block: improve handling of the magic discard payload

Instead of allocating a single unused biovec for discard requests, send
them down without any payload.  Instead we allow the driver to add a
"special" payload using a biovec embedded into struct request (unioned
over other fields never used while in the driver), and overloading
the number of segments for this case.

This has a couple of advantages:

 - we don't have to allocate the bio_vec
 - the amount of special casing for discard requests in the block
   layer is significantly reduced
 - using this same scheme for other request types is trivial,
   which will be important for implementing the new WRITE_ZEROES
   op on devices where it actually requires a payload (e.g. SCSI)
 - we can get rid of playing games with the request length, as
   we'll never touch it and completions will work just fine
 - it will allow us to support ranged discard operations in the
   future by merging non-contiguous discard bios into a single
   request
 - last but not least it removes a lot of code

This patch is the common base for my WIP series for ranges discards and to
remove discard_zeroes_data in favor of always using REQ_OP_WRITE_ZEROES,
so it would be good to get it in quickly.

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

drivers/scsi/scsi_lib.c

@@ -1007,8 +1007,8 @@ static int scsi_init_sgtable(struct request *req, struct scsi_data_buffer *sdb)
 	/*
 	 * If sg table allocation fails, requeue request later.
 	 */
-	if (unlikely(sg_alloc_table_chained(&sdb->table, req->nr_phys_segments,
-					sdb->table.sgl)))
+	if (unlikely(sg_alloc_table_chained(&sdb->table,
+			blk_rq_nr_phys_segments(req), sdb->table.sgl)))
 		return BLKPREP_DEFER;
 
 	/* 
@@ -1040,7 +1040,7 @@ int scsi_init_io(struct scsi_cmnd *cmd)
 	bool is_mq = (rq->mq_ctx != NULL);
 	int error;
 
-	BUG_ON(!rq->nr_phys_segments);
+	BUG_ON(!blk_rq_nr_phys_segments(rq));
 
 	error = scsi_init_sgtable(rq, &cmd->sdb);
 	if (error)

drivers/scsi/sd.c

@@ -716,7 +716,6 @@ static int sd_setup_discard_cmnd(struct scsi_cmnd *cmd)
 	struct scsi_disk *sdkp = scsi_disk(rq->rq_disk);
 	sector_t sector = blk_rq_pos(rq);
 	unsigned int nr_sectors = blk_rq_sectors(rq);
-	unsigned int nr_bytes = blk_rq_bytes(rq);
 	unsigned int len;
 	int ret;
 	char *buf;
@@ -772,24 +771,19 @@ static int sd_setup_discard_cmnd(struct scsi_cmnd *cmd)
 		goto out;
 	}
 
-	rq->completion_data = page;
 	rq->timeout = SD_TIMEOUT;
 
 	cmd->transfersize = len;
 	cmd->allowed = SD_MAX_RETRIES;
 
-	/*
-	 * Initially __data_len is set to the amount of data that needs to be
-	 * transferred to the target. This amount depends on whether WRITE SAME
-	 * or UNMAP is being used. After the scatterlist has been mapped by
-	 * scsi_init_io() we set __data_len to the size of the area to be
-	 * discarded on disk. This allows us to report completion on the full
-	 * amount of blocks described by the request.
-	 */
-	blk_add_request_payload(rq, page, 0, len);
-	ret = scsi_init_io(cmd);
-	rq->__data_len = nr_bytes;
+	rq->special_vec.bv_page = page;
+	rq->special_vec.bv_offset = 0;
+	rq->special_vec.bv_len = len;
 
+	rq->rq_flags |= RQF_SPECIAL_PAYLOAD;
+	rq->resid_len = len;
+
+	ret = scsi_init_io(cmd);
 out:
 	if (ret != BLKPREP_OK)
 		__free_page(page);
@@ -1182,8 +1176,8 @@ static void sd_uninit_command(struct scsi_cmnd *SCpnt)
 {
 	struct request *rq = SCpnt->request;
 
-	if (req_op(rq) == REQ_OP_DISCARD)
-		__free_page(rq->completion_data);
+	if (rq->rq_flags & RQF_SPECIAL_PAYLOAD)
+		__free_page(rq->special_vec.bv_page);
 
 	if (SCpnt->cmnd != rq->cmd) {
 		mempool_free(SCpnt->cmnd, sd_cdb_pool);