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 X-Spam-Level: X-Spam-Status: No, score=-8.3 required=3.0 tests=DKIM_SIGNED,DKIM_VALID, DKIM_VALID_AU,FREEMAIL_FORGED_FROMDOMAIN,FREEMAIL_FROM, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY, SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED,USER_AGENT_SANE_1 autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 4C1BFC43331 for ; Sun, 10 Nov 2019 13:37:21 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 07C0C20818 for ; Sun, 10 Nov 2019 13:37:20 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (2048-bit key) header.d=gmail.com header.i=@gmail.com header.b="uZVHwx6I" Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726390AbfKJNhT (ORCPT ); Sun, 10 Nov 2019 08:37:19 -0500 Received: from mail-wm1-f65.google.com ([209.85.128.65]:34005 "EHLO mail-wm1-f65.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726616AbfKJNhT (ORCPT ); Sun, 10 Nov 2019 08:37:19 -0500 Received: by mail-wm1-f65.google.com with SMTP id j18so1223434wmk.1 for ; Sun, 10 Nov 2019 05:37:16 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=to:references:from:autocrypt:subject:message-id:date:user-agent :mime-version:in-reply-to; bh=onOmLLZuKg3odcBHpnac/ZN75J/kVPxcS/iR/p/lsvA=; b=uZVHwx6IoapELwW+CfTkpFaWF6gr29TKaaoW01yqX/pJh5JQVzIA5X/th9GMawYLQ5 FIfx70G6PqXYOwYaOXEHyB+iGkej7qsKSyMFm6AY1GvsGg7rdrRRiWcqwlIBgFmJfdBm Muxo2qtM1uKOarFrnQTZMyommwQq9xIaJMPNi3EonlZpp7EfsMeoRQw1AwzJ/DD1d+yY S2tjiHBhbrLdw3JgqfzTlxQGfVTcphRBt6y+NZ24uiv6SzStc7qt81oULad7cFCnM6yH HmzBfrvU950M5FgrifYTM6zgysgeGONU83cDic+hWe7WKi+2+vsLfp6hLodFfTRrc2Ra mJ7w== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:to:references:from:autocrypt:subject:message-id :date:user-agent:mime-version:in-reply-to; bh=onOmLLZuKg3odcBHpnac/ZN75J/kVPxcS/iR/p/lsvA=; b=nc8EqkBLvrXDlqMsONFvh8XHCKsEZ5S+T5F8n+1awYs/cVXnlG2kbfpPek8BBCUg1N peWD/KxnPFYgALQULk/JKPJlvSFUFGKv2YGSXX22UwgysmjmxPvnOMRXxVZbUcDfXk+g 5zoQhKGy+ZUV2A3Q7CNCIrbmd2BpN6gkoSj55ORwOUnHP6OTPD9pJ0RTPACIc1J30Tyf qM7O7Zeh/3InibTlErJVg9JzEc+Md72Jie/gwqBs6GAjBovyTd5wjxRqd0NdQI0vn3Oz IdTi8En897zsNygJvr9zB7tIyNikZteBRSfCOreO9sYcfrePXZvKn/fPKRwDrIyI1vut PHYw== X-Gm-Message-State: APjAAAX45AqP3dgayxKsfD39Zs1wJXwGSJXQfy5jmm00WUq4OTlwjwJZ JK+e0yI2LjocbCWJoSAzVhPwohJi X-Google-Smtp-Source: APXvYqzFGZXrE+SWIsfGrFtvcXGtau9OaWGkzX7kcXn3eqiy5OHRiQA0ajTgd6FD4Y4FgfSccxXE5Q== X-Received: by 2002:a1c:f415:: with SMTP id z21mr16747173wma.140.1573393035570; Sun, 10 Nov 2019 05:37:15 -0800 (PST) Received: from [192.168.43.201] ([109.126.135.169]) by smtp.gmail.com with ESMTPSA id w81sm18403802wmg.5.2019.11.10.05.37.13 (version=TLS1_3 cipher=TLS_AES_128_GCM_SHA256 bits=128/128); Sun, 10 Nov 2019 05:37:14 -0800 (PST) To: Jens Axboe , io-uring@vger.kernel.org References: <8c57d694-6f73-4ac7-f390-ffbb3d780fea@kernel.dk> From: Pavel Begunkov Autocrypt: addr=asml.silence@gmail.com; prefer-encrypt=mutual; keydata= mQINBFmKBOQBEAC76ZFxLAKpDw0bKQ8CEiYJRGn8MHTUhURL02/7n1t0HkKQx2K1fCXClbps bdwSHrhOWdW61pmfMbDYbTj6ZvGRvhoLWfGkzujB2wjNcbNTXIoOzJEGISHaPf6E2IQx1ik9 6uqVkK1OMb7qRvKH0i7HYP4WJzYbEWVyLiAxUj611mC9tgd73oqZ2pLYzGTqF2j6a/obaqha +hXuWTvpDQXqcOZJXIW43atprH03G1tQs7VwR21Q1eq6Yvy2ESLdc38EqCszBfQRMmKy+cfp W3U9Mb1w0L680pXrONcnlDBCN7/sghGeMHjGKfNANjPc+0hzz3rApPxpoE7HC1uRiwC4et83 CKnncH1l7zgeBT9Oa3qEiBlaa1ZCBqrA4dY+z5fWJYjMpwI1SNp37RtF8fKXbKQg+JuUjAa9 Y6oXeyEvDHMyJYMcinl6xCqCBAXPHnHmawkMMgjr3BBRzODmMr+CPVvnYe7BFYfoajzqzq+h EyXSl3aBf0IDPTqSUrhbmjj5OEOYgRW5p+mdYtY1cXeK8copmd+fd/eTkghok5li58AojCba jRjp7zVOLOjDlpxxiKhuFmpV4yWNh5JJaTbwCRSd04sCcDNlJj+TehTr+o1QiORzc2t+N5iJ NbILft19Izdn8U39T5oWiynqa1qCLgbuFtnYx1HlUq/HvAm+kwARAQABtDFQYXZlbCBCZWd1 bmtvdiAoc2lsZW5jZSkgPGFzbWwuc2lsZW5jZUBnbWFpbC5jb20+iQJOBBMBCAA4FiEE+6Ju PTjTbx479o3OWt5b1Glr+6UFAlmKBOQCGwMFCwkIBwIGFQgJCgsCBBYCAwECHgECF4AACgkQ Wt5b1Glr+6WxZA//QueaKHzgdnOikJ7NA/Vq8FmhRlwgtP0+E+w93kL+ZGLzS/cUCIjn2f4Q Mcutj2Neg0CcYPX3b2nJiKr5Vn0rjJ/suiaOa1h1KzyNTOmxnsqE5fmxOf6C6x+NKE18I5Jy xzLQoktbdDVA7JfB1itt6iWSNoOTVcvFyvfe5ggy6FSCcP+m1RlR58XxVLH+qlAvxxOeEr/e aQfUzrs7gqdSd9zQGEZo0jtuBiB7k98t9y0oC9Jz0PJdvaj1NZUgtXG9pEtww3LdeXP/TkFl HBSxVflzeoFaj4UAuy8+uve7ya/ECNCc8kk0VYaEjoVrzJcYdKP583iRhOLlZA6HEmn/+Gh9 4orG67HNiJlbFiW3whxGizWsrtFNLsSP1YrEReYk9j1SoUHHzsu+ZtNfKuHIhK0sU07G1OPN 2rDLlzUWR9Jc22INAkhVHOogOcc5ajMGhgWcBJMLCoi219HlX69LIDu3Y34uIg9QPZIC2jwr 24W0kxmK6avJr7+n4o8m6sOJvhlumSp5TSNhRiKvAHB1I2JB8Q1yZCIPzx+w1ALxuoWiCdwV M/azguU42R17IuBzK0S3hPjXpEi2sK/k4pEPnHVUv9Cu09HCNnd6BRfFGjo8M9kZvw360gC1 reeMdqGjwQ68o9x0R7NBRrtUOh48TDLXCANAg97wjPoy37dQE7e5Ag0EWYoE5AEQAMWS+aBV IJtCjwtfCOV98NamFpDEjBMrCAfLm7wZlmXy5I6o7nzzCxEw06P2rhzp1hIqkaab1kHySU7g dkpjmQ7Jjlrf6KdMP87mC/Hx4+zgVCkTQCKkIxNE76Ff3O9uTvkWCspSh9J0qPYyCaVta2D1 Sq5HZ8WFcap71iVO1f2/FEHKJNz/YTSOS/W7dxJdXl2eoj3gYX2UZNfoaVv8OXKaWslZlgqN jSg9wsTv1K73AnQKt4fFhscN9YFxhtgD/SQuOldE5Ws4UlJoaFX/yCoJL3ky2kC0WFngzwRF Yo6u/KON/o28yyP+alYRMBrN0Dm60FuVSIFafSqXoJTIjSZ6olbEoT0u17Rag8BxnxryMrgR dkccq272MaSS0eOC9K2rtvxzddohRFPcy/8bkX+t2iukTDz75KSTKO+chce62Xxdg62dpkZX xK+HeDCZ7gRNZvAbDETr6XI63hPKi891GeZqvqQVYR8e+V2725w+H1iv3THiB1tx4L2bXZDI DtMKQ5D2RvCHNdPNcZeldEoJwKoA60yg6tuUquvsLvfCwtrmVI2rL2djYxRfGNmFMrUDN1Xq F3xozA91q3iZd9OYi9G+M/OA01husBdcIzj1hu0aL+MGg4Gqk6XwjoSxVd4YT41kTU7Kk+/I 5/Nf+i88ULt6HanBYcY/+Daeo/XFABEBAAGJAjYEGAEIACAWIQT7om49ONNvHjv2jc5a3lvU aWv7pQUCWYoE5AIbDAAKCRBa3lvUaWv7pfmcEACKTRQ28b1y5ztKuLdLr79+T+LwZKHjX++P 4wKjEOECCcB6KCv3hP+J2GCXDOPZvdg/ZYZafqP68Yy8AZqkfa4qPYHmIdpODtRzZSL48kM8 LRzV8Rl7J3ItvzdBRxf4T/Zseu5U6ELiQdCUkPGsJcPIJkgPjO2ROG/ZtYa9DvnShNWPlp+R uPwPccEQPWO/NP4fJl2zwC6byjljZhW5kxYswGMLBwb5cDUZAisIukyAa8Xshdan6C2RZcNs rB3L7vsg/R8UCehxOH0C+NypG2GqjVejNZsc7bgV49EOVltS+GmGyY+moIzxsuLmT93rqyII 5rSbbcTLe6KBYcs24XEoo49Zm9oDA3jYvNpeYD8rDcnNbuZh9kTgBwFN41JHOPv0W2FEEWqe JsCwQdcOQ56rtezdCJUYmRAt3BsfjN3Jn3N6rpodi4Dkdli8HylM5iq4ooeb5VkQ7UZxbCWt UVMKkOCdFhutRmYp0mbv2e87IK4erwNHQRkHUkzbsuym8RVpAZbLzLPIYK/J3RTErL6Z99N2 m3J6pjwSJY/zNwuFPs9zGEnRO4g0BUbwGdbuvDzaq6/3OJLKohr5eLXNU3JkT+3HezydWm3W OPhauth7W0db74Qd49HXK0xe/aPrK+Cp+kU1HRactyNtF8jZQbhMCC8vMGukZtWaAwpjWiiH bA== Subject: Re: [PATCH v4] io_uring: limit inflight IO Message-ID: Date: Sun, 10 Nov 2019 16:36:57 +0300 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.2.0 MIME-Version: 1.0 In-Reply-To: <8c57d694-6f73-4ac7-f390-ffbb3d780fea@kernel.dk> Content-Type: multipart/signed; micalg=pgp-sha256; protocol="application/pgp-signature"; boundary="kTmQx37tt2gm9W13KpPAHnP6yt7xGIykB" Sender: io-uring-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: io-uring@vger.kernel.org This is an OpenPGP/MIME signed message (RFC 4880 and 3156) --kTmQx37tt2gm9W13KpPAHnP6yt7xGIykB Content-Type: multipart/mixed; boundary="wcVroYOI6b7AdJ6TSD3AOrjjeqsOktS5J"; protected-headers="v1" From: Pavel Begunkov To: Jens Axboe , io-uring@vger.kernel.org Message-ID: Subject: Re: [PATCH v4] io_uring: limit inflight IO References: <8c57d694-6f73-4ac7-f390-ffbb3d780fea@kernel.dk> In-Reply-To: <8c57d694-6f73-4ac7-f390-ffbb3d780fea@kernel.dk> --wcVroYOI6b7AdJ6TSD3AOrjjeqsOktS5J Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: quoted-printable On 10/11/2019 00:01, Jens Axboe wrote: > With unbounded request times, we can potentially have a lot of IO > inflight. As we provide no real backpressure unless > IORING_SETUP_CQ_NODROP is set, and even there there's quite some delay > between overflows and backpressure being applied, let's put some safety= > in place to avoid going way overboard. >=20 > This limits the maximum number of inflight IO for any given io_ring_ctx= > to twice the CQ ring size. This is a losely managed limit, we only chec= k > for every SQ ring size number of events. That should be good enough to > achieve our goal, which is to prevent massively deep queues. If these > are async requests, they would just be waiting for an execution slot > anyway. >=20 > We return -EBUSY if we can't queue anymore IO. The caller should reap > some completions and retry the operation after that. Note that this is > a "should never hit this" kind of condition, as driving the depth into > CQ overflow situations is unreliable. >=20 > Signed-off-by: Jens Axboe >=20 > --- >=20 > Changes since v3: > - Fixup masking for when to check (Pavel) >=20 > diff --git a/fs/io_uring.c b/fs/io_uring.c > index 472bbc11688c..a2548a6dd195 100644 > --- a/fs/io_uring.c > +++ b/fs/io_uring.c > @@ -745,7 +745,7 @@ static struct io_kiocb *io_get_req(struct io_ring_c= tx *ctx, > struct io_kiocb *req; > =20 > if (!percpu_ref_tryget(&ctx->refs)) > - return NULL; > + return ERR_PTR(-ENXIO); > =20 > if (!state) { > req =3D kmem_cache_alloc(req_cachep, gfp); > @@ -791,7 +791,7 @@ static struct io_kiocb *io_get_req(struct io_ring_c= tx *ctx, > if (req) > goto got_it; > percpu_ref_put(&ctx->refs); > - return NULL; > + return ERR_PTR(-EBUSY); > } > =20 > static void io_free_req_many(struct io_ring_ctx *ctx, void **reqs, int= *nr) > @@ -2997,6 +2997,31 @@ static bool io_get_sqring(struct io_ring_ctx *ct= x, struct sqe_submit *s) > return false; > } > =20 > +static bool io_sq_over_limit(struct io_ring_ctx *ctx, unsigned to_subm= it) > +{ > + unsigned inflight; > + > + if (!list_empty(&ctx->cq_overflow_list)) { > + io_cqring_overflow_flush(ctx, false); > + return true; > + } > + > + /* > + * This doesn't need to be super precise, so only check every once > + * in a while. > + */ > + if ((ctx->cached_sq_head & ~ctx->sq_mask) !=3D > + ((ctx->cached_sq_head + to_submit) & ~ctx->sq_mask)) > + return false; Still not quite agree, with "!=3D" it will check almost every time. We want to continue with the checks only when we overrun sq_entries (i.e. changed higher bits). There is a simplified trace: let 1) mask =3D=3D 3, ~mask =3D=3D 4 2) head =3D=3D 0, 3) submitting by one 1: (0 & 4) !=3D (1 & 4): false // do checks 2: (1 & 4) !=3D (2 & 4): false // do checks 3: (2 & 4) !=3D (3 & 4): false // do checks 4: (3 & 4) !=3D (4 & 4): true // skip, return true The rest looks good, Reviewed-by: Pavel Begunkov > + > + /* > + * Limit us to 2x the CQ ring size > + */ > + inflight =3D ctx->cached_sq_head - > + (ctx->cached_cq_tail + atomic_read(&ctx->cached_cq_overflow)); > + return inflight > 2 * ctx->cq_entries; > +} > + > static int io_submit_sqes(struct io_ring_ctx *ctx, unsigned int nr, > struct file *ring_file, int ring_fd, > struct mm_struct **mm, bool async) > @@ -3007,10 +3032,8 @@ static int io_submit_sqes(struct io_ring_ctx *ct= x, unsigned int nr, > int i, submitted =3D 0; > bool mm_fault =3D false; > =20 > - if (!list_empty(&ctx->cq_overflow_list)) { > - io_cqring_overflow_flush(ctx, false); > + if (unlikely(io_sq_over_limit(ctx, nr))) > return -EBUSY; > - } > =20 > if (nr > IO_PLUG_THRESHOLD) { > io_submit_state_start(&state, ctx, nr); > @@ -3022,9 +3045,9 @@ static int io_submit_sqes(struct io_ring_ctx *ctx= , unsigned int nr, > unsigned int sqe_flags; > =20 > req =3D io_get_req(ctx, statep); > - if (unlikely(!req)) { > + if (unlikely(IS_ERR(req))) { > if (!submitted) > - submitted =3D -EAGAIN; > + submitted =3D PTR_ERR(req); > break; > } > if (!io_get_sqring(ctx, &req->submit)) { > @@ -3045,8 +3068,10 @@ static int io_submit_sqes(struct io_ring_ctx *ct= x, unsigned int nr, > if (link && (sqe_flags & IOSQE_IO_DRAIN)) { > if (!shadow_req) { > shadow_req =3D io_get_req(ctx, NULL); > - if (unlikely(!shadow_req)) > + if (unlikely(IS_ERR(shadow_req))) { > + shadow_req =3D NULL; > goto out; > + } > shadow_req->flags |=3D (REQ_F_IO_DRAIN | REQ_F_SHADOW_DRAIN); > refcount_dec(&shadow_req->refs); > } >=20 --=20 Pavel Begunkov --wcVroYOI6b7AdJ6TSD3AOrjjeqsOktS5J-- --kTmQx37tt2gm9W13KpPAHnP6yt7xGIykB Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- iQIzBAEBCAAdFiEE+6JuPTjTbx479o3OWt5b1Glr+6UFAl3IEn0ACgkQWt5b1Glr +6XjtQ//XG5yiIsdmU12vrIicDqORJ9cKOzGLbrF8dmgjGYqIIDPmtGAcoLm0J3d N+L/kv3KTQSWutxCCe/NYa7gtOLn39AI4mK0uBYXs69bBSixGR6bUR6tHuhM1cME FS/uKEEfHmQlh2/yI/neYJZENUax1tQjK1bTV9VWTfZW3385xex+INNviHc6xY05 l56q9f0lG906GMpnY/E4M0ocVL/jx8beIvISNtc5MWjeKGo5pphWqJHSOCUd6NaN AOlqOVKb0rE7uMex5M61TzETxK2cVuyNXPSppvBtADIQs3v50HrI/TcZFYeUiQhK 1yAQslnTCkfR4OOlZ/P6j672avOQq0PfIKEltKWJ+vhHsGkRukO93BaAWcauIqru Kc6gQ+eOkE6k5xDZ1+pPjBYBA3ikQTOPrU0YPk1RAhw7L4IgYA9VyNvENooyLDKP 9hukptlxQoEp+hHqo8ngbyHExJPlnhj3gAK8W8PwnkcQZHLj07Xrc7So4jciNtVU C3nN2OvvNY6/qghROy4gg/n71AuCSbbzmx1dzjA71+fWA8yCWrUIC1yTVh/ipqQu +R4KdOB7s2sk66KVg8H+J6LAE8XjTvlBJUpJlCdg2ISLnvfIDwvplR+1xgXJFcx8 lrg7OS3tpMt4NwLMScgdEHetwkW9lAFn/u9VWMcMCmLPhc26ym8= =OZiG -----END PGP SIGNATURE----- --kTmQx37tt2gm9W13KpPAHnP6yt7xGIykB--