Use uio for zvol_{read,write}

Since uio now supports bvec, we can convert bio into uio and reuse
dmu_{read,write}_uio. This way, we can remove some duplicate code.

Signed-off-by: Chunwei Chen <[email protected]>
Signed-off-by: Brian Behlendorf <[email protected]>
Closes #4078

4 files changed, 24 insertions, 172 deletions

diff --git a/include/linux/blkdev_compat.h b/include/linux/blkdev_compat.h
index 868b89c55..0cb07207e 100644
--- a/include/linux/blkdev_compat.h
+++ b/include/linux/blkdev_compat.h
@@ -118,6 +118,7 @@ get_disk_ro(struct gendisk *disk)
 #define	BIO_BI_SECTOR(bio)	(bio)->bi_iter.bi_sector
 #define	BIO_BI_SIZE(bio)	(bio)->bi_iter.bi_size
 #define	BIO_BI_IDX(bio)		(bio)->bi_iter.bi_idx
+#define	BIO_BI_SKIP(bio)	(bio)->bi_iter.bi_bvec_done
 #define	bio_for_each_segment4(bv, bvp, b, i)	\
 	bio_for_each_segment((bv), (b), (i))
 typedef struct bvec_iter bvec_iterator_t;
@@ -125,6 +126,7 @@ typedef struct bvec_iter bvec_iterator_t;
 #define	BIO_BI_SECTOR(bio)	(bio)->bi_sector
 #define	BIO_BI_SIZE(bio)	(bio)->bi_size
 #define	BIO_BI_IDX(bio)		(bio)->bi_idx
+#define	BIO_BI_SKIP(bio)	(0)
 #define	bio_for_each_segment4(bv, bvp, b, i)	\
 	bio_for_each_segment((bvp), (b), (i))
 typedef int bvec_iterator_t;
diff --git a/include/sys/dmu.h b/include/sys/dmu.h
index d9434db46..6bf51975e 100644
--- a/include/sys/dmu.h
+++ b/include/sys/dmu.h
@@ -710,9 +710,6 @@ void dmu_prealloc(objset_t *os, uint64_t object, uint64_t offset, uint64_t size,
 	dmu_tx_t *tx);
 #ifdef _KERNEL
 #include <linux/blkdev_compat.h>
-int dmu_read_bio(objset_t *os, uint64_t object, struct bio *bio);
-int dmu_write_bio(objset_t *os, uint64_t object, struct bio *bio,
-	dmu_tx_t *tx);
 int dmu_read_uio(objset_t *os, uint64_t object, struct uio *uio, uint64_t size);
 int dmu_read_uio_dbuf(dmu_buf_t *zdb, struct uio *uio, uint64_t size);
 int dmu_write_uio(objset_t *os, uint64_t object, struct uio *uio, uint64_t size,
diff --git a/module/zfs/dmu.c b/module/zfs/dmu.c
index 5e2a1db60..f4027af9c 100644
--- a/module/zfs/dmu.c
+++ b/module/zfs/dmu.c
@@ -1041,170 +1041,6 @@ xuio_stat_wbuf_nocopy()
 }
 
 #ifdef _KERNEL
-
-/*
- * Copy up to size bytes between arg_buf and req based on the data direction
- * described by the req.  If an entire req's data cannot be transfered in one
- * pass, you should pass in @req_offset to indicate where to continue. The
- * return value is the number of bytes successfully copied to arg_buf.
- */
-static int
-dmu_bio_copy(void *arg_buf, int size, struct bio *bio, size_t bio_offset)
-{
-	struct bio_vec bv, *bvp = &bv;
-	bvec_iterator_t iter;
-	char *bv_buf;
-	int tocpy, bv_len, bv_offset;
-	int offset = 0;
-
-	bio_for_each_segment4(bv, bvp, bio, iter) {
-
-		/*
-		 * Fully consumed the passed arg_buf. We use goto here because
-		 * rq_for_each_segment is a double loop
-		 */
-		ASSERT3S(offset, <=, size);
-		if (size == offset)
-			goto out;
-
-		/* Skip already copied bvp */
-		if (bio_offset >= bvp->bv_len) {
-			bio_offset -= bvp->bv_len;
-			continue;
-		}
-
-		bv_len = bvp->bv_len - bio_offset;
-		bv_offset = bvp->bv_offset + bio_offset;
-		bio_offset = 0;
-
-		tocpy = MIN(bv_len, size - offset);
-		ASSERT3S(tocpy, >=, 0);
-
-		bv_buf = page_address(bvp->bv_page) + bv_offset;
-		ASSERT3P(bv_buf, !=, NULL);
-
-		if (bio_data_dir(bio) == WRITE)
-			memcpy(arg_buf + offset, bv_buf, tocpy);
-		else
-			memcpy(bv_buf, arg_buf + offset, tocpy);
-
-		offset += tocpy;
-	}
-out:
-	return (offset);
-}
-
-int
-dmu_read_bio(objset_t *os, uint64_t object, struct bio *bio)
-{
-	uint64_t offset = BIO_BI_SECTOR(bio) << 9;
-	uint64_t size = BIO_BI_SIZE(bio);
-	dmu_buf_t **dbp;
-	int numbufs, i, err;
-	size_t bio_offset;
-
-	/*
-	 * NB: we could do this block-at-a-time, but it's nice
-	 * to be reading in parallel.
-	 */
-	err = dmu_buf_hold_array(os, object, offset, size, TRUE, FTAG,
-	    &numbufs, &dbp);
-	if (err)
-		return (err);
-
-	bio_offset = 0;
-	for (i = 0; i < numbufs; i++) {
-		uint64_t tocpy;
-		int64_t bufoff;
-		int didcpy;
-		dmu_buf_t *db = dbp[i];
-
-		bufoff = offset - db->db_offset;
-		ASSERT3S(bufoff, >=, 0);
-
-		tocpy = MIN(db->db_size - bufoff, size);
-		if (tocpy == 0)
-			break;
-
-		didcpy = dmu_bio_copy(db->db_data + bufoff, tocpy, bio,
-		    bio_offset);
-
-		if (didcpy < tocpy)
-			err = EIO;
-
-		if (err)
-			break;
-
-		size -= tocpy;
-		offset += didcpy;
-		bio_offset += didcpy;
-		err = 0;
-	}
-	dmu_buf_rele_array(dbp, numbufs, FTAG);
-
-	return (err);
-}
-
-int
-dmu_write_bio(objset_t *os, uint64_t object, struct bio *bio, dmu_tx_t *tx)
-{
-	uint64_t offset = BIO_BI_SECTOR(bio) << 9;
-	uint64_t size = BIO_BI_SIZE(bio);
-	dmu_buf_t **dbp;
-	int numbufs, i, err;
-	size_t bio_offset;
-
-	if (size == 0)
-		return (0);
-
-	err = dmu_buf_hold_array(os, object, offset, size, FALSE, FTAG,
-	    &numbufs, &dbp);
-	if (err)
-		return (err);
-
-	bio_offset = 0;
-	for (i = 0; i < numbufs; i++) {
-		uint64_t tocpy;
-		int64_t bufoff;
-		int didcpy;
-		dmu_buf_t *db = dbp[i];
-
-		bufoff = offset - db->db_offset;
-		ASSERT3S(bufoff, >=, 0);
-
-		tocpy = MIN(db->db_size - bufoff, size);
-		if (tocpy == 0)
-			break;
-
-		ASSERT(i == 0 || i == numbufs-1 || tocpy == db->db_size);
-
-		if (tocpy == db->db_size)
-			dmu_buf_will_fill(db, tx);
-		else
-			dmu_buf_will_dirty(db, tx);
-
-		didcpy = dmu_bio_copy(db->db_data + bufoff, tocpy, bio,
-		    bio_offset);
-
-		if (tocpy == db->db_size)
-			dmu_buf_fill_done(db, tx);
-
-		if (didcpy < tocpy)
-			err = EIO;
-
-		if (err)
-			break;
-
-		size -= tocpy;
-		offset += didcpy;
-		bio_offset += didcpy;
-		err = 0;
-	}
-
-	dmu_buf_rele_array(dbp, numbufs, FTAG);
-	return (err);
-}
-
 static int
 dmu_read_uio_dnode(dnode_t *dn, uio_t *uio, uint64_t size)
 {
diff --git a/module/zfs/zvol.c b/module/zfs/zvol.c
index 574d3c306..8208d29f3 100644
--- a/module/zfs/zvol.c
+++ b/module/zfs/zvol.c
@@ -597,6 +597,7 @@ zvol_write(struct bio *bio)
 	int error = 0;
 	dmu_tx_t *tx;
 	rl_t *rl;
+	uio_t uio;
 
 	if (bio->bi_rw & VDEV_REQ_FLUSH)
 		zil_commit(zv->zv_zilog, ZVOL_OBJ);
@@ -607,6 +608,14 @@ zvol_write(struct bio *bio)
 	if (size == 0)
 		goto out;
 
+	uio.uio_bvec = &bio->bi_io_vec[BIO_BI_IDX(bio)];
+	uio.uio_skip = BIO_BI_SKIP(bio);
+	uio.uio_resid = size;
+	uio.uio_iovcnt = bio->bi_vcnt - BIO_BI_IDX(bio);
+	uio.uio_loffset = offset;
+	uio.uio_limit = MAXOFFSET_T;
+	uio.uio_segflg = UIO_BVEC;
+
 	rl = zfs_range_lock(&zv->zv_znode, offset, size, RL_WRITER);
 
 	tx = dmu_tx_create(zv->zv_objset);
@@ -620,7 +629,7 @@ zvol_write(struct bio *bio)
 		goto out;
 	}
 
-	error = dmu_write_bio(zv->zv_objset, ZVOL_OBJ, bio, tx);
+	error = dmu_write_uio(zv->zv_objset, ZVOL_OBJ, &uio, size, tx);
 	if (error == 0)
 		zvol_log_write(zv, tx, offset, size,
 		    !!(bio->bi_rw & VDEV_REQ_FUA));
@@ -686,17 +695,25 @@ zvol_read(struct bio *bio)
 {
 	zvol_state_t *zv = bio->bi_bdev->bd_disk->private_data;
 	uint64_t offset = BIO_BI_SECTOR(bio) << 9;
-	uint64_t len = BIO_BI_SIZE(bio);
+	uint64_t size = BIO_BI_SIZE(bio);
 	int error;
 	rl_t *rl;
+	uio_t uio;
 
-	if (len == 0)
+	if (size == 0)
 		return (0);
 
+	uio.uio_bvec = &bio->bi_io_vec[BIO_BI_IDX(bio)];
+	uio.uio_skip = BIO_BI_SKIP(bio);
+	uio.uio_resid = size;
+	uio.uio_iovcnt = bio->bi_vcnt - BIO_BI_IDX(bio);
+	uio.uio_loffset = offset;
+	uio.uio_limit = MAXOFFSET_T;
+	uio.uio_segflg = UIO_BVEC;
 
-	rl = zfs_range_lock(&zv->zv_znode, offset, len, RL_READER);
+	rl = zfs_range_lock(&zv->zv_znode, offset, size, RL_READER);
 
-	error = dmu_read_bio(zv->zv_objset, ZVOL_OBJ, bio);
+	error = dmu_read_uio(zv->zv_objset, ZVOL_OBJ, &uio, size);
 
 	zfs_range_unlock(rl);