Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932106AbbETSLB (ORCPT ); Wed, 20 May 2015 14:11:01 -0400 Received: from mail.kernel.org ([198.145.29.136]:34917 "EHLO mail.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753303AbbETSK6 (ORCPT ); Wed, 20 May 2015 14:10:58 -0400 MIME-Version: 1.0 In-Reply-To: <20150520124817.GB15594@lst.de> References: <1430980461-5235-1-git-send-email-mlin@kernel.org> <20150520124817.GB15594@lst.de> Date: Wed, 20 May 2015 11:10:53 -0700 Message-ID: Subject: Re: [PATCH v3 00/11] simplify block layer based on immutable biovecs From: Ming Lin To: Christoph Hellwig Cc: Ming Lin , lkml , Jens Axboe , Kent Overstreet , Dongsu Park Content-Type: text/plain; charset=UTF-8 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 13967 Lines: 345 On Wed, May 20, 2015 at 5:48 AM, Christoph Hellwig wrote: > Passes test fine for me so far. You might also want to throw in the > following patch to make it more useful: I'll add it to next version. Thanks. > > --- > From a30035d4ae040723a6c94143db90231941d0caf7 Mon Sep 17 00:00:00 2001 > From: Kent Overstreet > Date: Tue, 19 May 2015 14:31:01 +0200 > Subject: block: remove bio_get_nr_vecs() > > We can always fill up the bio now, no need to estimate the possible > size based on queue parameters. > > Signed-off-by: Kent Overstreet > [hch: rebased and wrote a changelog] > Signed-off-by: Christoph Hellwig > --- > block/bio.c | 23 ----------------------- > drivers/md/dm-io.c | 2 +- > fs/btrfs/compression.c | 5 +---- > fs/btrfs/extent_io.c | 9 ++------- > fs/btrfs/inode.c | 3 +-- > fs/btrfs/scrub.c | 18 ++---------------- > fs/direct-io.c | 2 +- > fs/ext4/page-io.c | 3 +-- > fs/ext4/readpage.c | 2 +- > fs/gfs2/lops.c | 9 +-------- > fs/logfs/dev_bdev.c | 4 ++-- > fs/mpage.c | 4 ++-- > fs/nilfs2/segbuf.c | 2 +- > fs/xfs/xfs_aops.c | 3 +-- > include/linux/bio.h | 1 - > 15 files changed, 17 insertions(+), 73 deletions(-) > > diff --git a/block/bio.c b/block/bio.c > index ae31cdb..f59b647 100644 > --- a/block/bio.c > +++ b/block/bio.c > @@ -677,29 +677,6 @@ integrity_clone: > EXPORT_SYMBOL(bio_clone_bioset); > > /** > - * bio_get_nr_vecs - return approx number of vecs > - * @bdev: I/O target > - * > - * Return the approximate number of pages we can send to this target. > - * There's no guarantee that you will be able to fit this number of pages > - * into a bio, it does not account for dynamic restrictions that vary > - * on offset. > - */ > -int bio_get_nr_vecs(struct block_device *bdev) > -{ > - struct request_queue *q = bdev_get_queue(bdev); > - int nr_pages; > - > - nr_pages = min_t(unsigned, > - queue_max_segments(q), > - queue_max_sectors(q) / (PAGE_SIZE >> 9) + 1); > - > - return min_t(unsigned, nr_pages, BIO_MAX_PAGES); > - > -} > -EXPORT_SYMBOL(bio_get_nr_vecs); > - > -/** > * bio_add_pc_page - attempt to add page to bio > * @q: the target queue > * @bio: destination bio > diff --git a/drivers/md/dm-io.c b/drivers/md/dm-io.c > index 74adcd2..7d64272 100644 > --- a/drivers/md/dm-io.c > +++ b/drivers/md/dm-io.c > @@ -314,7 +314,7 @@ static void do_region(int rw, unsigned region, struct dm_io_region *where, > if ((rw & REQ_DISCARD) || (rw & REQ_WRITE_SAME)) > num_bvecs = 1; > else > - num_bvecs = min_t(int, bio_get_nr_vecs(where->bdev), > + num_bvecs = min_t(int, BIO_MAX_PAGES, > dm_sector_div_up(remaining, (PAGE_SIZE >> SECTOR_SHIFT))); > > bio = bio_alloc_bioset(GFP_NOIO, num_bvecs, io->client->bios); > diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c > index ce62324..449c752 100644 > --- a/fs/btrfs/compression.c > +++ b/fs/btrfs/compression.c > @@ -97,10 +97,7 @@ static inline int compressed_bio_size(struct btrfs_root *root, > static struct bio *compressed_bio_alloc(struct block_device *bdev, > u64 first_byte, gfp_t gfp_flags) > { > - int nr_vecs; > - > - nr_vecs = bio_get_nr_vecs(bdev); > - return btrfs_bio_alloc(bdev, first_byte >> 9, nr_vecs, gfp_flags); > + return btrfs_bio_alloc(bdev, first_byte >> 9, BIO_MAX_PAGES, gfp_flags); > } > > static int check_compressed_csum(struct inode *inode, > diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c > index 43af5a6..a37af69 100644 > --- a/fs/btrfs/extent_io.c > +++ b/fs/btrfs/extent_io.c > @@ -2799,9 +2799,7 @@ static int submit_extent_page(int rw, struct extent_io_tree *tree, > { > int ret = 0; > struct bio *bio; > - int nr; > int contig = 0; > - int this_compressed = bio_flags & EXTENT_BIO_COMPRESSED; > int old_compressed = prev_bio_flags & EXTENT_BIO_COMPRESSED; > size_t page_size = min_t(size_t, size, PAGE_CACHE_SIZE); > > @@ -2826,12 +2824,9 @@ static int submit_extent_page(int rw, struct extent_io_tree *tree, > return 0; > } > } > - if (this_compressed) > - nr = BIO_MAX_PAGES; > - else > - nr = bio_get_nr_vecs(bdev); > > - bio = btrfs_bio_alloc(bdev, sector, nr, GFP_NOFS | __GFP_HIGH); > + bio = btrfs_bio_alloc(bdev, sector, BIO_MAX_PAGES, > + GFP_NOFS | __GFP_HIGH); > if (!bio) > return -ENOMEM; > > diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c > index 8bb0136..eff9129 100644 > --- a/fs/btrfs/inode.c > +++ b/fs/btrfs/inode.c > @@ -7943,8 +7943,7 @@ out: > static struct bio *btrfs_dio_bio_alloc(struct block_device *bdev, > u64 first_sector, gfp_t gfp_flags) > { > - int nr_vecs = bio_get_nr_vecs(bdev); > - return btrfs_bio_alloc(bdev, first_sector, nr_vecs, gfp_flags); > + return btrfs_bio_alloc(bdev, first_sector, BIO_MAX_PAGES, gfp_flags); > } > > static inline int btrfs_lookup_and_bind_dio_csum(struct btrfs_root *root, > diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c > index ab58115..00b828a 100644 > --- a/fs/btrfs/scrub.c > +++ b/fs/btrfs/scrub.c > @@ -454,27 +454,14 @@ struct scrub_ctx *scrub_setup_ctx(struct btrfs_device *dev, int is_dev_replace) > struct scrub_ctx *sctx; > int i; > struct btrfs_fs_info *fs_info = dev->dev_root->fs_info; > - int pages_per_rd_bio; > int ret; > > - /* > - * the setting of pages_per_rd_bio is correct for scrub but might > - * be wrong for the dev_replace code where we might read from > - * different devices in the initial huge bios. However, that > - * code is able to correctly handle the case when adding a page > - * to a bio fails. > - */ > - if (dev->bdev) > - pages_per_rd_bio = min_t(int, SCRUB_PAGES_PER_RD_BIO, > - bio_get_nr_vecs(dev->bdev)); > - else > - pages_per_rd_bio = SCRUB_PAGES_PER_RD_BIO; > sctx = kzalloc(sizeof(*sctx), GFP_NOFS); > if (!sctx) > goto nomem; > atomic_set(&sctx->refs, 1); > sctx->is_dev_replace = is_dev_replace; > - sctx->pages_per_rd_bio = pages_per_rd_bio; > + sctx->pages_per_rd_bio = SCRUB_PAGES_PER_RD_BIO; > sctx->curr = -1; > sctx->dev_root = dev->dev_root; > for (i = 0; i < SCRUB_BIOS_PER_SCTX; ++i) { > @@ -3875,8 +3862,7 @@ static int scrub_setup_wr_ctx(struct scrub_ctx *sctx, > return 0; > > WARN_ON(!dev->bdev); > - wr_ctx->pages_per_wr_bio = min_t(int, SCRUB_PAGES_PER_WR_BIO, > - bio_get_nr_vecs(dev->bdev)); > + wr_ctx->pages_per_wr_bio = SCRUB_PAGES_PER_WR_BIO; > wr_ctx->tgtdev = dev; > atomic_set(&wr_ctx->flush_all_writes, 0); > return 0; > diff --git a/fs/direct-io.c b/fs/direct-io.c > index 745d234..89baebe 100644 > --- a/fs/direct-io.c > +++ b/fs/direct-io.c > @@ -653,7 +653,7 @@ static inline int dio_new_bio(struct dio *dio, struct dio_submit *sdio, > if (ret) > goto out; > sector = start_sector << (sdio->blkbits - 9); > - nr_pages = min(sdio->pages_in_io, bio_get_nr_vecs(map_bh->b_bdev)); > + nr_pages = min(sdio->pages_in_io, BIO_MAX_PAGES); > BUG_ON(nr_pages <= 0); > dio_bio_alloc(dio, sdio, map_bh->b_bdev, sector, nr_pages); > sdio->boundary = 0; > diff --git a/fs/ext4/page-io.c b/fs/ext4/page-io.c > index 5765f88..000682e 100644 > --- a/fs/ext4/page-io.c > +++ b/fs/ext4/page-io.c > @@ -376,10 +376,9 @@ void ext4_io_submit_init(struct ext4_io_submit *io, > static int io_submit_init_bio(struct ext4_io_submit *io, > struct buffer_head *bh) > { > - int nvecs = bio_get_nr_vecs(bh->b_bdev); > struct bio *bio; > > - bio = bio_alloc(GFP_NOIO, min(nvecs, BIO_MAX_PAGES)); > + bio = bio_alloc(GFP_NOIO, BIO_MAX_PAGES); > if (!bio) > return -ENOMEM; > bio->bi_iter.bi_sector = bh->b_blocknr * (bh->b_size >> 9); > diff --git a/fs/ext4/readpage.c b/fs/ext4/readpage.c > index 171b9ac..de1f96a 100644 > --- a/fs/ext4/readpage.c > +++ b/fs/ext4/readpage.c > @@ -284,7 +284,7 @@ int ext4_mpage_readpages(struct address_space *mapping, > goto set_error_page; > } > bio = bio_alloc(GFP_KERNEL, > - min_t(int, nr_pages, bio_get_nr_vecs(bdev))); > + min_t(int, nr_pages, BIO_MAX_PAGES)); > if (!bio) { > if (ctx) > ext4_release_crypto_ctx(ctx); > diff --git a/fs/gfs2/lops.c b/fs/gfs2/lops.c > index 2c1ae86..64d3116 100644 > --- a/fs/gfs2/lops.c > +++ b/fs/gfs2/lops.c > @@ -261,18 +261,11 @@ void gfs2_log_flush_bio(struct gfs2_sbd *sdp, int rw) > static struct bio *gfs2_log_alloc_bio(struct gfs2_sbd *sdp, u64 blkno) > { > struct super_block *sb = sdp->sd_vfs; > - unsigned nrvecs = bio_get_nr_vecs(sb->s_bdev); > struct bio *bio; > > BUG_ON(sdp->sd_log_bio); > > - while (1) { > - bio = bio_alloc(GFP_NOIO, nrvecs); > - if (likely(bio)) > - break; > - nrvecs = max(nrvecs/2, 1U); > - } > - > + bio = bio_alloc(GFP_NOIO, BIO_MAX_PAGES); > bio->bi_iter.bi_sector = blkno * (sb->s_blocksize >> 9); > bio->bi_bdev = sb->s_bdev; > bio->bi_end_io = gfs2_end_log_write; > diff --git a/fs/logfs/dev_bdev.c b/fs/logfs/dev_bdev.c > index 76279e1..fbb5f95 100644 > --- a/fs/logfs/dev_bdev.c > +++ b/fs/logfs/dev_bdev.c > @@ -83,7 +83,7 @@ static int __bdev_writeseg(struct super_block *sb, u64 ofs, pgoff_t index, > unsigned int max_pages; > int i; > > - max_pages = min(nr_pages, (size_t) bio_get_nr_vecs(super->s_bdev)); > + max_pages = min(nr_pages, BIO_MAX_PAGES); > > bio = bio_alloc(GFP_NOFS, max_pages); > BUG_ON(!bio); > @@ -175,7 +175,7 @@ static int do_erase(struct super_block *sb, u64 ofs, pgoff_t index, > unsigned int max_pages; > int i; > > - max_pages = min(nr_pages, (size_t) bio_get_nr_vecs(super->s_bdev)); > + max_pages = min(nr_pages, BIO_MAX_PAGES); > > bio = bio_alloc(GFP_NOFS, max_pages); > BUG_ON(!bio); > diff --git a/fs/mpage.c b/fs/mpage.c > index 3e79220..29bc6b7 100644 > --- a/fs/mpage.c > +++ b/fs/mpage.c > @@ -277,7 +277,7 @@ alloc_new: > goto out; > } > bio = mpage_alloc(bdev, blocks[0] << (blkbits - 9), > - min_t(int, nr_pages, bio_get_nr_vecs(bdev)), > + min_t(int, nr_pages, BIO_MAX_PAGES), > GFP_KERNEL); > if (bio == NULL) > goto confused; > @@ -602,7 +602,7 @@ alloc_new: > } > } > bio = mpage_alloc(bdev, blocks[0] << (blkbits - 9), > - bio_get_nr_vecs(bdev), GFP_NOFS|__GFP_HIGH); > + BIO_MAX_PAGES, GFP_NOFS|__GFP_HIGH); > if (bio == NULL) > goto confused; > } > diff --git a/fs/nilfs2/segbuf.c b/fs/nilfs2/segbuf.c > index dc3a9efd..9be5a79 100644 > --- a/fs/nilfs2/segbuf.c > +++ b/fs/nilfs2/segbuf.c > @@ -427,7 +427,7 @@ static void nilfs_segbuf_prepare_write(struct nilfs_segment_buffer *segbuf, > { > wi->bio = NULL; > wi->rest_blocks = segbuf->sb_sum.nblocks; > - wi->max_pages = bio_get_nr_vecs(wi->nilfs->ns_bdev); > + wi->max_pages = BIO_MAX_PAGES; > wi->nr_vecs = min(wi->max_pages, wi->rest_blocks); > wi->start = wi->end = 0; > wi->blocknr = segbuf->sb_pseg_start; > diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c > index a56960d..6339fe7 100644 > --- a/fs/xfs/xfs_aops.c > +++ b/fs/xfs/xfs_aops.c > @@ -383,8 +383,7 @@ STATIC struct bio * > xfs_alloc_ioend_bio( > struct buffer_head *bh) > { > - int nvecs = bio_get_nr_vecs(bh->b_bdev); > - struct bio *bio = bio_alloc(GFP_NOIO, nvecs); > + struct bio *bio = bio_alloc(GFP_NOIO, BIO_MAX_PAGES); > > ASSERT(bio->bi_private == NULL); > bio->bi_iter.bi_sector = bh->b_blocknr * (bh->b_size >> 9); > diff --git a/include/linux/bio.h b/include/linux/bio.h > index da3a127..015422a 100644 > --- a/include/linux/bio.h > +++ b/include/linux/bio.h > @@ -427,7 +427,6 @@ void bio_chain(struct bio *, struct bio *); > extern int bio_add_page(struct bio *, struct page *, unsigned int,unsigned int); > extern int bio_add_pc_page(struct request_queue *, struct bio *, struct page *, > unsigned int, unsigned int); > -extern int bio_get_nr_vecs(struct block_device *); > struct rq_map_data; > extern struct bio *bio_map_user_iov(struct request_queue *, > const struct iov_iter *, gfp_t); > -- > 1.9.1 > -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/