way IMO. > > > >>>> OK, it is the right way, but may not be the effective one. > > > >>>> > > > >>> Maybe, but I think we can try to optimize it. > > > >>> > > > >>>>> And with the vdpa framework, the > > > >>>>> userspace device can serve both virtual machines and containers= . > > > >>>> virtio is good for VM, but not sure it is good enough for other > > > >>>> cases. > > > >>>> > > > >>>>> Regarding the performance issue, actually I can't measure how m= uch of > > > >>>>> the performance loss is due to the difference between virtio ri= ng and > > > >>>>> iouring. But I think it should be very small. The main costs co= me from > > > >>>>> the two bottlenecks I mentioned before which could be mitigated= in the > > > >>>>> future. > > > >>>> Per my understanding, at least there are two places where virtio= ring is > > > >>>> less efficient than io_uring: > > > >>>> > > > >>> I might have misunderstood what you mean by virtio ring before. M= y > > > >>> previous understanding of the virtio ring does not include the > > > >>> virtio-blk driver. > > > >>> > > > >>>> 1) io_uring uses standalone submission queue(SQ) and completion = queue(CQ), > > > >>>> so no contention exists between submission and completion; but v= irtio queue > > > >>>> requires per-vq lock in both submission and completion. > > > >>>> > > > >>> Yes, this is the bottleneck of the virtio-blk driver, even in the= VM > > > >>> case. We are also trying to optimize this lock. > > > >>> > > > >>> One way to mitigate it is making submission and completion happen= in > > > >>> the same core. > > > >> QEMU sizes virtio-blk device num-queues to match the vCPU count. T= he > > > >> virtio-blk driver is a blk-mq driver, so submissions and completio= ns > > > >> for a given virtqueue should already be processed by the same vCPU= . > > > >> > > > >> Unless the device is misconfigured or the guest software chooses a > > > >> custom vq:vCPU mapping, there should be no vq lock contention betw= een > > > >> vCPUs. > > > >> > > > >> I can think of a reason why submission and completion require > > > >> coordination: descriptors are occupied until completion. The > > > >> submission logic chooses free descriptors from the table. The > > > >> completion logic returns free descriptors so they can be used in > > > >> future submissions. > > > >> > > > > Yes, we need to maintain a head pointer of the free descriptors in > > > > both submission and completion path. > > > > > > > > > Not necessarily after IN_ORDER? > > > > > > > Sounds like a good idea. > > Submission and completion are still not 100% independent with IN_ORDER > because descriptors are still in use until completion. It may not be > necessary to keep a freelist, but you cannot actually use the > descriptors for new submissions until existing requests complete. Is > that correct? > Yes. But we can get rid of the per-vq lock at least. > Anyway, independent submission and completion rings aren't perfect > either because independent submission introduces a new point of > communication: the device must tell the driver when submitted > descriptors have been processed. That means the driver must access a > hardware register on the device or the device must DMA to RAM. So it > involves extra bus traffic that is not necessary if descriptors are in > use until completion. io_uring gets away with it because the > io_uring_enter(2) syscall is synchronous and can therefore return the > number of consumed sq elements for free. > > There are ways to minimize that cost: > 1. The driver only needs to fetch the device's sq index when it has > run out of sq ring space. > 2. The device can include sq index updates with completions. This is > what NVMe does with the CQE SQ Head Pointer field, but the > disadvantage is that the driver has no way of determining the sq index > until a completion occurs. > It seems that the per-vq lock is still needed in this way if IN_ORDER is not supported. We still need to maintain a list of the free descriptors since out-of-order completion may occur. Thanks, Yongji