Merge branch 'for-linus' of git://git.kernel.dk/linux-block

Pull more block updates from Jens Axboe:
 "This is a followup for block changes, that didn't make the initial
  pull request. It's a bit of a mixed bag, this contains:

   - A followup pull request from Sagi for NVMe. Outside of fixups for
     NVMe, it also includes a series for ensuring that we properly
     quiesce hardware queues when browsing live tags.

   - Set of integrity fixes from Dmitry (mostly), fixing various issues
     for folks using DIF/DIX.

   - Fix for a bug introduced in cciss, with the req init changes. From
     Christoph.

   - Fix for a bug in BFQ, from Paolo.

   - Two followup fixes for lightnvm/pblk from Javier.

   - Depth fix from Ming for blk-mq-sched.

   - Also from Ming, performance fix for mtip32xx that was introduced
     with the dynamic initialization of commands"

* 'for-linus' of git://git.kernel.dk/linux-block: (44 commits)
  block: call bio_uninit in bio_endio
  nvmet: avoid unneeded assignment of submit_bio return value
  nvme-pci: add module parameter for io queue depth
  nvme-pci: compile warnings in nvme_alloc_host_mem()
  nvmet_fc: Accept variable pad lengths on Create Association LS
  nvme_fc/nvmet_fc: revise Create Association descriptor length
  lightnvm: pblk: remove unnecessary checks
  lightnvm: pblk: control I/O flow also on tear down
  cciss: initialize struct scsi_req
  null_blk: fix error flow for shared tags during module_init
  block: Fix __blkdev_issue_zeroout loop
  nvme-rdma: unconditionally recycle the request mr
  nvme: split nvme_uninit_ctrl into stop and uninit
  virtio_blk: quiesce/unquiesce live IO when entering PM states
  mtip32xx: quiesce request queues to make sure no submissions are inflight
  nbd: quiesce request queues to make sure no submissions are inflight
  nvme: kick requeue list when requeueing a request instead of when starting the queues
  nvme-pci: quiesce/unquiesce admin_q instead of start/stop its hw queues
  nvme-loop: quiesce/unquiesce admin_q instead of start/stop its hw queues
  nvme-fc: quiesce/unquiesce admin_q instead of start/stop its hw queues
  ...

Documentation/block/data-integrity.txt

@@ -192,7 +192,7 @@ will require extra work due to the application tag.
     supported by the block device.
 
 
-    int bio_integrity_prep(bio);
+    bool bio_integrity_prep(bio);
 
       To generate IMD for WRITE and to set up buffers for READ, the
       filesystem must call bio_integrity_prep(bio).
@@ -201,9 +201,7 @@ will require extra work due to the application tag.
       sector must be set, and the bio should have all data pages
       added.  It is up to the caller to ensure that the bio does not
       change while I/O is in progress.
-
+      Complete bio with error if prepare failed for some reson.
-      bio_integrity_prep() should only be called if
-      bio_integrity_enabled() returned 1.
 
 
 5.3 PASSING EXISTING INTEGRITY METADATA

block/bfq-iosched.c

@@ -3483,11 +3483,17 @@ static void bfq_update_wr_data(struct bfq_data *bfqd, struct bfq_queue *bfqq)
 			}
 		}
 	}
-	/* Update weight both if it must be raised and if it must be lowered */
+	/*
+	 * To improve latency (for this or other queues), immediately
+	 * update weight both if it must be raised and if it must be
+	 * lowered. Since, entity may be on some active tree here, and
+	 * might have a pending change of its ioprio class, invoke
+	 * next function with the last parameter unset (see the
+	 * comments on the function).
+	 */
 	if ((entity->weight > entity->orig_weight) != (bfqq->wr_coeff > 1))
-		__bfq_entity_update_weight_prio(
+		__bfq_entity_update_weight_prio(bfq_entity_service_tree(entity),
-			bfq_entity_service_tree(entity),
+						entity, false);
-			entity);
 }
 
 /*

block/bfq-iosched.h

@@ -892,7 +892,8 @@ void bfq_put_idle_entity(struct bfq_service_tree *st,
 			 struct bfq_entity *entity);
 struct bfq_service_tree *
 __bfq_entity_update_weight_prio(struct bfq_service_tree *old_st,
-				struct bfq_entity *entity);
+				struct bfq_entity *entity,
+				bool update_class_too);
 void bfq_bfqq_served(struct bfq_queue *bfqq, int served);
 void bfq_bfqq_charge_time(struct bfq_data *bfqd, struct bfq_queue *bfqq,
 			  unsigned long time_ms);

block/bfq-wf2q.c

@@ -694,10 +694,28 @@ struct bfq_service_tree *bfq_entity_service_tree(struct bfq_entity *entity)
 	return sched_data->service_tree + idx;
 }
 
-
+/*
+ * Update weight and priority of entity. If update_class_too is true,
+ * then update the ioprio_class of entity too.
+ *
+ * The reason why the update of ioprio_class is controlled through the
+ * last parameter is as follows. Changing the ioprio class of an
+ * entity implies changing the destination service trees for that
+ * entity. If such a change occurred when the entity is already on one
+ * of the service trees for its previous class, then the state of the
+ * entity would become more complex: none of the new possible service
+ * trees for the entity, according to bfq_entity_service_tree(), would
+ * match any of the possible service trees on which the entity
+ * is. Complex operations involving these trees, such as entity
+ * activations and deactivations, should take into account this
+ * additional complexity.  To avoid this issue, this function is
+ * invoked with update_class_too unset in the points in the code where
+ * entity may happen to be on some tree.
+ */
 struct bfq_service_tree *
 __bfq_entity_update_weight_prio(struct bfq_service_tree *old_st,
-				struct bfq_entity *entity)
+				struct bfq_entity *entity,
+				bool update_class_too)
 {
 	struct bfq_service_tree *new_st = old_st;
 
@@ -739,9 +757,15 @@ __bfq_entity_update_weight_prio(struct bfq_service_tree *old_st,
 				  bfq_weight_to_ioprio(entity->orig_weight);
 		}
 
-		if (bfqq)
+		if (bfqq && update_class_too)
 			bfqq->ioprio_class = bfqq->new_ioprio_class;
-		entity->prio_changed = 0;
+
+		/*
+		 * Reset prio_changed only if the ioprio_class change
+		 * is not pending any longer.
+		 */
+		if (!bfqq || bfqq->ioprio_class == bfqq->new_ioprio_class)
+			entity->prio_changed = 0;
 
 		/*
 		 * NOTE: here we may be changing the weight too early,
@@ -867,7 +891,12 @@ static void bfq_update_fin_time_enqueue(struct bfq_entity *entity,
 {
 	struct bfq_queue *bfqq = bfq_entity_to_bfqq(entity);
 
-	st = __bfq_entity_update_weight_prio(st, entity);
+	/*
+	 * When this function is invoked, entity is not in any service
+	 * tree, then it is safe to invoke next function with the last
+	 * parameter set (see the comments on the function).
+	 */
+	st = __bfq_entity_update_weight_prio(st, entity, true);
 	bfq_calc_finish(entity, entity->budget);
 
 	/*

block/bio-integrity.c

@@ -102,7 +102,7 @@ EXPORT_SYMBOL(bio_integrity_alloc);
  * Description: Used to free the integrity portion of a bio. Usually
  * called from bio_free().
  */
-void bio_integrity_free(struct bio *bio)
+static void bio_integrity_free(struct bio *bio)
 {
 	struct bio_integrity_payload *bip = bio_integrity(bio);
 	struct bio_set *bs = bio->bi_pool;
@@ -120,8 +120,8 @@ void bio_integrity_free(struct bio *bio)
 	}
 
 	bio->bi_integrity = NULL;
+	bio->bi_opf &= ~REQ_INTEGRITY;
 }
-EXPORT_SYMBOL(bio_integrity_free);
 
 /**
  * bio_integrity_add_page - Attach integrity metadata
@@ -159,44 +159,6 @@ int bio_integrity_add_page(struct bio *bio, struct page *page,
 }
 EXPORT_SYMBOL(bio_integrity_add_page);
 
-/**
- * bio_integrity_enabled - Check whether integrity can be passed
- * @bio:	bio to check
- *
- * Description: Determines whether bio_integrity_prep() can be called
- * on this bio or not.	bio data direction and target device must be
- * set prior to calling.  The functions honors the write_generate and
- * read_verify flags in sysfs.
- */
-bool bio_integrity_enabled(struct bio *bio)
-{
-	struct blk_integrity *bi = bdev_get_integrity(bio->bi_bdev);
-
-	if (bio_op(bio) != REQ_OP_READ && bio_op(bio) != REQ_OP_WRITE)
-		return false;
-
-	if (!bio_sectors(bio))
-		return false;
-
-	/* Already protected? */
-	if (bio_integrity(bio))
-		return false;
-
-	if (bi == NULL)
-		return false;
-
-	if (bio_data_dir(bio) == READ && bi->profile->verify_fn != NULL &&
-	    (bi->flags & BLK_INTEGRITY_VERIFY))
-		return true;
-
-	if (bio_data_dir(bio) == WRITE && bi->profile->generate_fn != NULL &&
-	    (bi->flags & BLK_INTEGRITY_GENERATE))
-		return true;
-
-	return false;
-}
-EXPORT_SYMBOL(bio_integrity_enabled);
-
 /**
  * bio_integrity_intervals - Return number of integrity intervals for a bio
  * @bi:		blk_integrity profile for device
@@ -222,10 +184,11 @@ static inline unsigned int bio_integrity_bytes(struct blk_integrity *bi,
 /**
  * bio_integrity_process - Process integrity metadata for a bio
  * @bio:	bio to generate/verify integrity metadata for
+ * @proc_iter:  iterator to process
  * @proc_fn:	Pointer to the relevant processing function
  */
 static blk_status_t bio_integrity_process(struct bio *bio,
-				 integrity_processing_fn *proc_fn)
+		struct bvec_iter *proc_iter, integrity_processing_fn *proc_fn)
 {
 	struct blk_integrity *bi = bdev_get_integrity(bio->bi_bdev);
 	struct blk_integrity_iter iter;
@@ -238,10 +201,10 @@ static blk_status_t bio_integrity_process(struct bio *bio,
 
 	iter.disk_name = bio->bi_bdev->bd_disk->disk_name;
 	iter.interval = 1 << bi->interval_exp;
-	iter.seed = bip_get_seed(bip);
+	iter.seed = proc_iter->bi_sector;
 	iter.prot_buf = prot_buf;
 
-	bio_for_each_segment(bv, bio, bviter) {
+	__bio_for_each_segment(bv, bio, bviter, *proc_iter) {
 		void *kaddr = kmap_atomic(bv.bv_page);
 
 		iter.data_buf = kaddr + bv.bv_offset;
@@ -262,14 +225,15 @@ static blk_status_t bio_integrity_process(struct bio *bio,
  * bio_integrity_prep - Prepare bio for integrity I/O
  * @bio:	bio to prepare
  *
- * Description: Allocates a buffer for integrity metadata, maps the
+ * Description:  Checks if the bio already has an integrity payload attached.
- * pages and attaches them to a bio.  The bio must have data
+ * If it does, the payload has been generated by another kernel subsystem,
- * direction, target device and start sector set priot to calling.  In
+ * and we just pass it through. Otherwise allocates integrity payload.
- * the WRITE case, integrity metadata will be generated using the
+ * The bio must have data direction, target device and start sector set priot
- * block device's integrity function.  In the READ case, the buffer
+ * to calling.  In the WRITE case, integrity metadata will be generated using
+ * the block device's integrity function.  In the READ case, the buffer
  * will be prepared for DMA and a suitable end_io handler set up.
  */
-int bio_integrity_prep(struct bio *bio)
+bool bio_integrity_prep(struct bio *bio)
 {
 	struct bio_integrity_payload *bip;
 	struct blk_integrity *bi;
@@ -279,20 +243,41 @@ int bio_integrity_prep(struct bio *bio)
 	unsigned int len, nr_pages;
 	unsigned int bytes, offset, i;
 	unsigned int intervals;
+	blk_status_t status;
 
 	bi = bdev_get_integrity(bio->bi_bdev);
 	q = bdev_get_queue(bio->bi_bdev);
-	BUG_ON(bi == NULL);
+	if (bio_op(bio) != REQ_OP_READ && bio_op(bio) != REQ_OP_WRITE)
-	BUG_ON(bio_integrity(bio));
+		return true;
 
+	if (!bio_sectors(bio))
+		return true;
+
+	/* Already protected? */
+	if (bio_integrity(bio))
+		return true;
+
+	if (bi == NULL)
+		return true;
+
+	if (bio_data_dir(bio) == READ) {
+		if (!bi->profile->verify_fn ||
+		    !(bi->flags & BLK_INTEGRITY_VERIFY))
+			return true;
+	} else {
+		if (!bi->profile->generate_fn ||
+		    !(bi->flags & BLK_INTEGRITY_GENERATE))
+			return true;
+	}
 	intervals = bio_integrity_intervals(bi, bio_sectors(bio));
 
 	/* Allocate kernel buffer for protection data */
 	len = intervals * bi->tuple_size;
 	buf = kmalloc(len, GFP_NOIO | q->bounce_gfp);
+	status = BLK_STS_RESOURCE;
 	if (unlikely(buf == NULL)) {
 		printk(KERN_ERR "could not allocate integrity buffer\n");
-		return -ENOMEM;
+		goto err_end_io;
 	}
 
 	end = (((unsigned long) buf) + len + PAGE_SIZE - 1) >> PAGE_SHIFT;
@@ -304,7 +289,8 @@ int bio_integrity_prep(struct bio *bio)
 	if (IS_ERR(bip)) {
 		printk(KERN_ERR "could not allocate data integrity bioset\n");
 		kfree(buf);
-		return PTR_ERR(bip);
+		status = BLK_STS_RESOURCE;
+		goto err_end_io;
 	}
 
 	bip->bip_flags |= BIP_BLOCK_INTEGRITY;
@@ -330,7 +316,7 @@ int bio_integrity_prep(struct bio *bio)
 					     bytes, offset);
 
 		if (ret == 0)
-			return 0;
+			return false;
 
 		if (ret < bytes)
 			break;
@@ -340,17 +326,18 @@ int bio_integrity_prep(struct bio *bio)
 		offset = 0;
 	}
 
-	/* Install custom I/O completion handler if read verify is enabled */
-	if (bio_data_dir(bio) == READ) {
-		bip->bip_end_io = bio->bi_end_io;
-		bio->bi_end_io = bio_integrity_endio;
-	}
-
 	/* Auto-generate integrity metadata if this is a write */
-	if (bio_data_dir(bio) == WRITE)
+	if (bio_data_dir(bio) == WRITE) {
-		bio_integrity_process(bio, bi->profile->generate_fn);
+		bio_integrity_process(bio, &bio->bi_iter,
+				      bi->profile->generate_fn);
+	}
+	return true;
+
+err_end_io:
+	bio->bi_status = status;
+	bio_endio(bio);
+	return false;
 
-	return 0;
 }
 EXPORT_SYMBOL(bio_integrity_prep);
 
@@ -368,16 +355,26 @@ static void bio_integrity_verify_fn(struct work_struct *work)
 		container_of(work, struct bio_integrity_payload, bip_work);
 	struct bio *bio = bip->bip_bio;
 	struct blk_integrity *bi = bdev_get_integrity(bio->bi_bdev);
+	struct bvec_iter iter = bio->bi_iter;
 
-	bio->bi_status = bio_integrity_process(bio, bi->profile->verify_fn);
+	/*
+	 * At the moment verify is called bio's iterator was advanced
+	 * during split and completion, we need to rewind iterator to
+	 * it's original position.
+	 */
+	if (bio_rewind_iter(bio, &iter, iter.bi_done)) {
+		bio->bi_status = bio_integrity_process(bio, &iter,
+						       bi->profile->verify_fn);
+	} else {
+		bio->bi_status = BLK_STS_IOERR;
+	}
 
-	/* Restore original bio completion handler */
+	bio_integrity_free(bio);
-	bio->bi_end_io = bip->bip_end_io;
 	bio_endio(bio);
 }
 
 /**
- * bio_integrity_endio - Integrity I/O completion function
+ * __bio_integrity_endio - Integrity I/O completion function
  * @bio:	Protected bio
  * @error:	Pointer to errno
  *
@@ -388,27 +385,19 @@ static void bio_integrity_verify_fn(struct work_struct *work)
  * in process context.	This function postpones completion
  * accordingly.
  */
-void bio_integrity_endio(struct bio *bio)
+bool __bio_integrity_endio(struct bio *bio)
 {
-	struct bio_integrity_payload *bip = bio_integrity(bio);
+	if (bio_op(bio) == REQ_OP_READ && !bio->bi_status) {
+		struct bio_integrity_payload *bip = bio_integrity(bio);
 
-	BUG_ON(bip->bip_bio != bio);
+		INIT_WORK(&bip->bip_work, bio_integrity_verify_fn);
-
+		queue_work(kintegrityd_wq, &bip->bip_work);
-	/* In case of an I/O error there is no point in verifying the
+		return false;
-	 * integrity metadata.  Restore original bio end_io handler
-	 * and run it.
-	 */
-	if (bio->bi_status) {
-		bio->bi_end_io = bip->bip_end_io;
-		bio_endio(bio);
-
-		return;
 	}
 
-	INIT_WORK(&bip->bip_work, bio_integrity_verify_fn);
+	bio_integrity_free(bio);
-	queue_work(kintegrityd_wq, &bip->bip_work);
+	return true;
 }
-EXPORT_SYMBOL(bio_integrity_endio);
 
 /**
  * bio_integrity_advance - Advance integrity vector
@@ -425,6 +414,7 @@ void bio_integrity_advance(struct bio *bio, unsigned int bytes_done)
 	struct blk_integrity *bi = bdev_get_integrity(bio->bi_bdev);
 	unsigned bytes = bio_integrity_bytes(bi, bytes_done >> 9);
 
+	bip->bip_iter.bi_sector += bytes_done >> 9;
 	bvec_iter_advance(bip->bip_vec, &bip->bip_iter, bytes);
 }
 EXPORT_SYMBOL(bio_integrity_advance);
@@ -432,22 +422,15 @@ EXPORT_SYMBOL(bio_integrity_advance);
 /**
  * bio_integrity_trim - Trim integrity vector
  * @bio:	bio whose integrity vector to update
- * @offset:	offset to first data sector
- * @sectors:	number of data sectors
  *
  * Description: Used to trim the integrity vector in a cloned bio.
- * The ivec will be advanced corresponding to 'offset' data sectors
- * and the length will be truncated corresponding to 'len' data
- * sectors.
  */
-void bio_integrity_trim(struct bio *bio, unsigned int offset,
+void bio_integrity_trim(struct bio *bio)
-			unsigned int sectors)
 {
 	struct bio_integrity_payload *bip = bio_integrity(bio);
 	struct blk_integrity *bi = bdev_get_integrity(bio->bi_bdev);
 
-	bio_integrity_advance(bio, offset << 9);
+	bip->bip_iter.bi_size = bio_integrity_bytes(bi, bio_sectors(bio));
-	bip->bip_iter.bi_size = bio_integrity_bytes(bi, sectors);
 }
 EXPORT_SYMBOL(bio_integrity_trim);
 

block/bio.c

@@ -243,9 +243,6 @@ fallback:
 void bio_uninit(struct bio *bio)
 {
 	bio_disassociate_task(bio);
-
-	if (bio_integrity(bio))
-		bio_integrity_free(bio);
 }
 EXPORT_SYMBOL(bio_uninit);
 
@@ -1813,6 +1810,8 @@ void bio_endio(struct bio *bio)
 again:
 	if (!bio_remaining_done(bio))
 		return;
+	if (!bio_integrity_endio(bio))
+		return;
 
 	/*
 	 * Need to have a real endio function for chained bios, otherwise
@@ -1834,6 +1833,8 @@ again:
 	}
 
 	blk_throtl_bio_endio(bio);
+	/* release cgroup info */
+	bio_uninit(bio);
 	if (bio->bi_end_io)
 		bio->bi_end_io(bio);
 }
@@ -1868,7 +1869,7 @@ struct bio *bio_split(struct bio *bio, int sectors,
 	split->bi_iter.bi_size = sectors << 9;
 
 	if (bio_integrity(split))
-		bio_integrity_trim(split, 0, sectors);
+		bio_integrity_trim(split);
 
 	bio_advance(bio, split->bi_iter.bi_size);
 
@@ -1900,6 +1901,10 @@ void bio_trim(struct bio *bio, int offset, int size)
 	bio_advance(bio, offset << 9);
 
 	bio->bi_iter.bi_size = size;
+
+	if (bio_integrity(bio))
+		bio_integrity_trim(bio);
+
 }
 EXPORT_SYMBOL_GPL(bio_trim);
 

block/blk-core.c

@@ -1787,11 +1787,8 @@ static blk_qc_t blk_queue_bio(struct request_queue *q, struct bio *bio)
 
 	blk_queue_split(q, &bio);
 
-	if (bio_integrity_enabled(bio) && bio_integrity_prep(bio)) {
+	if (!bio_integrity_prep(bio))
-		bio->bi_status = BLK_STS_IOERR;
-		bio_endio(bio);
 		return BLK_QC_T_NONE;
-	}
 
 	if (op_is_flush(bio->bi_opf)) {
 		spin_lock_irq(q->queue_lock);

block/blk-lib.c

@@ -261,6 +261,19 @@ static int __blkdev_issue_write_zeroes(struct block_device *bdev,
 	return 0;
 }
 
+/*
+ * Convert a number of 512B sectors to a number of pages.
+ * The result is limited to a number of pages that can fit into a BIO.
+ * Also make sure that the result is always at least 1 (page) for the cases
+ * where nr_sects is lower than the number of sectors in a page.
+ */
+static unsigned int __blkdev_sectors_to_bio_pages(sector_t nr_sects)
+{
+	sector_t bytes = (nr_sects << 9) + PAGE_SIZE - 1;
+
+	return min(bytes >> PAGE_SHIFT, (sector_t)BIO_MAX_PAGES);
+}
+
 /**
  * __blkdev_issue_zeroout - generate number of zero filed write bios
  * @bdev:	blockdev to issue
@@ -307,18 +320,18 @@ int __blkdev_issue_zeroout(struct block_device *bdev, sector_t sector,
 
 	ret = 0;
 	while (nr_sects != 0) {
-		bio = next_bio(bio, min(nr_sects, (sector_t)BIO_MAX_PAGES),
+		bio = next_bio(bio, __blkdev_sectors_to_bio_pages(nr_sects),
-				gfp_mask);
+			       gfp_mask);
 		bio->bi_iter.bi_sector = sector;
 		bio->bi_bdev   = bdev;
 		bio_set_op_attrs(bio, REQ_OP_WRITE, 0);
 
 		while (nr_sects != 0) {
-			sz = min((sector_t) PAGE_SIZE >> 9 , nr_sects);
+			sz = min((sector_t) PAGE_SIZE, nr_sects << 9);
-			bi_size = bio_add_page(bio, ZERO_PAGE(0), sz << 9, 0);
+			bi_size = bio_add_page(bio, ZERO_PAGE(0), sz, 0);
 			nr_sects -= bi_size >> 9;
 			sector += bi_size >> 9;
-			if (bi_size < (sz << 9))
+			if (bi_size < sz)
 				break;
 		}
 		cond_resched();

block/blk-mq-sched.c

@@ -515,10 +515,12 @@ int blk_mq_init_sched(struct request_queue *q, struct elevator_type *e)
 	}
 
 	/*
-	 * Default to 256, since we don't split into sync/async like the
+	 * Default to double of smaller one between hw queue_depth and 128,
-	 * old code did. Additionally, this is a per-hw queue depth.
+	 * since we don't split into sync/async like the old code did.
+	 * Additionally, this is a per-hw queue depth.
 	 */
-	q->nr_requests = 2 * BLKDEV_MAX_RQ;
+	q->nr_requests = 2 * min_t(unsigned int, q->tag_set->queue_depth,
+				   BLKDEV_MAX_RQ);
 
 	queue_for_each_hw_ctx(q, hctx, i) {
 		ret = blk_mq_sched_alloc_tags(q, hctx, i);

block/blk-mq.c

@@ -1547,10 +1547,8 @@ static blk_qc_t blk_mq_make_request(struct request_queue *q, struct bio *bio)
 
 	blk_queue_split(q, &bio);
 
-	if (bio_integrity_enabled(bio) && bio_integrity_prep(bio)) {
+	if (!bio_integrity_prep(bio))
-		bio_io_error(bio);
 		return BLK_QC_T_NONE;
-	}
 
 	if (!is_flush_fua && !blk_queue_nomerges(q) &&
 	    blk_attempt_plug_merge(q, bio, &request_count, &same_queue_rq))

block/blk.h

@@ -81,10 +81,21 @@ static inline void blk_queue_enter_live(struct request_queue *q)
 
 #ifdef CONFIG_BLK_DEV_INTEGRITY
 void blk_flush_integrity(void);
+bool __bio_integrity_endio(struct bio *);
+static inline bool bio_integrity_endio(struct bio *bio)
+{
+	if (bio_integrity(bio))
+		return __bio_integrity_endio(bio);
+	return true;
+}
 #else
 static inline void blk_flush_integrity(void)
 {
 }
+static inline bool bio_integrity_endio(struct bio *bio)
+{
+	return true;
+}
 #endif
 
 void blk_timeout_work(struct work_struct *work);

block/t10-pi.c

@@ -28,9 +28,6 @@
 
 typedef __be16 (csum_fn) (void *, unsigned int);
 
-static const __be16 APP_ESCAPE = (__force __be16) 0xffff;
-static const __be32 REF_ESCAPE = (__force __be32) 0xffffffff;
-
 static __be16 t10_pi_crc_fn(void *data, unsigned int len)
 {
 	return cpu_to_be16(crc_t10dif(data, len));
@@ -82,7 +79,7 @@ static blk_status_t t10_pi_verify(struct blk_integrity_iter *iter,
 		switch (type) {
 		case 1:
 		case 2:
-			if (pi->app_tag == APP_ESCAPE)
+			if (pi->app_tag == T10_PI_APP_ESCAPE)
 				goto next;
 
 			if (be32_to_cpu(pi->ref_tag) !=
@@ -95,8 +92,8 @@ static blk_status_t t10_pi_verify(struct blk_integrity_iter *iter,
 			}
 			break;
 		case 3:
-			if (pi->app_tag == APP_ESCAPE &&
+			if (pi->app_tag == T10_PI_APP_ESCAPE &&
-			    pi->ref_tag == REF_ESCAPE)
+			    pi->ref_tag == T10_PI_REF_ESCAPE)
 				goto next;
 			break;
 		}

drivers/block/cciss.c

@@ -1944,6 +1944,13 @@ static void cciss_get_serial_no(ctlr_info_t *h, int logvol,
 	return;
 }
 
+static void cciss_initialize_rq(struct request *rq)
+{
+	struct scsi_request *sreq = blk_mq_rq_to_pdu(rq);
+
+	scsi_req_init(sreq);
+}
+
 /*
  * cciss_add_disk sets up the block device queue for a logical drive
  */
@@ -1956,6 +1963,7 @@ static int cciss_add_disk(ctlr_info_t *h, struct gendisk *disk,
 
 	disk->queue->cmd_size = sizeof(struct scsi_request);
 	disk->queue->request_fn = do_cciss_request;
+	disk->queue->initialize_rq_fn = cciss_initialize_rq;
 	disk->queue->queue_lock = &h->lock;
 	queue_flag_set_unlocked(QUEUE_FLAG_SCSI_PASSTHROUGH, disk->queue);
 	if (blk_init_allocated_queue(disk->queue) < 0)

drivers/block/mtip32xx/mtip32xx.c

@@ -174,7 +174,6 @@ static void mtip_init_cmd_header(struct request *rq)
 {
 	struct driver_data *dd = rq->q->queuedata;
 	struct mtip_cmd *cmd = blk_mq_rq_to_pdu(rq);
-	u32 host_cap_64 = readl(dd->mmio + HOST_CAP) & HOST_CAP_64;
 
 	/* Point the command headers at the command tables. */
 	cmd->command_header = dd->port->command_list +
@@ -182,7 +181,7 @@ static void mtip_init_cmd_header(struct request *rq)
 	cmd->command_header_dma = dd->port->command_list_dma +
 				(sizeof(struct mtip_cmd_hdr) * rq->tag);
 
-	if (host_cap_64)
+	if (test_bit(MTIP_PF_HOST_CAP_64, &dd->port->flags))
 		cmd->command_header->ctbau = __force_bit2int cpu_to_le32((cmd->command_dma >> 16) >> 16);
 
 	cmd->command_header->ctba = __force_bit2int cpu_to_le32(cmd->command_dma & 0xFFFFFFFF);
@@ -386,6 +385,7 @@ static void mtip_init_port(struct mtip_port *port)
 			 port->mmio + PORT_LST_ADDR_HI);
 		writel((port->rxfis_dma >> 16) >> 16,
 			 port->mmio + PORT_FIS_ADDR_HI);
+		set_bit(MTIP_PF_HOST_CAP_64, &port->flags);
 	}
 
 	writel(port->command_list_dma & 0xFFFFFFFF,
@@ -950,7 +950,7 @@ static int mtip_quiesce_io(struct mtip_port *port, unsigned long timeout)
 	unsigned long to;
 	bool active = true;
 
-	blk_mq_stop_hw_queues(port->dd->queue);
+	blk_mq_quiesce_queue(port->dd->queue);
 
 	to = jiffies + msecs_to_jiffies(timeout);
 	do {
@@ -970,10 +970,10 @@ static int mtip_quiesce_io(struct mtip_port *port, unsigned long timeout)
 			break;
 	} while (time_before(jiffies, to));
 
-	blk_mq_start_stopped_hw_queues(port->dd->queue, true);
+	blk_mq_unquiesce_queue(port->dd->queue);
 	return active ? -EBUSY : 0;
 err_fault:
-	blk_mq_start_stopped_hw_queues(port->dd->queue, true);
+	blk_mq_unquiesce_queue(port->dd->queue);
 	return -EFAULT;
 }
 
@@ -2737,6 +2737,9 @@ static void mtip_abort_cmd(struct request *req, void *data,
 	struct mtip_cmd *cmd = blk_mq_rq_to_pdu(req);
 	struct driver_data *dd = data;
 
+	if (!blk_mq_request_started(req))
+		return;
+
 	dbg_printk(MTIP_DRV_NAME " Aborting request, tag = %d\n", req->tag);
 
 	clear_bit(req->tag, dd->port->cmds_to_issue);
@@ -2749,6 +2752,9 @@ static void mtip_queue_cmd(struct request *req, void *data,
 {
 	struct driver_data *dd = data;
 
+	if (!blk_mq_request_started(req))
+		return;
+
 	set_bit(req->tag, dd->port->cmds_to_issue);
 	blk_abort_request(req);
 }
@@ -2814,6 +2820,8 @@ restart_eh:
 				dev_warn(&dd->pdev->dev,
 					"Completion workers still active!");
 
+			blk_mq_quiesce_queue(dd->queue);
+
 			spin_lock(dd->queue->queue_lock);
 			blk_mq_tagset_busy_iter(&dd->tags,
 							mtip_queue_cmd, dd);
@@ -2826,6 +2834,8 @@ restart_eh:
 							mtip_abort_cmd, dd);
 
 			clear_bit(MTIP_PF_TO_ACTIVE_BIT, &dd->port->flags);
+
+			blk_mq_unquiesce_queue(dd->queue);
 		}
 
 		if (test_bit(MTIP_PF_ISSUE_CMDS_BIT, &port->flags)) {
@@ -3995,8 +4005,9 @@ static int mtip_block_remove(struct driver_data *dd)
 						dd->disk->disk_name);
 
 	blk_freeze_queue_start(dd->queue);
-	blk_mq_stop_hw_queues(dd->queue);
+	blk_mq_quiesce_queue(dd->queue);
 	blk_mq_tagset_busy_iter(&dd->tags, mtip_no_dev_cleanup, dd);
+	blk_mq_unquiesce_queue(dd->queue);
 
 	/*
 	 * Delete our gendisk structure. This also removes the device

drivers/block/mtip32xx/mtip32xx.h

@@ -140,6 +140,7 @@ enum {
 				(1 << MTIP_PF_SE_ACTIVE_BIT) |
 				(1 << MTIP_PF_DM_ACTIVE_BIT) |
 				(1 << MTIP_PF_TO_ACTIVE_BIT)),
+	MTIP_PF_HOST_CAP_64         = 10, /* cache HOST_CAP_64 */
 
 	MTIP_PF_SVC_THD_ACTIVE_BIT  = 4,
 	MTIP_PF_ISSUE_CMDS_BIT      = 5,

drivers/block/nbd.c

@@ -661,9 +661,9 @@ static void nbd_clear_req(struct request *req, void *data, bool reserved)
 
 static void nbd_clear_que(struct nbd_device *nbd)
 {
-	blk_mq_stop_hw_queues(nbd->disk->queue);
+	blk_mq_quiesce_queue(nbd->disk->queue);
 	blk_mq_tagset_busy_iter(&nbd->tag_set, nbd_clear_req, NULL);
-	blk_mq_start_hw_queues(nbd->disk->queue);
+	blk_mq_unquiesce_queue(nbd->disk->queue);
 	dev_dbg(disk_to_dev(nbd->disk), "queue cleared\n");
 }
 

drivers/block/null_blk.c

@@ -844,9 +844,6 @@ static int __init null_init(void)
 		queue_mode = NULL_Q_MQ;
 	}
 
-	if (queue_mode == NULL_Q_MQ && shared_tags)
-		null_init_tag_set(&tag_set);
-
 	if (queue_mode == NULL_Q_MQ && use_per_node_hctx) {
 		if (submit_queues < nr_online_nodes) {
 			pr_warn("null_blk: submit_queues param is set to %u.",
@@ -858,11 +855,19 @@ static int __init null_init(void)
 	else if (!submit_queues)
 		submit_queues = 1;
 
+	if (queue_mode == NULL_Q_MQ && shared_tags) {
+		ret = null_init_tag_set(&tag_set);
+		if (ret)
+			return ret;
+	}
+
 	mutex_init(&lock);
 
 	null_major = register_blkdev(0, "nullb");
-	if (null_major < 0)
+	if (null_major < 0) {
-		return null_major;
+		ret = null_major;
+		goto err_tagset;
+	}
 
 	if (use_lightnvm) {
 		ppa_cache = kmem_cache_create("ppa_cache", 64 * sizeof(u64),
@@ -891,6 +896,9 @@ err_dev:
 	kmem_cache_destroy(ppa_cache);
 err_ppa:
 	unregister_blkdev(null_major, "nullb");
+err_tagset:
+	if (queue_mode == NULL_Q_MQ && shared_tags)
+		blk_mq_free_tag_set(&tag_set);
 	return ret;
 }
 

drivers/block/virtio_blk.c

@@ -840,7 +840,7 @@ static int virtblk_freeze(struct virtio_device *vdev)
 	/* Make sure no work handler is accessing the device. */
 	flush_work(&vblk->config_work);
 
-	blk_mq_stop_hw_queues(vblk->disk->queue);
+	blk_mq_quiesce_queue(vblk->disk->queue);
 
 	vdev->config->del_vqs(vdev);
 	return 0;
@@ -857,7 +857,7 @@ static int virtblk_restore(struct virtio_device *vdev)
 
 	virtio_device_ready(vdev);
 
-	blk_mq_start_stopped_hw_queues(vblk->disk->queue, true);
+	blk_mq_unquiesce_queue(vblk->disk->queue);
 	return 0;
 }
 #endif

drivers/lightnvm/pblk-core.c

@@ -1670,13 +1670,10 @@ void pblk_line_run_ws(struct pblk *pblk, struct pblk_line *line, void *priv,
 	queue_work(wq, &line_ws->ws);
 }
 
-void pblk_down_rq(struct pblk *pblk, struct ppa_addr *ppa_list, int nr_ppas,
+static void __pblk_down_page(struct pblk *pblk, struct ppa_addr *ppa_list,
-		  unsigned long *lun_bitmap)
+			     int nr_ppas, int pos)
 {
-	struct nvm_tgt_dev *dev = pblk->dev;
+	struct pblk_lun *rlun = &pblk->luns[pos];
-	struct nvm_geo *geo = &dev->geo;
-	struct pblk_lun *rlun;
-	int pos = pblk_ppa_to_pos(geo, ppa_list[0]);
 	int ret;
 
 	/*
@@ -1690,14 +1687,8 @@ void pblk_down_rq(struct pblk *pblk, struct ppa_addr *ppa_list, int nr_ppas,
 		WARN_ON(ppa_list[0].g.lun != ppa_list[i].g.lun ||
 				ppa_list[0].g.ch != ppa_list[i].g.ch);
 #endif
-	/* If the LUN has been locked for this same request, do no attempt to
-	 * lock it again
-	 */
-	if (test_and_set_bit(pos, lun_bitmap))
-		return;
 
-	rlun = &pblk->luns[pos];
+	ret = down_timeout(&rlun->wr_sem, msecs_to_jiffies(30000));
-	ret = down_timeout(&rlun->wr_sem, msecs_to_jiffies(5000));
 	if (ret) {
 		switch (ret) {
 		case -ETIME:
@@ -1710,6 +1701,50 @@ void pblk_down_rq(struct pblk *pblk, struct ppa_addr *ppa_list, int nr_ppas,
 	}
 }
 
+void pblk_down_page(struct pblk *pblk, struct ppa_addr *ppa_list, int nr_ppas)
+{
+	struct nvm_tgt_dev *dev = pblk->dev;
+	struct nvm_geo *geo = &dev->geo;
+	int pos = pblk_ppa_to_pos(geo, ppa_list[0]);
+
+	__pblk_down_page(pblk, ppa_list, nr_ppas, pos);
+}
+
+void pblk_down_rq(struct pblk *pblk, struct ppa_addr *ppa_list, int nr_ppas,
+		  unsigned long *lun_bitmap)
+{
+	struct nvm_tgt_dev *dev = pblk->dev;
+	struct nvm_geo *geo = &dev->geo;
+	int pos = pblk_ppa_to_pos(geo, ppa_list[0]);
+
+	/* If the LUN has been locked for this same request, do no attempt to
+	 * lock it again
+	 */
+	if (test_and_set_bit(pos, lun_bitmap))
+		return;
+
+	__pblk_down_page(pblk, ppa_list, nr_ppas, pos);
+}
+
+void pblk_up_page(struct pblk *pblk, struct ppa_addr *ppa_list, int nr_ppas)
+{
+	struct nvm_tgt_dev *dev = pblk->dev;
+	struct nvm_geo *geo = &dev->geo;
+	struct pblk_lun *rlun;
+	int pos = pblk_ppa_to_pos(geo, ppa_list[0]);
+
+#ifdef CONFIG_NVM_DEBUG
+	int i;
+
+	for (i = 1; i < nr_ppas; i++)
+		WARN_ON(ppa_list[0].g.lun != ppa_list[i].g.lun ||
+				ppa_list[0].g.ch != ppa_list[i].g.ch);
+#endif
+
+	rlun = &pblk->luns[pos];
+	up(&rlun->wr_sem);
+}
+
 void pblk_up_rq(struct pblk *pblk, struct ppa_addr *ppa_list, int nr_ppas,
 		unsigned long *lun_bitmap)
 {

drivers/lightnvm/pblk-recovery.c

@@ -340,9 +340,14 @@ static void pblk_end_io_recov(struct nvm_rq *rqd)
 	struct pblk *pblk = pad_rq->pblk;
 	struct nvm_tgt_dev *dev = pblk->dev;
 
-	kref_put(&pad_rq->ref, pblk_recov_complete);
+	pblk_up_page(pblk, rqd->ppa_list, rqd->nr_ppas);
+
+	bio_put(rqd->bio);
 	nvm_dev_dma_free(dev->parent, rqd->meta_list, rqd->dma_meta_list);
 	pblk_free_rqd(pblk, rqd, WRITE);
+
+	atomic_dec(&pblk->inflight_io);
+	kref_put(&pad_rq->ref, pblk_recov_complete);
 }
 
 static int pblk_recov_pad_oob(struct pblk *pblk, struct pblk_line *line,
@@ -385,7 +390,7 @@ next_pad_rq:
 	rq_ppas = pblk_calc_secs(pblk, left_ppas, 0);
 	if (rq_ppas < pblk->min_write_pgs) {
 		pr_err("pblk: corrupted pad line %d\n", line->id);
-		goto free_rq;
+		goto fail_free_pad;
 	}
 
 	rq_len = rq_ppas * geo->sec_size;
@@ -393,7 +398,7 @@ next_pad_rq:
 	meta_list = nvm_dev_dma_alloc(dev->parent, GFP_KERNEL, &dma_meta_list);
 	if (!meta_list) {
 		ret = -ENOMEM;
-		goto free_data;
+		goto fail_free_pad;
 	}
 
 	ppa_list = (void *)(meta_list) + pblk_dma_meta_size;
@@ -404,9 +409,9 @@ next_pad_rq:
 		ret = PTR_ERR(rqd);
 		goto fail_free_meta;
 	}
-	memset(rqd, 0, pblk_w_rq_size);
 
-	bio = bio_map_kern(dev->q, data, rq_len, GFP_KERNEL);
+	bio = pblk_bio_map_addr(pblk, data, rq_ppas, rq_len,
+						PBLK_VMALLOC_META, GFP_KERNEL);
 	if (IS_ERR(bio)) {
 		ret = PTR_ERR(bio);
 		goto fail_free_rqd;
@@ -453,15 +458,15 @@ next_pad_rq:
 	}
 
 	kref_get(&pad_rq->ref);
+	pblk_down_page(pblk, rqd->ppa_list, rqd->nr_ppas);
 
 	ret = pblk_submit_io(pblk, rqd);
 	if (ret) {
 		pr_err("pblk: I/O submission failed: %d\n", ret);
-		goto free_data;
+		pblk_up_page(pblk, rqd->ppa_list, rqd->nr_ppas);
+		goto fail_free_bio;
 	}
 
-	atomic_dec(&pblk->inflight_io);
-
 	left_line_ppas -= rq_ppas;
 	left_ppas -= rq_ppas;
 	if (left_ppas && left_line_ppas)
@@ -475,17 +480,23 @@ next_pad_rq:
 		ret = -ETIME;
 	}
 
+	if (!pblk_line_is_full(line))
+		pr_err("pblk: corrupted padded line: %d\n", line->id);
+
+	vfree(data);
 free_rq:
 	kfree(pad_rq);
-free_data:
-	vfree(data);
 	return ret;
 
+fail_free_bio:
+	bio_put(bio);
 fail_free_rqd:
 	pblk_free_rqd(pblk, rqd, WRITE);
 fail_free_meta:
 	nvm_dev_dma_free(dev->parent, meta_list, dma_meta_list);
+fail_free_pad:
 	kfree(pad_rq);
+	vfree(data);
 	return ret;
 }
 

drivers/lightnvm/pblk-write.c

@@ -39,9 +39,7 @@ static unsigned long pblk_end_w_bio(struct pblk *pblk, struct nvm_rq *rqd,
 
 	ret = pblk_rb_sync_advance(&pblk->rwb, c_ctx->nr_valid);
 
-	if (rqd->meta_list)
+	nvm_dev_dma_free(dev->parent, rqd->meta_list, rqd->dma_meta_list);
-		nvm_dev_dma_free(dev->parent, rqd->meta_list,
-							rqd->dma_meta_list);
 
 	bio_put(rqd->bio);
 	pblk_free_rqd(pblk, rqd, WRITE);
@@ -178,15 +176,12 @@ static void pblk_end_io_write_meta(struct nvm_rq *rqd)
 {
 	struct pblk *pblk = rqd->private;
 	struct nvm_tgt_dev *dev = pblk->dev;
-	struct nvm_geo *geo = &dev->geo;
 	struct pblk_g_ctx *m_ctx = nvm_rq_to_pdu(rqd);
 	struct pblk_line *line = m_ctx->private;
 	struct pblk_emeta *emeta = line->emeta;
-	int pos = pblk_ppa_to_pos(geo, rqd->ppa_list[0]);
-	struct pblk_lun *rlun = &pblk->luns[pos];
 	int sync;
 
-	up(&rlun->wr_sem);
+	pblk_up_page(pblk, rqd->ppa_list, rqd->nr_ppas);
 
 	if (rqd->error) {
 		pblk_log_write_err(pblk, rqd);
@@ -203,6 +198,7 @@ static void pblk_end_io_write_meta(struct nvm_rq *rqd)
 								pblk->close_wq);
 
 	bio_put(rqd->bio);
+	nvm_dev_dma_free(dev->parent, rqd->meta_list, rqd->dma_meta_list);
 	pblk_free_rqd(pblk, rqd, READ);
 
 	atomic_dec(&pblk->inflight_io);
@@ -226,9 +222,6 @@ static int pblk_alloc_w_rq(struct pblk *pblk, struct nvm_rq *rqd,
 	if (!rqd->meta_list)
 		return -ENOMEM;
 
-	if (unlikely(nr_secs == 1))
-		return 0;
-
 	rqd->ppa_list = rqd->meta_list + pblk_dma_meta_size;
 	rqd->dma_ppa_list = rqd->dma_meta_list + pblk_dma_meta_size;
 
@@ -367,7 +360,6 @@ int pblk_submit_meta_io(struct pblk *pblk, struct pblk_line *meta_line)
 	struct pblk_line_meta *lm = &pblk->lm;
 	struct pblk_emeta *emeta = meta_line->emeta;
 	struct pblk_g_ctx *m_ctx;
-	struct pblk_lun *rlun;
 	struct bio *bio;
 	struct nvm_rq *rqd;
 	void *data;
@@ -411,13 +403,6 @@ int pblk_submit_meta_io(struct pblk *pblk, struct pblk_line *meta_line)
 			rqd->ppa_list[i] = addr_to_gen_ppa(pblk, paddr, id);
 	}
 
-	rlun = &pblk->luns[pblk_ppa_to_pos(geo, rqd->ppa_list[0])];
-	ret = down_timeout(&rlun->wr_sem, msecs_to_jiffies(5000));
-	if (ret) {
-		pr_err("pblk: lun semaphore timed out (%d)\n", ret);
-		goto fail_free_bio;
-	}
-
 	emeta->mem += rq_len;
 	if (emeta->mem >= lm->emeta_len[0]) {
 		spin_lock(&l_mg->close_lock);
@@ -427,6 +412,8 @@ int pblk_submit_meta_io(struct pblk *pblk, struct pblk_line *meta_line)
 		spin_unlock(&l_mg->close_lock);
 	}
 
+	pblk_down_page(pblk, rqd->ppa_list, rqd->nr_ppas);
+
 	ret = pblk_submit_io(pblk, rqd);
 	if (ret) {
 		pr_err("pblk: emeta I/O submission failed: %d\n", ret);
@@ -436,10 +423,13 @@ int pblk_submit_meta_io(struct pblk *pblk, struct pblk_line *meta_line)
 	return NVM_IO_OK;
 
 fail_rollback:
+	pblk_up_page(pblk, rqd->ppa_list, rqd->nr_ppas);
 	spin_lock(&l_mg->close_lock);
 	pblk_dealloc_page(pblk, meta_line, rq_ppas);
 	list_add(&meta_line->list, &meta_line->list);
 	spin_unlock(&l_mg->close_lock);
+
+	nvm_dev_dma_free(dev->parent, rqd->meta_list, rqd->dma_meta_list);
 fail_free_bio:
 	if (likely(l_mg->emeta_alloc_type == PBLK_VMALLOC_META))
 		bio_put(bio);

drivers/lightnvm/pblk.h

@@ -739,8 +739,10 @@ u64 pblk_alloc_page(struct pblk *pblk, struct pblk_line *line, int nr_secs);
 u64 __pblk_alloc_page(struct pblk *pblk, struct pblk_line *line, int nr_secs);
 int pblk_calc_secs(struct pblk *pblk, unsigned long secs_avail,
 		   unsigned long secs_to_flush);
+void pblk_up_page(struct pblk *pblk, struct ppa_addr *ppa_list, int nr_ppas);
 void pblk_down_rq(struct pblk *pblk, struct ppa_addr *ppa_list, int nr_ppas,
 		  unsigned long *lun_bitmap);
+void pblk_down_page(struct pblk *pblk, struct ppa_addr *ppa_list, int nr_ppas);
 void pblk_up_rq(struct pblk *pblk, struct ppa_addr *ppa_list, int nr_ppas,
 		unsigned long *lun_bitmap);
 void pblk_end_bio_sync(struct bio *bio);

drivers/md/dm.c

@@ -1279,7 +1279,7 @@ static int clone_bio(struct dm_target_io *tio, struct bio *bio,
 	clone->bi_iter.bi_size = to_bytes(len);
 
 	if (unlikely(bio_integrity(bio) != NULL))
-		bio_integrity_trim(clone, 0, len);
+		bio_integrity_trim(clone);
 
 	return 0;
 }

drivers/nvdimm/blk.c

@@ -106,7 +106,8 @@ static int nd_blk_rw_integrity(struct nd_namespace_blk *nsblk,
 
 		len -= cur_len;
 		dev_offset += cur_len;
-		bvec_iter_advance(bip->bip_vec, &bip->bip_iter, cur_len);
+		if (!bvec_iter_advance(bip->bip_vec, &bip->bip_iter, cur_len))
+			return -EIO;
 	}
 
 	return err;
@@ -179,16 +180,8 @@ static blk_qc_t nd_blk_make_request(struct request_queue *q, struct bio *bio)
 	int err = 0, rw;
 	bool do_acct;
 
-	/*
+	if (!bio_integrity_prep(bio))
-	 * bio_integrity_enabled also checks if the bio already has an
+		return BLK_QC_T_NONE;
-	 * integrity payload attached. If it does, we *don't* do a
-	 * bio_integrity_prep here - the payload has been generated by
-	 * another kernel subsystem, and we just pass it through.
-	 */
-	if (bio_integrity_enabled(bio) && bio_integrity_prep(bio)) {
-		bio->bi_status = BLK_STS_IOERR;
-		goto out;
-	}
 
 	bip = bio_integrity(bio);
 	nsblk = q->queuedata;
@@ -212,7 +205,6 @@ static blk_qc_t nd_blk_make_request(struct request_queue *q, struct bio *bio)
 	if (do_acct)
 		nd_iostat_end(bio, start);
 
- out:
 	bio_endio(bio);
 	return BLK_QC_T_NONE;
 }

drivers/nvdimm/btt.c

@@ -985,7 +985,8 @@ static int btt_rw_integrity(struct btt *btt, struct bio_integrity_payload *bip,
 
 		len -= cur_len;
 		meta_nsoff += cur_len;
-		bvec_iter_advance(bip->bip_vec, &bip->bip_iter, cur_len);
+		if (!bvec_iter_advance(bip->bip_vec, &bip->bip_iter, cur_len))
+			return -EIO;
 	}
 
 	return ret;
@@ -1203,16 +1204,8 @@ static blk_qc_t btt_make_request(struct request_queue *q, struct bio *bio)
 	int err = 0;
 	bool do_acct;
 
-	/*
+	if (!bio_integrity_prep(bio))
-	 * bio_integrity_enabled also checks if the bio already has an
+		return BLK_QC_T_NONE;
-	 * integrity payload attached. If it does, we *don't* do a
-	 * bio_integrity_prep here - the payload has been generated by
-	 * another kernel subsystem, and we just pass it through.
-	 */
-	if (bio_integrity_enabled(bio) && bio_integrity_prep(bio)) {
-		bio->bi_status = BLK_STS_IOERR;
-		goto out;
-	}
 
 	do_acct = nd_iostat_start(bio, &start);
 	bio_for_each_segment(bvec, bio, iter) {
@@ -1239,7 +1232,6 @@ static blk_qc_t btt_make_request(struct request_queue *q, struct bio *bio)
 	if (do_acct)
 		nd_iostat_end(bio, start);
 
-out:
 	bio_endio(bio);
 	return BLK_QC_T_NONE;
 }

drivers/nvme/host/core.c

@@ -131,7 +131,7 @@ void nvme_complete_rq(struct request *req)
 {
 	if (unlikely(nvme_req(req)->status && nvme_req_needs_retry(req))) {
 		nvme_req(req)->retries++;
-		blk_mq_requeue_request(req, !blk_mq_queue_stopped(req->q));
+		blk_mq_requeue_request(req, true);
 		return;
 	}
 
@@ -2591,12 +2591,29 @@ static void nvme_release_instance(struct nvme_ctrl *ctrl)
 	spin_unlock(&dev_list_lock);
 }
 
-void nvme_uninit_ctrl(struct nvme_ctrl *ctrl)
+void nvme_stop_ctrl(struct nvme_ctrl *ctrl)
 {
+	nvme_stop_keep_alive(ctrl);
 	flush_work(&ctrl->async_event_work);
 	flush_work(&ctrl->scan_work);
-	nvme_remove_namespaces(ctrl);
+}
+EXPORT_SYMBOL_GPL(nvme_stop_ctrl);
 
+void nvme_start_ctrl(struct nvme_ctrl *ctrl)
+{
+	if (ctrl->kato)
+		nvme_start_keep_alive(ctrl);
+
+	if (ctrl->queue_count > 1) {
+		nvme_queue_scan(ctrl);
+		nvme_queue_async_events(ctrl);
+		nvme_start_queues(ctrl);
+	}
+}
+EXPORT_SYMBOL_GPL(nvme_start_ctrl);
+
+void nvme_uninit_ctrl(struct nvme_ctrl *ctrl)
+{
 	device_destroy(nvme_class, MKDEV(nvme_char_major, ctrl->instance));
 
 	spin_lock(&dev_list_lock);
@@ -2694,9 +2711,6 @@ void nvme_kill_queues(struct nvme_ctrl *ctrl)
 	/* Forcibly unquiesce queues to avoid blocking dispatch */
 	blk_mq_unquiesce_queue(ctrl->admin_q);
 
-	/* Forcibly start all queues to avoid having stuck requests */
-	blk_mq_start_hw_queues(ctrl->admin_q);
-
 	list_for_each_entry(ns, &ctrl->namespaces, list) {
 		/*
 		 * Revalidating a dead namespace sets capacity to 0. This will
@@ -2709,16 +2723,6 @@ void nvme_kill_queues(struct nvme_ctrl *ctrl)
 
 		/* Forcibly unquiesce queues to avoid blocking dispatch */
 		blk_mq_unquiesce_queue(ns->queue);
-
-		/*
-		 * Forcibly start all queues to avoid having stuck requests.
-		 * Note that we must ensure the queues are not stopped
-		 * when the final removal happens.
-		 */
-		blk_mq_start_hw_queues(ns->queue);
-
-		/* draining requests in requeue list */
-		blk_mq_kick_requeue_list(ns->queue);
 	}
 	mutex_unlock(&ctrl->namespaces_mutex);
 }
@@ -2787,10 +2791,8 @@ void nvme_start_queues(struct nvme_ctrl *ctrl)
 	struct nvme_ns *ns;
 
 	mutex_lock(&ctrl->namespaces_mutex);
-	list_for_each_entry(ns, &ctrl->namespaces, list) {
+	list_for_each_entry(ns, &ctrl->namespaces, list)
 		blk_mq_unquiesce_queue(ns->queue);
-		blk_mq_kick_requeue_list(ns->queue);
-	}
 	mutex_unlock(&ctrl->namespaces_mutex);
 }
 EXPORT_SYMBOL_GPL(nvme_start_queues);

drivers/nvme/host/fc.c

@@ -148,13 +148,10 @@ struct nvme_fc_ctrl {
 	struct device		*dev;
 	struct nvme_fc_lport	*lport;
 	struct nvme_fc_rport	*rport;
-	u32			queue_count;
 	u32			cnum;
 
 	u64			association_id;
 
-	u64			cap;
-
 	struct list_head	ctrl_list;	/* rport->ctrl_list */
 
 	struct blk_mq_tag_set	admin_tag_set;
@@ -1614,7 +1611,7 @@ nvme_fc_free_io_queues(struct nvme_fc_ctrl *ctrl)
 {
 	int i;
 
-	for (i = 1; i < ctrl->queue_count; i++)
+	for (i = 1; i < ctrl->ctrl.queue_count; i++)
 		nvme_fc_free_queue(&ctrl->queues[i]);
 }
 
@@ -1635,10 +1632,10 @@ __nvme_fc_create_hw_queue(struct nvme_fc_ctrl *ctrl,
 static void
 nvme_fc_delete_hw_io_queues(struct nvme_fc_ctrl *ctrl)
 {
-	struct nvme_fc_queue *queue = &ctrl->queues[ctrl->queue_count - 1];
+	struct nvme_fc_queue *queue = &ctrl->queues[ctrl->ctrl.queue_count - 1];
 	int i;
 
-	for (i = ctrl->queue_count - 1; i >= 1; i--, queue--)
+	for (i = ctrl->ctrl.queue_count - 1; i >= 1; i--, queue--)
 		__nvme_fc_delete_hw_queue(ctrl, queue, i);
 }
 
@@ -1648,7 +1645,7 @@ nvme_fc_create_hw_io_queues(struct nvme_fc_ctrl *ctrl, u16 qsize)
 	struct nvme_fc_queue *queue = &ctrl->queues[1];
 	int i, ret;
 
-	for (i = 1; i < ctrl->queue_count; i++, queue++) {
+	for (i = 1; i < ctrl->ctrl.queue_count; i++, queue++) {
 		ret = __nvme_fc_create_hw_queue(ctrl, queue, i, qsize);
 		if (ret)
 			goto delete_queues;
@@ -1667,7 +1664,7 @@ nvme_fc_connect_io_queues(struct nvme_fc_ctrl *ctrl, u16 qsize)
 {
 	int i, ret = 0;
 
-	for (i = 1; i < ctrl->queue_count; i++) {
+	for (i = 1; i < ctrl->ctrl.queue_count; i++) {
 		ret = nvme_fc_connect_queue(ctrl, &ctrl->queues[i], qsize,
 					(qsize / 5));
 		if (ret)
@@ -1685,7 +1682,7 @@ nvme_fc_init_io_queues(struct nvme_fc_ctrl *ctrl)
 {
 	int i;
 
-	for (i = 1; i < ctrl->queue_count; i++)
+	for (i = 1; i < ctrl->ctrl.queue_count; i++)
 		nvme_fc_init_queue(ctrl, i, ctrl->ctrl.sqsize);
 }
 
@@ -1706,6 +1703,7 @@ nvme_fc_ctrl_free(struct kref *ref)
 	list_del(&ctrl->ctrl_list);
 	spin_unlock_irqrestore(&ctrl->rport->lock, flags);
 
+	blk_mq_unquiesce_queue(ctrl->ctrl.admin_q);
 	blk_cleanup_queue(ctrl->ctrl.admin_q);
 	blk_mq_free_tag_set(&ctrl->admin_tag_set);
 
@@ -1969,10 +1967,9 @@ nvme_fc_start_fcp_op(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue,
 		if (ret != -EBUSY)
 			return BLK_STS_IOERR;
 
-		if (op->rq) {
+		if (op->rq)
-			blk_mq_stop_hw_queues(op->rq->q);
+			blk_mq_delay_run_hw_queue(queue->hctx, NVMEFC_QUEUE_DELAY);
-			blk_mq_delay_queue(queue->hctx, NVMEFC_QUEUE_DELAY);
+
-		}
 		return BLK_STS_RESOURCE;
 	}
 
@@ -2178,17 +2175,20 @@ static int
 nvme_fc_create_io_queues(struct nvme_fc_ctrl *ctrl)
 {
 	struct nvmf_ctrl_options *opts = ctrl->ctrl.opts;
+	unsigned int nr_io_queues;
 	int ret;
 
-	ret = nvme_set_queue_count(&ctrl->ctrl, &opts->nr_io_queues);
+	nr_io_queues = min(min(opts->nr_io_queues, num_online_cpus()),
+				ctrl->lport->ops->max_hw_queues);
+	ret = nvme_set_queue_count(&ctrl->ctrl, &nr_io_queues);
 	if (ret) {
 		dev_info(ctrl->ctrl.device,
 			"set_queue_count failed: %d\n", ret);
 		return ret;
 	}
 
-	ctrl->queue_count = opts->nr_io_queues + 1;
+	ctrl->ctrl.queue_count = nr_io_queues + 1;
-	if (!opts->nr_io_queues)
+	if (!nr_io_queues)
 		return 0;
 
 	nvme_fc_init_io_queues(ctrl);
@@ -2204,7 +2204,7 @@ nvme_fc_create_io_queues(struct nvme_fc_ctrl *ctrl)
 						sizeof(struct scatterlist)) +
 					ctrl->lport->ops->fcprqst_priv_sz;
 	ctrl->tag_set.driver_data = ctrl;
-	ctrl->tag_set.nr_hw_queues = ctrl->queue_count - 1;
+	ctrl->tag_set.nr_hw_queues = ctrl->ctrl.queue_count - 1;
 	ctrl->tag_set.timeout = NVME_IO_TIMEOUT;
 
 	ret = blk_mq_alloc_tag_set(&ctrl->tag_set);
@@ -2232,7 +2232,6 @@ nvme_fc_create_io_queues(struct nvme_fc_ctrl *ctrl)
 out_delete_hw_queues:
 	nvme_fc_delete_hw_io_queues(ctrl);
 out_cleanup_blk_queue:
-	nvme_stop_keep_alive(&ctrl->ctrl);
 	blk_cleanup_queue(ctrl->ctrl.connect_q);
 out_free_tag_set:
 	blk_mq_free_tag_set(&ctrl->tag_set);
@@ -2248,17 +2247,21 @@ static int
 nvme_fc_reinit_io_queues(struct nvme_fc_ctrl *ctrl)
 {
 	struct nvmf_ctrl_options *opts = ctrl->ctrl.opts;
+	unsigned int nr_io_queues;
 	int ret;
 
-	ret = nvme_set_queue_count(&ctrl->ctrl, &opts->nr_io_queues);
+	nr_io_queues = min(min(opts->nr_io_queues, num_online_cpus()),
+				ctrl->lport->ops->max_hw_queues);
+	ret = nvme_set_queue_count(&ctrl->ctrl, &nr_io_queues);
 	if (ret) {
 		dev_info(ctrl->ctrl.device,
 			"set_queue_count failed: %d\n", ret);
 		return ret;
 	}
 
+	ctrl->ctrl.queue_count = nr_io_queues + 1;
 	/* check for io queues existing */
-	if (ctrl->queue_count == 1)
+	if (ctrl->ctrl.queue_count == 1)
 		return 0;
 
 	nvme_fc_init_io_queues(ctrl);
@@ -2275,6 +2278,8 @@ nvme_fc_reinit_io_queues(struct nvme_fc_ctrl *ctrl)
 	if (ret)
 		goto out_delete_hw_queues;
 
+	blk_mq_update_nr_hw_queues(&ctrl->tag_set, nr_io_queues);
+
 	return 0;
 
 out_delete_hw_queues:
@@ -2316,7 +2321,7 @@ nvme_fc_create_association(struct nvme_fc_ctrl *ctrl)
 		goto out_delete_hw_queue;
 
 	if (ctrl->ctrl.state != NVME_CTRL_NEW)
-		blk_mq_start_stopped_hw_queues(ctrl->ctrl.admin_q, true);
+		blk_mq_unquiesce_queue(ctrl->ctrl.admin_q);
 
 	ret = nvmf_connect_admin_queue(&ctrl->ctrl);
 	if (ret)
@@ -2329,7 +2334,7 @@ nvme_fc_create_association(struct nvme_fc_ctrl *ctrl)
 	 * prior connection values
 	 */
 
-	ret = nvmf_reg_read64(&ctrl->ctrl, NVME_REG_CAP, &ctrl->cap);
+	ret = nvmf_reg_read64(&ctrl->ctrl, NVME_REG_CAP, &ctrl->ctrl.cap);
 	if (ret) {
 		dev_err(ctrl->ctrl.device,
 			"prop_get NVME_REG_CAP failed\n");
@@ -2337,9 +2342,9 @@ nvme_fc_create_association(struct nvme_fc_ctrl *ctrl)
 	}
 
 	ctrl->ctrl.sqsize =
-		min_t(int, NVME_CAP_MQES(ctrl->cap) + 1, ctrl->ctrl.sqsize);
+		min_t(int, NVME_CAP_MQES(ctrl->ctrl.cap) + 1, ctrl->ctrl.sqsize);
 
-	ret = nvme_enable_ctrl(&ctrl->ctrl, ctrl->cap);
+	ret = nvme_enable_ctrl(&ctrl->ctrl, ctrl->ctrl.cap);
 	if (ret)
 		goto out_disconnect_admin_queue;
 
@@ -2360,8 +2365,6 @@ nvme_fc_create_association(struct nvme_fc_ctrl *ctrl)
 		goto out_disconnect_admin_queue;
 	}
 
-	nvme_start_keep_alive(&ctrl->ctrl);
-
 	/* FC-NVME supports normal SGL Data Block Descriptors */
 
 	if (opts->queue_size > ctrl->ctrl.maxcmd) {
@@ -2381,7 +2384,7 @@ nvme_fc_create_association(struct nvme_fc_ctrl *ctrl)
 	 * Create the io queues
 	 */
 
-	if (ctrl->queue_count > 1) {
+	if (ctrl->ctrl.queue_count > 1) {
 		if (ctrl->ctrl.state == NVME_CTRL_NEW)
 			ret = nvme_fc_create_io_queues(ctrl);
 		else
@@ -2395,17 +2398,12 @@ nvme_fc_create_association(struct nvme_fc_ctrl *ctrl)
 
 	ctrl->ctrl.nr_reconnects = 0;
 
-	if (ctrl->queue_count > 1) {
+	nvme_start_ctrl(&ctrl->ctrl);
-		nvme_start_queues(&ctrl->ctrl);
-		nvme_queue_scan(&ctrl->ctrl);
-		nvme_queue_async_events(&ctrl->ctrl);
-	}
 
 	return 0;	/* Success */
 
 out_term_aen_ops:
 	nvme_fc_term_aen_ops(ctrl);
-	nvme_stop_keep_alive(&ctrl->ctrl);
 out_disconnect_admin_queue:
 	/* send a Disconnect(association) LS to fc-nvme target */
 	nvme_fc_xmt_disconnect_assoc(ctrl);
@@ -2428,8 +2426,6 @@ nvme_fc_delete_association(struct nvme_fc_ctrl *ctrl)
 {
 	unsigned long flags;
 
-	nvme_stop_keep_alive(&ctrl->ctrl);
-
 	spin_lock_irqsave(&ctrl->lock, flags);
 	ctrl->flags |= FCCTRL_TERMIO;
 	ctrl->iocnt = 0;
@@ -2447,7 +2443,7 @@ nvme_fc_delete_association(struct nvme_fc_ctrl *ctrl)
 	 * io requests back to the block layer as part of normal completions
 	 * (but with error status).
 	 */
-	if (ctrl->queue_count > 1) {
+	if (ctrl->ctrl.queue_count > 1) {
 		nvme_stop_queues(&ctrl->ctrl);
 		blk_mq_tagset_busy_iter(&ctrl->tag_set,
 				nvme_fc_terminate_exchange, &ctrl->ctrl);
@@ -2470,7 +2466,7 @@ nvme_fc_delete_association(struct nvme_fc_ctrl *ctrl)
 	 * use blk_mq_tagset_busy_itr() and the transport routine to
 	 * terminate the exchanges.
 	 */
-	blk_mq_stop_hw_queues(ctrl->ctrl.admin_q);
+	blk_mq_quiesce_queue(ctrl->ctrl.admin_q);
 	blk_mq_tagset_busy_iter(&ctrl->admin_tag_set,
 				nvme_fc_terminate_exchange, &ctrl->ctrl);
 
@@ -2511,7 +2507,8 @@ nvme_fc_delete_ctrl_work(struct work_struct *work)
 
 	cancel_work_sync(&ctrl->ctrl.reset_work);
 	cancel_delayed_work_sync(&ctrl->connect_work);
-
+	nvme_stop_ctrl(&ctrl->ctrl);
+	nvme_remove_namespaces(&ctrl->ctrl);
 	/*
 	 * kill the association on the link side.  this will block
 	 * waiting for io to terminate
@@ -2606,6 +2603,7 @@ nvme_fc_reset_ctrl_work(struct work_struct *work)
 		container_of(work, struct nvme_fc_ctrl, ctrl.reset_work);
 	int ret;
 
+	nvme_stop_ctrl(&ctrl->ctrl);
 	/* will block will waiting for io to terminate */
 	nvme_fc_delete_association(ctrl);
 
@@ -2702,18 +2700,17 @@ nvme_fc_init_ctrl(struct device *dev, struct nvmf_ctrl_options *opts,
 	spin_lock_init(&ctrl->lock);
 
 	/* io queue count */
-	ctrl->queue_count = min_t(unsigned int,
+	ctrl->ctrl.queue_count = min_t(unsigned int,
 				opts->nr_io_queues,
 				lport->ops->max_hw_queues);
-	opts->nr_io_queues = ctrl->queue_count;	/* so opts has valid value */
+	ctrl->ctrl.queue_count++;	/* +1 for admin queue */
-	ctrl->queue_count++;	/* +1 for admin queue */
 
 	ctrl->ctrl.sqsize = opts->queue_size - 1;
 	ctrl->ctrl.kato = opts->kato;
 
 	ret = -ENOMEM;
-	ctrl->queues = kcalloc(ctrl->queue_count, sizeof(struct nvme_fc_queue),
+	ctrl->queues = kcalloc(ctrl->ctrl.queue_count,
-				GFP_KERNEL);
+				sizeof(struct nvme_fc_queue), GFP_KERNEL);
 	if (!ctrl->queues)
 		goto out_free_ida;
 

drivers/nvme/host/nvme.h

@@ -142,7 +142,9 @@ struct nvme_ctrl {
 	u16 cntlid;
 
 	u32 ctrl_config;
+	u32 queue_count;
 
+	u64 cap;
 	u32 page_size;
 	u32 max_hw_sectors;
 	u16 oncs;
@@ -278,6 +280,8 @@ int nvme_shutdown_ctrl(struct nvme_ctrl *ctrl);
 int nvme_init_ctrl(struct nvme_ctrl *ctrl, struct device *dev,
 		const struct nvme_ctrl_ops *ops, unsigned long quirks);
 void nvme_uninit_ctrl(struct nvme_ctrl *ctrl);
+void nvme_start_ctrl(struct nvme_ctrl *ctrl);
+void nvme_stop_ctrl(struct nvme_ctrl *ctrl);
 void nvme_put_ctrl(struct nvme_ctrl *ctrl);
 int nvme_init_identify(struct nvme_ctrl *ctrl);
 

drivers/nvme/host/pci.c

@@ -35,7 +35,6 @@
 
 #include "nvme.h"
 
-#define NVME_Q_DEPTH		1024
 #define SQ_SIZE(depth)		(depth * sizeof(struct nvme_command))
 #define CQ_SIZE(depth)		(depth * sizeof(struct nvme_completion))
 
@@ -57,6 +56,16 @@ module_param(max_host_mem_size_mb, uint, 0444);
 MODULE_PARM_DESC(max_host_mem_size_mb,
 	"Maximum Host Memory Buffer (HMB) size per controller (in MiB)");
 
+static int io_queue_depth_set(const char *val, const struct kernel_param *kp);
+static const struct kernel_param_ops io_queue_depth_ops = {
+	.set = io_queue_depth_set,
+	.get = param_get_int,
+};
+
+static int io_queue_depth = 1024;
+module_param_cb(io_queue_depth, &io_queue_depth_ops, &io_queue_depth, 0644);
+MODULE_PARM_DESC(io_queue_depth, "set io queue depth, should >= 2");
+
 struct nvme_dev;
 struct nvme_queue;
 
@@ -74,7 +83,6 @@ struct nvme_dev {
 	struct device *dev;
 	struct dma_pool *prp_page_pool;
 	struct dma_pool *prp_small_pool;
-	unsigned queue_count;
 	unsigned online_queues;
 	unsigned max_qid;
 	int q_depth;
@@ -105,6 +113,17 @@ struct nvme_dev {
 	void **host_mem_desc_bufs;
 };
 
+static int io_queue_depth_set(const char *val, const struct kernel_param *kp)
+{
+	int n = 0, ret;
+
+	ret = kstrtoint(val, 10, &n);
+	if (ret != 0 || n < 2)
+		return -EINVAL;
+
+	return param_set_int(val, kp);
+}
+
 static inline unsigned int sq_idx(unsigned int qid, u32 stride)
 {
 	return qid * 2 * stride;
@@ -1099,9 +1118,9 @@ static void nvme_free_queues(struct nvme_dev *dev, int lowest)
 {
 	int i;
 
-	for (i = dev->queue_count - 1; i >= lowest; i--) {
+	for (i = dev->ctrl.queue_count - 1; i >= lowest; i--) {
 		struct nvme_queue *nvmeq = dev->queues[i];
-		dev->queue_count--;
+		dev->ctrl.queue_count--;
 		dev->queues[i] = NULL;
 		nvme_free_queue(nvmeq);
 	}
@@ -1126,7 +1145,7 @@ static int nvme_suspend_queue(struct nvme_queue *nvmeq)
 	spin_unlock_irq(&nvmeq->q_lock);
 
 	if (!nvmeq->qid && nvmeq->dev->ctrl.admin_q)
-		blk_mq_stop_hw_queues(nvmeq->dev->ctrl.admin_q);
+		blk_mq_quiesce_queue(nvmeq->dev->ctrl.admin_q);
 
 	pci_free_irq(to_pci_dev(nvmeq->dev->dev), vector, nvmeq);
 
@@ -1145,8 +1164,7 @@ static void nvme_disable_admin_queue(struct nvme_dev *dev, bool shutdown)
 	if (shutdown)
 		nvme_shutdown_ctrl(&dev->ctrl);
 	else
-		nvme_disable_ctrl(&dev->ctrl, lo_hi_readq(
+		nvme_disable_ctrl(&dev->ctrl, dev->ctrl.cap);
-						dev->bar + NVME_REG_CAP));
 
 	spin_lock_irq(&nvmeq->q_lock);
 	nvme_process_cq(nvmeq);
@@ -1221,7 +1239,7 @@ static struct nvme_queue *nvme_alloc_queue(struct nvme_dev *dev, int qid,
 	nvmeq->qid = qid;
 	nvmeq->cq_vector = -1;
 	dev->queues[qid] = nvmeq;
-	dev->queue_count++;
+	dev->ctrl.queue_count++;
 
 	return nvmeq;
 
@@ -1317,7 +1335,7 @@ static void nvme_dev_remove_admin(struct nvme_dev *dev)
 		 * user requests may be waiting on a stopped queue. Start the
 		 * queue to flush these to completion.
 		 */
-		blk_mq_start_stopped_hw_queues(dev->ctrl.admin_q, true);
+		blk_mq_unquiesce_queue(dev->ctrl.admin_q);
 		blk_cleanup_queue(dev->ctrl.admin_q);
 		blk_mq_free_tag_set(&dev->admin_tagset);
 	}
@@ -1354,7 +1372,7 @@ static int nvme_alloc_admin_tags(struct nvme_dev *dev)
 			return -ENODEV;
 		}
 	} else
-		blk_mq_start_stopped_hw_queues(dev->ctrl.admin_q, true);
+		blk_mq_unquiesce_queue(dev->ctrl.admin_q);
 
 	return 0;
 }
@@ -1385,11 +1403,10 @@ static int nvme_remap_bar(struct nvme_dev *dev, unsigned long size)
 	return 0;
 }
 
-static int nvme_configure_admin_queue(struct nvme_dev *dev)
+static int nvme_pci_configure_admin_queue(struct nvme_dev *dev)
 {
 	int result;
 	u32 aqa;
-	u64 cap = lo_hi_readq(dev->bar + NVME_REG_CAP);
 	struct nvme_queue *nvmeq;
 
 	result = nvme_remap_bar(dev, db_bar_size(dev, 0));
@@ -1397,13 +1414,13 @@ static int nvme_configure_admin_queue(struct nvme_dev *dev)
 		return result;
 
 	dev->subsystem = readl(dev->bar + NVME_REG_VS) >= NVME_VS(1, 1, 0) ?
-						NVME_CAP_NSSRC(cap) : 0;
+				NVME_CAP_NSSRC(dev->ctrl.cap) : 0;
 
 	if (dev->subsystem &&
 	    (readl(dev->bar + NVME_REG_CSTS) & NVME_CSTS_NSSRO))
 		writel(NVME_CSTS_NSSRO, dev->bar + NVME_REG_CSTS);
 
-	result = nvme_disable_ctrl(&dev->ctrl, cap);
+	result = nvme_disable_ctrl(&dev->ctrl, dev->ctrl.cap);
 	if (result < 0)
 		return result;
 
@@ -1422,7 +1439,7 @@ static int nvme_configure_admin_queue(struct nvme_dev *dev)
 	lo_hi_writeq(nvmeq->sq_dma_addr, dev->bar + NVME_REG_ASQ);
 	lo_hi_writeq(nvmeq->cq_dma_addr, dev->bar + NVME_REG_ACQ);
 
-	result = nvme_enable_ctrl(&dev->ctrl, cap);
+	result = nvme_enable_ctrl(&dev->ctrl, dev->ctrl.cap);
 	if (result)
 		return result;
 
@@ -1441,7 +1458,7 @@ static int nvme_create_io_queues(struct nvme_dev *dev)
 	unsigned i, max;
 	int ret = 0;
 
-	for (i = dev->queue_count; i <= dev->max_qid; i++) {
+	for (i = dev->ctrl.queue_count; i <= dev->max_qid; i++) {
 		/* vector == qid - 1, match nvme_create_queue */
 		if (!nvme_alloc_queue(dev, i, dev->q_depth,
 		     pci_irq_get_node(to_pci_dev(dev->dev), i - 1))) {
@@ -1450,7 +1467,7 @@ static int nvme_create_io_queues(struct nvme_dev *dev)
 		}
 	}
 
-	max = min(dev->max_qid, dev->queue_count - 1);
+	max = min(dev->max_qid, dev->ctrl.queue_count - 1);
 	for (i = dev->online_queues; i <= max; i++) {
 		ret = nvme_create_queue(dev->queues[i], i);
 		if (ret)
@@ -1585,9 +1602,10 @@ static void nvme_free_host_mem(struct nvme_dev *dev)
 static int nvme_alloc_host_mem(struct nvme_dev *dev, u64 min, u64 preferred)
 {
 	struct nvme_host_mem_buf_desc *descs;
-	u32 chunk_size, max_entries, i = 0;
+	u32 chunk_size, max_entries;
+	int i = 0;
 	void **bufs;
-	u64 size, tmp;
+	u64 size = 0, tmp;
 
 	/* start big and work our way down */
 	chunk_size = min(preferred, (u64)PAGE_SIZE << MAX_ORDER);
@@ -1866,7 +1884,6 @@ static int nvme_dev_add(struct nvme_dev *dev)
 
 static int nvme_pci_enable(struct nvme_dev *dev)
 {
-	u64 cap;
 	int result = -ENOMEM;
 	struct pci_dev *pdev = to_pci_dev(dev->dev);
 
@@ -1893,10 +1910,11 @@ static int nvme_pci_enable(struct nvme_dev *dev)
 	if (result < 0)
 		return result;
 
-	cap = lo_hi_readq(dev->bar + NVME_REG_CAP);
+	dev->ctrl.cap = lo_hi_readq(dev->bar + NVME_REG_CAP);
 
-	dev->q_depth = min_t(int, NVME_CAP_MQES(cap) + 1, NVME_Q_DEPTH);
+	dev->q_depth = min_t(int, NVME_CAP_MQES(dev->ctrl.cap) + 1,
-	dev->db_stride = 1 << NVME_CAP_STRIDE(cap);
+				io_queue_depth);
+	dev->db_stride = 1 << NVME_CAP_STRIDE(dev->ctrl.cap);
 	dev->dbs = dev->bar + 4096;
 
 	/*
@@ -1908,6 +1926,12 @@ static int nvme_pci_enable(struct nvme_dev *dev)
 		dev_warn(dev->ctrl.device, "detected Apple NVMe controller, "
 			"set queue depth=%u to work around controller resets\n",
 			dev->q_depth);
+	} else if (pdev->vendor == PCI_VENDOR_ID_SAMSUNG &&
+		   (pdev->device == 0xa821 || pdev->device == 0xa822) &&
+		   NVME_CAP_MQES(dev->ctrl.cap) == 0) {
+		dev->q_depth = 64;
+		dev_err(dev->ctrl.device, "detected PM1725 NVMe controller, "
+                        "set queue depth=%u\n", dev->q_depth);
 	}
 
 	/*
@@ -1996,7 +2020,7 @@ static void nvme_dev_disable(struct nvme_dev *dev, bool shutdown)
 	nvme_stop_queues(&dev->ctrl);
 
 	queues = dev->online_queues - 1;
-	for (i = dev->queue_count - 1; i > 0; i--)
+	for (i = dev->ctrl.queue_count - 1; i > 0; i--)
 		nvme_suspend_queue(dev->queues[i]);
 
 	if (dead) {
@@ -2004,7 +2028,7 @@ static void nvme_dev_disable(struct nvme_dev *dev, bool shutdown)
 		 * probe, before the admin queue is configured. Thus,
 		 * queue_count can be 0 here.
 		 */
-		if (dev->queue_count)
+		if (dev->ctrl.queue_count)
 			nvme_suspend_queue(dev->queues[0]);
 	} else {
 		nvme_disable_io_queues(dev, queues);
@@ -2094,7 +2118,7 @@ static void nvme_reset_work(struct work_struct *work)
 	if (result)
 		goto out;
 
-	result = nvme_configure_admin_queue(dev);
+	result = nvme_pci_configure_admin_queue(dev);
 	if (result)
 		goto out;
 
@@ -2132,15 +2156,6 @@ static void nvme_reset_work(struct work_struct *work)
 	if (result)
 		goto out;
 
-	/*
-	 * A controller that can not execute IO typically requires user
-	 * intervention to correct. For such degraded controllers, the driver
-	 * should not submit commands the user did not request, so skip
-	 * registering for asynchronous event notification on this condition.
-	 */
-	if (dev->online_queues > 1)
-		nvme_queue_async_events(&dev->ctrl);
-
 	/*
 	 * Keep the controller around but remove all namespaces if we don't have
 	 * any working I/O queue.
@@ -2161,8 +2176,7 @@ static void nvme_reset_work(struct work_struct *work)
 		goto out;
 	}
 
-	if (dev->online_queues > 1)
+	nvme_start_ctrl(&dev->ctrl);
-		nvme_queue_scan(&dev->ctrl);
 	return;
 
  out:
@@ -2341,11 +2355,13 @@ static void nvme_remove(struct pci_dev *pdev)
 	}
 
 	flush_work(&dev->ctrl.reset_work);
-	nvme_uninit_ctrl(&dev->ctrl);
+	nvme_stop_ctrl(&dev->ctrl);
+	nvme_remove_namespaces(&dev->ctrl);
 	nvme_dev_disable(dev, true);
 	nvme_free_host_mem(dev);
 	nvme_dev_remove_admin(dev);
 	nvme_free_queues(dev, 0);
+	nvme_uninit_ctrl(&dev->ctrl);
 	nvme_release_prp_pools(dev);
 	nvme_dev_unmap(dev);
 	nvme_put_ctrl(&dev->ctrl);
@@ -2458,6 +2474,10 @@ static const struct pci_device_id nvme_id_table[] = {
 		.driver_data = NVME_QUIRK_DELAY_BEFORE_CHK_RDY, },
 	{ PCI_DEVICE(0x1c5f, 0x0540),	/* Memblaze Pblaze4 adapter */
 		.driver_data = NVME_QUIRK_DELAY_BEFORE_CHK_RDY, },
+	{ PCI_DEVICE(0x144d, 0xa821),   /* Samsung PM1725 */
+		.driver_data = NVME_QUIRK_DELAY_BEFORE_CHK_RDY, },
+	{ PCI_DEVICE(0x144d, 0xa822),   /* Samsung PM1725a */
+		.driver_data = NVME_QUIRK_DELAY_BEFORE_CHK_RDY, },
 	{ PCI_DEVICE_CLASS(PCI_CLASS_STORAGE_EXPRESS, 0xffffff) },
 	{ PCI_DEVICE(PCI_VENDOR_ID_APPLE, 0x2001) },
 	{ PCI_DEVICE(PCI_VENDOR_ID_APPLE, 0x2003) },

drivers/nvme/host/rdma.c

@@ -86,7 +86,7 @@ enum nvme_rdma_queue_flags {
 
 struct nvme_rdma_queue {
 	struct nvme_rdma_qe	*rsp_ring;
-	u8			sig_count;
+	atomic_t		sig_count;
 	int			queue_size;
 	size_t			cmnd_capsule_len;
 	struct nvme_rdma_ctrl	*ctrl;
@@ -103,7 +103,6 @@ struct nvme_rdma_queue {
 struct nvme_rdma_ctrl {
 	/* read only in the hot path */
 	struct nvme_rdma_queue	*queues;
-	u32			queue_count;
 
 	/* other member variables */
 	struct blk_mq_tag_set	tag_set;
@@ -119,7 +118,6 @@ struct nvme_rdma_ctrl {
 	struct blk_mq_tag_set	admin_tag_set;
 	struct nvme_rdma_device	*device;
 
-	u64			cap;
 	u32			max_fr_pages;
 
 	struct sockaddr_storage addr;
@@ -274,9 +272,6 @@ static int nvme_rdma_reinit_request(void *data, struct request *rq)
 	struct nvme_rdma_request *req = blk_mq_rq_to_pdu(rq);
 	int ret = 0;
 
-	if (!req->mr->need_inval)
-		goto out;
-
 	ib_dereg_mr(req->mr);
 
 	req->mr = ib_alloc_mr(dev->pd, IB_MR_TYPE_MEM_REG,
@@ -349,7 +344,7 @@ static int nvme_rdma_init_hctx(struct blk_mq_hw_ctx *hctx, void *data,
 	struct nvme_rdma_ctrl *ctrl = data;
 	struct nvme_rdma_queue *queue = &ctrl->queues[hctx_idx + 1];
 
-	BUG_ON(hctx_idx >= ctrl->queue_count);
+	BUG_ON(hctx_idx >= ctrl->ctrl.queue_count);
 
 	hctx->driver_data = queue;
 	return 0;
@@ -525,6 +520,7 @@ static int nvme_rdma_init_queue(struct nvme_rdma_ctrl *ctrl,
 		queue->cmnd_capsule_len = sizeof(struct nvme_command);
 
 	queue->queue_size = queue_size;
+	atomic_set(&queue->sig_count, 0);
 
 	queue->cm_id = rdma_create_id(&init_net, nvme_rdma_cm_handler, queue,
 			RDMA_PS_TCP, IB_QPT_RC);
@@ -587,7 +583,7 @@ static void nvme_rdma_free_io_queues(struct nvme_rdma_ctrl *ctrl)
 {
 	int i;
 
-	for (i = 1; i < ctrl->queue_count; i++)
+	for (i = 1; i < ctrl->ctrl.queue_count; i++)
 		nvme_rdma_stop_and_free_queue(&ctrl->queues[i]);
 }
 
@@ -595,7 +591,7 @@ static int nvme_rdma_connect_io_queues(struct nvme_rdma_ctrl *ctrl)
 {
 	int i, ret = 0;
 
-	for (i = 1; i < ctrl->queue_count; i++) {
+	for (i = 1; i < ctrl->ctrl.queue_count; i++) {
 		ret = nvmf_connect_io_queue(&ctrl->ctrl, i);
 		if (ret) {
 			dev_info(ctrl->ctrl.device,
@@ -623,14 +619,14 @@ static int nvme_rdma_init_io_queues(struct nvme_rdma_ctrl *ctrl)
 	if (ret)
 		return ret;
 
-	ctrl->queue_count = nr_io_queues + 1;
+	ctrl->ctrl.queue_count = nr_io_queues + 1;
-	if (ctrl->queue_count < 2)
+	if (ctrl->ctrl.queue_count < 2)
 		return 0;
 
 	dev_info(ctrl->ctrl.device,
 		"creating %d I/O queues.\n", nr_io_queues);
 
-	for (i = 1; i < ctrl->queue_count; i++) {
+	for (i = 1; i < ctrl->ctrl.queue_count; i++) {
 		ret = nvme_rdma_init_queue(ctrl, i,
 					   ctrl->ctrl.opts->queue_size);
 		if (ret) {
@@ -705,7 +701,7 @@ static void nvme_rdma_reconnect_ctrl_work(struct work_struct *work)
 
 	++ctrl->ctrl.nr_reconnects;
 
-	if (ctrl->queue_count > 1) {
+	if (ctrl->ctrl.queue_count > 1) {
 		nvme_rdma_free_io_queues(ctrl);
 
 		ret = blk_mq_reinit_tagset(&ctrl->tag_set);
@@ -729,13 +725,11 @@ static void nvme_rdma_reconnect_ctrl_work(struct work_struct *work)
 
 	set_bit(NVME_RDMA_Q_LIVE, &ctrl->queues[0].flags);
 
-	ret = nvme_enable_ctrl(&ctrl->ctrl, ctrl->cap);
+	ret = nvme_enable_ctrl(&ctrl->ctrl, ctrl->ctrl.cap);
 	if (ret)
 		goto requeue;
 
-	nvme_start_keep_alive(&ctrl->ctrl);
+	if (ctrl->ctrl.queue_count > 1) {
-
-	if (ctrl->queue_count > 1) {
 		ret = nvme_rdma_init_io_queues(ctrl);
 		if (ret)
 			goto requeue;
@@ -743,16 +737,16 @@ static void nvme_rdma_reconnect_ctrl_work(struct work_struct *work)
 		ret = nvme_rdma_connect_io_queues(ctrl);
 		if (ret)
 			goto requeue;
+
+		blk_mq_update_nr_hw_queues(&ctrl->tag_set,
+				ctrl->ctrl.queue_count - 1);
 	}
 
 	changed = nvme_change_ctrl_state(&ctrl->ctrl, NVME_CTRL_LIVE);
 	WARN_ON_ONCE(!changed);
 	ctrl->ctrl.nr_reconnects = 0;
 
-	if (ctrl->queue_count > 1) {
+	nvme_start_ctrl(&ctrl->ctrl);
-		nvme_queue_scan(&ctrl->ctrl);
-		nvme_queue_async_events(&ctrl->ctrl);
-	}
 
 	dev_info(ctrl->ctrl.device, "Successfully reconnected\n");
 
@@ -770,17 +764,17 @@ static void nvme_rdma_error_recovery_work(struct work_struct *work)
 			struct nvme_rdma_ctrl, err_work);
 	int i;
 
-	nvme_stop_keep_alive(&ctrl->ctrl);
+	nvme_stop_ctrl(&ctrl->ctrl);
 
-	for (i = 0; i < ctrl->queue_count; i++)
+	for (i = 0; i < ctrl->ctrl.queue_count; i++)
 		clear_bit(NVME_RDMA_Q_LIVE, &ctrl->queues[i].flags);
 
-	if (ctrl->queue_count > 1)
+	if (ctrl->ctrl.queue_count > 1)
 		nvme_stop_queues(&ctrl->ctrl);
-	blk_mq_stop_hw_queues(ctrl->ctrl.admin_q);
+	blk_mq_quiesce_queue(ctrl->ctrl.admin_q);
 
 	/* We must take care of fastfail/requeue all our inflight requests */
-	if (ctrl->queue_count > 1)
+	if (ctrl->ctrl.queue_count > 1)
 		blk_mq_tagset_busy_iter(&ctrl->tag_set,
 					nvme_cancel_request, &ctrl->ctrl);
 	blk_mq_tagset_busy_iter(&ctrl->admin_tag_set,
@@ -790,7 +784,7 @@ static void nvme_rdma_error_recovery_work(struct work_struct *work)
 	 * queues are not a live anymore, so restart the queues to fail fast
 	 * new IO
 	 */
-	blk_mq_start_stopped_hw_queues(ctrl->ctrl.admin_q, true);
+	blk_mq_unquiesce_queue(ctrl->ctrl.admin_q);
 	nvme_start_queues(&ctrl->ctrl);
 
 	nvme_rdma_reconnect_or_remove(ctrl);
@@ -1008,17 +1002,16 @@ static void nvme_rdma_send_done(struct ib_cq *cq, struct ib_wc *wc)
 		nvme_rdma_wr_error(cq, wc, "SEND");
 }
 
-static inline int nvme_rdma_queue_sig_limit(struct nvme_rdma_queue *queue)
+/*
+ * We want to signal completion at least every queue depth/2.  This returns the
+ * largest power of two that is not above half of (queue size + 1) to optimize
+ * (avoid divisions).
+ */
+static inline bool nvme_rdma_queue_sig_limit(struct nvme_rdma_queue *queue)
 {
-	int sig_limit;
+	int limit = 1 << ilog2((queue->queue_size + 1) / 2);
 
-	/*
+	return (atomic_inc_return(&queue->sig_count) & (limit - 1)) == 0;
-	 * We signal completion every queue depth/2 and also handle the
-	 * degenerated case of a  device with queue_depth=1, where we
-	 * would need to signal every message.
-	 */
-	sig_limit = max(queue->queue_size / 2, 1);
-	return (++queue->sig_count % sig_limit) == 0;
 }
 
 static int nvme_rdma_post_send(struct nvme_rdma_queue *queue,
@@ -1574,7 +1567,8 @@ static int nvme_rdma_configure_admin_queue(struct nvme_rdma_ctrl *ctrl)
 
 	set_bit(NVME_RDMA_Q_LIVE, &ctrl->queues[0].flags);
 
-	error = nvmf_reg_read64(&ctrl->ctrl, NVME_REG_CAP, &ctrl->cap);
+	error = nvmf_reg_read64(&ctrl->ctrl, NVME_REG_CAP,
+			&ctrl->ctrl.cap);
 	if (error) {
 		dev_err(ctrl->ctrl.device,
 			"prop_get NVME_REG_CAP failed\n");
@@ -1582,9 +1576,9 @@ static int nvme_rdma_configure_admin_queue(struct nvme_rdma_ctrl *ctrl)
 	}
 
 	ctrl->ctrl.sqsize =
-		min_t(int, NVME_CAP_MQES(ctrl->cap), ctrl->ctrl.sqsize);
+		min_t(int, NVME_CAP_MQES(ctrl->ctrl.cap), ctrl->ctrl.sqsize);
 
-	error = nvme_enable_ctrl(&ctrl->ctrl, ctrl->cap);
+	error = nvme_enable_ctrl(&ctrl->ctrl, ctrl->ctrl.cap);
 	if (error)
 		goto out_cleanup_queue;
 
@@ -1601,8 +1595,6 @@ static int nvme_rdma_configure_admin_queue(struct nvme_rdma_ctrl *ctrl)
 	if (error)
 		goto out_cleanup_queue;
 
-	nvme_start_keep_alive(&ctrl->ctrl);
-
 	return 0;
 
 out_cleanup_queue:
@@ -1620,11 +1612,10 @@ out_free_queue:
 
 static void nvme_rdma_shutdown_ctrl(struct nvme_rdma_ctrl *ctrl)
 {
-	nvme_stop_keep_alive(&ctrl->ctrl);
 	cancel_work_sync(&ctrl->err_work);
 	cancel_delayed_work_sync(&ctrl->reconnect_work);
 
-	if (ctrl->queue_count > 1) {
+	if (ctrl->ctrl.queue_count > 1) {
 		nvme_stop_queues(&ctrl->ctrl);
 		blk_mq_tagset_busy_iter(&ctrl->tag_set,
 					nvme_cancel_request, &ctrl->ctrl);
@@ -1634,18 +1625,21 @@ static void nvme_rdma_shutdown_ctrl(struct nvme_rdma_ctrl *ctrl)
 	if (test_bit(NVME_RDMA_Q_LIVE, &ctrl->queues[0].flags))
 		nvme_shutdown_ctrl(&ctrl->ctrl);
 
-	blk_mq_stop_hw_queues(ctrl->ctrl.admin_q);
+	blk_mq_quiesce_queue(ctrl->ctrl.admin_q);
 	blk_mq_tagset_busy_iter(&ctrl->admin_tag_set,
 				nvme_cancel_request, &ctrl->ctrl);
+	blk_mq_unquiesce_queue(ctrl->ctrl.admin_q);
 	nvme_rdma_destroy_admin_queue(ctrl);
 }
 
 static void __nvme_rdma_remove_ctrl(struct nvme_rdma_ctrl *ctrl, bool shutdown)
 {
-	nvme_uninit_ctrl(&ctrl->ctrl);
+	nvme_stop_ctrl(&ctrl->ctrl);
+	nvme_remove_namespaces(&ctrl->ctrl);
 	if (shutdown)
 		nvme_rdma_shutdown_ctrl(ctrl);
 
+	nvme_uninit_ctrl(&ctrl->ctrl);
 	if (ctrl->ctrl.tagset) {
 		blk_cleanup_queue(ctrl->ctrl.connect_q);
 		blk_mq_free_tag_set(&ctrl->tag_set);
@@ -1707,6 +1701,7 @@ static void nvme_rdma_reset_ctrl_work(struct work_struct *work)
 	int ret;
 	bool changed;
 
+	nvme_stop_ctrl(&ctrl->ctrl);
 	nvme_rdma_shutdown_ctrl(ctrl);
 
 	ret = nvme_rdma_configure_admin_queue(ctrl);
@@ -1716,7 +1711,7 @@ static void nvme_rdma_reset_ctrl_work(struct work_struct *work)
 		goto del_dead_ctrl;
 	}
 
-	if (ctrl->queue_count > 1) {
+	if (ctrl->ctrl.queue_count > 1) {
 		ret = blk_mq_reinit_tagset(&ctrl->tag_set);
 		if (ret)
 			goto del_dead_ctrl;
@@ -1728,16 +1723,15 @@ static void nvme_rdma_reset_ctrl_work(struct work_struct *work)
 		ret = nvme_rdma_connect_io_queues(ctrl);
 		if (ret)
 			goto del_dead_ctrl;
+
+		blk_mq_update_nr_hw_queues(&ctrl->tag_set,
+				ctrl->ctrl.queue_count - 1);
 	}
 
 	changed = nvme_change_ctrl_state(&ctrl->ctrl, NVME_CTRL_LIVE);
 	WARN_ON_ONCE(!changed);
 
-	if (ctrl->queue_count > 1) {
+	nvme_start_ctrl(&ctrl->ctrl);
-		nvme_start_queues(&ctrl->ctrl);
-		nvme_queue_scan(&ctrl->ctrl);
-		nvme_queue_async_events(&ctrl->ctrl);
-	}
 
 	return;
 
@@ -1785,7 +1779,7 @@ static int nvme_rdma_create_io_queues(struct nvme_rdma_ctrl *ctrl)
 	ctrl->tag_set.cmd_size = sizeof(struct nvme_rdma_request) +
 		SG_CHUNK_SIZE * sizeof(struct scatterlist);
 	ctrl->tag_set.driver_data = ctrl;
-	ctrl->tag_set.nr_hw_queues = ctrl->queue_count - 1;
+	ctrl->tag_set.nr_hw_queues = ctrl->ctrl.queue_count - 1;
 	ctrl->tag_set.timeout = NVME_IO_TIMEOUT;
 
 	ret = blk_mq_alloc_tag_set(&ctrl->tag_set);
@@ -1863,12 +1857,12 @@ static struct nvme_ctrl *nvme_rdma_create_ctrl(struct device *dev,
 	INIT_WORK(&ctrl->delete_work, nvme_rdma_del_ctrl_work);
 	INIT_WORK(&ctrl->ctrl.reset_work, nvme_rdma_reset_ctrl_work);
 
-	ctrl->queue_count = opts->nr_io_queues + 1; /* +1 for admin queue */
+	ctrl->ctrl.queue_count = opts->nr_io_queues + 1; /* +1 for admin queue */
 	ctrl->ctrl.sqsize = opts->queue_size - 1;
 	ctrl->ctrl.kato = opts->kato;
 
 	ret = -ENOMEM;
-	ctrl->queues = kcalloc(ctrl->queue_count, sizeof(*ctrl->queues),
+	ctrl->queues = kcalloc(ctrl->ctrl.queue_count, sizeof(*ctrl->queues),
 				GFP_KERNEL);
 	if (!ctrl->queues)
 		goto out_uninit_ctrl;
@@ -1925,15 +1919,11 @@ static struct nvme_ctrl *nvme_rdma_create_ctrl(struct device *dev,
 	list_add_tail(&ctrl->list, &nvme_rdma_ctrl_list);
 	mutex_unlock(&nvme_rdma_ctrl_mutex);
 
-	if (opts->nr_io_queues) {
+	nvme_start_ctrl(&ctrl->ctrl);
-		nvme_queue_scan(&ctrl->ctrl);
-		nvme_queue_async_events(&ctrl->ctrl);
-	}
 
 	return &ctrl->ctrl;
 
 out_remove_admin_queue:
-	nvme_stop_keep_alive(&ctrl->ctrl);
 	nvme_rdma_destroy_admin_queue(ctrl);
 out_kfree_queues:
 	kfree(ctrl->queues);

drivers/nvme/target/fc.c

@@ -1164,18 +1164,24 @@ nvmet_fc_ls_create_association(struct nvmet_fc_tgtport *tgtport,
 
 	memset(acc, 0, sizeof(*acc));
 
-	if (iod->rqstdatalen < sizeof(struct fcnvme_ls_cr_assoc_rqst))
+	/*
+	 * FC-NVME spec changes. There are initiators sending different
+	 * lengths as padding sizes for Create Association Cmd descriptor
+	 * was incorrect.
+	 * Accept anything of "minimum" length. Assume format per 1.15
+	 * spec (with HOSTID reduced to 16 bytes), ignore how long the
+	 * trailing pad length is.
+	 */
+	if (iod->rqstdatalen < FCNVME_LSDESC_CRA_RQST_MINLEN)
 		ret = VERR_CR_ASSOC_LEN;
-	else if (rqst->desc_list_len !=
+	else if (rqst->desc_list_len <
-			fcnvme_lsdesc_len(
+			cpu_to_be32(FCNVME_LSDESC_CRA_RQST_MIN_LISTLEN))
-				sizeof(struct fcnvme_ls_cr_assoc_rqst)))
 		ret = VERR_CR_ASSOC_RQST_LEN;
 	else if (rqst->assoc_cmd.desc_tag !=
 			cpu_to_be32(FCNVME_LSDESC_CREATE_ASSOC_CMD))
 		ret = VERR_CR_ASSOC_CMD;
-	else if (rqst->assoc_cmd.desc_len !=
+	else if (rqst->assoc_cmd.desc_len <
-			fcnvme_lsdesc_len(
+			cpu_to_be32(FCNVME_LSDESC_CRA_CMD_DESC_MIN_DESCLEN))
-				sizeof(struct fcnvme_lsdesc_cr_assoc_cmd)))
 		ret = VERR_CR_ASSOC_CMD_LEN;
 	else if (!rqst->assoc_cmd.ersp_ratio ||
 		 (be16_to_cpu(rqst->assoc_cmd.ersp_ratio) >=

drivers/nvme/target/io-cmd.c

@@ -85,7 +85,7 @@ static void nvmet_execute_rw(struct nvmet_req *req)
 			bio_set_op_attrs(bio, op, op_flags);
 
 			bio_chain(bio, prev);
-			cookie = submit_bio(prev);
+			submit_bio(prev);
 		}
 
 		sector += sg->length >> 9;

drivers/nvme/target/loop.c

@@ -44,12 +44,10 @@ struct nvme_loop_iod {
 
 struct nvme_loop_ctrl {
 	struct nvme_loop_queue	*queues;
-	u32			queue_count;
 
 	struct blk_mq_tag_set	admin_tag_set;
 
 	struct list_head	list;
-	u64			cap;
 	struct blk_mq_tag_set	tag_set;
 	struct nvme_loop_iod	async_event_iod;
 	struct nvme_ctrl	ctrl;
@@ -241,7 +239,7 @@ static int nvme_loop_init_hctx(struct blk_mq_hw_ctx *hctx, void *data,
 	struct nvme_loop_ctrl *ctrl = data;
 	struct nvme_loop_queue *queue = &ctrl->queues[hctx_idx + 1];
 
-	BUG_ON(hctx_idx >= ctrl->queue_count);
+	BUG_ON(hctx_idx >= ctrl->ctrl.queue_count);
 
 	hctx->driver_data = queue;
 	return 0;
@@ -307,7 +305,7 @@ static void nvme_loop_destroy_io_queues(struct nvme_loop_ctrl *ctrl)
 {
 	int i;
 
-	for (i = 1; i < ctrl->queue_count; i++)
+	for (i = 1; i < ctrl->ctrl.queue_count; i++)
 		nvmet_sq_destroy(&ctrl->queues[i].nvme_sq);
 }
 
@@ -330,7 +328,7 @@ static int nvme_loop_init_io_queues(struct nvme_loop_ctrl *ctrl)
 		if (ret)
 			goto out_destroy_queues;
 
-		ctrl->queue_count++;
+		ctrl->ctrl.queue_count++;
 	}
 
 	return 0;
@@ -344,7 +342,7 @@ static int nvme_loop_connect_io_queues(struct nvme_loop_ctrl *ctrl)
 {
 	int i, ret;
 
-	for (i = 1; i < ctrl->queue_count; i++) {
+	for (i = 1; i < ctrl->ctrl.queue_count; i++) {
 		ret = nvmf_connect_io_queue(&ctrl->ctrl, i);
 		if (ret)
 			return ret;
@@ -372,7 +370,7 @@ static int nvme_loop_configure_admin_queue(struct nvme_loop_ctrl *ctrl)
 	error = nvmet_sq_init(&ctrl->queues[0].nvme_sq);
 	if (error)
 		return error;
-	ctrl->queue_count = 1;
+	ctrl->ctrl.queue_count = 1;
 
 	error = blk_mq_alloc_tag_set(&ctrl->admin_tag_set);
 	if (error)
@@ -388,7 +386,7 @@ static int nvme_loop_configure_admin_queue(struct nvme_loop_ctrl *ctrl)
 	if (error)
 		goto out_cleanup_queue;
 
-	error = nvmf_reg_read64(&ctrl->ctrl, NVME_REG_CAP, &ctrl->cap);
+	error = nvmf_reg_read64(&ctrl->ctrl, NVME_REG_CAP, &ctrl->ctrl.cap);
 	if (error) {
 		dev_err(ctrl->ctrl.device,
 			"prop_get NVME_REG_CAP failed\n");
@@ -396,9 +394,9 @@ static int nvme_loop_configure_admin_queue(struct nvme_loop_ctrl *ctrl)
 	}
 
 	ctrl->ctrl.sqsize =
-		min_t(int, NVME_CAP_MQES(ctrl->cap), ctrl->ctrl.sqsize);
+		min_t(int, NVME_CAP_MQES(ctrl->ctrl.cap), ctrl->ctrl.sqsize);
 
-	error = nvme_enable_ctrl(&ctrl->ctrl, ctrl->cap);
+	error = nvme_enable_ctrl(&ctrl->ctrl, ctrl->ctrl.cap);
 	if (error)
 		goto out_cleanup_queue;
 
@@ -409,8 +407,6 @@ static int nvme_loop_configure_admin_queue(struct nvme_loop_ctrl *ctrl)
 	if (error)
 		goto out_cleanup_queue;
 
-	nvme_start_keep_alive(&ctrl->ctrl);
-
 	return 0;
 
 out_cleanup_queue:
@@ -424,9 +420,7 @@ out_free_sq:
 
 static void nvme_loop_shutdown_ctrl(struct nvme_loop_ctrl *ctrl)
 {
-	nvme_stop_keep_alive(&ctrl->ctrl);
+	if (ctrl->ctrl.queue_count > 1) {
-
-	if (ctrl->queue_count > 1) {
 		nvme_stop_queues(&ctrl->ctrl);
 		blk_mq_tagset_busy_iter(&ctrl->tag_set,
 					nvme_cancel_request, &ctrl->ctrl);
@@ -436,9 +430,10 @@ static void nvme_loop_shutdown_ctrl(struct nvme_loop_ctrl *ctrl)
 	if (ctrl->ctrl.state == NVME_CTRL_LIVE)
 		nvme_shutdown_ctrl(&ctrl->ctrl);
 
-	blk_mq_stop_hw_queues(ctrl->ctrl.admin_q);
+	blk_mq_quiesce_queue(ctrl->ctrl.admin_q);
 	blk_mq_tagset_busy_iter(&ctrl->admin_tag_set,
 				nvme_cancel_request, &ctrl->ctrl);
+	blk_mq_unquiesce_queue(ctrl->ctrl.admin_q);
 	nvme_loop_destroy_admin_queue(ctrl);
 }
 
@@ -447,8 +442,10 @@ static void nvme_loop_del_ctrl_work(struct work_struct *work)
 	struct nvme_loop_ctrl *ctrl = container_of(work,
 				struct nvme_loop_ctrl, delete_work);
 
-	nvme_uninit_ctrl(&ctrl->ctrl);
+	nvme_stop_ctrl(&ctrl->ctrl);
+	nvme_remove_namespaces(&ctrl->ctrl);
 	nvme_loop_shutdown_ctrl(ctrl);
+	nvme_uninit_ctrl(&ctrl->ctrl);
 	nvme_put_ctrl(&ctrl->ctrl);
 }
 
@@ -496,6 +493,7 @@ static void nvme_loop_reset_ctrl_work(struct work_struct *work)
 	bool changed;
 	int ret;
 
+	nvme_stop_ctrl(&ctrl->ctrl);
 	nvme_loop_shutdown_ctrl(ctrl);
 
 	ret = nvme_loop_configure_admin_queue(ctrl);
@@ -510,13 +508,13 @@ static void nvme_loop_reset_ctrl_work(struct work_struct *work)
 	if (ret)
 		goto out_destroy_io;
 
+	blk_mq_update_nr_hw_queues(&ctrl->tag_set,
+			ctrl->ctrl.queue_count - 1);
+
 	changed = nvme_change_ctrl_state(&ctrl->ctrl, NVME_CTRL_LIVE);
 	WARN_ON_ONCE(!changed);
 
-	nvme_queue_scan(&ctrl->ctrl);
+	nvme_start_ctrl(&ctrl->ctrl);
-	nvme_queue_async_events(&ctrl->ctrl);
-
-	nvme_start_queues(&ctrl->ctrl);
 
 	return;
 
@@ -559,7 +557,7 @@ static int nvme_loop_create_io_queues(struct nvme_loop_ctrl *ctrl)
 	ctrl->tag_set.cmd_size = sizeof(struct nvme_loop_iod) +
 		SG_CHUNK_SIZE * sizeof(struct scatterlist);
 	ctrl->tag_set.driver_data = ctrl;
-	ctrl->tag_set.nr_hw_queues = ctrl->queue_count - 1;
+	ctrl->tag_set.nr_hw_queues = ctrl->ctrl.queue_count - 1;
 	ctrl->tag_set.timeout = NVME_IO_TIMEOUT;
 	ctrl->ctrl.tagset = &ctrl->tag_set;
 
@@ -651,10 +649,7 @@ static struct nvme_ctrl *nvme_loop_create_ctrl(struct device *dev,
 	list_add_tail(&ctrl->list, &nvme_loop_ctrl_list);
 	mutex_unlock(&nvme_loop_ctrl_mutex);
 
-	if (opts->nr_io_queues) {
+	nvme_start_ctrl(&ctrl->ctrl);
-		nvme_queue_scan(&ctrl->ctrl);
-		nvme_queue_async_events(&ctrl->ctrl);
-	}
 
 	return &ctrl->ctrl;
 

drivers/scsi/lpfc/lpfc_scsi.c

@@ -26,6 +26,7 @@
 #include <linux/export.h>
 #include <linux/delay.h>
 #include <asm/unaligned.h>
+#include <linux/t10-pi.h>
 #include <linux/crc-t10dif.h>
 #include <net/checksum.h>
 
@@ -2934,8 +2935,8 @@ lpfc_calc_bg_err(struct lpfc_hba *phba, struct lpfc_scsi_buf *lpfc_cmd)
 				 * First check to see if a protection data
 				 * check is valid
 				 */
-				if ((src->ref_tag == 0xffffffff) ||
+				if ((src->ref_tag == T10_PI_REF_ESCAPE) ||
-				    (src->app_tag == 0xffff)) {
+				    (src->app_tag == T10_PI_APP_ESCAPE)) {
 					start_ref_tag++;
 					goto skipit;
 				}

drivers/scsi/qla2xxx/qla_isr.c

@@ -2040,9 +2040,9 @@ qla2x00_handle_dif_error(srb_t *sp, struct sts_entry_24xx *sts24)
 	 * For type     3: ref & app tag is all 'f's
 	 * For type 0,1,2: app tag is all 'f's
 	 */
-	if ((a_app_tag == 0xffff) &&
+	if ((a_app_tag == T10_PI_APP_ESCAPE) &&
 	    ((scsi_get_prot_type(cmd) != SCSI_PROT_DIF_TYPE3) ||
-	     (a_ref_tag == 0xffffffff))) {
+	     (a_ref_tag == T10_PI_REF_ESCAPE))) {
 		uint32_t blocks_done, resid;
 		sector_t lba_s = scsi_get_lba(cmd);
 
@@ -2084,9 +2084,9 @@ qla2x00_handle_dif_error(srb_t *sp, struct sts_entry_24xx *sts24)
 			spt = page_address(sg_page(sg)) + sg->offset;
 			spt += j;
 
-			spt->app_tag = 0xffff;
+			spt->app_tag = T10_PI_APP_ESCAPE;
 			if (scsi_get_prot_type(cmd) == SCSI_PROT_DIF_TYPE3)
-				spt->ref_tag = 0xffffffff;
+				spt->ref_tag = T10_PI_REF_ESCAPE;
 		}
 
 		return 0;

drivers/target/target_core_sbc.c

@@ -1450,7 +1450,7 @@ sbc_dif_verify(struct se_cmd *cmd, sector_t start, unsigned int sectors,
 				 (unsigned long long)sector, sdt->guard_tag,
 				 sdt->app_tag, be32_to_cpu(sdt->ref_tag));
 
-			if (sdt->app_tag == cpu_to_be16(0xffff)) {
+			if (sdt->app_tag == T10_PI_APP_ESCAPE) {
 				dsg_off += block_size;
 				goto next;
 			}

include/linux/bio.h

@@ -165,10 +165,27 @@ static inline void bio_advance_iter(struct bio *bio, struct bvec_iter *iter,
 {
 	iter->bi_sector += bytes >> 9;
 
-	if (bio_no_advance_iter(bio))
+	if (bio_no_advance_iter(bio)) {
 		iter->bi_size -= bytes;
-	else
+		iter->bi_done += bytes;
+	} else {
 		bvec_iter_advance(bio->bi_io_vec, iter, bytes);
+		/* TODO: It is reasonable to complete bio with error here. */
+	}
+}
+
+static inline bool bio_rewind_iter(struct bio *bio, struct bvec_iter *iter,
+		unsigned int bytes)
+{
+	iter->bi_sector -= bytes >> 9;
+
+	if (bio_no_advance_iter(bio)) {
+		iter->bi_size += bytes;
+		iter->bi_done -= bytes;
+		return true;
+	}
+
+	return bvec_iter_rewind(bio->bi_io_vec, iter, bytes);
 }
 
 #define __bio_for_each_segment(bvl, bio, iter, start)			\
@@ -303,8 +320,6 @@ struct bio_integrity_payload {
 
 	struct bvec_iter	bip_iter;
 
-	bio_end_io_t		*bip_end_io;	/* saved I/O completion fn */
-
 	unsigned short		bip_slab;	/* slab the bip came from */
 	unsigned short		bip_vcnt;	/* # of integrity bio_vecs */
 	unsigned short		bip_max_vcnt;	/* integrity bio_vec slots */
@@ -722,13 +737,10 @@ struct biovec_slab {
 		bip_for_each_vec(_bvl, _bio->bi_integrity, _iter)
 
 extern struct bio_integrity_payload *bio_integrity_alloc(struct bio *, gfp_t, unsigned int);
-extern void bio_integrity_free(struct bio *);
 extern int bio_integrity_add_page(struct bio *, struct page *, unsigned int, unsigned int);
-extern bool bio_integrity_enabled(struct bio *bio);
+extern bool bio_integrity_prep(struct bio *);
-extern int bio_integrity_prep(struct bio *);
-extern void bio_integrity_endio(struct bio *);
 extern void bio_integrity_advance(struct bio *, unsigned int);
-extern void bio_integrity_trim(struct bio *, unsigned int, unsigned int);
+extern void bio_integrity_trim(struct bio *);
 extern int bio_integrity_clone(struct bio *, struct bio *, gfp_t);
 extern int bioset_integrity_create(struct bio_set *, int);
 extern void bioset_integrity_free(struct bio_set *);
@@ -741,11 +753,6 @@ static inline void *bio_integrity(struct bio *bio)
 	return NULL;
 }
 
-static inline bool bio_integrity_enabled(struct bio *bio)
-{
-	return false;
-}
-
 static inline int bioset_integrity_create(struct bio_set *bs, int pool_size)
 {
 	return 0;
@@ -756,14 +763,9 @@ static inline void bioset_integrity_free (struct bio_set *bs)
 	return;
 }
 
-static inline int bio_integrity_prep(struct bio *bio)
+static inline bool bio_integrity_prep(struct bio *bio)
 {
-	return 0;
+	return true;
-}
-
-static inline void bio_integrity_free(struct bio *bio)
-{
-	return;
 }
 
 static inline int bio_integrity_clone(struct bio *bio, struct bio *bio_src,
@@ -778,8 +780,7 @@ static inline void bio_integrity_advance(struct bio *bio,
 	return;
 }
 
-static inline void bio_integrity_trim(struct bio *bio, unsigned int offset,
+static inline void bio_integrity_trim(struct bio *bio)
-				      unsigned int sectors)
 {
 	return;
 }

include/linux/bvec.h

@@ -22,6 +22,7 @@
 
 #include <linux/kernel.h>
 #include <linux/bug.h>
+#include <linux/errno.h>
 
 /*
  * was unsigned short, but we might as well be ready for > 64kB I/O pages
@@ -39,6 +40,8 @@ struct bvec_iter {
 
 	unsigned int		bi_idx;		/* current index into bvl_vec */
 
+	unsigned int            bi_done;	/* number of bytes completed */
+
 	unsigned int            bi_bvec_done;	/* number of bytes completed in
 						   current bvec */
 };
@@ -66,12 +69,14 @@ struct bvec_iter {
 	.bv_offset	= bvec_iter_offset((bvec), (iter)),	\
 })
 
-static inline void bvec_iter_advance(const struct bio_vec *bv,
+static inline bool bvec_iter_advance(const struct bio_vec *bv,
-				     struct bvec_iter *iter,
+		struct bvec_iter *iter, unsigned bytes)
-				     unsigned bytes)
 {
-	WARN_ONCE(bytes > iter->bi_size,
+	if (WARN_ONCE(bytes > iter->bi_size,
-		  "Attempted to advance past end of bvec iter\n");
+		     "Attempted to advance past end of bvec iter\n")) {
+		iter->bi_size = 0;
+		return false;
+	}
 
 	while (bytes) {
 		unsigned iter_len = bvec_iter_len(bv, *iter);
@@ -80,12 +85,38 @@ static inline void bvec_iter_advance(const struct bio_vec *bv,
 		bytes -= len;
 		iter->bi_size -= len;
 		iter->bi_bvec_done += len;
+		iter->bi_done += len;
 
 		if (iter->bi_bvec_done == __bvec_iter_bvec(bv, *iter)->bv_len) {
 			iter->bi_bvec_done = 0;
 			iter->bi_idx++;
 		}
 	}
+	return true;
+}
+
+static inline bool bvec_iter_rewind(const struct bio_vec *bv,
+				     struct bvec_iter *iter,
+				     unsigned int bytes)
+{
+	while (bytes) {
+		unsigned len = min(bytes, iter->bi_bvec_done);
+
+		if (iter->bi_bvec_done == 0) {
+			if (WARN_ONCE(iter->bi_idx == 0,
+				      "Attempted to rewind iter beyond "
+				      "bvec's boundaries\n")) {
+				return false;
+			}
+			iter->bi_idx--;
+			iter->bi_bvec_done = __bvec_iter_bvec(bv, *iter)->bv_len;
+			continue;
+		}
+		bytes -= len;
+		iter->bi_size += len;
+		iter->bi_bvec_done -= len;
+	}
+	return true;
 }
 
 #define for_each_bvec(bvl, bio_vec, iter, start)			\

include/linux/nvme-fc.h

@@ -17,6 +17,7 @@
 
 /*
  * This file contains definitions relative to FC-NVME r1.14 (16-020vB).
+ * The fcnvme_lsdesc_cr_assoc_cmd struct reflects expected r1.16 content.
  */
 
 #ifndef _NVME_FC_H
@@ -193,9 +194,21 @@ struct fcnvme_lsdesc_cr_assoc_cmd {
 	uuid_t	hostid;
 	u8	hostnqn[FCNVME_ASSOC_HOSTNQN_LEN];
 	u8	subnqn[FCNVME_ASSOC_SUBNQN_LEN];
-	u8	rsvd632[384];
+	__be32	rsvd584[108];		/* pad to 1016 bytes,
+					 * which makes overall LS rqst
+					 * payload 1024 bytes
+					 */
 };
 
+#define FCNVME_LSDESC_CRA_CMD_DESC_MINLEN	\
+		offsetof(struct fcnvme_lsdesc_cr_assoc_cmd, rsvd584)
+
+#define FCNVME_LSDESC_CRA_CMD_DESC_MIN_DESCLEN	\
+		(FCNVME_LSDESC_CRA_CMD_DESC_MINLEN - \
+		 offsetof(struct fcnvme_lsdesc_cr_assoc_cmd, ersp_ratio))
+
+
+
 /* FCNVME_LSDESC_CREATE_CONN_CMD */
 struct fcnvme_lsdesc_cr_conn_cmd {
 	__be32	desc_tag;		/* FCNVME_LSDESC_xxx */
@@ -273,6 +286,14 @@ struct fcnvme_ls_cr_assoc_rqst {
 	struct fcnvme_lsdesc_cr_assoc_cmd	assoc_cmd;
 };
 
+#define FCNVME_LSDESC_CRA_RQST_MINLEN	\
+		(offsetof(struct fcnvme_ls_cr_assoc_rqst, assoc_cmd) + \
+			FCNVME_LSDESC_CRA_CMD_DESC_MINLEN)
+
+#define FCNVME_LSDESC_CRA_RQST_MIN_LISTLEN	\
+		FCNVME_LSDESC_CRA_CMD_DESC_MINLEN
+
+
 struct fcnvme_ls_cr_assoc_acc {
 	struct fcnvme_ls_acc_hdr		hdr;
 	struct fcnvme_lsdesc_assoc_id		associd;

include/linux/t10-pi.h

@@ -33,6 +33,8 @@ struct t10_pi_tuple {
 	__be32 ref_tag;		/* Target LBA or indirect LBA */
 };
 
+#define T10_PI_APP_ESCAPE cpu_to_be16(0xffff)
+#define T10_PI_REF_ESCAPE cpu_to_be32(0xffffffff)
 
 extern const struct blk_integrity_profile t10_pi_type1_crc;
 extern const struct blk_integrity_profile t10_pi_type1_ip;