From patchwork Thu Sep 22 18:28:01 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Jens Axboe X-Patchwork-Id: 608537 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 77A1AC6FA82 for ; Thu, 22 Sep 2022 18:31:25 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S231984AbiIVSbY (ORCPT ); Thu, 22 Sep 2022 14:31:24 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:51088 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S232705AbiIVSax (ORCPT ); Thu, 22 Sep 2022 14:30:53 -0400 Received: from mail-io1-xd35.google.com (mail-io1-xd35.google.com [IPv6:2607:f8b0:4864:20::d35]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 007F810B58E for ; Thu, 22 Sep 2022 11:28:08 -0700 (PDT) Received: by mail-io1-xd35.google.com with SMTP id 138so8433934iou.9 for ; Thu, 22 Sep 2022 11:28:08 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=kernel-dk.20210112.gappssmtp.com; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date; bh=KOdha0xZ2/QYkK5YkfhyDv4n+7NP1qIjesr4vuNTao8=; b=I3ozfRpCO+8sfk1czHqJkuud5fbhOBBjBPNani3SoWO5IZrr+n/xykd4BYCj6iF/yZ YUgMhkihNpdD90pGDqffGeTVe1Lb9NSQnHeRoD8cgg2gFQ/EmHGBP46WBRkepSKacBe9 0/BDSuaDvLg45rQ8CzKBSAt8/NtYunQdrfrc0/qYd/SHmQXOTUhVkzxgGGcPxLM8Q6yG Pnxpp3iPWxoszKqZD6piQndRMfFxGNQXSl6LEm2shcEG/V1F+VDWoYsayPrD5S/hJS3w /DgQMp06IIBo+G6qQyJy0W2qJ9hcQXFC9VwcVse6SFsPWkJYASV0uNPVG/clm0ofomyV cjPQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date; bh=KOdha0xZ2/QYkK5YkfhyDv4n+7NP1qIjesr4vuNTao8=; b=ZNMB/JM3vZcw0cwo1r1efU+UDnK04mmkoQlQKY++xs33E7G2J0sFXzgAUpCJo196fo FTMs6wQqbhnWiz6g0sbv8pgOZ+kdIZefU4wq6esbhspZWzdqx4Mf+Q1ir5XBa8BcETu/ 9RoZY0jXaml1XgBoA9zAn3NeTUqoqrQo5J8UXlKhfTAMS2pc1/BsSnkk6GpNbSqCF7SJ m/2z/SEevL8temPPur+hOe3wF0DFgADTEDkrO7Src4OWciJ/QMq/xkX9uJOyBHy5WQ9J /rsCxw9x3NFheWxfeiB1UHpQ2xTZMpJzAfvCfmA6zbcBqPpsnkKoIybcD9EbQQMAtaeC a4wA== X-Gm-Message-State: ACrzQf0AD17H2QIqnc62yB7C1bqmE03SQEMIJVImfmAJgn/vA8fcUXSB yq1U+BBIftQatMDa+Y2RjQIEaQ== X-Google-Smtp-Source: AMsMyM6q1U2zUDZ6fq1ZHALNlDos6d1f2IhUDrAzg+XF/fY3dxFEtLxvEozxHX6pNVp/47zfuhfqkg== X-Received: by 2002:a05:6602:2d09:b0:688:f387:aab5 with SMTP id c9-20020a0566022d0900b00688f387aab5mr2274193iow.107.1663871288335; Thu, 22 Sep 2022 11:28:08 -0700 (PDT) Received: from m1max.localdomain ([207.135.234.126]) by smtp.gmail.com with ESMTPSA id q20-20020a05663810d400b0035a468b7fbesm2440646jad.71.2022.09.22.11.28.07 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Thu, 22 Sep 2022 11:28:07 -0700 (PDT) From: Jens Axboe To: linux-block@vger.kernel.org Cc: linux-scsi@vger.kernel.org, linux-nvme@lists.infradead.org, Jens Axboe Subject: [PATCH 1/5] block: enable batched allocation for blk_mq_alloc_request() Date: Thu, 22 Sep 2022 12:28:01 -0600 Message-Id: <20220922182805.96173-2-axboe@kernel.dk> X-Mailer: git-send-email 2.35.1 In-Reply-To: <20220922182805.96173-1-axboe@kernel.dk> References: <20220922182805.96173-1-axboe@kernel.dk> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-scsi@vger.kernel.org The filesystem IO path can take advantage of allocating batches of requests, if the underlying submitter tells the block layer about it through the blk_plug. For passthrough IO, the exported API is the blk_mq_alloc_request() helper, and that one does not allow for request caching. Wire up request caching for blk_mq_alloc_request(), which is generally done without having a bio available upfront. Signed-off-by: Jens Axboe --- block/blk-mq.c | 80 ++++++++++++++++++++++++++++++++++++++++++++------ 1 file changed, 71 insertions(+), 9 deletions(-) diff --git a/block/blk-mq.c b/block/blk-mq.c index c11949d66163..d3a9f8b9c7ee 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -510,25 +510,87 @@ static struct request *__blk_mq_alloc_requests(struct blk_mq_alloc_data *data) alloc_time_ns); } -struct request *blk_mq_alloc_request(struct request_queue *q, blk_opf_t opf, - blk_mq_req_flags_t flags) +static struct request *blk_mq_rq_cache_fill(struct request_queue *q, + struct blk_plug *plug, + blk_opf_t opf, + blk_mq_req_flags_t flags) { struct blk_mq_alloc_data data = { .q = q, .flags = flags, .cmd_flags = opf, - .nr_tags = 1, + .nr_tags = plug->nr_ios, + .cached_rq = &plug->cached_rq, }; struct request *rq; - int ret; - ret = blk_queue_enter(q, flags); - if (ret) - return ERR_PTR(ret); + if (blk_queue_enter(q, flags)) + return NULL; + + plug->nr_ios = 1; rq = __blk_mq_alloc_requests(&data); - if (!rq) - goto out_queue_exit; + if (unlikely(!rq)) + blk_queue_exit(q); + return rq; +} + +static struct request *blk_mq_alloc_cached_request(struct request_queue *q, + blk_opf_t opf, + blk_mq_req_flags_t flags) +{ + struct blk_plug *plug = current->plug; + struct request *rq; + + if (!plug) + return NULL; + if (rq_list_empty(plug->cached_rq)) { + if (plug->nr_ios == 1) + return NULL; + rq = blk_mq_rq_cache_fill(q, plug, opf, flags); + if (rq) + goto got_it; + return NULL; + } + rq = rq_list_peek(&plug->cached_rq); + if (!rq || rq->q != q) + return NULL; + + if (blk_mq_get_hctx_type(opf) != rq->mq_hctx->type) + return NULL; + if (op_is_flush(rq->cmd_flags) != op_is_flush(opf)) + return NULL; + + plug->cached_rq = rq_list_next(rq); +got_it: + rq->cmd_flags = opf; + INIT_LIST_HEAD(&rq->queuelist); + return rq; +} + +struct request *blk_mq_alloc_request(struct request_queue *q, blk_opf_t opf, + blk_mq_req_flags_t flags) +{ + struct request *rq; + + rq = blk_mq_alloc_cached_request(q, opf, flags); + if (!rq) { + struct blk_mq_alloc_data data = { + .q = q, + .flags = flags, + .cmd_flags = opf, + .nr_tags = 1, + }; + int ret; + + ret = blk_queue_enter(q, flags); + if (ret) + return ERR_PTR(ret); + + rq = __blk_mq_alloc_requests(&data); + if (!rq) + goto out_queue_exit; + } rq->__data_len = 0; rq->__sector = (sector_t) -1; rq->bio = rq->biotail = NULL; From patchwork Thu Sep 22 18:28:02 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Jens Axboe X-Patchwork-Id: 609090 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 9A63FC54EE9 for ; Thu, 22 Sep 2022 18:31:36 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S232690AbiIVSbf (ORCPT ); Thu, 22 Sep 2022 14:31:35 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:51154 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S231601AbiIVSaz (ORCPT ); Thu, 22 Sep 2022 14:30:55 -0400 Received: from mail-io1-xd2b.google.com (mail-io1-xd2b.google.com [IPv6:2607:f8b0:4864:20::d2b]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id DDB0B50537 for ; Thu, 22 Sep 2022 11:28:09 -0700 (PDT) Received: by mail-io1-xd2b.google.com with SMTP id h194so8460361iof.4 for ; Thu, 22 Sep 2022 11:28:09 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=kernel-dk.20210112.gappssmtp.com; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date; bh=Ez71/bwQQ37IExTa9EvzsotH0lbr+r0Fmd6a9LpY5qI=; b=NPeWa/A8vJTdvLPo5gTvnowSY+ovLKq9+O5essJvGBWSFKDs/bbcVDtK+P822U3bk/ ZNGfB7tuXwoLIVKs/rJSage5JEj64FhD6qb09xtp5TYr6Za/xmaJ9w2X+Y8vlaey0L/I 5XIm8k1dSJQZp+s0tLzeCUtRrkO7zRb4LQy0N4whqnWB7/0LmZHDJDgI7kJB5+IgMvaj uf2hFnBsAH+n8a20Uhfi1rmoRev2yy3aHwqMkb/zhsimS3TvyLSJWDaXvbHd36oDEkRV 9PWGeLtwpLr6HWeBK7+RotxcLkDHSWM8OosyShM/uVI1HTU2T0eTBoemyZTyhHH6Rrq8 8GEw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date; bh=Ez71/bwQQ37IExTa9EvzsotH0lbr+r0Fmd6a9LpY5qI=; b=gTti7hIAi4//3hbhPcrsk3O+niN6D0xThZTzIFwqryAqxFmH3X4BMeLqvmf0biDPHr TrTkWIFWDqoQZanIMh2H5OTfXOvOcuIjk12VWygKcUz/PCS+LJr0U9yDlfP136OYhQtd 6YvCivm416fA0pLlKOps72MAugbV9egijyAPhRnCei3DlO1qm/AUYNgBWvaiOqmpfUgN My3ow6kckoioY4c18gknY4wslwApRNRX1iaWvz3uIrxl4GUOzQEWmpbOR0MUevxU/Jpz F2rboRUr0w1igBM7l6C2SyRRl0gpFpQ3ZjRx8b1ImOa2cN6vQTVyjUB1FuvFBFN+FnCK V0Pw== X-Gm-Message-State: ACrzQf0dO0dv35NfmKgx3oXhBMmEt81PEpcnugBjdCNlRmfQxBTnBRfV GbKYPuPFy3unlGUJrT+a9y6gKzDO1avT9w== X-Google-Smtp-Source: AMsMyM42NywJNwfvNyR1a/wX27JlMMmlhhibOrM/rQEm/dt2kbH+j3rv301Vk80NEhuuMdu8j8j2EQ== X-Received: by 2002:a05:6638:138e:b0:357:a6c7:e972 with SMTP id w14-20020a056638138e00b00357a6c7e972mr2758853jad.40.1663871289185; Thu, 22 Sep 2022 11:28:09 -0700 (PDT) Received: from m1max.localdomain ([207.135.234.126]) by smtp.gmail.com with ESMTPSA id q20-20020a05663810d400b0035a468b7fbesm2440646jad.71.2022.09.22.11.28.08 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Thu, 22 Sep 2022 11:28:08 -0700 (PDT) From: Jens Axboe To: linux-block@vger.kernel.org Cc: linux-scsi@vger.kernel.org, linux-nvme@lists.infradead.org, Jens Axboe Subject: [PATCH 2/5] block: change request end_io handler to pass back a return value Date: Thu, 22 Sep 2022 12:28:02 -0600 Message-Id: <20220922182805.96173-3-axboe@kernel.dk> X-Mailer: git-send-email 2.35.1 In-Reply-To: <20220922182805.96173-1-axboe@kernel.dk> References: <20220922182805.96173-1-axboe@kernel.dk> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-scsi@vger.kernel.org Everything is just converted to returning RQ_END_IO_NONE, and there should be no functional changes with this patch. In preparation for allowing the end_io handler to pass ownership back to the block layer, rather than retain ownership of the request. Signed-off-by: Jens Axboe --- block/blk-flush.c | 10 +++++++--- block/blk-mq.c | 14 +++++++++----- drivers/md/dm-rq.c | 4 +++- drivers/nvme/host/core.c | 6 ++++-- drivers/nvme/host/ioctl.c | 5 ++++- drivers/nvme/host/pci.c | 12 ++++++++---- drivers/nvme/target/passthru.c | 5 +++-- drivers/scsi/scsi_error.c | 4 +++- drivers/scsi/sg.c | 9 +++++---- drivers/scsi/st.c | 4 +++- drivers/target/target_core_pscsi.c | 6 ++++-- drivers/ufs/core/ufshpb.c | 8 ++++++-- include/linux/blk-mq.h | 7 ++++++- 13 files changed, 65 insertions(+), 29 deletions(-) diff --git a/block/blk-flush.c b/block/blk-flush.c index d20a0c6b2c66..ac850f4d9c4c 100644 --- a/block/blk-flush.c +++ b/block/blk-flush.c @@ -218,7 +218,8 @@ static void blk_flush_complete_seq(struct request *rq, blk_kick_flush(q, fq, cmd_flags); } -static void flush_end_io(struct request *flush_rq, blk_status_t error) +static enum rq_end_io_ret flush_end_io(struct request *flush_rq, + blk_status_t error) { struct request_queue *q = flush_rq->q; struct list_head *running; @@ -232,7 +233,7 @@ static void flush_end_io(struct request *flush_rq, blk_status_t error) if (!req_ref_put_and_test(flush_rq)) { fq->rq_status = error; spin_unlock_irqrestore(&fq->mq_flush_lock, flags); - return; + return RQ_END_IO_NONE; } blk_account_io_flush(flush_rq); @@ -269,6 +270,7 @@ static void flush_end_io(struct request *flush_rq, blk_status_t error) } spin_unlock_irqrestore(&fq->mq_flush_lock, flags); + return RQ_END_IO_NONE; } bool is_flush_rq(struct request *rq) @@ -354,7 +356,8 @@ static void blk_kick_flush(struct request_queue *q, struct blk_flush_queue *fq, blk_flush_queue_rq(flush_rq, false); } -static void mq_flush_data_end_io(struct request *rq, blk_status_t error) +static enum rq_end_io_ret mq_flush_data_end_io(struct request *rq, + blk_status_t error) { struct request_queue *q = rq->q; struct blk_mq_hw_ctx *hctx = rq->mq_hctx; @@ -376,6 +379,7 @@ static void mq_flush_data_end_io(struct request *rq, blk_status_t error) spin_unlock_irqrestore(&fq->mq_flush_lock, flags); blk_mq_sched_restart(hctx); + return RQ_END_IO_NONE; } /** diff --git a/block/blk-mq.c b/block/blk-mq.c index d3a9f8b9c7ee..a4e018c82b7c 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -1001,7 +1001,8 @@ inline void __blk_mq_end_request(struct request *rq, blk_status_t error) if (rq->end_io) { rq_qos_done(rq->q, rq); - rq->end_io(rq, error); + if (rq->end_io(rq, error) == RQ_END_IO_FREE) + blk_mq_free_request(rq); } else { blk_mq_free_request(rq); } @@ -1287,12 +1288,13 @@ struct blk_rq_wait { blk_status_t ret; }; -static void blk_end_sync_rq(struct request *rq, blk_status_t ret) +static enum rq_end_io_ret blk_end_sync_rq(struct request *rq, blk_status_t ret) { struct blk_rq_wait *wait = rq->end_io_data; wait->ret = ret; complete(&wait->done); + return RQ_END_IO_NONE; } bool blk_rq_is_poll(struct request *rq) @@ -1526,10 +1528,12 @@ static bool blk_mq_req_expired(struct request *rq, unsigned long *next) void blk_mq_put_rq_ref(struct request *rq) { - if (is_flush_rq(rq)) - rq->end_io(rq, 0); - else if (req_ref_put_and_test(rq)) + if (is_flush_rq(rq)) { + if (rq->end_io(rq, 0) == RQ_END_IO_FREE) + blk_mq_free_request(rq); + } else if (req_ref_put_and_test(rq)) { __blk_mq_free_request(rq); + } } static bool blk_mq_check_expired(struct request *rq, void *priv) diff --git a/drivers/md/dm-rq.c b/drivers/md/dm-rq.c index 4f49bbcce4f1..3001b10a3fbf 100644 --- a/drivers/md/dm-rq.c +++ b/drivers/md/dm-rq.c @@ -292,11 +292,13 @@ static void dm_kill_unmapped_request(struct request *rq, blk_status_t error) dm_complete_request(rq, error); } -static void end_clone_request(struct request *clone, blk_status_t error) +static enum rq_end_io_ret end_clone_request(struct request *clone, + blk_status_t error) { struct dm_rq_target_io *tio = clone->end_io_data; dm_complete_request(tio->orig, error); + return RQ_END_IO_NONE; } static int dm_rq_bio_constructor(struct bio *bio, struct bio *bio_orig, diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c index 13080a017ecf..f946f85e7a66 100644 --- a/drivers/nvme/host/core.c +++ b/drivers/nvme/host/core.c @@ -1177,7 +1177,8 @@ static void nvme_queue_keep_alive_work(struct nvme_ctrl *ctrl) queue_delayed_work(nvme_wq, &ctrl->ka_work, ctrl->kato * HZ / 2); } -static void nvme_keep_alive_end_io(struct request *rq, blk_status_t status) +static enum rq_end_io_ret nvme_keep_alive_end_io(struct request *rq, + blk_status_t status) { struct nvme_ctrl *ctrl = rq->end_io_data; unsigned long flags; @@ -1189,7 +1190,7 @@ static void nvme_keep_alive_end_io(struct request *rq, blk_status_t status) dev_err(ctrl->device, "failed nvme_keep_alive_end_io error=%d\n", status); - return; + return RQ_END_IO_NONE; } ctrl->comp_seen = false; @@ -1200,6 +1201,7 @@ static void nvme_keep_alive_end_io(struct request *rq, blk_status_t status) spin_unlock_irqrestore(&ctrl->lock, flags); if (startka) nvme_queue_keep_alive_work(ctrl); + return RQ_END_IO_NONE; } static void nvme_keep_alive_work(struct work_struct *work) diff --git a/drivers/nvme/host/ioctl.c b/drivers/nvme/host/ioctl.c index 548aca8b5b9f..c80b3ecca5c8 100644 --- a/drivers/nvme/host/ioctl.c +++ b/drivers/nvme/host/ioctl.c @@ -385,7 +385,8 @@ static void nvme_uring_task_cb(struct io_uring_cmd *ioucmd) io_uring_cmd_done(ioucmd, status, result); } -static void nvme_uring_cmd_end_io(struct request *req, blk_status_t err) +static enum rq_end_io_ret nvme_uring_cmd_end_io(struct request *req, + blk_status_t err) { struct io_uring_cmd *ioucmd = req->end_io_data; struct nvme_uring_cmd_pdu *pdu = nvme_uring_cmd_pdu(ioucmd); @@ -404,6 +405,8 @@ static void nvme_uring_cmd_end_io(struct request *req, blk_status_t err) nvme_uring_task_cb(ioucmd); else io_uring_cmd_complete_in_task(ioucmd, nvme_uring_task_cb); + + return RQ_END_IO_NONE; } static int nvme_uring_cmd_io(struct nvme_ctrl *ctrl, struct nvme_ns *ns, diff --git a/drivers/nvme/host/pci.c b/drivers/nvme/host/pci.c index 09b5d62f342b..361f09f23648 100644 --- a/drivers/nvme/host/pci.c +++ b/drivers/nvme/host/pci.c @@ -1268,7 +1268,7 @@ static int adapter_delete_sq(struct nvme_dev *dev, u16 sqid) return adapter_delete_queue(dev, nvme_admin_delete_sq, sqid); } -static void abort_endio(struct request *req, blk_status_t error) +static enum rq_end_io_ret abort_endio(struct request *req, blk_status_t error) { struct nvme_queue *nvmeq = req->mq_hctx->driver_data; @@ -1276,6 +1276,7 @@ static void abort_endio(struct request *req, blk_status_t error) "Abort status: 0x%x", nvme_req(req)->status); atomic_inc(&nvmeq->dev->ctrl.abort_limit); blk_mq_free_request(req); + return RQ_END_IO_NONE; } static bool nvme_should_reset(struct nvme_dev *dev, u32 csts) @@ -2447,22 +2448,25 @@ static int nvme_setup_io_queues(struct nvme_dev *dev) return result; } -static void nvme_del_queue_end(struct request *req, blk_status_t error) +static enum rq_end_io_ret nvme_del_queue_end(struct request *req, + blk_status_t error) { struct nvme_queue *nvmeq = req->end_io_data; blk_mq_free_request(req); complete(&nvmeq->delete_done); + return RQ_END_IO_NONE; } -static void nvme_del_cq_end(struct request *req, blk_status_t error) +static enum rq_end_io_ret nvme_del_cq_end(struct request *req, + blk_status_t error) { struct nvme_queue *nvmeq = req->end_io_data; if (error) set_bit(NVMEQ_DELETE_ERROR, &nvmeq->flags); - nvme_del_queue_end(req, error); + return nvme_del_queue_end(req, error); } static int nvme_delete_queue(struct nvme_queue *nvmeq, u8 opcode) diff --git a/drivers/nvme/target/passthru.c b/drivers/nvme/target/passthru.c index 6f39a29828b1..4876ccaac55b 100644 --- a/drivers/nvme/target/passthru.c +++ b/drivers/nvme/target/passthru.c @@ -240,14 +240,15 @@ static void nvmet_passthru_execute_cmd_work(struct work_struct *w) blk_mq_free_request(rq); } -static void nvmet_passthru_req_done(struct request *rq, - blk_status_t blk_status) +static enum rq_end_io_ret nvmet_passthru_req_done(struct request *rq, + blk_status_t blk_status) { struct nvmet_req *req = rq->end_io_data; req->cqe->result = nvme_req(rq)->result; nvmet_req_complete(req, nvme_req(rq)->status); blk_mq_free_request(rq); + return RQ_END_IO_NONE; } static int nvmet_passthru_map_sg(struct nvmet_req *req, struct request *rq) diff --git a/drivers/scsi/scsi_error.c b/drivers/scsi/scsi_error.c index 448748e3fba5..786fb963cf3f 100644 --- a/drivers/scsi/scsi_error.c +++ b/drivers/scsi/scsi_error.c @@ -2004,9 +2004,11 @@ enum scsi_disposition scsi_decide_disposition(struct scsi_cmnd *scmd) } } -static void eh_lock_door_done(struct request *req, blk_status_t status) +static enum rq_end_io_ret eh_lock_door_done(struct request *req, + blk_status_t status) { blk_mq_free_request(req); + return RQ_END_IO_NONE; } /** diff --git a/drivers/scsi/sg.c b/drivers/scsi/sg.c index 340b050ad28d..94c5e9a9309c 100644 --- a/drivers/scsi/sg.c +++ b/drivers/scsi/sg.c @@ -177,7 +177,7 @@ typedef struct sg_device { /* holds the state of each scsi generic device */ } Sg_device; /* tasklet or soft irq callback */ -static void sg_rq_end_io(struct request *rq, blk_status_t status); +static enum rq_end_io_ret sg_rq_end_io(struct request *rq, blk_status_t status); static int sg_start_req(Sg_request *srp, unsigned char *cmd); static int sg_finish_rem_req(Sg_request * srp); static int sg_build_indirect(Sg_scatter_hold * schp, Sg_fd * sfp, int buff_size); @@ -1311,7 +1311,7 @@ sg_rq_end_io_usercontext(struct work_struct *work) * This function is a "bottom half" handler that is called by the mid * level when a command is completed (or has failed). */ -static void +static enum rq_end_io_ret sg_rq_end_io(struct request *rq, blk_status_t status) { struct scsi_cmnd *scmd = blk_mq_rq_to_pdu(rq); @@ -1324,11 +1324,11 @@ sg_rq_end_io(struct request *rq, blk_status_t status) int result, resid, done = 1; if (WARN_ON(srp->done != 0)) - return; + return RQ_END_IO_NONE; sfp = srp->parentfp; if (WARN_ON(sfp == NULL)) - return; + return RQ_END_IO_NONE; sdp = sfp->parentdp; if (unlikely(atomic_read(&sdp->detaching))) @@ -1406,6 +1406,7 @@ sg_rq_end_io(struct request *rq, blk_status_t status) INIT_WORK(&srp->ew.work, sg_rq_end_io_usercontext); schedule_work(&srp->ew.work); } + return RQ_END_IO_NONE; } static const struct file_operations sg_fops = { diff --git a/drivers/scsi/st.c b/drivers/scsi/st.c index 850172a2b8f1..55e7c07ebe4c 100644 --- a/drivers/scsi/st.c +++ b/drivers/scsi/st.c @@ -512,7 +512,8 @@ static void st_do_stats(struct scsi_tape *STp, struct request *req) atomic64_dec(&STp->stats->in_flight); } -static void st_scsi_execute_end(struct request *req, blk_status_t status) +static enum rq_end_io_ret st_scsi_execute_end(struct request *req, + blk_status_t status) { struct scsi_cmnd *scmd = blk_mq_rq_to_pdu(req); struct st_request *SRpnt = req->end_io_data; @@ -532,6 +533,7 @@ static void st_scsi_execute_end(struct request *req, blk_status_t status) blk_rq_unmap_user(tmp); blk_mq_free_request(req); + return RQ_END_IO_NONE; } static int st_scsi_execute(struct st_request *SRpnt, const unsigned char *cmd, diff --git a/drivers/target/target_core_pscsi.c b/drivers/target/target_core_pscsi.c index e6a967ddc08c..8a7306e5e133 100644 --- a/drivers/target/target_core_pscsi.c +++ b/drivers/target/target_core_pscsi.c @@ -39,7 +39,7 @@ static inline struct pscsi_dev_virt *PSCSI_DEV(struct se_device *dev) } static sense_reason_t pscsi_execute_cmd(struct se_cmd *cmd); -static void pscsi_req_done(struct request *, blk_status_t); +static enum rq_end_io_ret pscsi_req_done(struct request *, blk_status_t); /* pscsi_attach_hba(): * @@ -1002,7 +1002,8 @@ static sector_t pscsi_get_blocks(struct se_device *dev) return 0; } -static void pscsi_req_done(struct request *req, blk_status_t status) +static enum rq_end_io_ret pscsi_req_done(struct request *req, + blk_status_t status) { struct se_cmd *cmd = req->end_io_data; struct scsi_cmnd *scmd = blk_mq_rq_to_pdu(req); @@ -1029,6 +1030,7 @@ static void pscsi_req_done(struct request *req, blk_status_t status) } blk_mq_free_request(req); + return RQ_END_IO_NONE; } static const struct target_backend_ops pscsi_ops = { diff --git a/drivers/ufs/core/ufshpb.c b/drivers/ufs/core/ufshpb.c index a1a7a1175a5a..3d69a81c5b17 100644 --- a/drivers/ufs/core/ufshpb.c +++ b/drivers/ufs/core/ufshpb.c @@ -613,14 +613,17 @@ static void ufshpb_activate_subregion(struct ufshpb_lu *hpb, srgn->srgn_state = HPB_SRGN_VALID; } -static void ufshpb_umap_req_compl_fn(struct request *req, blk_status_t error) +static enum rq_end_io_ret ufshpb_umap_req_compl_fn(struct request *req, + blk_status_t error) { struct ufshpb_req *umap_req = (struct ufshpb_req *)req->end_io_data; ufshpb_put_req(umap_req->hpb, umap_req); + return RQ_END_IO_NONE; } -static void ufshpb_map_req_compl_fn(struct request *req, blk_status_t error) +static enum rq_end_io_ret ufshpb_map_req_compl_fn(struct request *req, + blk_status_t error) { struct ufshpb_req *map_req = (struct ufshpb_req *) req->end_io_data; struct ufshpb_lu *hpb = map_req->hpb; @@ -636,6 +639,7 @@ static void ufshpb_map_req_compl_fn(struct request *req, blk_status_t error) spin_unlock_irqrestore(&hpb->rgn_state_lock, flags); ufshpb_put_map_req(map_req->hpb, map_req); + return RQ_END_IO_NONE; } static void ufshpb_set_unmap_cmd(unsigned char *cdb, struct ufshpb_region *rgn) diff --git a/include/linux/blk-mq.h b/include/linux/blk-mq.h index 00a15808c137..e6fa49dd6196 100644 --- a/include/linux/blk-mq.h +++ b/include/linux/blk-mq.h @@ -14,7 +14,12 @@ struct blk_flush_queue; #define BLKDEV_MIN_RQ 4 #define BLKDEV_DEFAULT_RQ 128 -typedef void (rq_end_io_fn)(struct request *, blk_status_t); +enum rq_end_io_ret { + RQ_END_IO_NONE, + RQ_END_IO_FREE, +}; + +typedef enum rq_end_io_ret (rq_end_io_fn)(struct request *, blk_status_t); /* * request flags */ From patchwork Thu Sep 22 18:28:03 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Jens Axboe X-Patchwork-Id: 608536 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 1628EECAAD8 for ; Thu, 22 Sep 2022 18:31:39 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S232142AbiIVSbg (ORCPT ); Thu, 22 Sep 2022 14:31:36 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:51638 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S232528AbiIVSa4 (ORCPT ); Thu, 22 Sep 2022 14:30:56 -0400 Received: from mail-io1-xd2c.google.com (mail-io1-xd2c.google.com [IPv6:2607:f8b0:4864:20::d2c]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 41CA010BB2D for ; Thu, 22 Sep 2022 11:28:11 -0700 (PDT) Received: by mail-io1-xd2c.google.com with SMTP id b23so8471816iof.2 for ; Thu, 22 Sep 2022 11:28:11 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=kernel-dk.20210112.gappssmtp.com; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date; bh=Or6y/P+wCgw7ERLdnKk9k3MTRgFakhY9VTpCDMSU5NU=; b=DdaRIUVcXAC0PQL+apg/G+AjI2f67CbJodcek5ww9nFAO2BwNSHNCCxyA69h3aCoaM c8dFUovgd1Drm3MhsG5dCul89+q4n+Khw1JhESY3iHgB4ShRhp9VEMVUp57yjQm5v28v oSJnU2H3CgZTZONneZYK83VfiSHtzbfpxiz1QfnQqD0bs1K+5tmyijIqbnz6upfKVw3x lt4B+nt231UmP2KsK7u07z92rMmNfhf8WijKOrp2pSBhiq43clWLg1ATlroeeG6Z4C9l qkGtMckF44rjRNra0MuENekYeSNfiaLmxzoJ9wF2FaIxyxduFLHw9EDTkoOIEt05q2eq KFdA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date; bh=Or6y/P+wCgw7ERLdnKk9k3MTRgFakhY9VTpCDMSU5NU=; b=CpxZIw6dxGZa4yPoh06QyrNaOo3MfbML8q1hyui7FPcnlc2iB44iClue6KbnWXyNeS ZNLxe5C2deHbE2BM/kXbAM+XuwnwKFnmJp7vKi2pNtXU/9fBhBAK2Pl3jfgaqrBk1HkY OgwMqWUmZBGZtz/HmYteuL5esY/PasIZsY5NDlSKx5fHMW25a7tsYQFA0oU1GADGjB2Y SoP2YC5B/lXcEqWLiDf9EQFmzW2ExZAIKSMslk7SvM2AhBM2EEV4VlIJJfkhl3J5iVMN EHVz4HlqXme6Q2QRWSSpwvEegzRf4NUbPTESnJmiiKki+KPBji+xC3OYh1rdB002lKCF 3Z4A== X-Gm-Message-State: ACrzQf2f95gzaCAA/j4GXZ7CkrQd/aqzFF5ZuwA6ViJJUiXyBtzroF/0 2fNZSN73URV5Bt4zACXU+rOq7A== X-Google-Smtp-Source: AMsMyM7+Xxg8rN7mARF8H/zFcQ5I90sT+zv1UiDqHoKrC8RjleUl2bwORbdfdB60pONvnVH5WKhLUQ== X-Received: by 2002:a05:6602:2a42:b0:678:84be:c9ec with SMTP id k2-20020a0566022a4200b0067884bec9ecmr2211543iov.64.1663871290172; Thu, 22 Sep 2022 11:28:10 -0700 (PDT) Received: from m1max.localdomain ([207.135.234.126]) by smtp.gmail.com with ESMTPSA id q20-20020a05663810d400b0035a468b7fbesm2440646jad.71.2022.09.22.11.28.09 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Thu, 22 Sep 2022 11:28:09 -0700 (PDT) From: Jens Axboe To: linux-block@vger.kernel.org Cc: linux-scsi@vger.kernel.org, linux-nvme@lists.infradead.org, Jens Axboe , Stefan Roesch Subject: [PATCH 3/5] block: allow end_io based requests in the completion batch handling Date: Thu, 22 Sep 2022 12:28:03 -0600 Message-Id: <20220922182805.96173-4-axboe@kernel.dk> X-Mailer: git-send-email 2.35.1 In-Reply-To: <20220922182805.96173-1-axboe@kernel.dk> References: <20220922182805.96173-1-axboe@kernel.dk> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-scsi@vger.kernel.org With end_io handlers now being able to potentially pass ownership of the request upon completion, we can allow requests with end_io handlers in the batch completion handling. Co-developed-by: Stefan Roesch Signed-off-by: Jens Axboe --- block/blk-mq.c | 13 +++++++++++-- include/linux/blk-mq.h | 3 ++- 2 files changed, 13 insertions(+), 3 deletions(-) diff --git a/block/blk-mq.c b/block/blk-mq.c index a4e018c82b7c..523a201b9acf 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -823,8 +823,10 @@ static void blk_complete_request(struct request *req) * can find how many bytes remain in the request * later. */ - req->bio = NULL; - req->__data_len = 0; + if (!req->end_io) { + req->bio = NULL; + req->__data_len = 0; + }; } /** @@ -1055,6 +1057,13 @@ void blk_mq_end_request_batch(struct io_comp_batch *iob) rq_qos_done(rq->q, rq); + /* + * If end_io handler returns NONE, then it still has + * ownership of the request. + */ + if (rq->end_io && rq->end_io(rq, 0) == RQ_END_IO_NONE) + continue; + WRITE_ONCE(rq->state, MQ_RQ_IDLE); if (!req_ref_put_and_test(rq)) continue; diff --git a/include/linux/blk-mq.h b/include/linux/blk-mq.h index e6fa49dd6196..50811d0fb143 100644 --- a/include/linux/blk-mq.h +++ b/include/linux/blk-mq.h @@ -853,8 +853,9 @@ static inline bool blk_mq_add_to_batch(struct request *req, struct io_comp_batch *iob, int ioerror, void (*complete)(struct io_comp_batch *)) { - if (!iob || (req->rq_flags & RQF_ELV) || req->end_io || ioerror) + if (!iob || (req->rq_flags & RQF_ELV) || ioerror) return false; + if (!iob->complete) iob->complete = complete; else if (iob->complete != complete) From patchwork Thu Sep 22 18:28:04 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Jens Axboe X-Patchwork-Id: 609089 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 8F798C54EE9 for ; Thu, 22 Sep 2022 18:31:42 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S231294AbiIVSbj (ORCPT ); Thu, 22 Sep 2022 14:31:39 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:51710 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S231564AbiIVSa6 (ORCPT ); Thu, 22 Sep 2022 14:30:58 -0400 Received: from mail-io1-xd2e.google.com (mail-io1-xd2e.google.com [IPv6:2607:f8b0:4864:20::d2e]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 32CE610CA41 for ; Thu, 22 Sep 2022 11:28:12 -0700 (PDT) Received: by mail-io1-xd2e.google.com with SMTP id p3so8428894iof.13 for ; Thu, 22 Sep 2022 11:28:12 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=kernel-dk.20210112.gappssmtp.com; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date; bh=8xAQy7aZ7BYH1DPQ+8tvyApZhb4BnQnHW544FcJ7XiE=; b=aR9k5LYhnfTw7XXQzoO5WAz8DP39KvgUKyrz1rGsakd/XddUjw6vg3NHGqsum9PG+V QoYoKbT2kP5nhO/G7Sy/qccemyc0E3pmzD1crhRZt6HqXTJEmz/xelocvDqKLNuVcNQL Sl0HAFLPpFWkeQlEmS28RX7jV9+hMHFzbx3TOKoTuYtLcwK9Es9RpjUt11hx8enBYeBL f8ky37PViY/csWh7e3o3oS3HydD07wSBCxDqzwz1zuVf+k0iAXXg38nE5IjnxNq1FTab lrWaeH8Gb5fnrrASwQQHIZ/F9tg1KlBE+kuCAi0dR8ZXyjP4xFMf/mRUd8VVQV7PK8FO ew0w== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date; bh=8xAQy7aZ7BYH1DPQ+8tvyApZhb4BnQnHW544FcJ7XiE=; b=VmThlZq/97AZ4/qY6bFfNh0TVF1MsJoUtpFXUTdzp6Cte/GUC8iYLspzfUkTUte8GF i0FGiidvJ62xoKBMTqM6z1LihMKSgDbDmY83S1OrAwCsiE87E10oAe6Dyylx+H0s9x0A rAh+0HIM/viPWclThvNJGezCDWwlNayRnukSlzQImzVGgNpkJCVetdXMDYyhf4vKtGae elkm+p4s7eu6vf76Kgtx99pXNOqCCWV6idV3XhHxVnPFwhXzlqchTTUQEDMy7zYWiEfw G0HFrN0/rIJf/WfuuJA76gJby8jirby9/lmp/N5EBCm6/P6/c27FnMy2pJduOQjy0Yi8 WhUw== X-Gm-Message-State: ACrzQf0bUgofij4VG8f33EYrApPH2I49ICh514ZEIbRCmspUiplc4h02 X+IIVG08dGsgia8/l4E80brudA== X-Google-Smtp-Source: AMsMyM4SGqW5jyi20l7Smo7p+cJHNnri3FkeeL9qvvrk+1n6md6nK0KgGtTcdgUAs/H34Ch0+VvUPg== X-Received: by 2002:a6b:ba05:0:b0:68a:8a2f:8fdc with SMTP id k5-20020a6bba05000000b0068a8a2f8fdcmr2255996iof.148.1663871291224; Thu, 22 Sep 2022 11:28:11 -0700 (PDT) Received: from m1max.localdomain ([207.135.234.126]) by smtp.gmail.com with ESMTPSA id q20-20020a05663810d400b0035a468b7fbesm2440646jad.71.2022.09.22.11.28.10 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Thu, 22 Sep 2022 11:28:10 -0700 (PDT) From: Jens Axboe To: linux-block@vger.kernel.org Cc: linux-scsi@vger.kernel.org, linux-nvme@lists.infradead.org, Jens Axboe , Stefan Roesch Subject: [PATCH 4/5] nvme: split out metadata vs non metadata end_io uring_cmd completions Date: Thu, 22 Sep 2022 12:28:04 -0600 Message-Id: <20220922182805.96173-5-axboe@kernel.dk> X-Mailer: git-send-email 2.35.1 In-Reply-To: <20220922182805.96173-1-axboe@kernel.dk> References: <20220922182805.96173-1-axboe@kernel.dk> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-scsi@vger.kernel.org By splitting up the metadata and non-metadata end_io handling, we can remove any request dependencies on the normal non-metadata IO path. This is in preparation for enabling the normal IO passthrough path to pass the ownership of the request back to the block layer. Co-developed-by: Stefan Roesch Signed-off-by: Jens Axboe --- drivers/nvme/host/ioctl.c | 82 +++++++++++++++++++++++++++++++-------- 1 file changed, 66 insertions(+), 16 deletions(-) diff --git a/drivers/nvme/host/ioctl.c b/drivers/nvme/host/ioctl.c index c80b3ecca5c8..1ccc9dd6d434 100644 --- a/drivers/nvme/host/ioctl.c +++ b/drivers/nvme/host/ioctl.c @@ -349,9 +349,18 @@ struct nvme_uring_cmd_pdu { struct bio *bio; struct request *req; }; - void *meta; /* kernel-resident buffer */ - void __user *meta_buffer; u32 meta_len; + union { + struct { + void *meta; /* kernel-resident buffer */ + void __user *meta_buffer; + }; + struct { + u32 nvme_flags; + u32 nvme_status; + u64 result; + }; + }; }; static inline struct nvme_uring_cmd_pdu *nvme_uring_cmd_pdu( @@ -360,11 +369,10 @@ static inline struct nvme_uring_cmd_pdu *nvme_uring_cmd_pdu( return (struct nvme_uring_cmd_pdu *)&ioucmd->pdu; } -static void nvme_uring_task_cb(struct io_uring_cmd *ioucmd) +static void nvme_uring_task_meta_cb(struct io_uring_cmd *ioucmd) { struct nvme_uring_cmd_pdu *pdu = nvme_uring_cmd_pdu(ioucmd); struct request *req = pdu->req; - struct bio *bio = req->bio; int status; u64 result; @@ -375,27 +383,43 @@ static void nvme_uring_task_cb(struct io_uring_cmd *ioucmd) result = le64_to_cpu(nvme_req(req)->result.u64); - if (pdu->meta) + if (pdu->meta_len) status = nvme_finish_user_metadata(req, pdu->meta_buffer, pdu->meta, pdu->meta_len, status); - if (bio) - blk_rq_unmap_user(bio); + if (req->bio) + blk_rq_unmap_user(req->bio); blk_mq_free_request(req); io_uring_cmd_done(ioucmd, status, result); } +static void nvme_uring_task_cb(struct io_uring_cmd *ioucmd) +{ + struct nvme_uring_cmd_pdu *pdu = nvme_uring_cmd_pdu(ioucmd); + int status; + + if (pdu->nvme_flags & NVME_REQ_CANCELLED) + status = -EINTR; + else + status = pdu->nvme_status; + + if (pdu->bio) + blk_rq_unmap_user(pdu->bio); + + io_uring_cmd_done(ioucmd, status, pdu->result); +} + static enum rq_end_io_ret nvme_uring_cmd_end_io(struct request *req, blk_status_t err) { struct io_uring_cmd *ioucmd = req->end_io_data; struct nvme_uring_cmd_pdu *pdu = nvme_uring_cmd_pdu(ioucmd); - /* extract bio before reusing the same field for request */ - struct bio *bio = pdu->bio; void *cookie = READ_ONCE(ioucmd->cookie); - pdu->req = req; - req->bio = bio; + req->bio = pdu->bio; + pdu->nvme_flags = nvme_req(req)->flags; + pdu->nvme_status = nvme_req(req)->status; + pdu->result = le64_to_cpu(nvme_req(req)->result.u64); /* * For iopoll, complete it directly. @@ -406,6 +430,29 @@ static enum rq_end_io_ret nvme_uring_cmd_end_io(struct request *req, else io_uring_cmd_complete_in_task(ioucmd, nvme_uring_task_cb); + blk_mq_free_request(req); + return RQ_END_IO_NONE; +} + +static enum rq_end_io_ret nvme_uring_cmd_end_io_meta(struct request *req, + blk_status_t err) +{ + struct io_uring_cmd *ioucmd = req->end_io_data; + struct nvme_uring_cmd_pdu *pdu = nvme_uring_cmd_pdu(ioucmd); + void *cookie = READ_ONCE(ioucmd->cookie); + + req->bio = pdu->bio; + pdu->req = req; + + /* + * For iopoll, complete it directly. + * Otherwise, move the completion to task work. + */ + if (cookie != NULL && blk_rq_is_poll(req)) + nvme_uring_task_meta_cb(ioucmd); + else + io_uring_cmd_complete_in_task(ioucmd, nvme_uring_task_meta_cb); + return RQ_END_IO_NONE; } @@ -467,8 +514,6 @@ static int nvme_uring_cmd_io(struct nvme_ctrl *ctrl, struct nvme_ns *ns, blk_flags); if (IS_ERR(req)) return PTR_ERR(req); - req->end_io = nvme_uring_cmd_end_io; - req->end_io_data = ioucmd; if (issue_flags & IO_URING_F_IOPOLL && rq_flags & REQ_POLLED) { if (unlikely(!req->bio)) { @@ -483,10 +528,15 @@ static int nvme_uring_cmd_io(struct nvme_ctrl *ctrl, struct nvme_ns *ns, } /* to free bio on completion, as req->bio will be null at that time */ pdu->bio = req->bio; - pdu->meta = meta; - pdu->meta_buffer = nvme_to_user_ptr(d.metadata); pdu->meta_len = d.metadata_len; - + req->end_io_data = ioucmd; + if (pdu->meta_len) { + pdu->meta = meta; + pdu->meta_buffer = nvme_to_user_ptr(d.metadata); + req->end_io = nvme_uring_cmd_end_io_meta; + } else { + req->end_io = nvme_uring_cmd_end_io; + } blk_execute_rq_nowait(req, false); return -EIOCBQUEUED; } From patchwork Thu Sep 22 18:28:05 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Jens Axboe X-Patchwork-Id: 608535 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 3D51BC6FA82 for ; Thu, 22 Sep 2022 18:31:43 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S232558AbiIVSbm (ORCPT ); Thu, 22 Sep 2022 14:31:42 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:51124 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S232380AbiIVSa6 (ORCPT ); Thu, 22 Sep 2022 14:30:58 -0400 Received: from mail-io1-xd35.google.com (mail-io1-xd35.google.com [IPv6:2607:f8b0:4864:20::d35]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 0F18C10CA49 for ; Thu, 22 Sep 2022 11:28:13 -0700 (PDT) Received: by mail-io1-xd35.google.com with SMTP id c4so8469006iof.3 for ; Thu, 22 Sep 2022 11:28:13 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=kernel-dk.20210112.gappssmtp.com; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date; bh=kQJ1C1SjENmgCWeZQvNCbMqCIh5TAvtoZAf25z/EVT4=; b=pDHEDqNu7gIG1tZwKjXkijlAv3FbeAirIAJbQJQAChrlUlJo/pXRxqh4kPVNmZHRN0 ho2TBXNKrJRs0X0PuTx/XSnV1N/3gEMyvAD+PKLBJABAtrtukOHMYmJlaoXTrGEBJZYu Wz9Bnx81LWQKMEHsI3dQbaXqmEUiiUrrFm3n9mgBmkwa1hddM0Bnm5cAAjO9D4tR+m7F xGfFlpuxYbyEkJe3z68GYesMmaq9y6k6UwVXEc1/rc68JCnvYnX+blLE1vt0O+IbMk3C 63jF4C+tW0MfbGRhjK3il/WqkrNi9L+bhW9W4BEXr3cIS/fKD6C/TMNWY747CKcKVQhe r35g== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date; bh=kQJ1C1SjENmgCWeZQvNCbMqCIh5TAvtoZAf25z/EVT4=; b=a3XiXczf3SfhQ2aaFJ2y2ZcP0fj3xCUifyOiaTImG3czQiRgpKM8sSoLW3xKHJ2Iw6 V/Y+4U3OKzV1a9/uyGozJ3U/SrChZG9XYD8JXcrSsm6rI2YjKC2orYL/HeijDR+p4fV9 11sjqTzL+drd1xdoVX0bmANoC55Dvkf4X/WUDZ5RD3mntQX/WovNzaIptqPmPqdb+m1O iA0uUnZzZN5DbdxR4ZWuttrN1sVl41OMx0MFzEHadEMBjlm09sRZjChhDdcV6UsNkEKs g0cMD8jSbrJ+hxXdlsb5E9uBGmnz2HOdIDuZ/lxHIE9KktuFDGY8dWW3YXp3414DjiNo s+UQ== X-Gm-Message-State: ACrzQf2e8ZbfIMCuLTG/C/7/eNIGkgT5DPaf7ONrvDDKdnO2cXOVbqEd YAQR8kST0JXB5ab2iHVbhZltbQ== X-Google-Smtp-Source: AMsMyM7wM97v1Oqw5RrX8wIN6VT4gEUijfOCbtCcdhJgZwn4LHZr9RZRcQ17ytyb3H0/RwWNoOeDOw== X-Received: by 2002:a05:6602:1352:b0:69d:e793:abd with SMTP id i18-20020a056602135200b0069de7930abdmr2341538iov.172.1663871292289; Thu, 22 Sep 2022 11:28:12 -0700 (PDT) Received: from m1max.localdomain ([207.135.234.126]) by smtp.gmail.com with ESMTPSA id q20-20020a05663810d400b0035a468b7fbesm2440646jad.71.2022.09.22.11.28.11 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Thu, 22 Sep 2022 11:28:11 -0700 (PDT) From: Jens Axboe To: linux-block@vger.kernel.org Cc: linux-scsi@vger.kernel.org, linux-nvme@lists.infradead.org, Jens Axboe , Stefan Roesch Subject: [PATCH 5/5] nvme: enable batched completions of passthrough IO Date: Thu, 22 Sep 2022 12:28:05 -0600 Message-Id: <20220922182805.96173-6-axboe@kernel.dk> X-Mailer: git-send-email 2.35.1 In-Reply-To: <20220922182805.96173-1-axboe@kernel.dk> References: <20220922182805.96173-1-axboe@kernel.dk> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-scsi@vger.kernel.org Now that the normal passthrough end_io path doesn't need the request anymore, we can kill the explicit blk_mq_free_request() and just pass back RQ_END_IO_FREE instead. This enables the batched completion from freeing batches of requests at the time. This brings passthrough IO performance at least on par with bdev based O_DIRECT with io_uring. With this and batche allocations, peak performance goes from 110M IOPS to 122M IOPS. For IRQ based, passthrough is now also about 10% faster than previously, going from ~61M to ~67M IOPS. Co-developed-by: Stefan Roesch Signed-off-by: Jens Axboe --- drivers/nvme/host/ioctl.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/nvme/host/ioctl.c b/drivers/nvme/host/ioctl.c index 1ccc9dd6d434..25f2f6df1602 100644 --- a/drivers/nvme/host/ioctl.c +++ b/drivers/nvme/host/ioctl.c @@ -430,8 +430,7 @@ static enum rq_end_io_ret nvme_uring_cmd_end_io(struct request *req, else io_uring_cmd_complete_in_task(ioucmd, nvme_uring_task_cb); - blk_mq_free_request(req); - return RQ_END_IO_NONE; + return RQ_END_IO_FREE; } static enum rq_end_io_ret nvme_uring_cmd_end_io_meta(struct request *req,