blk-mq: add a flags parameter to blk_mq_alloc_request

We already have the reserved flag, and a nowait flag awkwardly encoded as
a gfp_t.  Add a real flags argument to make the scheme more extensible and
allow for a nicer calling convention.

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

block/blk-core.c

@@ -630,7 +630,7 @@ struct request_queue *blk_alloc_queue(gfp_t gfp_mask)
 }
 EXPORT_SYMBOL(blk_alloc_queue);
 
-int blk_queue_enter(struct request_queue *q, gfp_t gfp)
+int blk_queue_enter(struct request_queue *q, bool nowait)
 {
 	while (true) {
 		int ret;
@@ -638,7 +638,7 @@ int blk_queue_enter(struct request_queue *q, gfp_t gfp)
 		if (percpu_ref_tryget_live(&q->q_usage_counter))
 			return 0;
 
-		if (!gfpflags_allow_blocking(gfp))
+		if (nowait)
 			return -EBUSY;
 
 		ret = wait_event_interruptible(q->mq_freeze_wq,
@@ -1276,7 +1276,9 @@ static struct request *blk_old_get_request(struct request_queue *q, int rw,
 struct request *blk_get_request(struct request_queue *q, int rw, gfp_t gfp_mask)
 {
 	if (q->mq_ops)
-		return blk_mq_alloc_request(q, rw, gfp_mask, false);
+		return blk_mq_alloc_request(q, rw,
+			(gfp_mask & __GFP_DIRECT_RECLAIM) ?
+				0 : BLK_MQ_REQ_NOWAIT);
 	else
 		return blk_old_get_request(q, rw, gfp_mask);
 }
@@ -2044,8 +2046,7 @@ blk_qc_t generic_make_request(struct bio *bio)
 	do {
 		struct request_queue *q = bdev_get_queue(bio->bi_bdev);
 
-		if (likely(blk_queue_enter(q, __GFP_DIRECT_RECLAIM) == 0)) {
+		if (likely(blk_queue_enter(q, false) == 0)) {
-
 			ret = q->make_request_fn(q, bio);
 
 			blk_queue_exit(q);

block/blk-mq-tag.c

@@ -268,7 +268,7 @@ static int bt_get(struct blk_mq_alloc_data *data,
 	if (tag != -1)
 		return tag;
 
-	if (!gfpflags_allow_blocking(data->gfp))
+	if (data->flags & BLK_MQ_REQ_NOWAIT)
 		return -1;
 
 	bs = bt_wait_ptr(bt, hctx);
@@ -303,7 +303,7 @@ static int bt_get(struct blk_mq_alloc_data *data,
 		data->ctx = blk_mq_get_ctx(data->q);
 		data->hctx = data->q->mq_ops->map_queue(data->q,
 				data->ctx->cpu);
-		if (data->reserved) {
+		if (data->flags & BLK_MQ_REQ_RESERVED) {
 			bt = &data->hctx->tags->breserved_tags;
 		} else {
 			last_tag = &data->ctx->last_tag;
@@ -349,10 +349,9 @@ static unsigned int __blk_mq_get_reserved_tag(struct blk_mq_alloc_data *data)
 
 unsigned int blk_mq_get_tag(struct blk_mq_alloc_data *data)
 {
-	if (!data->reserved)
+	if (data->flags & BLK_MQ_REQ_RESERVED)
-		return __blk_mq_get_tag(data);
+		return __blk_mq_get_reserved_tag(data);
-
+	return __blk_mq_get_tag(data);
-	return __blk_mq_get_reserved_tag(data);
 }
 
 static struct bt_wait_state *bt_wake_ptr(struct blk_mq_bitmap_tags *bt)

block/blk-mq.c

@@ -229,8 +229,8 @@ __blk_mq_alloc_request(struct blk_mq_alloc_data *data, int rw)
 	return NULL;
 }
 
-struct request *blk_mq_alloc_request(struct request_queue *q, int rw, gfp_t gfp,
+struct request *blk_mq_alloc_request(struct request_queue *q, int rw,
-		bool reserved)
+		unsigned int flags)
 {
 	struct blk_mq_ctx *ctx;
 	struct blk_mq_hw_ctx *hctx;
@@ -238,24 +238,22 @@ struct request *blk_mq_alloc_request(struct request_queue *q, int rw, gfp_t gfp,
 	struct blk_mq_alloc_data alloc_data;
 	int ret;
 
-	ret = blk_queue_enter(q, gfp);
+	ret = blk_queue_enter(q, flags & BLK_MQ_REQ_NOWAIT);
 	if (ret)
 		return ERR_PTR(ret);
 
 	ctx = blk_mq_get_ctx(q);
 	hctx = q->mq_ops->map_queue(q, ctx->cpu);
-	blk_mq_set_alloc_data(&alloc_data, q, gfp & ~__GFP_DIRECT_RECLAIM,
+	blk_mq_set_alloc_data(&alloc_data, q, flags, ctx, hctx);
-			reserved, ctx, hctx);
 
 	rq = __blk_mq_alloc_request(&alloc_data, rw);
-	if (!rq && (gfp & __GFP_DIRECT_RECLAIM)) {
+	if (!rq && !(flags & BLK_MQ_REQ_NOWAIT)) {
 		__blk_mq_run_hw_queue(hctx);
 		blk_mq_put_ctx(ctx);
 
 		ctx = blk_mq_get_ctx(q);
 		hctx = q->mq_ops->map_queue(q, ctx->cpu);
-		blk_mq_set_alloc_data(&alloc_data, q, gfp, reserved, ctx,
+		blk_mq_set_alloc_data(&alloc_data, q, flags, ctx, hctx);
-				hctx);
 		rq =  __blk_mq_alloc_request(&alloc_data, rw);
 		ctx = alloc_data.ctx;
 	}
@@ -1175,8 +1173,7 @@ static struct request *blk_mq_map_request(struct request_queue *q,
 		rw |= REQ_SYNC;
 
 	trace_block_getrq(q, bio, rw);
-	blk_mq_set_alloc_data(&alloc_data, q, GFP_ATOMIC, false, ctx,
+	blk_mq_set_alloc_data(&alloc_data, q, BLK_MQ_REQ_NOWAIT, ctx, hctx);
-			hctx);
 	rq = __blk_mq_alloc_request(&alloc_data, rw);
 	if (unlikely(!rq)) {
 		__blk_mq_run_hw_queue(hctx);
@@ -1185,8 +1182,7 @@ static struct request *blk_mq_map_request(struct request_queue *q,
 
 		ctx = blk_mq_get_ctx(q);
 		hctx = q->mq_ops->map_queue(q, ctx->cpu);
-		blk_mq_set_alloc_data(&alloc_data, q,
+		blk_mq_set_alloc_data(&alloc_data, q, 0, ctx, hctx);
-				__GFP_RECLAIM|__GFP_HIGH, false, ctx, hctx);
 		rq = __blk_mq_alloc_request(&alloc_data, rw);
 		ctx = alloc_data.ctx;
 		hctx = alloc_data.hctx;

block/blk-mq.h

@@ -96,8 +96,7 @@ static inline void blk_mq_put_ctx(struct blk_mq_ctx *ctx)
 struct blk_mq_alloc_data {
 	/* input parameter */
 	struct request_queue *q;
-	gfp_t gfp;
+	unsigned int flags;
-	bool reserved;
 
 	/* input & output parameter */
 	struct blk_mq_ctx *ctx;
@@ -105,13 +104,11 @@ struct blk_mq_alloc_data {
 };
 
 static inline void blk_mq_set_alloc_data(struct blk_mq_alloc_data *data,
-		struct request_queue *q, gfp_t gfp, bool reserved,
+		struct request_queue *q, unsigned int flags,
-		struct blk_mq_ctx *ctx,
+		struct blk_mq_ctx *ctx, struct blk_mq_hw_ctx *hctx)
-		struct blk_mq_hw_ctx *hctx)
 {
 	data->q = q;
-	data->gfp = gfp;
+	data->flags = flags;
-	data->reserved = reserved;
 	data->ctx = ctx;
 	data->hctx = hctx;
 }

drivers/block/mtip32xx/mtip32xx.c

@@ -173,7 +173,7 @@ static struct mtip_cmd *mtip_get_int_command(struct driver_data *dd)
 {
 	struct request *rq;
 
-	rq = blk_mq_alloc_request(dd->queue, 0, __GFP_RECLAIM, true);
+	rq = blk_mq_alloc_request(dd->queue, 0, BLK_MQ_REQ_RESERVED);
 	return blk_mq_rq_to_pdu(rq);
 }
 

drivers/block/null_blk.c

@@ -464,7 +464,7 @@ static int null_lnvm_submit_io(struct request_queue *q, struct nvm_rq *rqd)
 	struct request *rq;
 	struct bio *bio = rqd->bio;
 
-	rq = blk_mq_alloc_request(q, bio_rw(bio), GFP_KERNEL, 0);
+	rq = blk_mq_alloc_request(q, bio_rw(bio), 0);
 	if (IS_ERR(rq))
 		return -ENOMEM;
 

drivers/nvme/host/lightnvm.c

@@ -470,7 +470,7 @@ static int nvme_nvm_submit_io(struct request_queue *q, struct nvm_rq *rqd)
 	struct bio *bio = rqd->bio;
 	struct nvme_nvm_command *cmd;
 
-	rq = blk_mq_alloc_request(q, bio_rw(bio), GFP_KERNEL, 0);
+	rq = blk_mq_alloc_request(q, bio_rw(bio), 0);
 	if (IS_ERR(rq))
 		return -ENOMEM;
 

drivers/nvme/host/pci.c

@@ -1041,7 +1041,7 @@ int __nvme_submit_sync_cmd(struct request_queue *q, struct nvme_command *cmd,
 	struct request *req;
 	int ret;
 
-	req = blk_mq_alloc_request(q, write, GFP_KERNEL, false);
+	req = blk_mq_alloc_request(q, write, 0);
 	if (IS_ERR(req))
 		return PTR_ERR(req);
 
@@ -1094,7 +1094,8 @@ static int nvme_submit_async_admin_req(struct nvme_dev *dev)
 	struct nvme_cmd_info *cmd_info;
 	struct request *req;
 
-	req = blk_mq_alloc_request(dev->admin_q, WRITE, GFP_ATOMIC, true);
+	req = blk_mq_alloc_request(dev->admin_q, WRITE,
+			BLK_MQ_REQ_NOWAIT | BLK_MQ_REQ_RESERVED);
 	if (IS_ERR(req))
 		return PTR_ERR(req);
 
@@ -1119,7 +1120,7 @@ static int nvme_submit_admin_async_cmd(struct nvme_dev *dev,
 	struct request *req;
 	struct nvme_cmd_info *cmd_rq;
 
-	req = blk_mq_alloc_request(dev->admin_q, WRITE, GFP_KERNEL, false);
+	req = blk_mq_alloc_request(dev->admin_q, WRITE, 0);
 	if (IS_ERR(req))
 		return PTR_ERR(req);
 
@@ -1320,8 +1321,8 @@ static void nvme_abort_req(struct request *req)
 	if (!dev->abort_limit)
 		return;
 
-	abort_req = blk_mq_alloc_request(dev->admin_q, WRITE, GFP_ATOMIC,
+	abort_req = blk_mq_alloc_request(dev->admin_q, WRITE,
-									false);
+			BLK_MQ_REQ_NOWAIT);
 	if (IS_ERR(abort_req))
 		return;
 

fs/block_dev.c

@@ -395,7 +395,7 @@ int bdev_read_page(struct block_device *bdev, sector_t sector,
 	if (!ops->rw_page || bdev_get_integrity(bdev))
 		return result;
 
-	result = blk_queue_enter(bdev->bd_queue, GFP_KERNEL);
+	result = blk_queue_enter(bdev->bd_queue, false);
 	if (result)
 		return result;
 	result = ops->rw_page(bdev, sector + get_start_sect(bdev), page, READ);
@@ -432,7 +432,7 @@ int bdev_write_page(struct block_device *bdev, sector_t sector,
 
 	if (!ops->rw_page || bdev_get_integrity(bdev))
 		return -EOPNOTSUPP;
-	result = blk_queue_enter(bdev->bd_queue, GFP_KERNEL);
+	result = blk_queue_enter(bdev->bd_queue, false);
 	if (result)
 		return result;
 

include/linux/blk-mq.h

@@ -188,8 +188,14 @@ void blk_mq_insert_request(struct request *, bool, bool, bool);
 void blk_mq_free_request(struct request *rq);
 void blk_mq_free_hctx_request(struct blk_mq_hw_ctx *, struct request *rq);
 bool blk_mq_can_queue(struct blk_mq_hw_ctx *);
+
+enum {
+	BLK_MQ_REQ_NOWAIT	= (1 << 0), /* return when out of requests */
+	BLK_MQ_REQ_RESERVED	= (1 << 1), /* allocate from reserved pool */
+};
+
 struct request *blk_mq_alloc_request(struct request_queue *q, int rw,
-		gfp_t gfp, bool reserved);
+		unsigned int flags);
 struct request *blk_mq_tag_to_rq(struct blk_mq_tags *tags, unsigned int tag);
 struct cpumask *blk_mq_tags_cpumask(struct blk_mq_tags *tags);
 

include/linux/blkdev.h

@@ -794,7 +794,7 @@ extern int scsi_cmd_ioctl(struct request_queue *, struct gendisk *, fmode_t,
 extern int sg_scsi_ioctl(struct request_queue *, struct gendisk *, fmode_t,
 			 struct scsi_ioctl_command __user *);
 
-extern int blk_queue_enter(struct request_queue *q, gfp_t gfp);
+extern int blk_queue_enter(struct request_queue *q, bool nowait);
 extern void blk_queue_exit(struct request_queue *q);
 extern void blk_start_queue(struct request_queue *q);
 extern void blk_stop_queue(struct request_queue *q);