From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org [10.30.226.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 82B691ED38 for ; Tue, 25 Jul 2023 11:26:15 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 078F8C433C7; Tue, 25 Jul 2023 11:26:14 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linuxfoundation.org; s=korg; t=1690284375; bh=iO9QVdD47x8v4TjKLcrILY8ClzTFgslHn+QJsF9+AQE=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=ePRn+xDKpZ9LIOTspr3nd60dPfWx5T/qUVUge2AbTzUGguUtAtx7O6Un5oIx94TdK TpvOMwQ4THMVm9fXCEBD+U6RLPo1R6lLyzfzWtyqcmP//ts8EjwDgdOnPG1fdjDRbJ gpW0LfOmq0MwMgGXdcoewYK8ZL73itWbjPfEwA0s= From: Greg Kroah-Hartman To: stable@vger.kernel.org Cc: Greg Kroah-Hartman , patches@lists.linux.dev, Pavel Begunkov , io-uring@vger.kernel.org, linux-kernel@vger.kernel.org, Andres Freund , Jens Axboe Subject: [PATCH 5.10 333/509] io_uring: Use io_schedule* in cqring wait Date: Tue, 25 Jul 2023 12:44:32 +0200 Message-ID: <20230725104608.950442148@linuxfoundation.org> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230725104553.588743331@linuxfoundation.org> References: <20230725104553.588743331@linuxfoundation.org> User-Agent: quilt/0.67 Precedence: bulk X-Mailing-List: patches@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit From: Andres Freund Commit 8a796565cec3601071cbbd27d6304e202019d014 upstream. I observed poor performance of io_uring compared to synchronous IO. That turns out to be caused by deeper CPU idle states entered with io_uring, due to io_uring using plain schedule(), whereas synchronous IO uses io_schedule(). The losses due to this are substantial. On my cascade lake workstation, t/io_uring from the fio repository e.g. yields regressions between 20% and 40% with the following command: ./t/io_uring -r 5 -X0 -d 1 -s 1 -c 1 -p 0 -S$use_sync -R 0 /mnt/t2/fio/write.0.0 This is repeatable with different filesystems, using raw block devices and using different block devices. Use io_schedule_prepare() / io_schedule_finish() in io_cqring_wait_schedule() to address the difference. After that using io_uring is on par or surpassing synchronous IO (using registered files etc makes it reliably win, but arguably is a less fair comparison). There are other calls to schedule() in io_uring/, but none immediately jump out to be similarly situated, so I did not touch them. Similarly, it's possible that mutex_lock_io() should be used, but it's not clear if there are cases where that matters. Cc: stable@vger.kernel.org # 5.10+ Cc: Pavel Begunkov Cc: io-uring@vger.kernel.org Cc: linux-kernel@vger.kernel.org Signed-off-by: Andres Freund Link: https://lore.kernel.org/r/20230707162007.194068-1-andres@anarazel.de [axboe: minor style fixup] Signed-off-by: Jens Axboe Signed-off-by: Greg Kroah-Hartman --- io_uring/io_uring.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) --- a/io_uring/io_uring.c +++ b/io_uring/io_uring.c @@ -7625,7 +7625,7 @@ static inline int io_cqring_wait_schedul struct io_wait_queue *iowq, ktime_t *timeout) { - int ret; + int token, ret; /* make sure we run task_work before checking for signals */ ret = io_run_task_work_sig(); @@ -7635,9 +7635,17 @@ static inline int io_cqring_wait_schedul if (test_bit(0, &ctx->check_cq_overflow)) return 1; + /* + * Use io_schedule_prepare/finish, so cpufreq can take into account + * that the task is waiting for IO - turns out to be important for low + * QD IO. + */ + token = io_schedule_prepare(); + ret = 1; if (!schedule_hrtimeout(timeout, HRTIMER_MODE_ABS)) - return -ETIME; - return 1; + ret = -ETIME; + io_schedule_finish(token); + return ret; } /*