From patchwork Fri Mar 26 05:58:22 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Christoph Hellwig X-Patchwork-Id: 409855 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=-16.8 required=3.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID, HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_CR_TRAILER, INCLUDES_PATCH, MAILING_LIST_MULTI, SPF_HELO_NONE, SPF_PASS, URIBL_BLOCKED, USER_AGENT_GIT 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 C754DC433F1 for ; Fri, 26 Mar 2021 05:59:48 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id AE5B361A49 for ; Fri, 26 Mar 2021 05:59:48 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S229982AbhCZF7T (ORCPT ); Fri, 26 Mar 2021 01:59:19 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:54192 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229913AbhCZF64 (ORCPT ); Fri, 26 Mar 2021 01:58:56 -0400 Received: from bombadil.infradead.org (bombadil.infradead.org [IPv6:2607:7c80:54:e::133]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 808FCC0613AA; Thu, 25 Mar 2021 22:58:56 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=infradead.org; s=bombadil.20210309; h=Content-Transfer-Encoding: MIME-Version:References:In-Reply-To:Message-Id:Date:Subject:Cc:To:From:Sender :Reply-To:Content-Type:Content-ID:Content-Description; bh=H4Dtrbj4H6gmQ0NyhNB1Ua+zBbZw+6zj61B3DLC3WTA=; b=Tp9j3ZgUT/dVCc6zx9s9CaIfro H5YfnnTITzUlc8q/H0kb5cvq6ieEET+NqnajF8RhWBR/FT9hx9Hr9a/9ZEo4bQIh6facueEnocGdD 59GwMlSGKOQ2Tho9X9kQlCQQuBImE1gNZ6plc5hR/wHUJbd3il74y/QpUzaj1Vs1X86/JAbu5scxD sJrF4M660dxrri6dAp99IWOfdpcRRjhkCZuPJ8quSyINJl4KjtO3aPgCo9RRy/3zvFlGK7WghTV0s uYGCQVhpCFznRIKRCF9fMqv422lusbViAyZzjFRyKvUw5W6B0rDunhj5Jct4xH3egIQpH0KBc8Nik QgsBYbeg==; Received: from [2001:4bb8:191:f692:97ff:1e47:aee2:c7e5] (helo=localhost) by bombadil.infradead.org with esmtpsa (Exim 4.94 #2 (Red Hat Linux)) id 1lPfUa-005AP9-KW; Fri, 26 Mar 2021 05:58:45 +0000 From: Christoph Hellwig To: Jens Axboe , Khalid Aziz , "Martin K. Petersen" , Matthew Wilcox , Hannes Reinecke , Ondrej Zary Cc: linux-block@vger.kernel.org, linux-scsi@vger.kernel.org Subject: [PATCH 8/8] block: stop calling blk_queue_bounce for passthrough requests Date: Fri, 26 Mar 2021 06:58:22 +0100 Message-Id: <20210326055822.1437471-9-hch@lst.de> X-Mailer: git-send-email 2.30.1 In-Reply-To: <20210326055822.1437471-1-hch@lst.de> References: <20210326055822.1437471-1-hch@lst.de> MIME-Version: 1.0 X-SRS-Rewrite: SMTP reverse-path rewritten from by bombadil.infradead.org. See http://www.infradead.org/rpr.html Precedence: bulk List-ID: X-Mailing-List: linux-scsi@vger.kernel.org Instead of overloading the passthrough fast path with the deprecated block layer bounce buffering let the users that combine an old undermaintained driver with a highmem system pay the price by always falling back to copies in that case. Signed-off-by: Christoph Hellwig Reviewed-by: Hannes Reinecke --- block/blk-map.c | 116 ++++++++--------------------- block/bounce.c | 11 +-- drivers/nvme/host/lightnvm.c | 2 +- drivers/target/target_core_pscsi.c | 4 +- include/linux/blkdev.h | 2 +- 5 files changed, 36 insertions(+), 99 deletions(-) diff --git a/block/blk-map.c b/block/blk-map.c index b62b52dcb61d97..dac78376acc899 100644 --- a/block/blk-map.c +++ b/block/blk-map.c @@ -123,7 +123,6 @@ static int bio_uncopy_user(struct bio *bio) bio_free_pages(bio); } kfree(bmd); - bio_put(bio); return ret; } @@ -132,7 +131,7 @@ static int bio_copy_user_iov(struct request *rq, struct rq_map_data *map_data, { struct bio_map_data *bmd; struct page *page; - struct bio *bio, *bounce_bio; + struct bio *bio; int i = 0, ret; int nr_pages; unsigned int len = iter->count; @@ -218,16 +217,9 @@ static int bio_copy_user_iov(struct request *rq, struct rq_map_data *map_data, bio->bi_private = bmd; - bounce_bio = bio; - ret = blk_rq_append_bio(rq, &bounce_bio); + ret = blk_rq_append_bio(rq, bio); if (ret) goto cleanup; - - /* - * We link the bounce buffer in and could have to traverse it later, so - * we have to get a ref to prevent it from being freed - */ - bio_get(bounce_bio); return 0; cleanup: if (!map_data) @@ -242,7 +234,7 @@ static int bio_map_user_iov(struct request *rq, struct iov_iter *iter, gfp_t gfp_mask) { unsigned int max_sectors = queue_max_hw_sectors(rq->q); - struct bio *bio, *bounce_bio; + struct bio *bio; int ret; int j; @@ -304,49 +296,17 @@ static int bio_map_user_iov(struct request *rq, struct iov_iter *iter, break; } - /* - * Subtle: if we end up needing to bounce a bio, it would normally - * disappear when its bi_end_io is run. However, we need the original - * bio for the unmap, so grab an extra reference to it - */ - bio_get(bio); - - bounce_bio = bio; - ret = blk_rq_append_bio(rq, &bounce_bio); + ret = blk_rq_append_bio(rq, bio); if (ret) - goto out_put_orig; - - /* - * We link the bounce buffer in and could have to traverse it - * later, so we have to get a ref to prevent it from being freed - */ - bio_get(bounce_bio); + goto out_unmap; return 0; - out_put_orig: - bio_put(bio); out_unmap: bio_release_pages(bio, false); bio_put(bio); return ret; } -/** - * bio_unmap_user - unmap a bio - * @bio: the bio being unmapped - * - * Unmap a bio previously mapped by bio_map_user_iov(). Must be called from - * process context. - * - * bio_unmap_user() may sleep. - */ -static void bio_unmap_user(struct bio *bio) -{ - bio_release_pages(bio, bio_data_dir(bio) == READ); - bio_put(bio); - bio_put(bio); -} - static void bio_invalidate_vmalloc_pages(struct bio *bio) { #ifdef ARCH_HAS_FLUSH_KERNEL_DCACHE_PAGE @@ -519,33 +479,27 @@ static struct bio *bio_copy_kern(struct request_queue *q, void *data, * Append a bio to a passthrough request. Only works if the bio can be merged * into the request based on the driver constraints. */ -int blk_rq_append_bio(struct request *rq, struct bio **bio) +int blk_rq_append_bio(struct request *rq, struct bio *bio) { - struct bio *orig_bio = *bio; struct bvec_iter iter; struct bio_vec bv; unsigned int nr_segs = 0; - blk_queue_bounce(rq->q, bio); + if (WARN_ON_ONCE(rq->q->limits.bounce != BLK_BOUNCE_NONE)) + return -EINVAL; - bio_for_each_bvec(bv, *bio, iter) + bio_for_each_bvec(bv, bio, iter) nr_segs++; if (!rq->bio) { - blk_rq_bio_prep(rq, *bio, nr_segs); + blk_rq_bio_prep(rq, bio, nr_segs); } else { - if (!ll_back_merge_fn(rq, *bio, nr_segs)) { - if (orig_bio != *bio) { - bio_put(*bio); - *bio = orig_bio; - } + if (!ll_back_merge_fn(rq, bio, nr_segs)) return -EINVAL; - } - - rq->biotail->bi_next = *bio; - rq->biotail = *bio; - rq->__data_len += (*bio)->bi_iter.bi_size; - bio_crypt_free_ctx(*bio); + rq->biotail->bi_next = bio; + rq->biotail = bio; + rq->__data_len += (bio)->bi_iter.bi_size; + bio_crypt_free_ctx(bio); } return 0; @@ -566,12 +520,6 @@ EXPORT_SYMBOL(blk_rq_append_bio); * * A matching blk_rq_unmap_user() must be issued at the end of I/O, while * still in process context. - * - * Note: The mapped bio may need to be bounced through blk_queue_bounce() - * before being submitted to the device, as pages mapped may be out of - * reach. It's the callers responsibility to make sure this happens. The - * original bio must be passed back in to blk_rq_unmap_user() for proper - * unmapping. */ int blk_rq_map_user_iov(struct request_queue *q, struct request *rq, struct rq_map_data *map_data, @@ -588,6 +536,8 @@ int blk_rq_map_user_iov(struct request_queue *q, struct request *rq, if (map_data) copy = true; + else if (blk_queue_may_bounce(q)) + copy = true; else if (iov_iter_alignment(iter) & align) copy = true; else if (queue_virt_boundary(q)) @@ -641,25 +591,21 @@ EXPORT_SYMBOL(blk_rq_map_user); */ int blk_rq_unmap_user(struct bio *bio) { - struct bio *mapped_bio; + struct bio *next_bio; int ret = 0, ret2; while (bio) { - mapped_bio = bio; - if (unlikely(bio_flagged(bio, BIO_BOUNCED))) - mapped_bio = bio->bi_private; - if (bio->bi_private) { - ret2 = bio_uncopy_user(mapped_bio); + ret2 = bio_uncopy_user(bio); if (ret2 && !ret) ret = ret2; } else { - bio_unmap_user(mapped_bio); + bio_release_pages(bio, bio_data_dir(bio) == READ); } - mapped_bio = bio; + next_bio = bio; bio = bio->bi_next; - bio_put(mapped_bio); + bio_put(next_bio); } return ret; @@ -684,7 +630,7 @@ int blk_rq_map_kern(struct request_queue *q, struct request *rq, void *kbuf, { int reading = rq_data_dir(rq) == READ; unsigned long addr = (unsigned long) kbuf; - struct bio *bio, *orig_bio; + struct bio *bio; int ret; if (len > (queue_max_hw_sectors(q) << 9)) @@ -692,7 +638,8 @@ int blk_rq_map_kern(struct request_queue *q, struct request *rq, void *kbuf, if (!len || !kbuf) return -EINVAL; - if (!blk_rq_aligned(q, addr, len) || object_is_on_stack(kbuf)) + if (!blk_rq_aligned(q, addr, len) || object_is_on_stack(kbuf) || + blk_queue_may_bounce(q)) bio = bio_copy_kern(q, kbuf, len, gfp_mask, reading); else bio = bio_map_kern(q, kbuf, len, gfp_mask); @@ -703,14 +650,9 @@ int blk_rq_map_kern(struct request_queue *q, struct request *rq, void *kbuf, bio->bi_opf &= ~REQ_OP_MASK; bio->bi_opf |= req_op(rq); - orig_bio = bio; - ret = blk_rq_append_bio(rq, &bio); - if (unlikely(ret)) { - /* request is too big */ - bio_put(orig_bio); - return ret; - } - - return 0; + ret = blk_rq_append_bio(rq, bio); + if (unlikely(ret)) + bio_put(bio); + return ret; } EXPORT_SYMBOL(blk_rq_map_kern); diff --git a/block/bounce.c b/block/bounce.c index 6bafc0d1f867a1..94081e013c58cc 100644 --- a/block/bounce.c +++ b/block/bounce.c @@ -180,12 +180,8 @@ static struct bio *bounce_clone_bio(struct bio *bio_src) * asking for trouble and would force extra work on * __bio_clone_fast() anyways. */ - if (bio_is_passthrough(bio_src)) - bio = bio_kmalloc(GFP_NOIO | __GFP_NOFAIL, - bio_segments(bio_src)); - else - bio = bio_alloc_bioset(GFP_NOIO, bio_segments(bio_src), - &bounce_bio_set); + bio = bio_alloc_bioset(GFP_NOIO, bio_segments(bio_src), + &bounce_bio_set); bio->bi_bdev = bio_src->bi_bdev; if (bio_flagged(bio_src, BIO_REMAPPED)) bio_set_flag(bio, BIO_REMAPPED); @@ -245,8 +241,7 @@ void __blk_queue_bounce(struct request_queue *q, struct bio **bio_orig) if (!bounce) return; - if (!bio_is_passthrough(*bio_orig) && - sectors < bio_sectors(*bio_orig)) { + if (sectors < bio_sectors(*bio_orig)) { bio = bio_split(*bio_orig, sectors, GFP_NOIO, &bounce_bio_split); bio_chain(bio, *bio_orig); submit_bio_noacct(*bio_orig); diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c index b705988629f224..f6ca2fbb711e98 100644 --- a/drivers/nvme/host/lightnvm.c +++ b/drivers/nvme/host/lightnvm.c @@ -660,7 +660,7 @@ static struct request *nvme_nvm_alloc_request(struct request_queue *q, rq->cmd_flags &= ~REQ_FAILFAST_DRIVER; if (rqd->bio) - blk_rq_append_bio(rq, &rqd->bio); + blk_rq_append_bio(rq, rqd->bio); else rq->ioprio = IOPRIO_PRIO_VALUE(IOPRIO_CLASS_BE, IOPRIO_NORM); diff --git a/drivers/target/target_core_pscsi.c b/drivers/target/target_core_pscsi.c index 3cbc074992bc86..7df4a9c9c7ffaa 100644 --- a/drivers/target/target_core_pscsi.c +++ b/drivers/target/target_core_pscsi.c @@ -911,7 +911,7 @@ pscsi_map_sg(struct se_cmd *cmd, struct scatterlist *sgl, u32 sgl_nents, " %d i: %d bio: %p, allocating another" " bio\n", bio->bi_vcnt, i, bio); - rc = blk_rq_append_bio(req, &bio); + rc = blk_rq_append_bio(req, bio); if (rc) { pr_err("pSCSI: failed to append bio\n"); goto fail; @@ -930,7 +930,7 @@ pscsi_map_sg(struct se_cmd *cmd, struct scatterlist *sgl, u32 sgl_nents, } if (bio) { - rc = blk_rq_append_bio(req, &bio); + rc = blk_rq_append_bio(req, bio); if (rc) { pr_err("pSCSI: failed to append bio\n"); goto fail; diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 55cc8b96c84427..d5d320da51f8bf 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -909,7 +909,7 @@ extern int blk_rq_prep_clone(struct request *rq, struct request *rq_src, extern void blk_rq_unprep_clone(struct request *rq); extern blk_status_t blk_insert_cloned_request(struct request_queue *q, struct request *rq); -extern int blk_rq_append_bio(struct request *rq, struct bio **bio); +int blk_rq_append_bio(struct request *rq, struct bio *bio); extern void blk_queue_split(struct bio **); extern int scsi_verify_blk_ioctl(struct block_device *, unsigned int); extern int scsi_cmd_blk_ioctl(struct block_device *, fmode_t,