block: replace bi_bdev with a gendisk pointer and partitions index

This way we don't need a block_device structure to submit I/O.  The
block_device has different life time rules from the gendisk and
request_queue and is usually only available when the block device node
is open.  Other callers need to explicitly create one (e.g. the lightnvm
passthrough code, or the new nvme multipathing code).

For the actual I/O path all that we need is the gendisk, which exists
once per block device.  But given that the block layer also does
partition remapping we additionally need a partition index, which is
used for said remapping in generic_make_request.

Note that all the block drivers generally want request_queue or
sometimes the gendisk, so this removes a layer of indirection all
over the stack.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>

5 files changed, 32 insertions, 13 deletions

diff --git a/include/linux/bio.h b/include/linux/bio.h
index 9276788a9b24..a8fe7935332f 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -494,6 +494,24 @@ extern struct bio_vec *bvec_alloc(gfp_t, int, unsigned long *, mempool_t *);
 extern void bvec_free(mempool_t *, struct bio_vec *, unsigned int);
 extern unsigned int bvec_nr_vecs(unsigned short idx);
 
+#define bio_set_dev(bio, bdev) 			\
+do {						\
+	(bio)->bi_disk = (bdev)->bd_disk;	\
+	(bio)->bi_partno = (bdev)->bd_partno;	\
+} while (0)
+
+#define bio_copy_dev(dst, src)			\
+do {						\
+	(dst)->bi_disk = (src)->bi_disk;	\
+	(dst)->bi_partno = (src)->bi_partno;	\
+} while (0)
+
+#define bio_dev(bio) \
+	disk_devt((bio)->bi_disk)
+
+#define bio_devname(bio, buf) \
+	__bdevname(bio_dev(bio), (buf))
+
 #ifdef CONFIG_BLK_CGROUP
 int bio_associate_blkcg(struct bio *bio, struct cgroup_subsys_state *blkcg_css);
 int bio_associate_current(struct bio *bio);
diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h
index d2eb87c84d82..a2d2aa709cef 100644
--- a/include/linux/blk_types.h
+++ b/include/linux/blk_types.h
@@ -48,7 +48,8 @@ struct blk_issue_stat {
  */
 struct bio {
 	struct bio		*bi_next;	/* request queue link */
-	struct block_device	*bi_bdev;
+	struct gendisk		*bi_disk;
+	u8			bi_partno;
 	blk_status_t		bi_status;
 	unsigned int		bi_opf;		/* bottom bits req flags,
 						 * top bits REQ_OP. Use
diff --git a/include/trace/events/bcache.h b/include/trace/events/bcache.h
index df3e9ae5ad8d..daf749138ff8 100644
--- a/include/trace/events/bcache.h
+++ b/include/trace/events/bcache.h
@@ -21,7 +21,7 @@ DECLARE_EVENT_CLASS(bcache_request,
 	),
 
 	TP_fast_assign(
-		__entry->dev		= bio->bi_bdev->bd_dev;
+		__entry->dev		= bio_dev(bio);
 		__entry->orig_major	= d->disk->major;
 		__entry->orig_minor	= d->disk->first_minor;
 		__entry->sector		= bio->bi_iter.bi_sector;
@@ -98,7 +98,7 @@ DECLARE_EVENT_CLASS(bcache_bio,
 	),
 
 	TP_fast_assign(
-		__entry->dev		= bio->bi_bdev->bd_dev;
+		__entry->dev		= bio_dev(bio);
 		__entry->sector		= bio->bi_iter.bi_sector;
 		__entry->nr_sector	= bio->bi_iter.bi_size >> 9;
 		blk_fill_rwbs(__entry->rwbs, bio->bi_opf, bio->bi_iter.bi_size);
@@ -133,7 +133,7 @@ TRACE_EVENT(bcache_read,
 	),
 
 	TP_fast_assign(
-		__entry->dev		= bio->bi_bdev->bd_dev;
+		__entry->dev		= bio_dev(bio);
 		__entry->sector		= bio->bi_iter.bi_sector;
 		__entry->nr_sector	= bio->bi_iter.bi_size >> 9;
 		blk_fill_rwbs(__entry->rwbs, bio->bi_opf, bio->bi_iter.bi_size);
diff --git a/include/trace/events/block.h b/include/trace/events/block.h
index d0dbe60d8a6d..f815aaaef755 100644
--- a/include/trace/events/block.h
+++ b/include/trace/events/block.h
@@ -236,8 +236,7 @@ TRACE_EVENT(block_bio_bounce,
 	),
 
 	TP_fast_assign(
-		__entry->dev		= bio->bi_bdev ?
-					  bio->bi_bdev->bd_dev : 0;
+		__entry->dev		= bio_dev(bio);
 		__entry->sector		= bio->bi_iter.bi_sector;
 		__entry->nr_sector	= bio_sectors(bio);
 		blk_fill_rwbs(__entry->rwbs, bio->bi_opf, bio->bi_iter.bi_size);
@@ -274,7 +273,7 @@ TRACE_EVENT(block_bio_complete,
 	),
 
 	TP_fast_assign(
-		__entry->dev		= bio->bi_bdev->bd_dev;
+		__entry->dev		= bio_dev(bio);
 		__entry->sector		= bio->bi_iter.bi_sector;
 		__entry->nr_sector	= bio_sectors(bio);
 		__entry->error		= error;
@@ -302,7 +301,7 @@ DECLARE_EVENT_CLASS(block_bio_merge,
 	),
 
 	TP_fast_assign(
-		__entry->dev		= bio->bi_bdev->bd_dev;
+		__entry->dev		= bio_dev(bio);
 		__entry->sector		= bio->bi_iter.bi_sector;
 		__entry->nr_sector	= bio_sectors(bio);
 		blk_fill_rwbs(__entry->rwbs, bio->bi_opf, bio->bi_iter.bi_size);
@@ -369,7 +368,7 @@ TRACE_EVENT(block_bio_queue,
 	),
 
 	TP_fast_assign(
-		__entry->dev		= bio->bi_bdev->bd_dev;
+		__entry->dev		= bio_dev(bio);
 		__entry->sector		= bio->bi_iter.bi_sector;
 		__entry->nr_sector	= bio_sectors(bio);
 		blk_fill_rwbs(__entry->rwbs, bio->bi_opf, bio->bi_iter.bi_size);
@@ -397,7 +396,8 @@ DECLARE_EVENT_CLASS(block_get_rq,
         ),
 
 	TP_fast_assign(
-		__entry->dev		= bio ? bio->bi_bdev->bd_dev : 0;
+		__entry->dev		= bio ? bio_dev(bio) : 0;
+		__entry->dev		= bio_dev(bio);
 		__entry->sector		= bio ? bio->bi_iter.bi_sector : 0;
 		__entry->nr_sector	= bio ? bio_sectors(bio) : 0;
 		blk_fill_rwbs(__entry->rwbs,
@@ -532,7 +532,7 @@ TRACE_EVENT(block_split,
 	),
 
 	TP_fast_assign(
-		__entry->dev		= bio->bi_bdev->bd_dev;
+		__entry->dev		= bio_dev(bio);
 		__entry->sector		= bio->bi_iter.bi_sector;
 		__entry->new_sector	= new_sector;
 		blk_fill_rwbs(__entry->rwbs, bio->bi_opf, bio->bi_iter.bi_size);
@@ -573,7 +573,7 @@ TRACE_EVENT(block_bio_remap,
 	),
 
 	TP_fast_assign(
-		__entry->dev		= bio->bi_bdev->bd_dev;
+		__entry->dev		= bio_dev(bio);
 		__entry->sector		= bio->bi_iter.bi_sector;
 		__entry->nr_sector	= bio_sectors(bio);
 		__entry->old_dev	= dev;
diff --git a/include/trace/events/f2fs.h b/include/trace/events/f2fs.h
index 6f77a2755abb..bc4dd7837e4c 100644
--- a/include/trace/events/f2fs.h
+++ b/include/trace/events/f2fs.h
@@ -829,7 +829,7 @@ DECLARE_EVENT_CLASS(f2fs__bio,
 
 	TP_fast_assign(
 		__entry->dev		= sb->s_dev;
-		__entry->target		= bio->bi_bdev->bd_dev;
+		__entry->target		= bio_dev(bio);
 		__entry->op		= bio_op(bio);
 		__entry->op_flags	= bio->bi_opf;
 		__entry->type		= type;