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 lists.gnu.org (lists.gnu.org [209.51.188.17]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id 58A60EA4FBA for ; Mon, 23 Feb 2026 14:22:11 +0000 (UTC) Received: from localhost ([::1] helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1vuWoJ-0004JP-19; Mon, 23 Feb 2026 09:21:19 -0500 Received: from eggs.gnu.org ([2001:470:142:3::10]) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1vuWnu-00049z-EN for qemu-devel@nongnu.org; Mon, 23 Feb 2026 09:20:55 -0500 Received: from us-smtp-delivery-124.mimecast.com ([170.10.129.124]) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1vuWns-0005X1-0G for qemu-devel@nongnu.org; Mon, 23 Feb 2026 09:20:54 -0500 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1771856450; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: in-reply-to:in-reply-to:references:references; bh=6Cpp2uOtTuAerYkO3V2kjH5f8ubujF6oq4kL+eaS50s=; b=KRqsqob8urXgVIErg8P3m57dfFdrKIvBrUzHIbVPoMh6yXXLHU9EzbrE7Q3ZjySHnw9Rs4 oPnTWVJqfh5JW0GMlg7LQ2ElAmXIpNi/7YnB8NpyFqmtogh1wskfnaXcBxLS8TePMu9OnK l6PAGaSOKGYaN8ISpV0S/56jdPjtRKo= Received: from mx-prod-mc-06.mail-002.prod.us-west-2.aws.redhat.com (ec2-35-165-154-97.us-west-2.compute.amazonaws.com [35.165.154.97]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-653-XvZxzZesP4G8IluxpSf9Tw-1; Mon, 23 Feb 2026 09:20:46 -0500 X-MC-Unique: XvZxzZesP4G8IluxpSf9Tw-1 X-Mimecast-MFC-AGG-ID: XvZxzZesP4G8IluxpSf9Tw_1771856445 Received: from mx-prod-int-01.mail-002.prod.us-west-2.aws.redhat.com (mx-prod-int-01.mail-002.prod.us-west-2.aws.redhat.com [10.30.177.4]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by mx-prod-mc-06.mail-002.prod.us-west-2.aws.redhat.com (Postfix) with ESMTPS id A90FD18005B7; Mon, 23 Feb 2026 14:20:44 +0000 (UTC) Received: from localhost (unknown [10.2.16.160]) by mx-prod-int-01.mail-002.prod.us-west-2.aws.redhat.com (Postfix) with ESMTP id 1E94730001BE; Mon, 23 Feb 2026 14:20:42 +0000 (UTC) Date: Mon, 23 Feb 2026 08:53:49 -0500 From: Stefan Hajnoczi To: Kevin Wolf Cc: Jens Axboe , Fiona Ebner , qemu-block@nongnu.org, qemu-devel@nongnu.org, fam@euphon.net Subject: Re: [PATCH v2] aio-posix: notify main loop when SQEs are queued Message-ID: <20260223135349.GA1162700@fedora> References: <20260213143225.161043-1-axboe@kernel.dk> <20260213143225.161043-2-axboe@kernel.dk> <20260218161127.GC587447@fedora> <07d701b9-3039-4f9b-99a2-abeae51146a5@kernel.dk> MIME-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha512; protocol="application/pgp-signature"; boundary="oN9CSbfbIWir6E0A" Content-Disposition: inline In-Reply-To: X-Scanned-By: MIMEDefang 3.4.1 on 10.30.177.4 Received-SPF: pass client-ip=170.10.129.124; envelope-from=stefanha@redhat.com; helo=us-smtp-delivery-124.mimecast.com X-Spam_score_int: -1 X-Spam_score: -0.2 X-Spam_bar: / X-Spam_report: (-0.2 / 5.0 requ) BAYES_00=-1.9, DKIMWL_WL_HIGH=-0.001, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, RCVD_IN_DNSWL_NONE=-0.0001, RCVD_IN_MSPIKE_H4=0.001, RCVD_IN_MSPIKE_WL=0.001, RCVD_IN_VALIDITY_CERTIFIED_BLOCKED=1.179, RCVD_IN_VALIDITY_RPBL_BLOCKED=0.717, SPF_HELO_PASS=-0.001, SPF_PASS=-0.001 autolearn=no autolearn_force=no X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: qemu development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-devel-bounces+qemu-devel=archiver.kernel.org@nongnu.org Sender: qemu-devel-bounces+qemu-devel=archiver.kernel.org@nongnu.org --oN9CSbfbIWir6E0A Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Thu, Feb 19, 2026 at 04:49:11PM +0100, Kevin Wolf wrote: > Am 18.02.2026 um 17:41 hat Jens Axboe geschrieben: > > On 2/18/26 9:19 AM, Jens Axboe wrote: > > > On 2/18/26 9:11 AM, Stefan Hajnoczi wrote: > > >> On Wed, Feb 18, 2026 at 10:57:02AM +0100, Fiona Ebner wrote: > > >>> Am 13.02.26 um 5:05 PM schrieb Kevin Wolf: > > >>>> Am 13.02.2026 um 15:26 hat Jens Axboe geschrieben: > > >>>>> When a vCPU thread handles MMIO (holding BQL), aio_co_enter() run= s the > > >>>>> block I/O coroutine inline on the vCPU thread because > > >>>>> qemu_get_current_aio_context() returns the main AioContext when B= QL is > > >>>>> held. The coroutine calls luring_co_submit() which queues an SQE = via > > >>>>> fdmon_io_uring_add_sqe(), but the actual io_uring_submit() only h= appens > > >>>>> in gsource_prepare() on the main loop thread. > > >>>> > > >>>> Ouch! Yes, looks like we completely missed I/O submitted in vCPU t= hreads > > >>>> in the recent changes (or I guess worker threads in theory, but I = don't > > >>>> think there any that actually make use of aio_add_sqe()). > > >>>> > > >>>>> Since the coroutine ran inline (not via aio_co_schedule()), no BH= is > > >>>>> scheduled and aio_notify() is never called. The main loop remains= asleep > > >>>>> in ppoll() with up to a 499ms timeout, leaving the SQE unsubmitte= d until > > >>>>> the next timer fires. > > >>>>> > > >>>>> Fix this by calling aio_notify() after queuing the SQE. This wake= s the > > >>>>> main loop via the eventfd so it can run gsource_prepare() and sub= mit the > > >>>>> pending SQE promptly. > > >>>>> > > >>>>> This is a generic fix that benefits all devices using aio=3Dio_ur= ing. > > >>>>> Without it, AHCI/SATA devices see MUCH worse I/O latency since th= ey use > > >>>>> MMIO (not ioeventfd like virtio) and have no other mechanism to w= ake the > > >>>>> main loop after queuing block I/O. > > >>>>> > > >>>>> This is usually a bit hard to detect, as it also relies on the pp= oll > > >>>>> loop not waking up for other activity, and micro benchmarks tend = not to > > >>>>> see it because they don't have any real processing time. With a > > >>>>> synthetic test case that has a few usleep() to simulate processin= g of > > >>>>> read data, it's very noticeable. The below example reads 128MB wi= th > > >>>>> O_DIRECT in 128KB chunks in batches of 16, and has a 1ms delay be= fore > > >>>>> each batch submit, and a 1ms delay after processing each completi= on. > > >>>>> Running it on /dev/sda yields: > > >>>>> > > >>>>> time sudo ./iotest /dev/sda > > >>>>> > > >>>>> ________________________________________________________ > > >>>>> Executed in 25.76 secs fish external > > >>>>> usr time 6.19 millis 783.00 micros 5.41 millis > > >>>>> sys time 12.43 millis 642.00 micros 11.79 millis > > >>>>> > > >>>>> while on a virtio-blk or NVMe device we get: > > >>>>> > > >>>>> time sudo ./iotest /dev/vdb > > >>>>> > > >>>>> ________________________________________________________ > > >>>>> Executed in 1.25 secs fish external > > >>>>> usr time 1.40 millis 0.30 millis 1.10 millis > > >>>>> sys time 17.61 millis 1.43 millis 16.18 millis > > >>>>> > > >>>>> time sudo ./iotest /dev/nvme0n1 > > >>>>> > > >>>>> ________________________________________________________ > > >>>>> Executed in 1.26 secs fish external > > >>>>> usr time 6.11 millis 0.52 millis 5.59 millis > > >>>>> sys time 13.94 millis 1.50 millis 12.43 millis > > >>>>> > > >>>>> where the latter are consistent. If we run the same test but keep= the > > >>>>> socket for the ssh connection active by having activity there, th= en > > >>>>> the sda test looks as follows: > > >>>>> > > >>>>> time sudo ./iotest /dev/sda > > >>>>> > > >>>>> ________________________________________________________ > > >>>>> Executed in 1.23 secs fish external > > >>>>> usr time 2.70 millis 39.00 micros 2.66 millis > > >>>>> sys time 4.97 millis 977.00 micros 3.99 millis > > >>>>> > > >>>>> as now the ppoll loop is woken all the time anyway. > > >>>>> > > >>>>> After this fix, on an idle system: > > >>>>> > > >>>>> time sudo ./iotest /dev/sda > > >>>>> > > >>>>> ________________________________________________________ > > >>>>> Executed in 1.30 secs fish external > > >>>>> usr time 2.14 millis 0.14 millis 2.00 millis > > >>>>> sys time 16.93 millis 1.16 millis 15.76 millis > > >>>>> > > >>>>> Signed-off-by: Jens Axboe > > >>>>> --- > > >>>>> util/fdmon-io_uring.c | 8 ++++++++ > > >>>>> 1 file changed, 8 insertions(+) > > >>>>> > > >>>>> diff --git a/util/fdmon-io_uring.c b/util/fdmon-io_uring.c > > >>>>> index d0b56127c670..96392876b490 100644 > > >>>>> --- a/util/fdmon-io_uring.c > > >>>>> +++ b/util/fdmon-io_uring.c > > >>>>> @@ -181,6 +181,14 @@ static void fdmon_io_uring_add_sqe(AioContex= t *ctx, > > >>>>> =20 > > >>>>> trace_fdmon_io_uring_add_sqe(ctx, opaque, sqe->opcode, sqe->= fd, sqe->off, > > >>>>> cqe_handler); > > >>>>> + > > >>>>> + /* > > >>>>> + * Wake the main loop if it is sleeping in ppoll(). When a = vCPU thread > > >>>>> + * runs a coroutine inline (holding BQL), it queues SQEs her= e but the > > >>>>> + * actual io_uring_submit() only happens in gsource_prepare(= ). Without > > >>>>> + * this notify, ppoll() can sleep up to 499ms before submitt= ing. > > >>>>> + */ > > >>>>> + aio_notify(ctx); > > >>>>> } > > >>>> > > >>>> Makes sense to me. > > >>>> > > >>>> At first I wondered if we should use defer_call() for the aio_noti= fy() > > >>>> to batch the submission, but of course holding the BQL will alread= y take > > >>>> care of that. And in iothreads where there is no BQL, the aio_noti= fy() > > >>>> shouldn't make a difference anyway because we're already in the ri= ght > > >>>> thread. > > >>>> > > >>>> I suppose the other variation could be have another io_uring_enter= () > > >>>> call here (but then probably really through defer_call()) to avoid > > >>>> waiting for another CPU to submit the request in its main loop. Bu= t I > > >>>> don't really have an intuition if that would make things better or= worse > > >>>> in the common case. > > >>>> > > >>>> Fiona, does this fix your case, too? > > >>> > > >>> Yes, it does fix my issue [0] and the second patch gives another sm= all > > >>> improvement :) > > >>> > > >>> Would it be slightly cleaner to have aio_add_sqe() call aio_notify() > > >>> itself? Since aio-posix.c calls downwards into fdmon-io_uring.c, it > > >>> would feel nicer to me to not have fdmon-io_uring.c call "back up".= I > > >>> guess it also depends on whether we expect another future fdmon > > >>> implementation with .add_sqe() to also benefit from it. > > >> > > >> Calling aio_notify() from aio-posix.c:aio_add_sqe() sounds better to= me > > >> because fdmon-io_uring.c has to be careful about calling aio_*() API= s to > > >> avoid loops. > > >=20 > > > Would anyone care to make that edit? I'm on a plane and gone for a bi= t, > > > so won't get back to this for the next week. But I would love to see a > > > fix go in, as this issue has been plaguing me with test timeouts for > > > quite a while on the CI front. And seems like I'm not alone, if the > > > patches fix Fiona's issues as well. > >=20 > > Still on a plane but tested this one and it works for me too. Does seem > > like a better approach, rather than stuff it in the fdmon part. > >=20 > > Feel free to run with this one and also to update the commit message if > > you want. Thanks! > >=20 > >=20 > > commit a8a94e7a05964d470b8fba50c9d4769489c21752 > > Author: Jens Axboe > > Date: Fri Feb 13 06:52:14 2026 -0700 > >=20 > > aio-posix: notify main loop when SQEs are queued > > =20 > > When a vCPU thread handles MMIO (holding BQL), aio_co_enter() runs = the > > block I/O coroutine inline on the vCPU thread because > > qemu_get_current_aio_context() returns the main AioContext when BQL= is > > held. The coroutine calls luring_co_submit() which queues an SQE via > > fdmon_io_uring_add_sqe(), but the actual io_uring_submit() only hap= pens > > in gsource_prepare() on the main loop thread. > > =20 > > Since the coroutine ran inline (not via aio_co_schedule()), no BH is > > scheduled and aio_notify() is never called. The main loop remains a= sleep > > in ppoll() with up to a 499ms timeout, leaving the SQE unsubmitted = until > > the next timer fires. > > =20 > > Fix this by calling aio_notify() after queuing the SQE. This wakes = the > > main loop via the eventfd so it can run gsource_prepare() and submi= t the > > pending SQE promptly. > > =20 > > This is a generic fix that benefits all devices using aio=3Dio_urin= g. > > Without it, AHCI/SATA devices see MUCH worse I/O latency since they= use > > MMIO (not ioeventfd like virtio) and have no other mechanism to wak= e the > > main loop after queuing block I/O. > > =20 > > This is usually a bit hard to detect, as it also relies on the ppoll > > loop not waking up for other activity, and micro benchmarks tend no= t to > > see it because they don't have any real processing time. With a > > synthetic test case that has a few usleep() to simulate processing = of > > read data, it's very noticeable. The below example reads 128MB with > > O_DIRECT in 128KB chunks in batches of 16, and has a 1ms delay befo= re > > each batch submit, and a 1ms delay after processing each completion. > > Running it on /dev/sda yields: > > =20 > > time sudo ./iotest /dev/sda > > =20 > > ________________________________________________________ > > Executed in 25.76 secs fish external > > usr time 6.19 millis 783.00 micros 5.41 millis > > sys time 12.43 millis 642.00 micros 11.79 millis > > =20 > > while on a virtio-blk or NVMe device we get: > > =20 > > time sudo ./iotest /dev/vdb > > =20 > > ________________________________________________________ > > Executed in 1.25 secs fish external > > usr time 1.40 millis 0.30 millis 1.10 millis > > sys time 17.61 millis 1.43 millis 16.18 millis > > =20 > > time sudo ./iotest /dev/nvme0n1 > > =20 > > ________________________________________________________ > > Executed in 1.26 secs fish external > > usr time 6.11 millis 0.52 millis 5.59 millis > > sys time 13.94 millis 1.50 millis 12.43 millis > > =20 > > where the latter are consistent. If we run the same test but keep t= he > > socket for the ssh connection active by having activity there, then > > the sda test looks as follows: > > =20 > > time sudo ./iotest /dev/sda > > =20 > > ________________________________________________________ > > Executed in 1.23 secs fish external > > usr time 2.70 millis 39.00 micros 2.66 millis > > sys time 4.97 millis 977.00 micros 3.99 millis > > =20 > > as now the ppoll loop is woken all the time anyway. > > =20 > > After this fix, on an idle system: > > =20 > > time sudo ./iotest /dev/sda > > =20 > > ________________________________________________________ > > Executed in 1.30 secs fish external > > usr time 2.14 millis 0.14 millis 2.00 millis > > sys time 16.93 millis 1.16 millis 15.76 millis > > =20 > > Signed-off-by: Jens Axboe > >=20 > > diff --git a/util/aio-posix.c b/util/aio-posix.c > > index e24b955fd91a..8c7b3795c82d 100644 > > --- a/util/aio-posix.c > > +++ b/util/aio-posix.c > > @@ -813,5 +813,13 @@ void aio_add_sqe(void (*prep_sqe)(struct io_uring_= sqe *sqe, void *opaque), > > { > > AioContext *ctx =3D qemu_get_current_aio_context(); > > ctx->fdmon_ops->add_sqe(ctx, prep_sqe, opaque, cqe_handler); > > + > > + /* > > + * Wake the main loop if it is sleeping in ppoll(). When a vCPU t= hread > > + * runs a coroutine inline (holding BQL), it queues SQEs here but = the >=20 > I think the comment could even be more generic here. This is not > specific to coroutines, but the scenario is just that a vCPU thread > holding the BQL performs I/O. Good idea, I generalized the comment when merging the patch. Stefan --oN9CSbfbIWir6E0A Content-Type: application/pgp-signature; name=signature.asc -----BEGIN PGP SIGNATURE----- iQEzBAEBCgAdFiEEhpWov9P5fNqsNXdanKSrs4Grc8gFAmmcW+0ACgkQnKSrs4Gr c8jt0wgAsQ8OupU5OeTv1+MhURm4bN2XXaYOj/VqvFQQ5UMmopbnI9gjZJLf3GjM qjebH+NwpLLfVnHjtbiajRcJdTsGZFRs/stjn0d/1MYF0UdzPfAvoEkFipjCCrKR q5H9QwXVewiSszDv6hjn1e63+R5bek4N6ugbJ9HToV7iKTk0aWC3XUJtn8R4xVqz 8RceHvRmd+F3p49fZpPbf4HAvX6n08UBUeUGlCjs0gey+53bacEpVANn+SjU/qYL G2PXkypYvInkBm+WMvp+9OY4LanTlOQYxt8UgSH8HIQ1bAKE9MIncf8IUIxNJRQ5 sOi6c2DdcaichNNmpolqaBHa5+d0Aw== =7VGg -----END PGP SIGNATURE----- --oN9CSbfbIWir6E0A--