misc bio allocation cleanups

misc bio allocation cleanups

[Christoph Hellwig]

Hi Jens,

this series contains various cleanups for how bios are allocated or
initialized plus related fallout.

Diffstat:
 Documentation/filesystems/f2fs.rst |    1 
 block/bio.c                        |  167 ++++++++++++++++++-------------------
 block/blk-crypto-fallback.c        |    2 
 block/blk-flush.c                  |   17 +--
 drivers/block/drbd/drbd_actlog.c   |    2 
 drivers/block/drbd/drbd_bitmap.c   |    2 
 drivers/block/drbd/drbd_int.h      |    2 
 drivers/block/drbd/drbd_main.c     |   13 --
 drivers/block/drbd/drbd_req.c      |    5 -
 drivers/block/drbd/drbd_req.h      |   12 --
 drivers/block/drbd/drbd_worker.c   |    5 -
 drivers/md/dm-clone-target.c       |   14 ---
 drivers/md/dm-zoned-metadata.c     |    6 -
 drivers/md/md.c                    |   48 +++-------
 drivers/md/md.h                    |    2 
 drivers/md/raid1.c                 |    2 
 drivers/md/raid10.c                |    2 
 drivers/md/raid5-ppl.c             |    2 
 drivers/md/raid5.c                 |  108 +++++++++--------------
 drivers/nvme/target/io-cmd-bdev.c  |    2 
 fs/block_dev.c                     |    2 
 fs/btrfs/volumes.c                 |    2 
 fs/exfat/file.c                    |    2 
 fs/ext4/fast_commit.c              |    4 
 fs/ext4/fsync.c                    |    2 
 fs/ext4/ialloc.c                   |    2 
 fs/ext4/super.c                    |    2 
 fs/f2fs/data.c                     |   28 ------
 fs/f2fs/f2fs.h                     |    2 
 fs/f2fs/segment.c                  |   12 --
 fs/f2fs/super.c                    |    1 
 fs/fat/file.c                      |    2 
 fs/hfsplus/inode.c                 |    2 
 fs/hfsplus/super.c                 |    2 
 fs/jbd2/checkpoint.c               |    2 
 fs/jbd2/commit.c                   |    4 
 fs/jbd2/recovery.c                 |    2 
 fs/libfs.c                         |    2 
 fs/nfs/blocklayout/blocklayout.c   |    5 -
 fs/nilfs2/segbuf.c                 |    4 
 fs/nilfs2/the_nilfs.h              |    2 
 fs/ocfs2/file.c                    |    2 
 fs/reiserfs/file.c                 |    2 
 fs/xfs/xfs_super.c                 |    2 
 fs/zonefs/super.c                  |    4 
 include/linux/bio.h                |    6 -
 include/linux/blkdev.h             |    4 
 include/linux/swap.h               |    1 
 mm/page_io.c                       |   45 ++-------
 mm/swapfile.c                      |   10 --
 50 files changed, 213 insertions(+), 363 deletions(-)

[PATCH 01/17] zonefs: use bio_alloc in zonefs_file_dio_append

[Christoph Hellwig]

Use bio_alloc instead of open coding it.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 fs/zonefs/super.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/zonefs/super.c b/fs/zonefs/super.c
index bec47f2d074beb..faea2ed34b4a37 100644
--- a/fs/zonefs/super.c
+++ b/fs/zonefs/super.c
@@ -678,7 +678,7 @@ static ssize_t zonefs_file_dio_append(struct kiocb *iocb, struct iov_iter *from)
 	if (!nr_pages)
 		return 0;
 
-	bio = bio_alloc_bioset(GFP_NOFS, nr_pages, &fs_bio_set);
+	bio = bio_alloc(GFP_NOFS, nr_pages);
 	if (!bio)
 		return -ENOMEM;
 
-- 
2.29.2

[PATCH 02/17] btrfs: use bio_kmalloc in __alloc_device

[Christoph Hellwig]

Use bio_kmalloc instead of open coding it.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 fs/btrfs/volumes.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index 0a6de859eb2226..584ba093cf4966 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -421,7 +421,7 @@ static struct btrfs_device *__alloc_device(struct btrfs_fs_info *fs_info)
 	 * Preallocate a bio that's always going to be used for flushing device
 	 * barriers and matches the device lifespan
 	 */
-	dev->flush_bio = bio_alloc_bioset(GFP_KERNEL, 0, NULL);
+	dev->flush_bio = bio_kmalloc(GFP_KERNEL, 0);
 	if (!dev->flush_bio) {
 		kfree(dev);
 		return ERR_PTR(-ENOMEM);
-- 
2.29.2

[PATCH 03/17] blk-crypto: use bio_kmalloc in blk_crypto_clone_bio

[Christoph Hellwig]

Use bio_kmalloc instead of open coding it.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-crypto-fallback.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/block/blk-crypto-fallback.c b/block/blk-crypto-fallback.c
index 50c225398e4d60..e8327c50d7c9f4 100644
--- a/block/blk-crypto-fallback.c
+++ b/block/blk-crypto-fallback.c
@@ -164,7 +164,7 @@ static struct bio *blk_crypto_clone_bio(struct bio *bio_src)
 	struct bio_vec bv;
 	struct bio *bio;
 
-	bio = bio_alloc_bioset(GFP_NOIO, bio_segments(bio_src), NULL);
+	bio = bio_kmalloc(GFP_NOIO, bio_segments(bio_src));
 	if (!bio)
 		return NULL;
 	bio->bi_bdev		= bio_src->bi_bdev;
-- 
2.29.2

[PATCH 04/17] block: split bio_kmalloc from bio_alloc_bioset

[Christoph Hellwig]

bio_kmalloc shares almost no logic with the bio_set based fast path
in bio_alloc_bioset.  Split it into an entirely separate implementation.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/bio.c         | 167 ++++++++++++++++++++++----------------------
 include/linux/bio.h |   6 +-
 2 files changed, 86 insertions(+), 87 deletions(-)

diff --git a/block/bio.c b/block/bio.c
index dfd7740a32300a..d4375619348c52 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -396,123 +396,101 @@ static void punt_bios_to_rescuer(struct bio_set *bs)
  * @nr_iovecs:	number of iovecs to pre-allocate
  * @bs:		the bio_set to allocate from.
  *
- * Description:
- *   If @bs is NULL, uses kmalloc() to allocate the bio; else the allocation is
- *   backed by the @bs's mempool.
+ * Allocate a bio from the mempools in @bs.
  *
- *   When @bs is not NULL, if %__GFP_DIRECT_RECLAIM is set then bio_alloc will
- *   always be able to allocate a bio. This is due to the mempool guarantees.
- *   To make this work, callers must never allocate more than 1 bio at a time
- *   from this pool. Callers that need to allocate more than 1 bio must always
- *   submit the previously allocated bio for IO before attempting to allocate
- *   a new one. Failure to do so can cause deadlocks under memory pressure.
+ * If %__GFP_DIRECT_RECLAIM is set then bio_alloc will always be able to
+ * allocate a bio.  This is due to the mempool guarantees.  To make this work,
+ * callers must never allocate more than 1 bio at a time from the general pool.
+ * Callers that need to allocate more than 1 bio must always submit the
+ * previously allocated bio for IO before attempting to allocate a new one.
+ * Failure to do so can cause deadlocks under memory pressure.
  *
- *   Note that when running under submit_bio_noacct() (i.e. any block
- *   driver), bios are not submitted until after you return - see the code in
- *   submit_bio_noacct() that converts recursion into iteration, to prevent
- *   stack overflows.
+ * Note that when running under submit_bio_noacct() (i.e. any block driver),
+ * bios are not submitted until after you return - see the code in
+ * submit_bio_noacct() that converts recursion into iteration, to prevent
+ * stack overflows.
  *
- *   This would normally mean allocating multiple bios under
- *   submit_bio_noacct() would be susceptible to deadlocks, but we have
- *   deadlock avoidance code that resubmits any blocked bios from a rescuer
- *   thread.
+ * This would normally mean allocating multiple bios under submit_bio_noacct()
+ * would be susceptible to deadlocks, but we have
+ * deadlock avoidance code that resubmits any blocked bios from a rescuer
+ * thread.
  *
- *   However, we do not guarantee forward progress for allocations from other
- *   mempools. Doing multiple allocations from the same mempool under
- *   submit_bio_noacct() should be avoided - instead, use bio_set's front_pad
- *   for per bio allocations.
+ * However, we do not guarantee forward progress for allocations from other
+ * mempools. Doing multiple allocations from the same mempool under
+ * submit_bio_noacct() should be avoided - instead, use bio_set's front_pad
+ * for per bio allocations.
  *
- *   RETURNS:
- *   Pointer to new bio on success, NULL on failure.
+ * Returns: Pointer to new bio on success, NULL on failure.
  */
 struct bio *bio_alloc_bioset(gfp_t gfp_mask, unsigned int nr_iovecs,
 			     struct bio_set *bs)
 {
 	gfp_t saved_gfp = gfp_mask;
-	unsigned front_pad;
-	unsigned inline_vecs;
-	struct bio_vec *bvl = NULL;
 	struct bio *bio;
 	void *p;
 
-	if (!bs) {
-		if (nr_iovecs > UIO_MAXIOV)
-			return NULL;
-
-		p = kmalloc(struct_size(bio, bi_inline_vecs, nr_iovecs), gfp_mask);
-		front_pad = 0;
-		inline_vecs = nr_iovecs;
-	} else {
-		/* should not use nobvec bioset for nr_iovecs > 0 */
-		if (WARN_ON_ONCE(!mempool_initialized(&bs->bvec_pool) &&
-				 nr_iovecs > 0))
-			return NULL;
-		/*
-		 * submit_bio_noacct() converts recursion to iteration; this
-		 * means if we're running beneath it, any bios we allocate and
-		 * submit will not be submitted (and thus freed) until after we
-		 * return.
-		 *
-		 * This exposes us to a potential deadlock if we allocate
-		 * multiple bios from the same bio_set() while running
-		 * underneath submit_bio_noacct(). If we were to allocate
-		 * multiple bios (say a stacking block driver that was splitting
-		 * bios), we would deadlock if we exhausted the mempool's
-		 * reserve.
-		 *
-		 * We solve this, and guarantee forward progress, with a rescuer
-		 * workqueue per bio_set. If we go to allocate and there are
-		 * bios on current->bio_list, we first try the allocation
-		 * without __GFP_DIRECT_RECLAIM; if that fails, we punt those
-		 * bios we would be blocking to the rescuer workqueue before
-		 * we retry with the original gfp_flags.
-		 */
-
-		if (current->bio_list &&
-		    (!bio_list_empty(&current->bio_list[0]) ||
-		     !bio_list_empty(&current->bio_list[1])) &&
-		    bs->rescue_workqueue)
-			gfp_mask &= ~__GFP_DIRECT_RECLAIM;
+	/* should not use nobvec bioset for nr_iovecs > 0 */
+	if (WARN_ON_ONCE(!mempool_initialized(&bs->bvec_pool) && nr_iovecs > 0))
+		return NULL;
 
+	/*
+	 * submit_bio_noacct() converts recursion to iteration; this means if
+	 * we're running beneath it, any bios we allocate and submit will not be
+	 * submitted (and thus freed) until after we return.
+	 *
+	 * This exposes us to a potential deadlock if we allocate multiple bios
+	 * from the same bio_set() while running underneath submit_bio_noacct().
+	 * If we were to allocate multiple bios (say a stacking block driver
+	 * that was splitting bios), we would deadlock if we exhausted the
+	 * mempool's reserve.
+	 *
+	 * We solve this, and guarantee forward progress, with a rescuer
+	 * workqueue per bio_set. If we go to allocate and there are bios on
+	 * current->bio_list, we first try the allocation without
+	 * __GFP_DIRECT_RECLAIM; if that fails, we punt those bios we would be
+	 * blocking to the rescuer workqueue before we retry with the original
+	 * gfp_flags.
+	 */
+	if (current->bio_list &&
+	    (!bio_list_empty(&current->bio_list[0]) ||
+	     !bio_list_empty(&current->bio_list[1])) &&
+	    bs->rescue_workqueue)
+		gfp_mask &= ~__GFP_DIRECT_RECLAIM;
+
+	p = mempool_alloc(&bs->bio_pool, gfp_mask);
+	if (!p && gfp_mask != saved_gfp) {
+		punt_bios_to_rescuer(bs);
+		gfp_mask = saved_gfp;
 		p = mempool_alloc(&bs->bio_pool, gfp_mask);
-		if (!p && gfp_mask != saved_gfp) {
-			punt_bios_to_rescuer(bs);
-			gfp_mask = saved_gfp;
-			p = mempool_alloc(&bs->bio_pool, gfp_mask);
-		}
-
-		front_pad = bs->front_pad;
-		inline_vecs = BIO_INLINE_VECS;
 	}
-
 	if (unlikely(!p))
 		return NULL;
 
-	bio = p + front_pad;
-	bio_init(bio, NULL, 0);
-
-	if (nr_iovecs > inline_vecs) {
+	bio = p + bs->front_pad;
+	if (nr_iovecs > BIO_INLINE_VECS) {
 		unsigned long idx = 0;
+		struct bio_vec *bvl = NULL;
 
 		bvl = bvec_alloc(gfp_mask, nr_iovecs, &idx, &bs->bvec_pool);
 		if (!bvl && gfp_mask != saved_gfp) {
 			punt_bios_to_rescuer(bs);
 			gfp_mask = saved_gfp;
-			bvl = bvec_alloc(gfp_mask, nr_iovecs, &idx, &bs->bvec_pool);
+			bvl = bvec_alloc(gfp_mask, nr_iovecs, &idx,
+					 &bs->bvec_pool);
 		}
 
 		if (unlikely(!bvl))
 			goto err_free;
 
 		bio->bi_flags |= idx << BVEC_POOL_OFFSET;
-		bio->bi_max_vecs = bvec_nr_vecs(idx);
+		bio_init(bio, bvl, bvec_nr_vecs(idx));
 	} else if (nr_iovecs) {
-		bvl = bio->bi_inline_vecs;
-		bio->bi_max_vecs = inline_vecs;
+		bio_init(bio, bio->bi_inline_vecs, BIO_INLINE_VECS);
+	} else {
+		bio_init(bio, NULL, 0);
 	}
 
 	bio->bi_pool = bs;
-	bio->bi_io_vec = bvl;
 	return bio;
 
 err_free:
@@ -521,6 +499,31 @@ struct bio *bio_alloc_bioset(gfp_t gfp_mask, unsigned int nr_iovecs,
 }
 EXPORT_SYMBOL(bio_alloc_bioset);
 
+/**
+ * bio_kmalloc - kmalloc a bio for I/O
+ * @gfp_mask:   the GFP_* mask given to the slab allocator
+ * @nr_iovecs:	number of iovecs to pre-allocate
+ *
+ * Use kmalloc to allocate and initialize a bio.
+ *
+ * Returns: Pointer to new bio on success, NULL on failure.
+ */
+struct bio *bio_kmalloc(gfp_t gfp_mask, unsigned int nr_iovecs)
+{
+	struct bio *bio;
+
+	if (nr_iovecs > UIO_MAXIOV)
+		return NULL;
+
+	bio = kmalloc(struct_size(bio, bi_inline_vecs, nr_iovecs), gfp_mask);
+	if (unlikely(!bio))
+		return NULL;
+	bio_init(bio, nr_iovecs ? bio->bi_inline_vecs : NULL, nr_iovecs);
+	bio->bi_pool = NULL;
+	return bio;
+}
+EXPORT_SYMBOL(bio_kmalloc);
+
 void zero_fill_bio_iter(struct bio *bio, struct bvec_iter start)
 {
 	unsigned long flags;
diff --git a/include/linux/bio.h b/include/linux/bio.h
index 676870b2c88d80..c74857cf12528c 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -408,6 +408,7 @@ extern int biovec_init_pool(mempool_t *pool, int pool_entries);
 extern int bioset_init_from_src(struct bio_set *bs, struct bio_set *src);
 
 extern struct bio *bio_alloc_bioset(gfp_t, unsigned int, struct bio_set *);
+struct bio *bio_kmalloc(gfp_t gfp_mask, unsigned int nr_iovecs);
 extern void bio_put(struct bio *);
 
 extern void __bio_clone_fast(struct bio *, struct bio *);
@@ -420,11 +421,6 @@ static inline struct bio *bio_alloc(gfp_t gfp_mask, unsigned int nr_iovecs)
 	return bio_alloc_bioset(gfp_mask, nr_iovecs, &fs_bio_set);
 }
 
-static inline struct bio *bio_kmalloc(gfp_t gfp_mask, unsigned int nr_iovecs)
-{
-	return bio_alloc_bioset(gfp_mask, nr_iovecs, NULL);
-}
-
 extern blk_qc_t submit_bio(struct bio *);
 
 extern void bio_endio(struct bio *);
-- 
2.29.2

[PATCH 05/17] block: use an on-stack bio in blkdev_issue_flush

[Christoph Hellwig]

There is no point in allocating memory for a synchronous flush.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-flush.c                 | 17 ++++++-----------
 drivers/md/dm-zoned-metadata.c    |  6 +++---
 drivers/md/raid5-ppl.c            |  2 +-
 drivers/nvme/target/io-cmd-bdev.c |  2 +-
 fs/block_dev.c                    |  2 +-
 fs/exfat/file.c                   |  2 +-
 fs/ext4/fast_commit.c             |  4 ++--
 fs/ext4/fsync.c                   |  2 +-
 fs/ext4/ialloc.c                  |  2 +-
 fs/ext4/super.c                   |  2 +-
 fs/fat/file.c                     |  2 +-
 fs/hfsplus/inode.c                |  2 +-
 fs/hfsplus/super.c                |  2 +-
 fs/jbd2/checkpoint.c              |  2 +-
 fs/jbd2/commit.c                  |  4 ++--
 fs/jbd2/recovery.c                |  2 +-
 fs/libfs.c                        |  2 +-
 fs/nilfs2/the_nilfs.h             |  2 +-
 fs/ocfs2/file.c                   |  2 +-
 fs/reiserfs/file.c                |  2 +-
 fs/xfs/xfs_super.c                |  2 +-
 fs/zonefs/super.c                 |  2 +-
 include/linux/blkdev.h            |  4 ++--
 23 files changed, 33 insertions(+), 38 deletions(-)

diff --git a/block/blk-flush.c b/block/blk-flush.c
index 76c1624cb06c05..7942ca6ed3211c 100644
--- a/block/blk-flush.c
+++ b/block/blk-flush.c
@@ -432,23 +432,18 @@ void blk_insert_flush(struct request *rq)
 /**
  * blkdev_issue_flush - queue a flush
  * @bdev:	blockdev to issue flush for
- * @gfp_mask:	memory allocation flags (for bio_alloc)
  *
  * Description:
  *    Issue a flush for the block device in question.
  */
-int blkdev_issue_flush(struct block_device *bdev, gfp_t gfp_mask)
+int blkdev_issue_flush(struct block_device *bdev)
 {
-	struct bio *bio;
-	int ret = 0;
+	struct bio bio;
 
-	bio = bio_alloc(gfp_mask, 0);
-	bio_set_dev(bio, bdev);
-	bio->bi_opf = REQ_OP_WRITE | REQ_PREFLUSH;
-
-	ret = submit_bio_wait(bio);
-	bio_put(bio);
-	return ret;
+	bio_init(&bio, NULL, 0);
+	bio_set_dev(&bio, bdev);
+	bio.bi_opf = REQ_OP_WRITE | REQ_PREFLUSH;
+	return submit_bio_wait(&bio);
 }
 EXPORT_SYMBOL(blkdev_issue_flush);
 
diff --git a/drivers/md/dm-zoned-metadata.c b/drivers/md/dm-zoned-metadata.c
index b298fefb022eb9..039d17b2893849 100644
--- a/drivers/md/dm-zoned-metadata.c
+++ b/drivers/md/dm-zoned-metadata.c
@@ -819,7 +819,7 @@ static int dmz_write_sb(struct dmz_metadata *zmd, unsigned int set)
 	ret = dmz_rdwr_block(dev, REQ_OP_WRITE, zmd->sb[set].block,
 			     mblk->page);
 	if (ret == 0)
-		ret = blkdev_issue_flush(dev->bdev, GFP_NOIO);
+		ret = blkdev_issue_flush(dev->bdev);
 
 	return ret;
 }
@@ -862,7 +862,7 @@ static int dmz_write_dirty_mblocks(struct dmz_metadata *zmd,
 
 	/* Flush drive cache (this will also sync data) */
 	if (ret == 0)
-		ret = blkdev_issue_flush(dev->bdev, GFP_NOIO);
+		ret = blkdev_issue_flush(dev->bdev);
 
 	return ret;
 }
@@ -933,7 +933,7 @@ int dmz_flush_metadata(struct dmz_metadata *zmd)
 
 	/* If there are no dirty metadata blocks, just flush the device cache */
 	if (list_empty(&write_list)) {
-		ret = blkdev_issue_flush(dev->bdev, GFP_NOIO);
+		ret = blkdev_issue_flush(dev->bdev);
 		goto err;
 	}
 
diff --git a/drivers/md/raid5-ppl.c b/drivers/md/raid5-ppl.c
index d0f540296fe91e..e8c118e05dfd46 100644
--- a/drivers/md/raid5-ppl.c
+++ b/drivers/md/raid5-ppl.c
@@ -1037,7 +1037,7 @@ static int ppl_recover(struct ppl_log *log, struct ppl_header *pplhdr,
 	}
 
 	/* flush the disk cache after recovery if necessary */
-	ret = blkdev_issue_flush(rdev->bdev, GFP_KERNEL);
+	ret = blkdev_issue_flush(rdev->bdev);
 out:
 	__free_page(page);
 	return ret;
diff --git a/drivers/nvme/target/io-cmd-bdev.c b/drivers/nvme/target/io-cmd-bdev.c
index 125dde3f410ee7..bf6e0ac9ad2890 100644
--- a/drivers/nvme/target/io-cmd-bdev.c
+++ b/drivers/nvme/target/io-cmd-bdev.c
@@ -333,7 +333,7 @@ static void nvmet_bdev_execute_flush(struct nvmet_req *req)
 
 u16 nvmet_bdev_flush(struct nvmet_req *req)
 {
-	if (blkdev_issue_flush(req->ns->bdev, GFP_KERNEL))
+	if (blkdev_issue_flush(req->ns->bdev))
 		return NVME_SC_INTERNAL | NVME_SC_DNR;
 	return 0;
 }
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 6f5bd9950baf45..bc957ef7cebaae 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -681,7 +681,7 @@ int blkdev_fsync(struct file *filp, loff_t start, loff_t end, int datasync)
 	 * i_mutex and doing so causes performance issues with concurrent
 	 * O_SYNC writers to a block device.
 	 */
-	error = blkdev_issue_flush(bdev, GFP_KERNEL);
+	error = blkdev_issue_flush(bdev);
 	if (error == -EOPNOTSUPP)
 		error = 0;
 
diff --git a/fs/exfat/file.c b/fs/exfat/file.c
index a92478eabfa4e4..183ffdf4d43c5f 100644
--- a/fs/exfat/file.c
+++ b/fs/exfat/file.c
@@ -361,7 +361,7 @@ int exfat_file_fsync(struct file *filp, loff_t start, loff_t end, int datasync)
 	if (err)
 		return err;
 
-	return blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL);
+	return blkdev_issue_flush(inode->i_sb->s_bdev);
 }
 
 const struct file_operations exfat_file_operations = {
diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c
index 0a14a7c87bf82e..6e8208acfc62ac 100644
--- a/fs/ext4/fast_commit.c
+++ b/fs/ext4/fast_commit.c
@@ -1076,7 +1076,7 @@ static int ext4_fc_perform_commit(journal_t *journal)
 	 * flush before we start writing fast commit blocks.
 	 */
 	if (journal->j_fs_dev != journal->j_dev)
-		blkdev_issue_flush(journal->j_fs_dev, GFP_NOFS);
+		blkdev_issue_flush(journal->j_fs_dev);
 
 	blk_start_plug(&plug);
 	if (sbi->s_fc_bytes == 0) {
@@ -1535,7 +1535,7 @@ static int ext4_fc_replay_inode(struct super_block *sb, struct ext4_fc_tl *tl)
 out:
 	iput(inode);
 	if (!ret)
-		blkdev_issue_flush(sb->s_bdev, GFP_KERNEL);
+		blkdev_issue_flush(sb->s_bdev);
 
 	return 0;
 }
diff --git a/fs/ext4/fsync.c b/fs/ext4/fsync.c
index 113bfb023a4a09..027a7d7037a07b 100644
--- a/fs/ext4/fsync.c
+++ b/fs/ext4/fsync.c
@@ -174,7 +174,7 @@ int ext4_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 		ret = ext4_fsync_journal(inode, datasync, &needs_barrier);
 
 	if (needs_barrier) {
-		err = blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL);
+		err = blkdev_issue_flush(inode->i_sb->s_bdev);
 		if (!ret)
 			ret = err;
 	}
diff --git a/fs/ext4/ialloc.c b/fs/ext4/ialloc.c
index b215c564bc318a..20f2fcb799f55f 100644
--- a/fs/ext4/ialloc.c
+++ b/fs/ext4/ialloc.c
@@ -1583,7 +1583,7 @@ int ext4_init_inode_table(struct super_block *sb, ext4_group_t group,
 	if (ret < 0)
 		goto err_out;
 	if (barrier)
-		blkdev_issue_flush(sb->s_bdev, GFP_NOFS);
+		blkdev_issue_flush(sb->s_bdev);
 
 skip_zeroout:
 	ext4_lock_group(sb, group);
diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index 9a6f9875aa3499..fb5985102c1db7 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -5709,7 +5709,7 @@ static int ext4_sync_fs(struct super_block *sb, int wait)
 		needs_barrier = true;
 	if (needs_barrier) {
 		int err;
-		err = blkdev_issue_flush(sb->s_bdev, GFP_KERNEL);
+		err = blkdev_issue_flush(sb->s_bdev);
 		if (!ret)
 			ret = err;
 	}
diff --git a/fs/fat/file.c b/fs/fat/file.c
index f9ee27cf4d7c2f..5fee74f1ad611d 100644
--- a/fs/fat/file.c
+++ b/fs/fat/file.c
@@ -195,7 +195,7 @@ int fat_file_fsync(struct file *filp, loff_t start, loff_t end, int datasync)
 	if (err)
 		return err;
 
-	return blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL);
+	return blkdev_issue_flush(inode->i_sb->s_bdev);
 }
 
 
diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c
index e3da9e96b83578..ca464328b79cc5 100644
--- a/fs/hfsplus/inode.c
+++ b/fs/hfsplus/inode.c
@@ -340,7 +340,7 @@ int hfsplus_file_fsync(struct file *file, loff_t start, loff_t end,
 	}
 
 	if (!test_bit(HFSPLUS_SB_NOBARRIER, &sbi->flags))
-		blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL);
+		blkdev_issue_flush(inode->i_sb->s_bdev);
 
 	inode_unlock(inode);
 
diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c
index 807119ae5adf73..b9e3db3f855f96 100644
--- a/fs/hfsplus/super.c
+++ b/fs/hfsplus/super.c
@@ -239,7 +239,7 @@ static int hfsplus_sync_fs(struct super_block *sb, int wait)
 	mutex_unlock(&sbi->vh_mutex);
 
 	if (!test_bit(HFSPLUS_SB_NOBARRIER, &sbi->flags))
-		blkdev_issue_flush(sb->s_bdev, GFP_KERNEL);
+		blkdev_issue_flush(sb->s_bdev);
 
 	return error;
 }
diff --git a/fs/jbd2/checkpoint.c b/fs/jbd2/checkpoint.c
index 472932b9e6bca9..63b526d44886d0 100644
--- a/fs/jbd2/checkpoint.c
+++ b/fs/jbd2/checkpoint.c
@@ -416,7 +416,7 @@ int jbd2_cleanup_journal_tail(journal_t *journal)
 	 * jbd2_cleanup_journal_tail() doesn't get called all that often.
 	 */
 	if (journal->j_flags & JBD2_BARRIER)
-		blkdev_issue_flush(journal->j_fs_dev, GFP_NOFS);
+		blkdev_issue_flush(journal->j_fs_dev);
 
 	return __jbd2_update_log_tail(journal, first_tid, blocknr);
 }
diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c
index b121d7d434c675..3cc4ab2ba7f4f2 100644
--- a/fs/jbd2/commit.c
+++ b/fs/jbd2/commit.c
@@ -825,7 +825,7 @@ void jbd2_journal_commit_transaction(journal_t *journal)
 	if (commit_transaction->t_need_data_flush &&
 	    (journal->j_fs_dev != journal->j_dev) &&
 	    (journal->j_flags & JBD2_BARRIER))
-		blkdev_issue_flush(journal->j_fs_dev, GFP_NOFS);
+		blkdev_issue_flush(journal->j_fs_dev);
 
 	/* Done it all: now write the commit record asynchronously. */
 	if (jbd2_has_feature_async_commit(journal)) {
@@ -932,7 +932,7 @@ void jbd2_journal_commit_transaction(journal_t *journal)
 	stats.run.rs_blocks_logged++;
 	if (jbd2_has_feature_async_commit(journal) &&
 	    journal->j_flags & JBD2_BARRIER) {
-		blkdev_issue_flush(journal->j_dev, GFP_NOFS);
+		blkdev_issue_flush(journal->j_dev);
 	}
 
 	if (err)
diff --git a/fs/jbd2/recovery.c b/fs/jbd2/recovery.c
index dc0694fcfcd123..69f18fe2092369 100644
--- a/fs/jbd2/recovery.c
+++ b/fs/jbd2/recovery.c
@@ -326,7 +326,7 @@ int jbd2_journal_recover(journal_t *journal)
 		err = err2;
 	/* Make sure all replayed data is on permanent storage */
 	if (journal->j_flags & JBD2_BARRIER) {
-		err2 = blkdev_issue_flush(journal->j_fs_dev, GFP_KERNEL);
+		err2 = blkdev_issue_flush(journal->j_fs_dev);
 		if (!err)
 			err = err2;
 	}
diff --git a/fs/libfs.c b/fs/libfs.c
index d1c3bade9f30dd..8398a0efb401c6 100644
--- a/fs/libfs.c
+++ b/fs/libfs.c
@@ -1117,7 +1117,7 @@ int generic_file_fsync(struct file *file, loff_t start, loff_t end,
 	err = __generic_file_fsync(file, start, end, datasync);
 	if (err)
 		return err;
-	return blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL);
+	return blkdev_issue_flush(inode->i_sb->s_bdev);
 }
 EXPORT_SYMBOL(generic_file_fsync);
 
diff --git a/fs/nilfs2/the_nilfs.h b/fs/nilfs2/the_nilfs.h
index b55cdeb4d16991..987c8ab02aeee6 100644
--- a/fs/nilfs2/the_nilfs.h
+++ b/fs/nilfs2/the_nilfs.h
@@ -375,7 +375,7 @@ static inline int nilfs_flush_device(struct the_nilfs *nilfs)
 	 */
 	smp_wmb();
 
-	err = blkdev_issue_flush(nilfs->ns_bdev, GFP_KERNEL);
+	err = blkdev_issue_flush(nilfs->ns_bdev);
 	if (err != -EIO)
 		err = 0;
 	return err;
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
index 85979e2214b39d..df6d709d2ae3b6 100644
--- a/fs/ocfs2/file.c
+++ b/fs/ocfs2/file.c
@@ -194,7 +194,7 @@ static int ocfs2_sync_file(struct file *file, loff_t start, loff_t end,
 		needs_barrier = true;
 	err = jbd2_complete_transaction(journal, commit_tid);
 	if (needs_barrier) {
-		ret = blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL);
+		ret = blkdev_issue_flush(inode->i_sb->s_bdev);
 		if (!err)
 			err = ret;
 	}
diff --git a/fs/reiserfs/file.c b/fs/reiserfs/file.c
index 0b641ae694f123..1db0254bc38b26 100644
--- a/fs/reiserfs/file.c
+++ b/fs/reiserfs/file.c
@@ -159,7 +159,7 @@ static int reiserfs_sync_file(struct file *filp, loff_t start, loff_t end,
 	barrier_done = reiserfs_commit_for_inode(inode);
 	reiserfs_write_unlock(inode->i_sb);
 	if (barrier_done != 1 && reiserfs_barrier_flush(inode->i_sb))
-		blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL);
+		blkdev_issue_flush(inode->i_sb->s_bdev);
 	inode_unlock(inode);
 	if (barrier_done < 0)
 		return barrier_done;
diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c
index 813be879a5e51d..c3e32789829f90 100644
--- a/fs/xfs/xfs_super.c
+++ b/fs/xfs/xfs_super.c
@@ -342,7 +342,7 @@ void
 xfs_blkdev_issue_flush(
 	xfs_buftarg_t		*buftarg)
 {
-	blkdev_issue_flush(buftarg->bt_bdev, GFP_NOFS);
+	blkdev_issue_flush(buftarg->bt_bdev);
 }
 
 STATIC void
diff --git a/fs/zonefs/super.c b/fs/zonefs/super.c
index faea2ed34b4a37..ab68e27bb322e0 100644
--- a/fs/zonefs/super.c
+++ b/fs/zonefs/super.c
@@ -541,7 +541,7 @@ static int zonefs_file_fsync(struct file *file, loff_t start, loff_t end,
 	if (ZONEFS_I(inode)->i_ztype == ZONEFS_ZTYPE_CNV)
 		ret = file_write_and_wait_range(file, start, end);
 	if (!ret)
-		ret = blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL);
+		ret = blkdev_issue_flush(inode->i_sb->s_bdev);
 
 	if (ret)
 		zonefs_io_error(inode, true);
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 20f3706b6b2e68..09f7ac5507d146 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -1288,7 +1288,7 @@ static inline bool blk_needs_flush_plug(struct task_struct *tsk)
 		 !list_empty(&plug->cb_list));
 }
 
-int blkdev_issue_flush(struct block_device *, gfp_t);
+int blkdev_issue_flush(struct block_device *bdev);
 long nr_blockdev_pages(void);
 #else /* CONFIG_BLOCK */
 struct blk_plug {
@@ -1316,7 +1316,7 @@ static inline bool blk_needs_flush_plug(struct task_struct *tsk)
 	return false;
 }
 
-static inline int blkdev_issue_flush(struct block_device *bdev, gfp_t gfp_mask)
+static inline int blkdev_issue_flush(struct block_device *bdev)
 {
 	return 0;
 }
-- 
2.29.2

[PATCH 06/17] dm-clone: use blkdev_issue_flush in commit_metadata

[Christoph Hellwig]

Use blkdev_issue_flush instead of open coding it.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/md/dm-clone-target.c | 14 +-------------
 1 file changed, 1 insertion(+), 13 deletions(-)

diff --git a/drivers/md/dm-clone-target.c b/drivers/md/dm-clone-target.c
index bdb255edc20043..a90bdf9b2ca6bd 100644
--- a/drivers/md/dm-clone-target.c
+++ b/drivers/md/dm-clone-target.c
@@ -85,12 +85,6 @@ struct clone {
 
 	struct dm_clone_metadata *cmd;
 
-	/*
-	 * bio used to flush the destination device, before committing the
-	 * metadata.
-	 */
-	struct bio flush_bio;
-
 	/* Region hydration hash table */
 	struct hash_table_bucket *ht;
 
@@ -1155,11 +1149,7 @@ static int commit_metadata(struct clone *clone, bool *dest_dev_flushed)
 		goto out;
 	}
 
-	bio_reset(&clone->flush_bio);
-	bio_set_dev(&clone->flush_bio, clone->dest_dev->bdev);
-	clone->flush_bio.bi_opf = REQ_OP_WRITE | REQ_PREFLUSH;
-
-	r = submit_bio_wait(&clone->flush_bio);
+	r = blkdev_issue_flush(clone->dest_dev->bdev);
 	if (unlikely(r)) {
 		__metadata_operation_failed(clone, "flush destination device", r);
 		goto out;
@@ -1886,7 +1876,6 @@ static int clone_ctr(struct dm_target *ti, unsigned int argc, char **argv)
 	bio_list_init(&clone->deferred_flush_completions);
 	clone->hydration_offset = 0;
 	atomic_set(&clone->hydrations_in_flight, 0);
-	bio_init(&clone->flush_bio, NULL, 0);
 
 	clone->wq = alloc_workqueue("dm-" DM_MSG_PREFIX, WQ_MEM_RECLAIM, 0);
 	if (!clone->wq) {
@@ -1958,7 +1947,6 @@ static void clone_dtr(struct dm_target *ti)
 	struct clone *clone = ti->private;
 
 	mutex_destroy(&clone->commit_lock);
-	bio_uninit(&clone->flush_bio);
 
 	for (i = 0; i < clone->nr_ctr_args; i++)
 		kfree(clone->ctr_args[i]);
-- 
2.29.2

[PATCH 07/17] f2fs: use blkdev_issue_flush in __submit_flush_wait

[Christoph Hellwig]

Use the blkdev_issue_flush helper instead of duplicating it.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 fs/f2fs/data.c    |  3 ++-
 fs/f2fs/f2fs.h    |  1 -
 fs/f2fs/segment.c | 12 +-----------
 3 files changed, 3 insertions(+), 13 deletions(-)

diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
index 8cbf0315975228..0cf0c605992431 100644
--- a/fs/f2fs/data.c
+++ b/fs/f2fs/data.c
@@ -56,7 +56,8 @@ static inline struct bio *__f2fs_bio_alloc(gfp_t gfp_mask,
 	return bio_alloc_bioset(gfp_mask, nr_iovecs, &f2fs_bioset);
 }
 
-struct bio *f2fs_bio_alloc(struct f2fs_sb_info *sbi, int npages, bool noio)
+static struct bio *f2fs_bio_alloc(struct f2fs_sb_info *sbi, int npages,
+		bool noio)
 {
 	if (noio) {
 		/* No failure on bio allocation */
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index bb11759191dcc9..902bd3267c03e1 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -3424,7 +3424,6 @@ void f2fs_destroy_checkpoint_caches(void);
  */
 int __init f2fs_init_bioset(void);
 void f2fs_destroy_bioset(void);
-struct bio *f2fs_bio_alloc(struct f2fs_sb_info *sbi, int npages, bool noio);
 int f2fs_init_bio_entry_cache(void);
 void f2fs_destroy_bio_entry_cache(void);
 void f2fs_submit_bio(struct f2fs_sb_info *sbi,
diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index deca74cb17dfd8..c495f170ee400b 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -566,17 +566,7 @@ void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg)
 static int __submit_flush_wait(struct f2fs_sb_info *sbi,
 				struct block_device *bdev)
 {
-	struct bio *bio;
-	int ret;
-
-	bio = f2fs_bio_alloc(sbi, 0, false);
-	if (!bio)
-		return -ENOMEM;
-
-	bio->bi_opf = REQ_OP_WRITE | REQ_SYNC | REQ_PREFLUSH;
-	bio_set_dev(bio, bdev);
-	ret = submit_bio_wait(bio);
-	bio_put(bio);
+	int ret = blkdev_issue_flush(bdev);
 
 	trace_f2fs_issue_flush(bdev, test_opt(sbi, NOBARRIER),
 				test_opt(sbi, FLUSH_MERGE), ret);
-- 
2.29.2

[PATCH 08/17] f2fs: remove FAULT_ALLOC_BIO

[Christoph Hellwig]

Sleeping bio allocations do not fail, which means that injecting an error
into sleeping bio allocations is a little silly.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 Documentation/filesystems/f2fs.rst |  1 -
 fs/f2fs/data.c                     | 29 ++++-------------------------
 fs/f2fs/f2fs.h                     |  1 -
 fs/f2fs/super.c                    |  1 -
 4 files changed, 4 insertions(+), 28 deletions(-)

diff --git a/Documentation/filesystems/f2fs.rst b/Documentation/filesystems/f2fs.rst
index dae15c96e659e2..624f5f3ed93e86 100644
--- a/Documentation/filesystems/f2fs.rst
+++ b/Documentation/filesystems/f2fs.rst
@@ -179,7 +179,6 @@ fault_type=%d		 Support configuring fault injection type, should be
 			 FAULT_KVMALLOC		  0x000000002
 			 FAULT_PAGE_ALLOC	  0x000000004
 			 FAULT_PAGE_GET		  0x000000008
-			 FAULT_ALLOC_BIO	  0x000000010
 			 FAULT_ALLOC_NID	  0x000000020
 			 FAULT_ORPHAN		  0x000000040
 			 FAULT_BLOCK		  0x000000080
diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
index 0cf0c605992431..9fb6be65592b1f 100644
--- a/fs/f2fs/data.c
+++ b/fs/f2fs/data.c
@@ -50,28 +50,6 @@ void f2fs_destroy_bioset(void)
 	bioset_exit(&f2fs_bioset);
 }
 
-static inline struct bio *__f2fs_bio_alloc(gfp_t gfp_mask,
-						unsigned int nr_iovecs)
-{
-	return bio_alloc_bioset(gfp_mask, nr_iovecs, &f2fs_bioset);
-}
-
-static struct bio *f2fs_bio_alloc(struct f2fs_sb_info *sbi, int npages,
-		bool noio)
-{
-	if (noio) {
-		/* No failure on bio allocation */
-		return __f2fs_bio_alloc(GFP_NOIO, npages);
-	}
-
-	if (time_to_inject(sbi, FAULT_ALLOC_BIO)) {
-		f2fs_show_injection_info(sbi, FAULT_ALLOC_BIO);
-		return NULL;
-	}
-
-	return __f2fs_bio_alloc(GFP_KERNEL, npages);
-}
-
 static bool __is_cp_guaranteed(struct page *page)
 {
 	struct address_space *mapping = page->mapping;
@@ -433,7 +411,7 @@ static struct bio *__bio_alloc(struct f2fs_io_info *fio, int npages)
 	struct f2fs_sb_info *sbi = fio->sbi;
 	struct bio *bio;
 
-	bio = f2fs_bio_alloc(sbi, npages, true);
+	bio = bio_alloc_bioset(GFP_NOIO, npages, &f2fs_bioset);
 
 	f2fs_target_device(sbi, fio->new_blkaddr, bio);
 	if (is_read_io(fio->op)) {
@@ -1029,8 +1007,9 @@ static struct bio *f2fs_grab_read_bio(struct inode *inode, block_t blkaddr,
 	struct bio_post_read_ctx *ctx;
 	unsigned int post_read_steps = 0;
 
-	bio = f2fs_bio_alloc(sbi, min_t(int, nr_pages, BIO_MAX_PAGES),
-								for_write);
+	bio = bio_alloc_bioset(for_write ? GFP_NOIO : GFP_KERNEL,
+			       min_t(int, nr_pages, BIO_MAX_PAGES),
+			       &f2fs_bioset);
 	if (!bio)
 		return ERR_PTR(-ENOMEM);
 
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index 902bd3267c03e1..6c78365d80ceb5 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -43,7 +43,6 @@ enum {
 	FAULT_KVMALLOC,
 	FAULT_PAGE_ALLOC,
 	FAULT_PAGE_GET,
-	FAULT_ALLOC_BIO,
 	FAULT_ALLOC_NID,
 	FAULT_ORPHAN,
 	FAULT_BLOCK,
diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index b4a07fe62d1a58..3a312642907e86 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -45,7 +45,6 @@ const char *f2fs_fault_name[FAULT_MAX] = {
 	[FAULT_KVMALLOC]	= "kvmalloc",
 	[FAULT_PAGE_ALLOC]	= "page alloc",
 	[FAULT_PAGE_GET]	= "page get",
-	[FAULT_ALLOC_BIO]	= "alloc bio",
 	[FAULT_ALLOC_NID]	= "alloc nid",
 	[FAULT_ORPHAN]		= "orphan",
 	[FAULT_BLOCK]		= "no more block",
-- 
2.29.2

[PATCH 09/17] drbd: remove bio_alloc_drbd

[Christoph Hellwig]

Given that drbd_md_io_bio_set is initialized during module initialization
and the module fails to load if the initialization fails there is no need
to fall back to plain bio_alloc.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/drbd/drbd_actlog.c |  2 +-
 drivers/block/drbd/drbd_bitmap.c |  2 +-
 drivers/block/drbd/drbd_int.h    |  2 --
 drivers/block/drbd/drbd_main.c   | 13 -------------
 4 files changed, 2 insertions(+), 17 deletions(-)

diff --git a/drivers/block/drbd/drbd_actlog.c b/drivers/block/drbd/drbd_actlog.c
index 7227fc7ab8ed1e..72cf7603d51fc7 100644
--- a/drivers/block/drbd/drbd_actlog.c
+++ b/drivers/block/drbd/drbd_actlog.c
@@ -138,7 +138,7 @@ static int _drbd_md_sync_page_io(struct drbd_device *device,
 		op_flags |= REQ_FUA | REQ_PREFLUSH;
 	op_flags |= REQ_SYNC;
 
-	bio = bio_alloc_drbd(GFP_NOIO);
+	bio = bio_alloc_bioset(GFP_NOIO, 1, &drbd_md_io_bio_set);
 	bio_set_dev(bio, bdev->md_bdev);
 	bio->bi_iter.bi_sector = sector;
 	err = -EIO;
diff --git a/drivers/block/drbd/drbd_bitmap.c b/drivers/block/drbd/drbd_bitmap.c
index df53dca5d02c7e..c1f816f896a89a 100644
--- a/drivers/block/drbd/drbd_bitmap.c
+++ b/drivers/block/drbd/drbd_bitmap.c
@@ -976,7 +976,7 @@ static void drbd_bm_endio(struct bio *bio)
 
 static void bm_page_io_async(struct drbd_bm_aio_ctx *ctx, int page_nr) __must_hold(local)
 {
-	struct bio *bio = bio_alloc_drbd(GFP_NOIO);
+	struct bio *bio = bio_alloc_bioset(GFP_NOIO, 1, &drbd_md_io_bio_set);
 	struct drbd_device *device = ctx->device;
 	struct drbd_bitmap *b = device->bitmap;
 	struct page *page;
diff --git a/drivers/block/drbd/drbd_int.h b/drivers/block/drbd/drbd_int.h
index b2c93a29c251fd..02db50d7e4c668 100644
--- a/drivers/block/drbd/drbd_int.h
+++ b/drivers/block/drbd/drbd_int.h
@@ -1422,8 +1422,6 @@ extern mempool_t drbd_md_io_page_pool;
 /* We also need to make sure we get a bio
  * when we need it for housekeeping purposes */
 extern struct bio_set drbd_md_io_bio_set;
-/* to allocate from that set */
-extern struct bio *bio_alloc_drbd(gfp_t gfp_mask);
 
 /* And a bio_set for cloning */
 extern struct bio_set drbd_io_bio_set;
diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
index 1c8c18b2a25f33..788dd97e6026b8 100644
--- a/drivers/block/drbd/drbd_main.c
+++ b/drivers/block/drbd/drbd_main.c
@@ -138,19 +138,6 @@ static const struct block_device_operations drbd_ops = {
 	.release	= drbd_release,
 };
 
-struct bio *bio_alloc_drbd(gfp_t gfp_mask)
-{
-	struct bio *bio;
-
-	if (!bioset_initialized(&drbd_md_io_bio_set))
-		return bio_alloc(gfp_mask, 1);
-
-	bio = bio_alloc_bioset(gfp_mask, 1, &drbd_md_io_bio_set);
-	if (!bio)
-		return NULL;
-	return bio;
-}
-
 #ifdef __CHECKER__
 /* When checking with sparse, and this is an inline function, sparse will
    give tons of false positives. When this is a real functions sparse works.
-- 
2.29.2

[PATCH 10/17] drbd: remove drbd_req_make_private_bio

[Christoph Hellwig]

Open code drbd_req_make_private_bio in the two callers to prepare
for further changes.  Also don't bother to initialize bi_next as the
bio code already does that that.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/drbd/drbd_req.c    |  5 ++++-
 drivers/block/drbd/drbd_req.h    | 12 ------------
 drivers/block/drbd/drbd_worker.c |  5 ++++-
 3 files changed, 8 insertions(+), 14 deletions(-)

diff --git a/drivers/block/drbd/drbd_req.c b/drivers/block/drbd/drbd_req.c
index ea0f31ab334361..9dbb660a7d7c8e 100644
--- a/drivers/block/drbd/drbd_req.c
+++ b/drivers/block/drbd/drbd_req.c
@@ -30,7 +30,10 @@ static struct drbd_request *drbd_req_new(struct drbd_device *device, struct bio
 		return NULL;
 	memset(req, 0, sizeof(*req));
 
-	drbd_req_make_private_bio(req, bio_src);
+	req->private_bio = bio_clone_fast(bio_src, GFP_NOIO, &drbd_io_bio_set);
+	req->private_bio->bi_private = req;
+	req->private_bio->bi_end_io = drbd_request_endio;
+
 	req->rq_state = (bio_data_dir(bio_src) == WRITE ? RQ_WRITE : 0)
 		      | (bio_op(bio_src) == REQ_OP_WRITE_SAME ? RQ_WSAME : 0)
 		      | (bio_op(bio_src) == REQ_OP_WRITE_ZEROES ? RQ_ZEROES : 0)
diff --git a/drivers/block/drbd/drbd_req.h b/drivers/block/drbd/drbd_req.h
index 55bb0f8721faa3..511f39a08de453 100644
--- a/drivers/block/drbd/drbd_req.h
+++ b/drivers/block/drbd/drbd_req.h
@@ -256,18 +256,6 @@ enum drbd_req_state_bits {
 #define MR_WRITE       1
 #define MR_READ        2
 
-static inline void drbd_req_make_private_bio(struct drbd_request *req, struct bio *bio_src)
-{
-	struct bio *bio;
-	bio = bio_clone_fast(bio_src, GFP_NOIO, &drbd_io_bio_set);
-
-	req->private_bio = bio;
-
-	bio->bi_private  = req;
-	bio->bi_end_io   = drbd_request_endio;
-	bio->bi_next     = NULL;
-}
-
 /* Short lived temporary struct on the stack.
  * We could squirrel the error to be returned into
  * bio->bi_iter.bi_size, or similar. But that would be too ugly. */
diff --git a/drivers/block/drbd/drbd_worker.c b/drivers/block/drbd/drbd_worker.c
index 02044ab7f767d5..64563bfdf0da02 100644
--- a/drivers/block/drbd/drbd_worker.c
+++ b/drivers/block/drbd/drbd_worker.c
@@ -1523,8 +1523,11 @@ int w_restart_disk_io(struct drbd_work *w, int cancel)
 	if (bio_data_dir(req->master_bio) == WRITE && req->rq_state & RQ_IN_ACT_LOG)
 		drbd_al_begin_io(device, &req->i);
 
-	drbd_req_make_private_bio(req, req->master_bio);
+	req->private_bio = bio_clone_fast(req->master_bio, GFP_NOIO,
+					  &drbd_io_bio_set);
 	bio_set_dev(req->private_bio, device->ldev->backing_bdev);
+	req->private_bio->bi_private = req;
+	req->private_bio->bi_end_io = drbd_request_endio;
 	submit_bio_noacct(req->private_bio);
 
 	return 0;
-- 
2.29.2

[PATCH 11/17] md: remove bio_alloc_mddev

[Christoph Hellwig]

bio_alloc_mddev is never called with a NULL mddev, and ->bio_set is
initialized in md_run, so it always must be initialized as well.  Just
open code the remaining call to bio_alloc_bioset.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/md/md.c     | 12 +-----------
 drivers/md/md.h     |  2 --
 drivers/md/raid1.c  |  2 +-
 drivers/md/raid10.c |  2 +-
 4 files changed, 3 insertions(+), 15 deletions(-)

diff --git a/drivers/md/md.c b/drivers/md/md.c
index 7d1bb24add3107..e2b9dbb6e888f6 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -340,16 +340,6 @@ static int start_readonly;
  */
 static bool create_on_open = true;
 
-struct bio *bio_alloc_mddev(gfp_t gfp_mask, int nr_iovecs,
-			    struct mddev *mddev)
-{
-	if (!mddev || !bioset_initialized(&mddev->bio_set))
-		return bio_alloc(gfp_mask, nr_iovecs);
-
-	return bio_alloc_bioset(gfp_mask, nr_iovecs, &mddev->bio_set);
-}
-EXPORT_SYMBOL_GPL(bio_alloc_mddev);
-
 static struct bio *md_bio_alloc_sync(struct mddev *mddev)
 {
 	if (!mddev || !bioset_initialized(&mddev->sync_set))
@@ -613,7 +603,7 @@ static void submit_flushes(struct work_struct *ws)
 			atomic_inc(&rdev->nr_pending);
 			atomic_inc(&rdev->nr_pending);
 			rcu_read_unlock();
-			bi = bio_alloc_mddev(GFP_NOIO, 0, mddev);
+			bi = bio_alloc_bioset(GFP_NOIO, 0, &mddev->bio_set);
 			bi->bi_end_io = md_end_flush;
 			bi->bi_private = rdev;
 			bio_set_dev(bi, rdev->bdev);
diff --git a/drivers/md/md.h b/drivers/md/md.h
index f13290ccc1c248..bcbba1b5ec4a71 100644
--- a/drivers/md/md.h
+++ b/drivers/md/md.h
@@ -742,8 +742,6 @@ extern void md_rdev_clear(struct md_rdev *rdev);
 extern void md_handle_request(struct mddev *mddev, struct bio *bio);
 extern void mddev_suspend(struct mddev *mddev);
 extern void mddev_resume(struct mddev *mddev);
-extern struct bio *bio_alloc_mddev(gfp_t gfp_mask, int nr_iovecs,
-				   struct mddev *mddev);
 
 extern void md_reload_sb(struct mddev *mddev, int raid_disk);
 extern void md_update_sb(struct mddev *mddev, int force);
diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c
index 3b19141cdb4bc2..d2378765dc154f 100644
--- a/drivers/md/raid1.c
+++ b/drivers/md/raid1.c
@@ -1104,7 +1104,7 @@ static void alloc_behind_master_bio(struct r1bio *r1_bio,
 	int i = 0;
 	struct bio *behind_bio = NULL;
 
-	behind_bio = bio_alloc_mddev(GFP_NOIO, vcnt, r1_bio->mddev);
+	behind_bio = bio_alloc_bioset(GFP_NOIO, vcnt, &r1_bio->mddev->bio_set);
 	if (!behind_bio)
 		return;
 
diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
index be8f14afb6d143..e1eefbec15d444 100644
--- a/drivers/md/raid10.c
+++ b/drivers/md/raid10.c
@@ -4531,7 +4531,7 @@ static sector_t reshape_request(struct mddev *mddev, sector_t sector_nr,
 		return sectors_done;
 	}
 
-	read_bio = bio_alloc_mddev(GFP_KERNEL, RESYNC_PAGES, mddev);
+	read_bio = bio_alloc_bioset(GFP_KERNEL, RESYNC_PAGES, &mddev->bio_set);
 
 	bio_set_dev(read_bio, rdev->bdev);
 	read_bio->bi_iter.bi_sector = (r10_bio->devs[r10_bio->read_slot].addr
-- 
2.29.2

[PATCH 12/17] md: simplify sync_page_io

[Christoph Hellwig]

Use an on-stack bio and biovec for the single page synchronous I/O.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/md/md.c | 26 +++++++++++++-------------
 1 file changed, 13 insertions(+), 13 deletions(-)

diff --git a/drivers/md/md.c b/drivers/md/md.c
index e2b9dbb6e888f6..6a27f52007c871 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -1021,29 +1021,29 @@ int md_super_wait(struct mddev *mddev)
 int sync_page_io(struct md_rdev *rdev, sector_t sector, int size,
 		 struct page *page, int op, int op_flags, bool metadata_op)
 {
-	struct bio *bio = md_bio_alloc_sync(rdev->mddev);
-	int ret;
+	struct bio bio;
+	struct bio_vec bvec;
+
+	bio_init(&bio, &bvec, 1);
 
 	if (metadata_op && rdev->meta_bdev)
-		bio_set_dev(bio, rdev->meta_bdev);
+		bio_set_dev(&bio, rdev->meta_bdev);
 	else
-		bio_set_dev(bio, rdev->bdev);
-	bio_set_op_attrs(bio, op, op_flags);
+		bio_set_dev(&bio, rdev->bdev);
+	bio.bi_opf = op | op_flags;
 	if (metadata_op)
-		bio->bi_iter.bi_sector = sector + rdev->sb_start;
+		bio.bi_iter.bi_sector = sector + rdev->sb_start;
 	else if (rdev->mddev->reshape_position != MaxSector &&
 		 (rdev->mddev->reshape_backwards ==
 		  (sector >= rdev->mddev->reshape_position)))
-		bio->bi_iter.bi_sector = sector + rdev->new_data_offset;
+		bio.bi_iter.bi_sector = sector + rdev->new_data_offset;
 	else
-		bio->bi_iter.bi_sector = sector + rdev->data_offset;
-	bio_add_page(bio, page, size, 0);
+		bio.bi_iter.bi_sector = sector + rdev->data_offset;
+	bio_add_page(&bio, page, size, 0);
 
-	submit_bio_wait(bio);
+	submit_bio_wait(&bio);
 
-	ret = !bio->bi_status;
-	bio_put(bio);
-	return ret;
+	return !bio.bi_status;
 }
 EXPORT_SYMBOL_GPL(sync_page_io);
 
-- 
2.29.2

[PATCH 13/17] md: remove md_bio_alloc_sync

[Christoph Hellwig]

md_bio_alloc_sync is never called with a NULL mddev, and ->sync_set is
initialized in md_run, so it always must be initialized as well.  Just
open code the remaining call to bio_alloc_bioset.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/md/md.c | 10 +---------
 1 file changed, 1 insertion(+), 9 deletions(-)

diff --git a/drivers/md/md.c b/drivers/md/md.c
index 6a27f52007c871..399c81bddc1ae1 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -340,14 +340,6 @@ static int start_readonly;
  */
 static bool create_on_open = true;
 
-static struct bio *md_bio_alloc_sync(struct mddev *mddev)
-{
-	if (!mddev || !bioset_initialized(&mddev->sync_set))
-		return bio_alloc(GFP_NOIO, 1);
-
-	return bio_alloc_bioset(GFP_NOIO, 1, &mddev->sync_set);
-}
-
 /*
  * We have a system wide 'event count' that is incremented
  * on any 'interesting' event, and readers of /proc/mdstat
@@ -989,7 +981,7 @@ void md_super_write(struct mddev *mddev, struct md_rdev *rdev,
 	if (test_bit(Faulty, &rdev->flags))
 		return;
 
-	bio = md_bio_alloc_sync(mddev);
+	bio = bio_alloc_bioset(GFP_NOIO, 1, &mddev->sync_set);
 
 	atomic_inc(&rdev->nr_pending);
 
-- 
2.29.2

[PATCH 14/17] md/raid6: refactor raid5_read_one_chunk

[Christoph Hellwig]

Refactor raid5_read_one_chunk so that all simple checks are done
before allocating the bio.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/md/raid5.c | 108 +++++++++++++++++++--------------------------
 1 file changed, 45 insertions(+), 63 deletions(-)

diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index f411b9e5c332f4..a348b2adf2a9f9 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -5393,90 +5393,72 @@ static void raid5_align_endio(struct bio *bi)
 static int raid5_read_one_chunk(struct mddev *mddev, struct bio *raid_bio)
 {
 	struct r5conf *conf = mddev->private;
-	int dd_idx;
-	struct bio* align_bi;
+	struct bio *align_bio;
 	struct md_rdev *rdev;
-	sector_t end_sector;
+	sector_t sector, end_sector, first_bad;
+	int bad_sectors, dd_idx;
 
 	if (!in_chunk_boundary(mddev, raid_bio)) {
 		pr_debug("%s: non aligned\n", __func__);
 		return 0;
 	}
-	/*
-	 * use bio_clone_fast to make a copy of the bio
-	 */
-	align_bi = bio_clone_fast(raid_bio, GFP_NOIO, &mddev->bio_set);
-	if (!align_bi)
-		return 0;
-	/*
-	 *   set bi_end_io to a new function, and set bi_private to the
-	 *     original bio.
-	 */
-	align_bi->bi_end_io  = raid5_align_endio;
-	align_bi->bi_private = raid_bio;
-	/*
-	 *	compute position
-	 */
-	align_bi->bi_iter.bi_sector =
-		raid5_compute_sector(conf, raid_bio->bi_iter.bi_sector,
-				     0, &dd_idx, NULL);
 
-	end_sector = bio_end_sector(align_bi);
+	sector = raid5_compute_sector(conf, raid_bio->bi_iter.bi_sector, 0,
+				      &dd_idx, NULL);
+	end_sector = bio_end_sector(raid_bio);
+
 	rcu_read_lock();
+	if (r5c_big_stripe_cached(conf, sector))
+		goto out_rcu_unlock;
+
 	rdev = rcu_dereference(conf->disks[dd_idx].replacement);
 	if (!rdev || test_bit(Faulty, &rdev->flags) ||
 	    rdev->recovery_offset < end_sector) {
 		rdev = rcu_dereference(conf->disks[dd_idx].rdev);
-		if (rdev &&
-		    (test_bit(Faulty, &rdev->flags) ||
+		if (!rdev)
+			goto out_rcu_unlock;
+		if (test_bit(Faulty, &rdev->flags) ||
 		    !(test_bit(In_sync, &rdev->flags) ||
-		      rdev->recovery_offset >= end_sector)))
-			rdev = NULL;
+		      rdev->recovery_offset >= end_sector))
+			goto out_rcu_unlock;
 	}
 
-	if (r5c_big_stripe_cached(conf, align_bi->bi_iter.bi_sector)) {
-		rcu_read_unlock();
-		bio_put(align_bi);
+	atomic_inc(&rdev->nr_pending);
+	rcu_read_unlock();
+
+	align_bio = bio_clone_fast(raid_bio, GFP_NOIO, &mddev->bio_set);
+	bio_set_dev(align_bio, rdev->bdev);
+	align_bio->bi_end_io = raid5_align_endio;
+	align_bio->bi_private = raid_bio;
+	align_bio->bi_iter.bi_sector = sector;
+
+	raid_bio->bi_next = (void *)rdev;
+
+	if (is_badblock(rdev, sector, bio_sectors(align_bio), &first_bad,
+			&bad_sectors)) {
+		bio_put(align_bio);
+		rdev_dec_pending(rdev, mddev);
 		return 0;
 	}
 
-	if (rdev) {
-		sector_t first_bad;
-		int bad_sectors;
-
-		atomic_inc(&rdev->nr_pending);
-		rcu_read_unlock();
-		raid_bio->bi_next = (void*)rdev;
-		bio_set_dev(align_bi, rdev->bdev);
-
-		if (is_badblock(rdev, align_bi->bi_iter.bi_sector,
-				bio_sectors(align_bi),
-				&first_bad, &bad_sectors)) {
-			bio_put(align_bi);
-			rdev_dec_pending(rdev, mddev);
-			return 0;
-		}
+	/* No reshape active, so we can trust rdev->data_offset */
+	align_bio->bi_iter.bi_sector += rdev->data_offset;
 
-		/* No reshape active, so we can trust rdev->data_offset */
-		align_bi->bi_iter.bi_sector += rdev->data_offset;
+	spin_lock_irq(&conf->device_lock);
+	wait_event_lock_irq(conf->wait_for_quiescent, conf->quiesce == 0,
+			    conf->device_lock);
+	atomic_inc(&conf->active_aligned_reads);
+	spin_unlock_irq(&conf->device_lock);
 
-		spin_lock_irq(&conf->device_lock);
-		wait_event_lock_irq(conf->wait_for_quiescent,
-				    conf->quiesce == 0,
-				    conf->device_lock);
-		atomic_inc(&conf->active_aligned_reads);
-		spin_unlock_irq(&conf->device_lock);
+	if (mddev->gendisk)
+		trace_block_bio_remap(align_bio, disk_devt(mddev->gendisk),
+				      raid_bio->bi_iter.bi_sector);
+	submit_bio_noacct(align_bio);
+	return 1;
 
-		if (mddev->gendisk)
-			trace_block_bio_remap(align_bi, disk_devt(mddev->gendisk),
-					      raid_bio->bi_iter.bi_sector);
-		submit_bio_noacct(align_bi);
-		return 1;
-	} else {
-		rcu_read_unlock();
-		bio_put(align_bi);
-		return 0;
-	}
+out_rcu_unlock:
+	rcu_read_unlock();
+	return 0;
 }
 
 static struct bio *chunk_aligned_read(struct mddev *mddev, struct bio *raid_bio)
-- 
2.29.2

[PATCH 15/17] nfs/blocklayout: remove cruft in bl_alloc_init_bio

[Christoph Hellwig]

bio_alloc never returns NULL when it can sleep.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 fs/nfs/blocklayout/blocklayout.c | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/fs/nfs/blocklayout/blocklayout.c b/fs/nfs/blocklayout/blocklayout.c
index 3be6836074ae92..1a96ce28efb026 100644
--- a/fs/nfs/blocklayout/blocklayout.c
+++ b/fs/nfs/blocklayout/blocklayout.c
@@ -123,11 +123,6 @@ bl_alloc_init_bio(int npg, struct block_device *bdev, sector_t disk_sector,
 
 	npg = min(npg, BIO_MAX_PAGES);
 	bio = bio_alloc(GFP_NOIO, npg);
-	if (!bio && (current->flags & PF_MEMALLOC)) {
-		while (!bio && (npg /= 2))
-			bio = bio_alloc(GFP_NOIO, npg);
-	}
-
 	if (bio) {
 		bio->bi_iter.bi_sector = disk_sector;
 		bio_set_dev(bio, bdev);
-- 
2.29.2

[PATCH 16/17] nilfs2: remove cruft in nilfs_alloc_seg_bio

[Christoph Hellwig]

bio_alloc never returns NULL when it can sleep.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 fs/nilfs2/segbuf.c | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/fs/nilfs2/segbuf.c b/fs/nilfs2/segbuf.c
index 1a8729eded8b14..1e75417bfe6e52 100644
--- a/fs/nilfs2/segbuf.c
+++ b/fs/nilfs2/segbuf.c
@@ -386,10 +386,6 @@ static struct bio *nilfs_alloc_seg_bio(struct the_nilfs *nilfs, sector_t start,
 	struct bio *bio;
 
 	bio = bio_alloc(GFP_NOIO, nr_vecs);
-	if (bio == NULL) {
-		while (!bio && (nr_vecs >>= 1))
-			bio = bio_alloc(GFP_NOIO, nr_vecs);
-	}
 	if (likely(bio)) {
 		bio_set_dev(bio, nilfs->ns_bdev);
 		bio->bi_iter.bi_sector =
-- 
2.29.2

[PATCH 17/17] mm: remove get_swap_bio

[Christoph Hellwig]

Just reuse the block_device and sector from the swap_info structure,
just as used by the SWP_SYNCHRONOUS path.  Also remove the checks for
NULL returns from bio_alloc as that can't happen for sleeping
allocations.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 include/linux/swap.h |  1 -
 mm/page_io.c         | 45 +++++++++++++-------------------------------
 mm/swapfile.c        | 10 ----------
 3 files changed, 13 insertions(+), 43 deletions(-)

diff --git a/include/linux/swap.h b/include/linux/swap.h
index 596bc2f4d9b03e..3f1f7ae0fbe9e1 100644
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -468,7 +468,6 @@ extern int free_swap_and_cache(swp_entry_t);
 int swap_type_of(dev_t device, sector_t offset);
 int find_first_swap(dev_t *device);
 extern unsigned int count_swap_pages(int, int);
-extern sector_t map_swap_page(struct page *, struct block_device **);
 extern sector_t swapdev_block(int, pgoff_t);
 extern int page_swapcount(struct page *);
 extern int __swap_count(swp_entry_t entry);
diff --git a/mm/page_io.c b/mm/page_io.c
index a75f35464a4e73..92f7941c6d018b 100644
--- a/mm/page_io.c
+++ b/mm/page_io.c
@@ -26,25 +26,6 @@
 #include <linux/uio.h>
 #include <linux/sched/task.h>
 
-static struct bio *get_swap_bio(gfp_t gfp_flags,
-				struct page *page, bio_end_io_t end_io)
-{
-	struct bio *bio;
-
-	bio = bio_alloc(gfp_flags, 1);
-	if (bio) {
-		struct block_device *bdev;
-
-		bio->bi_iter.bi_sector = map_swap_page(page, &bdev);
-		bio_set_dev(bio, bdev);
-		bio->bi_iter.bi_sector <<= PAGE_SHIFT - 9;
-		bio->bi_end_io = end_io;
-
-		bio_add_page(bio, page, thp_size(page), 0);
-	}
-	return bio;
-}
-
 void end_swap_bio_write(struct bio *bio)
 {
 	struct page *page = bio_first_page_all(bio);
@@ -361,13 +342,13 @@ int __swap_writepage(struct page *page, struct writeback_control *wbc,
 		return 0;
 	}
 
-	bio = get_swap_bio(GFP_NOIO, page, end_write_func);
-	if (bio == NULL) {
-		set_page_dirty(page);
-		unlock_page(page);
-		return -ENOMEM;
-	}
+	bio = bio_alloc(GFP_NOIO, 1);
+	bio_set_dev(bio, sis->bdev);
+	bio->bi_iter.bi_sector = swap_page_sector(page);
 	bio->bi_opf = REQ_OP_WRITE | REQ_SWAP | wbc_to_write_flags(wbc);
+	bio->bi_end_io = end_write_func;
+	bio_add_page(bio, page, thp_size(page), 0);
+
 	bio_associate_blkg_from_page(bio, page);
 	count_swpout_vm_event(page);
 	set_page_writeback(page);
@@ -427,18 +408,18 @@ int swap_readpage(struct page *page, bool synchronous)
 	}
 
 	ret = 0;
-	bio = get_swap_bio(GFP_KERNEL, page, end_swap_bio_read);
-	if (bio == NULL) {
-		unlock_page(page);
-		ret = -ENOMEM;
-		goto out;
-	}
+	bio = bio_alloc(GFP_KERNEL, 1);
+	bio_set_dev(bio, sis->bdev);
+	bio->bi_opf = REQ_OP_READ;
+	bio->bi_iter.bi_sector = swap_page_sector(page);
+	bio->bi_end_io = end_swap_bio_read;
+	bio_add_page(bio, page, thp_size(page), 0);
+
 	disk = bio->bi_bdev->bd_disk;
 	/*
 	 * Keep this task valid during swap readpage because the oom killer may
 	 * attempt to access it in the page fault retry time check.
 	 */
-	bio_set_op_attrs(bio, REQ_OP_READ, 0);
 	if (synchronous) {
 		bio->bi_opf |= REQ_HIPRI;
 		get_task_struct(current);
diff --git a/mm/swapfile.c b/mm/swapfile.c
index 9fffc5af29d1b9..bfa9e8b0c2ef61 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -2301,16 +2301,6 @@ static sector_t map_swap_entry(swp_entry_t entry, struct block_device **bdev)
 	return se->start_block + (offset - se->start_page);
 }
 
-/*
- * Returns the page offset into bdev for the specified page's swap entry.
- */
-sector_t map_swap_page(struct page *page, struct block_device **bdev)
-{
-	swp_entry_t entry;
-	entry.val = page_private(page);
-	return map_swap_entry(entry, bdev);
-}
-
 /*
  * Free all of a swapdev's extent information
  */
-- 
2.29.2

Re: [PATCH 02/17] btrfs: use bio_kmalloc in __alloc_device

[Josef Bacik]

On 1/26/21 9:52 AM, Christoph Hellwig wrote:
> Use bio_kmalloc instead of open coding it.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Reviewed-by: Josef Bacik <josef@toxicpanda.com>

yay I contributed,

Josef

Re: [PATCH 01/17] zonefs: use bio_alloc in zonefs_file_dio_append

[Johannes Thumshirn]

On 26/01/2021 16:01, Christoph Hellwig wrote:
> Use bio_alloc instead of open coding it.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  fs/zonefs/super.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/zonefs/super.c b/fs/zonefs/super.c
> index bec47f2d074beb..faea2ed34b4a37 100644
> --- a/fs/zonefs/super.c
> +++ b/fs/zonefs/super.c
> @@ -678,7 +678,7 @@ static ssize_t zonefs_file_dio_append(struct kiocb *iocb, struct iov_iter *from)
>  	if (!nr_pages)
>  		return 0;
>  
> -	bio = bio_alloc_bioset(GFP_NOFS, nr_pages, &fs_bio_set);
> +	bio = bio_alloc(GFP_NOFS, nr_pages);
>  	if (!bio)
>  		return -ENOMEM;
>  
> 

Whoopsie my bad,
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>

Re: [PATCH 01/17] zonefs: use bio_alloc in zonefs_file_dio_append

[Chaitanya Kulkarni]

On 1/26/21 7:01 AM, Christoph Hellwig wrote:
> Use bio_alloc instead of open coding it.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Ha Ha I was going send out a patch for this :P.

Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>

Re: [PATCH 02/17] btrfs: use bio_kmalloc in __alloc_device

[Chaitanya Kulkarni]

On 1/26/21 7:04 AM, Christoph Hellwig wrote:
> Use bio_kmalloc instead of open coding it.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>
Looks good.

Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>

Re: [PATCH 05/17] block: use an on-stack bio in blkdev_issue_flush

[Chaitanya Kulkarni]

On 1/26/21 7:07 AM, Christoph Hellwig wrote:
> There is no point in allocating memory for a synchronous flush.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>

True, looks good.

Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>

Re: [PATCH 16/17] nilfs2: remove cruft in nilfs_alloc_seg_bio

[Chaitanya Kulkarni]

On 1/26/21 7:27 AM, Christoph Hellwig wrote:
> bio_alloc never returns NULL when it can sleep.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
Looks good.

Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>

Re: [PATCH 03/17] blk-crypto: use bio_kmalloc in blk_crypto_clone_bio

[Eric Biggers]

On Tue, Jan 26, 2021 at 03:52:33PM +0100, Christoph Hellwig wrote:
> Use bio_kmalloc instead of open coding it.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/blk-crypto-fallback.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/block/blk-crypto-fallback.c b/block/blk-crypto-fallback.c
> index 50c225398e4d60..e8327c50d7c9f4 100644
> --- a/block/blk-crypto-fallback.c
> +++ b/block/blk-crypto-fallback.c
> @@ -164,7 +164,7 @@ static struct bio *blk_crypto_clone_bio(struct bio *bio_src)
>  	struct bio_vec bv;
>  	struct bio *bio;
>  
> -	bio = bio_alloc_bioset(GFP_NOIO, bio_segments(bio_src), NULL);
> +	bio = bio_kmalloc(GFP_NOIO, bio_segments(bio_src));
>  	if (!bio)
>  		return NULL;
>  	bio->bi_bdev		= bio_src->bi_bdev;
> -- 

Looks good,

Reviewed-by: Eric Biggers <ebiggers@google.com>

Re: [dm-devel] [PATCH 01/17] zonefs: use bio_alloc in zonefs_file_dio_append

[Damien Le Moal]

On 2021/01/26 23:58, Christoph Hellwig wrote:
> Use bio_alloc instead of open coding it.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  fs/zonefs/super.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/zonefs/super.c b/fs/zonefs/super.c
> index bec47f2d074beb..faea2ed34b4a37 100644
> --- a/fs/zonefs/super.c
> +++ b/fs/zonefs/super.c
> @@ -678,7 +678,7 @@ static ssize_t zonefs_file_dio_append(struct kiocb *iocb, struct iov_iter *from)
>  	if (!nr_pages)
>  		return 0;
>  
> -	bio = bio_alloc_bioset(GFP_NOFS, nr_pages, &fs_bio_set);
> +	bio = bio_alloc(GFP_NOFS, nr_pages);
>  	if (!bio)
>  		return -ENOMEM;
>  
> 

Acked-by: Damien Le Moal <damien.lemoal@wdc.com>

-- 
Damien Le Moal
Western Digital Research

Re: [PATCH 14/17] md/raid6: refactor raid5_read_one_chunk

[Song Liu]

On Tue, Jan 26, 2021 at 7:19 AM Christoph Hellwig <hch@lst.de> wrote:
>
> Refactor raid5_read_one_chunk so that all simple checks are done
> before allocating the bio.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Acked-by: Song Liu <song@kernel.org>

Thanks for the clean-up!


> ---
>  drivers/md/raid5.c | 108 +++++++++++++++++++--------------------------
>  1 file changed, 45 insertions(+), 63 deletions(-)
>
> diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
> index f411b9e5c332f4..a348b2adf2a9f9 100644
> --- a/drivers/md/raid5.c
> +++ b/drivers/md/raid5.c
> @@ -5393,90 +5393,72 @@ static void raid5_align_endio(struct bio *bi)
>  static int raid5_read_one_chunk(struct mddev *mddev, struct bio *raid_bio)
>  {
>         struct r5conf *conf = mddev->private;
> -       int dd_idx;
> -       struct bio* align_bi;
> +       struct bio *align_bio;
>         struct md_rdev *rdev;
> -       sector_t end_sector;
> +       sector_t sector, end_sector, first_bad;
> +       int bad_sectors, dd_idx;
>
>         if (!in_chunk_boundary(mddev, raid_bio)) {
>                 pr_debug("%s: non aligned\n", __func__);
>                 return 0;
>         }
> -       /*
> -        * use bio_clone_fast to make a copy of the bio
> -        */
> -       align_bi = bio_clone_fast(raid_bio, GFP_NOIO, &mddev->bio_set);
> -       if (!align_bi)
> -               return 0;
> -       /*
> -        *   set bi_end_io to a new function, and set bi_private to the
> -        *     original bio.
> -        */
> -       align_bi->bi_end_io  = raid5_align_endio;
> -       align_bi->bi_private = raid_bio;
> -       /*
> -        *      compute position
> -        */
> -       align_bi->bi_iter.bi_sector =
> -               raid5_compute_sector(conf, raid_bio->bi_iter.bi_sector,
> -                                    0, &dd_idx, NULL);
>
> -       end_sector = bio_end_sector(align_bi);
> +       sector = raid5_compute_sector(conf, raid_bio->bi_iter.bi_sector, 0,
> +                                     &dd_idx, NULL);
> +       end_sector = bio_end_sector(raid_bio);
> +
>         rcu_read_lock();
> +       if (r5c_big_stripe_cached(conf, sector))
> +               goto out_rcu_unlock;
> +
>         rdev = rcu_dereference(conf->disks[dd_idx].replacement);
>         if (!rdev || test_bit(Faulty, &rdev->flags) ||
>             rdev->recovery_offset < end_sector) {
>                 rdev = rcu_dereference(conf->disks[dd_idx].rdev);
> -               if (rdev &&
> -                   (test_bit(Faulty, &rdev->flags) ||
> +               if (!rdev)
> +                       goto out_rcu_unlock;
> +               if (test_bit(Faulty, &rdev->flags) ||
>                     !(test_bit(In_sync, &rdev->flags) ||
> -                     rdev->recovery_offset >= end_sector)))
> -                       rdev = NULL;
> +                     rdev->recovery_offset >= end_sector))
> +                       goto out_rcu_unlock;
>         }
>
> -       if (r5c_big_stripe_cached(conf, align_bi->bi_iter.bi_sector)) {
> -               rcu_read_unlock();
> -               bio_put(align_bi);
> +       atomic_inc(&rdev->nr_pending);
> +       rcu_read_unlock();
> +
> +       align_bio = bio_clone_fast(raid_bio, GFP_NOIO, &mddev->bio_set);
> +       bio_set_dev(align_bio, rdev->bdev);
> +       align_bio->bi_end_io = raid5_align_endio;
> +       align_bio->bi_private = raid_bio;
> +       align_bio->bi_iter.bi_sector = sector;
> +
> +       raid_bio->bi_next = (void *)rdev;
> +
> +       if (is_badblock(rdev, sector, bio_sectors(align_bio), &first_bad,
> +                       &bad_sectors)) {
> +               bio_put(align_bio);
> +               rdev_dec_pending(rdev, mddev);
>                 return 0;
>         }
>
> -       if (rdev) {
> -               sector_t first_bad;
> -               int bad_sectors;
> -
> -               atomic_inc(&rdev->nr_pending);
> -               rcu_read_unlock();
> -               raid_bio->bi_next = (void*)rdev;
> -               bio_set_dev(align_bi, rdev->bdev);
> -
> -               if (is_badblock(rdev, align_bi->bi_iter.bi_sector,
> -                               bio_sectors(align_bi),
> -                               &first_bad, &bad_sectors)) {
> -                       bio_put(align_bi);
> -                       rdev_dec_pending(rdev, mddev);
> -                       return 0;
> -               }
> +       /* No reshape active, so we can trust rdev->data_offset */
> +       align_bio->bi_iter.bi_sector += rdev->data_offset;
>
> -               /* No reshape active, so we can trust rdev->data_offset */
> -               align_bi->bi_iter.bi_sector += rdev->data_offset;
> +       spin_lock_irq(&conf->device_lock);
> +       wait_event_lock_irq(conf->wait_for_quiescent, conf->quiesce == 0,
> +                           conf->device_lock);
> +       atomic_inc(&conf->active_aligned_reads);
> +       spin_unlock_irq(&conf->device_lock);
>
> -               spin_lock_irq(&conf->device_lock);
> -               wait_event_lock_irq(conf->wait_for_quiescent,
> -                                   conf->quiesce == 0,
> -                                   conf->device_lock);
> -               atomic_inc(&conf->active_aligned_reads);
> -               spin_unlock_irq(&conf->device_lock);
> +       if (mddev->gendisk)
> +               trace_block_bio_remap(align_bio, disk_devt(mddev->gendisk),
> +                                     raid_bio->bi_iter.bi_sector);
> +       submit_bio_noacct(align_bio);
> +       return 1;
>
> -               if (mddev->gendisk)
> -                       trace_block_bio_remap(align_bi, disk_devt(mddev->gendisk),
> -                                             raid_bio->bi_iter.bi_sector);
> -               submit_bio_noacct(align_bi);
> -               return 1;
> -       } else {
> -               rcu_read_unlock();
> -               bio_put(align_bi);
> -               return 0;
> -       }
> +out_rcu_unlock:
> +       rcu_read_unlock();
> +       return 0;
>  }
>
>  static struct bio *chunk_aligned_read(struct mddev *mddev, struct bio *raid_bio)
> --
> 2.29.2
>

Re: [PATCH 11/17] md: remove bio_alloc_mddev

[Song Liu]

On Tue, Jan 26, 2021 at 7:12 AM Christoph Hellwig <hch@lst.de> wrote:
>
> bio_alloc_mddev is never called with a NULL mddev, and ->bio_set is
> initialized in md_run, so it always must be initialized as well.  Just
> open code the remaining call to bio_alloc_bioset.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Acked-by: Song Liu <song@kernel.org>

> ---
>  drivers/md/md.c     | 12 +-----------
>  drivers/md/md.h     |  2 --
>  drivers/md/raid1.c  |  2 +-
>  drivers/md/raid10.c |  2 +-
>  4 files changed, 3 insertions(+), 15 deletions(-)
>
> diff --git a/drivers/md/md.c b/drivers/md/md.c
> index 7d1bb24add3107..e2b9dbb6e888f6 100644
> --- a/drivers/md/md.c
> +++ b/drivers/md/md.c
> @@ -340,16 +340,6 @@ static int start_readonly;
>   */
>  static bool create_on_open = true;
>
> -struct bio *bio_alloc_mddev(gfp_t gfp_mask, int nr_iovecs,
> -                           struct mddev *mddev)
> -{
> -       if (!mddev || !bioset_initialized(&mddev->bio_set))
> -               return bio_alloc(gfp_mask, nr_iovecs);
> -
> -       return bio_alloc_bioset(gfp_mask, nr_iovecs, &mddev->bio_set);
> -}
> -EXPORT_SYMBOL_GPL(bio_alloc_mddev);
> -
>  static struct bio *md_bio_alloc_sync(struct mddev *mddev)
>  {
>         if (!mddev || !bioset_initialized(&mddev->sync_set))
> @@ -613,7 +603,7 @@ static void submit_flushes(struct work_struct *ws)
>                         atomic_inc(&rdev->nr_pending);
>                         atomic_inc(&rdev->nr_pending);
>                         rcu_read_unlock();
> -                       bi = bio_alloc_mddev(GFP_NOIO, 0, mddev);
> +                       bi = bio_alloc_bioset(GFP_NOIO, 0, &mddev->bio_set);
>                         bi->bi_end_io = md_end_flush;
>                         bi->bi_private = rdev;
>                         bio_set_dev(bi, rdev->bdev);
> diff --git a/drivers/md/md.h b/drivers/md/md.h
> index f13290ccc1c248..bcbba1b5ec4a71 100644
> --- a/drivers/md/md.h
> +++ b/drivers/md/md.h
> @@ -742,8 +742,6 @@ extern void md_rdev_clear(struct md_rdev *rdev);
>  extern void md_handle_request(struct mddev *mddev, struct bio *bio);
>  extern void mddev_suspend(struct mddev *mddev);
>  extern void mddev_resume(struct mddev *mddev);
> -extern struct bio *bio_alloc_mddev(gfp_t gfp_mask, int nr_iovecs,
> -                                  struct mddev *mddev);
>
>  extern void md_reload_sb(struct mddev *mddev, int raid_disk);
>  extern void md_update_sb(struct mddev *mddev, int force);
> diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c
> index 3b19141cdb4bc2..d2378765dc154f 100644
> --- a/drivers/md/raid1.c
> +++ b/drivers/md/raid1.c
> @@ -1104,7 +1104,7 @@ static void alloc_behind_master_bio(struct r1bio *r1_bio,
>         int i = 0;
>         struct bio *behind_bio = NULL;
>
> -       behind_bio = bio_alloc_mddev(GFP_NOIO, vcnt, r1_bio->mddev);
> +       behind_bio = bio_alloc_bioset(GFP_NOIO, vcnt, &r1_bio->mddev->bio_set);
>         if (!behind_bio)
>                 return;
>
> diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
> index be8f14afb6d143..e1eefbec15d444 100644
> --- a/drivers/md/raid10.c
> +++ b/drivers/md/raid10.c
> @@ -4531,7 +4531,7 @@ static sector_t reshape_request(struct mddev *mddev, sector_t sector_nr,
>                 return sectors_done;
>         }
>
> -       read_bio = bio_alloc_mddev(GFP_KERNEL, RESYNC_PAGES, mddev);
> +       read_bio = bio_alloc_bioset(GFP_KERNEL, RESYNC_PAGES, &mddev->bio_set);
>
>         bio_set_dev(read_bio, rdev->bdev);
>         read_bio->bi_iter.bi_sector = (r10_bio->devs[r10_bio->read_slot].addr
> --
> 2.29.2
>

Re: [PATCH 12/17] md: simplify sync_page_io

[Song Liu]

On Tue, Jan 26, 2021 at 7:14 AM Christoph Hellwig <hch@lst.de> wrote:
>
> Use an on-stack bio and biovec for the single page synchronous I/O.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Acked-by: Song Liu <song@kernel.org>

> ---
>  drivers/md/md.c | 26 +++++++++++++-------------
>  1 file changed, 13 insertions(+), 13 deletions(-)
>
> diff --git a/drivers/md/md.c b/drivers/md/md.c
> index e2b9dbb6e888f6..6a27f52007c871 100644
> --- a/drivers/md/md.c
> +++ b/drivers/md/md.c
> @@ -1021,29 +1021,29 @@ int md_super_wait(struct mddev *mddev)
>  int sync_page_io(struct md_rdev *rdev, sector_t sector, int size,
>                  struct page *page, int op, int op_flags, bool metadata_op)
>  {
> -       struct bio *bio = md_bio_alloc_sync(rdev->mddev);
> -       int ret;
> +       struct bio bio;
> +       struct bio_vec bvec;
> +
> +       bio_init(&bio, &bvec, 1);
>
>         if (metadata_op && rdev->meta_bdev)
> -               bio_set_dev(bio, rdev->meta_bdev);
> +               bio_set_dev(&bio, rdev->meta_bdev);
>         else
> -               bio_set_dev(bio, rdev->bdev);
> -       bio_set_op_attrs(bio, op, op_flags);
> +               bio_set_dev(&bio, rdev->bdev);
> +       bio.bi_opf = op | op_flags;
>         if (metadata_op)
> -               bio->bi_iter.bi_sector = sector + rdev->sb_start;
> +               bio.bi_iter.bi_sector = sector + rdev->sb_start;
>         else if (rdev->mddev->reshape_position != MaxSector &&
>                  (rdev->mddev->reshape_backwards ==
>                   (sector >= rdev->mddev->reshape_position)))
> -               bio->bi_iter.bi_sector = sector + rdev->new_data_offset;
> +               bio.bi_iter.bi_sector = sector + rdev->new_data_offset;
>         else
> -               bio->bi_iter.bi_sector = sector + rdev->data_offset;
> -       bio_add_page(bio, page, size, 0);
> +               bio.bi_iter.bi_sector = sector + rdev->data_offset;
> +       bio_add_page(&bio, page, size, 0);
>
> -       submit_bio_wait(bio);
> +       submit_bio_wait(&bio);
>
> -       ret = !bio->bi_status;
> -       bio_put(bio);
> -       return ret;
> +       return !bio.bi_status;
>  }
>  EXPORT_SYMBOL_GPL(sync_page_io);
>
> --
> 2.29.2
>

Re: [PATCH 13/17] md: remove md_bio_alloc_sync

[Song Liu]

On Tue, Jan 26, 2021 at 7:17 AM Christoph Hellwig <hch@lst.de> wrote:
>
> md_bio_alloc_sync is never called with a NULL mddev, and ->sync_set is
> initialized in md_run, so it always must be initialized as well.  Just
> open code the remaining call to bio_alloc_bioset.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Acked-by: Song Liu <song@kernel.org>

> ---
>  drivers/md/md.c | 10 +---------
>  1 file changed, 1 insertion(+), 9 deletions(-)
>
> diff --git a/drivers/md/md.c b/drivers/md/md.c
> index 6a27f52007c871..399c81bddc1ae1 100644
> --- a/drivers/md/md.c
> +++ b/drivers/md/md.c
> @@ -340,14 +340,6 @@ static int start_readonly;
>   */
>  static bool create_on_open = true;
>
> -static struct bio *md_bio_alloc_sync(struct mddev *mddev)
> -{
> -       if (!mddev || !bioset_initialized(&mddev->sync_set))
> -               return bio_alloc(GFP_NOIO, 1);
> -
> -       return bio_alloc_bioset(GFP_NOIO, 1, &mddev->sync_set);
> -}
> -
>  /*
>   * We have a system wide 'event count' that is incremented
>   * on any 'interesting' event, and readers of /proc/mdstat
> @@ -989,7 +981,7 @@ void md_super_write(struct mddev *mddev, struct md_rdev *rdev,
>         if (test_bit(Faulty, &rdev->flags))
>                 return;
>
> -       bio = md_bio_alloc_sync(mddev);
> +       bio = bio_alloc_bioset(GFP_NOIO, 1, &mddev->sync_set);
>
>         atomic_inc(&rdev->nr_pending);
>
> --
> 2.29.2
>

Re: misc bio allocation cleanups

[Jens Axboe]

On 1/26/21 7:52 AM, Christoph Hellwig wrote:
> Hi Jens,
> 
> this series contains various cleanups for how bios are allocated or
> initialized plus related fallout.

Applied, thanks.

-- 
Jens Axboe

Re: [PATCH 08/17] f2fs: remove FAULT_ALLOC_BIO

[Jaegeuk Kim]

On 01/26, Christoph Hellwig wrote:
> Sleeping bio allocations do not fail, which means that injecting an error
> into sleeping bio allocations is a little silly.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Already merged tho.

Acked-by: Jaegeuk Kim <jaegeuk@kernel.org>

> ---
>  Documentation/filesystems/f2fs.rst |  1 -
>  fs/f2fs/data.c                     | 29 ++++-------------------------
>  fs/f2fs/f2fs.h                     |  1 -
>  fs/f2fs/super.c                    |  1 -
>  4 files changed, 4 insertions(+), 28 deletions(-)
> 
> diff --git a/Documentation/filesystems/f2fs.rst b/Documentation/filesystems/f2fs.rst
> index dae15c96e659e2..624f5f3ed93e86 100644
> --- a/Documentation/filesystems/f2fs.rst
> +++ b/Documentation/filesystems/f2fs.rst
> @@ -179,7 +179,6 @@ fault_type=%d		 Support configuring fault injection type, should be
>  			 FAULT_KVMALLOC		  0x000000002
>  			 FAULT_PAGE_ALLOC	  0x000000004
>  			 FAULT_PAGE_GET		  0x000000008
> -			 FAULT_ALLOC_BIO	  0x000000010
>  			 FAULT_ALLOC_NID	  0x000000020
>  			 FAULT_ORPHAN		  0x000000040
>  			 FAULT_BLOCK		  0x000000080
> diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
> index 0cf0c605992431..9fb6be65592b1f 100644
> --- a/fs/f2fs/data.c
> +++ b/fs/f2fs/data.c
> @@ -50,28 +50,6 @@ void f2fs_destroy_bioset(void)
>  	bioset_exit(&f2fs_bioset);
>  }
>  
> -static inline struct bio *__f2fs_bio_alloc(gfp_t gfp_mask,
> -						unsigned int nr_iovecs)
> -{
> -	return bio_alloc_bioset(gfp_mask, nr_iovecs, &f2fs_bioset);
> -}
> -
> -static struct bio *f2fs_bio_alloc(struct f2fs_sb_info *sbi, int npages,
> -		bool noio)
> -{
> -	if (noio) {
> -		/* No failure on bio allocation */
> -		return __f2fs_bio_alloc(GFP_NOIO, npages);
> -	}
> -
> -	if (time_to_inject(sbi, FAULT_ALLOC_BIO)) {
> -		f2fs_show_injection_info(sbi, FAULT_ALLOC_BIO);
> -		return NULL;
> -	}
> -
> -	return __f2fs_bio_alloc(GFP_KERNEL, npages);
> -}
> -
>  static bool __is_cp_guaranteed(struct page *page)
>  {
>  	struct address_space *mapping = page->mapping;
> @@ -433,7 +411,7 @@ static struct bio *__bio_alloc(struct f2fs_io_info *fio, int npages)
>  	struct f2fs_sb_info *sbi = fio->sbi;
>  	struct bio *bio;
>  
> -	bio = f2fs_bio_alloc(sbi, npages, true);
> +	bio = bio_alloc_bioset(GFP_NOIO, npages, &f2fs_bioset);
>  
>  	f2fs_target_device(sbi, fio->new_blkaddr, bio);
>  	if (is_read_io(fio->op)) {
> @@ -1029,8 +1007,9 @@ static struct bio *f2fs_grab_read_bio(struct inode *inode, block_t blkaddr,
>  	struct bio_post_read_ctx *ctx;
>  	unsigned int post_read_steps = 0;
>  
> -	bio = f2fs_bio_alloc(sbi, min_t(int, nr_pages, BIO_MAX_PAGES),
> -								for_write);
> +	bio = bio_alloc_bioset(for_write ? GFP_NOIO : GFP_KERNEL,
> +			       min_t(int, nr_pages, BIO_MAX_PAGES),
> +			       &f2fs_bioset);
>  	if (!bio)
>  		return ERR_PTR(-ENOMEM);
>  
> diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
> index 902bd3267c03e1..6c78365d80ceb5 100644
> --- a/fs/f2fs/f2fs.h
> +++ b/fs/f2fs/f2fs.h
> @@ -43,7 +43,6 @@ enum {
>  	FAULT_KVMALLOC,
>  	FAULT_PAGE_ALLOC,
>  	FAULT_PAGE_GET,
> -	FAULT_ALLOC_BIO,
>  	FAULT_ALLOC_NID,
>  	FAULT_ORPHAN,
>  	FAULT_BLOCK,
> diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
> index b4a07fe62d1a58..3a312642907e86 100644
> --- a/fs/f2fs/super.c
> +++ b/fs/f2fs/super.c
> @@ -45,7 +45,6 @@ const char *f2fs_fault_name[FAULT_MAX] = {
>  	[FAULT_KVMALLOC]	= "kvmalloc",
>  	[FAULT_PAGE_ALLOC]	= "page alloc",
>  	[FAULT_PAGE_GET]	= "page get",
> -	[FAULT_ALLOC_BIO]	= "alloc bio",
>  	[FAULT_ALLOC_NID]	= "alloc nid",
>  	[FAULT_ORPHAN]		= "orphan",
>  	[FAULT_BLOCK]		= "no more block",
> -- 
> 2.29.2

Re: [f2fs-dev] [PATCH 07/17] f2fs: use blkdev_issue_flush in __submit_flush_wait

[Chao Yu]

On 2021/1/26 22:52, Christoph Hellwig wrote:
> Use the blkdev_issue_flush helper instead of duplicating it.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Acked-by: Chao Yu <yuchao0@huawei.com>

Thanks,

Re: [f2fs-dev] [PATCH 08/17] f2fs: remove FAULT_ALLOC_BIO

[Chao Yu]

On 2021/1/26 22:52, Christoph Hellwig wrote:
> Sleeping bio allocations do not fail, which means that injecting an error
> into sleeping bio allocations is a little silly.
> 
> Signed-off-by: Christoph Hellwig<hch@lst.de>

Acked-by: Chao Yu <yuchao0@huawei.com>

Thanks,

Re: [PATCH 04/17] block: split bio_kmalloc from bio_alloc_bioset

[Matthew Wilcox]

On Tue, Jan 26, 2021 at 03:52:34PM +0100, Christoph Hellwig wrote:
> bio_kmalloc shares almost no logic with the bio_set based fast path
> in bio_alloc_bioset.  Split it into an entirely separate implementation.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/bio.c         | 167 ++++++++++++++++++++++----------------------
>  include/linux/bio.h |   6 +-
>  2 files changed, 86 insertions(+), 87 deletions(-)

This patch causes current linux-next to OOM for me when running xfstests
after about ten minutes.  Haven't looked into why yet, this is just the
results of a git bisect.

The qemu command line is:

qemu-system-x86_64 -nodefaults -nographic -cpu host -machine accel=kvm,nvdimm -m 2G,slots=8,maxmem=1T -smp 6 -kernel /home/willy/kernel/folio/.build_test_kernel-x86_64/kpgk/vmlinuz -append console=hvc0 root=/dev/sda rw log_buf_len=8M ktest.dir=/home/willy/kernel/ktest ktest.env=/tmp/build-test-kernel-nJO6QgxOmo/env quiet systemd.show_status=0 systemd.log-target=journal crashkernel=128M no_console_suspend -device virtio-serial -chardev stdio,id=console -device virtconsole,chardev=console -serial unix:/tmp/build-test-kernel-nJO6QgxOmo/vm-kgdb,server,nowait -monitor unix:/tmp/build-test-kernel-nJO6QgxOmo/vm-mon,server,nowait -gdb unix:/tmp/build-test-kernel-nJO6QgxOmo/vm-gdb,server,nowait -device virtio-rng-pci -virtfs local,path=/,mount_tag=host,security_model=none -device virtio-scsi-pci,id=hba -nic user,model=virtio,hostfwd=tcp:127.0.0.1:24674-:22 -drive if=none,format=raw,id=disk0,file=/var/lib/ktest/root.amd64,snapshot=on -device scsi-hd,bus=hba.0,drive=disk0 -drive if=none,format=raw,id=disk1,file=/tmp/build-test-kernel-nJO6QgxOmo/dev-1,cache=unsafe -device scsi-hd,bus=hba.0,drive=disk1 -drive if=none,format=raw,id=disk2,file=/tmp/build-test-kernel-nJO6QgxOmo/dev-2,cache=unsafe -device scsi-hd,bus=hba.0,drive=disk2 -drive if=none,format=raw,id=disk3,file=/tmp/build-test-kernel-nJO6QgxOmo/dev-3,cache=unsafe -device scsi-hd,bus=hba.0,drive=disk3

Re: misc bio allocation cleanups

[Guoqing Jiang]

Hi Christoph,

I tried with latest for-5.12/block branch, and got below issue when 
trying to create raid1 array.

[   80.605832] BUG: kernel NULL pointer dereference, address: 
0000000000000018
[   80.606159] #PF: supervisor read access in kernel mode
[   80.606159] #PF: error_code(0x0000) - not-present page
[   80.606159] PGD 0 P4D 0
[   80.606159] Oops: 0000 [#1] SMP NOPTI
[   80.606159] CPU: 1 PID: 207 Comm: mdadm Not tainted 5.11.0-rc5+ #37
[   80.606159] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), 
BIOS 1.10.2-1ubuntu1 04/01/2014
[   80.606159] RIP: 0010:bdev_read_only+0x0/0x30
[   80.606159] Code: 89 df 5b 5d e9 71 21 e7 ff 81 e6 ff ff 0f 00 48 c7 
c7 40 99 d9 9e e8 0f e5 04 00 eb a0 0f 1f 00 66 2e 0f 1f 84 00 00 00 00 
00 <80> 7f 18 00 b8 01 00 00 00 75 11 48 8b 97 60 03 00 00 48 8b 4a 40
[   80.606159] RSP: 0018:ffffb14f405dfd38 EFLAGS: 00000246
[   80.606159] RAX: 0000000000000000 RBX: ffff9b0e0863b018 RCX: 
ffff9b0e01edc000
[   80.606159] RDX: ffff9b0e01b54c00 RSI: ffff9b0e0863b000 RDI: 
0000000000000000
[   80.606159] RBP: ffff9b0e01272200 R08: 0000000000000000 R09: 
0000000000000000
[   80.606159] R10: ffffee978004da40 R11: 0000000000000cc0 R12: 
ffff9b0e0863b000
[   80.606159] R13: ffff9b0e01f2cc00 R14: ffff9b0e0863b000 R15: 
0000000000000000
[   80.606159] FS:  00007f522ec65740(0000) GS:ffff9b0e7bc80000(0000) 
knlGS:0000000000000000
[   80.606159] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   80.606159] CR2: 0000000000000018 CR3: 0000000001364000 CR4: 
00000000000006e0
[   80.606159] Call Trace:
[   80.606159]  bind_rdev_to_array+0x2f7/0x380
[   80.606159]  ? cred_has_capability+0x80/0x120
[   80.606159]  md_add_new_disk+0x204/0x630
[   80.606159]  ? security_capable+0x33/0x50
[   80.606159]  md_ioctl+0xee7/0x1690
[   80.606159]  ? selinux_file_ioctl+0x143/0x200
[   80.606159]  blkdev_ioctl+0x1ff/0x240
[   80.606159]  block_ioctl+0x34/0x40
[   80.606159]  __x64_sys_ioctl+0x89/0xc0
[   80.606159]  do_syscall_64+0x33/0x40
[   80.606159]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[   80.606159] RIP: 0033:0x7f522e564317
[   80.606159] Code: b3 66 90 48 8b 05 71 4b 2d 00 64 c7 00 26 00 00 00 
48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f 
05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 41 4b 2d 00 f7 d8 64 89 01 48
[   80.606159] RSP: 002b:00007ffc3abd96f8 EFLAGS: 00000246 ORIG_RAX: 
0000000000000010
[   80.606159] RAX: ffffffffffffffda RBX: 00005653ee690350 RCX: 
00007f522e564317
[   80.606159] RDX: 00005653ee694058 RSI: 0000000040140921 RDI: 
0000000000000004
[   80.606159] RBP: 00005653ee690410 R08: 00007f522e839db0 R09: 
0000000000000000
[   80.606159] R10: 00005653ee690010 R11: 0000000000000246 R12: 
0000000000000000
[   80.606159] R13: 0000000000000000 R14: 0000000000000000 R15: 
00005653ee694010
[   80.606159] Modules linked in:
[   80.606159] CR2: 0000000000000018
[   80.622996] ---[ end trace 22144b856a3c1001 ]---
[   80.623285] RIP: 0010:bdev_read_only+0x0/0x30
[   80.623501] Code: 89 df 5b 5d e9 71 21 e7 ff 81 e6 ff ff 0f 00 48 c7 
c7 40 99 d9 9e e8 0f e5 04 00 eb a0 0f 1f 00 66 2e 0f 1f 84 00 00 00 00 
00 <80> 7f 18 00 b8 01 00 00 00 75 11 48 8b 97 60 03 00 00 48 8b 4a 40
[   80.624544] RSP: 0018:ffffb14f405dfd38 EFLAGS: 00000246
[   80.624788] RAX: 0000000000000000 RBX: ffff9b0e0863b018 RCX: 
ffff9b0e01edc000
[   80.625136] RDX: ffff9b0e01b54c00 RSI: ffff9b0e0863b000 RDI: 
0000000000000000
[   80.625449] RBP: ffff9b0e01272200 R08: 0000000000000000 R09: 
0000000000000000
[   80.625761] R10: ffffee978004da40 R11: 0000000000000cc0 R12: 
ffff9b0e0863b000
[   80.626112] R13: ffff9b0e01f2cc00 R14: ffff9b0e0863b000 R15: 
0000000000000000
[   80.626429] FS:  00007f522ec65740(0000) GS:ffff9b0e7bc80000(0000) 
knlGS:0000000000000000
[   80.626784] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   80.627035] CR2: 0000000000000018 CR3: 0000000001364000 CR4: 
00000000000006e0
Killed


Thanks,
Guoqing

Re: [PATCH 04/17] block: split bio_kmalloc from bio_alloc_bioset

[Christoph Hellwig]

On Sat, Jan 30, 2021 at 03:56:46AM +0000, Matthew Wilcox wrote:
> On Tue, Jan 26, 2021 at 03:52:34PM +0100, Christoph Hellwig wrote:
> > bio_kmalloc shares almost no logic with the bio_set based fast path
> > in bio_alloc_bioset.  Split it into an entirely separate implementation.
> > 
> > Signed-off-by: Christoph Hellwig <hch@lst.de>
> > ---
> >  block/bio.c         | 167 ++++++++++++++++++++++----------------------
> >  include/linux/bio.h |   6 +-
> >  2 files changed, 86 insertions(+), 87 deletions(-)
> 
> This patch causes current linux-next to OOM for me when running xfstests
> after about ten minutes.  Haven't looked into why yet, this is just the
> results of a git bisect.

I've run tests on linux-next all weekend and could not reproduce
the issue.  Can you share your .config?