Merge branch 'for-4.16/block' of git://git.kernel.dk/linux-block

Pull block updates from Jens Axboe:
 "This is the main pull request for block IO related changes for the
  4.16 kernel. Nothing major in this pull request, but a good amount of
  improvements and fixes all over the map. This contains:

   - BFQ improvements, fixes, and cleanups from Angelo, Chiara, and
     Paolo.

   - Support for SMR zones for deadline and mq-deadline from Damien and
     Christoph.

   - Set of fixes for bcache by way of Michael Lyle, including fixes
     from himself, Kent, Rui, Tang, and Coly.

   - Series from Matias for lightnvm with fixes from Hans Holmberg,
     Javier, and Matias. Mostly centered around pblk, and the removing
     rrpc 1.2 in preparation for supporting 2.0.

   - A couple of NVMe pull requests from Christoph. Nothing major in
     here, just fixes and cleanups, and support for command tracing from
     Johannes.

   - Support for blk-throttle for tracking reads and writes separately.
     From Joseph Qi. A few cleanups/fixes also for blk-throttle from
     Weiping.

   - Series from Mike Snitzer that enables dm to register its queue more
     logically, something that's alwways been problematic on dm since
     it's a stacked device.

   - Series from Ming cleaning up some of the bio accessor use, in
     preparation for supporting multipage bvecs.

   - Various fixes from Ming closing up holes around queue mapping and
     quiescing.

   - BSD partition fix from Richard Narron, fixing a problem where we
     can't mount newer (10/11) FreeBSD partitions.

   - Series from Tejun reworking blk-mq timeout handling. The previous
     scheme relied on atomic bits, but it had races where we would think
     a request had timed out if it to reused at the wrong time.

   - null_blk now supports faking timeouts, to enable us to better
     exercise and test that functionality separately. From me.

   - Kill the separate atomic poll bit in the request struct. After
     this, we don't use the atomic bits on blk-mq anymore at all. From
     me.

   - sgl_alloc/free helpers from Bart.

   - Heavily contended tag case scalability improvement from me.

   - Various little fixes and cleanups from Arnd, Bart, Corentin,
     Douglas, Eryu, Goldwyn, and myself"

* 'for-4.16/block' of git://git.kernel.dk/linux-block: (186 commits)
  block: remove smart1,2.h
  nvme: add tracepoint for nvme_complete_rq
  nvme: add tracepoint for nvme_setup_cmd
  nvme-pci: introduce RECONNECTING state to mark initializing procedure
  nvme-rdma: remove redundant boolean for inline_data
  nvme: don't free uuid pointer before printing it
  nvme-pci: Suspend queues after deleting them
  bsg: use pr_debug instead of hand crafted macros
  blk-mq-debugfs: don't allow write on attributes with seq_operations set
  nvme-pci: Fix queue double allocations
  block: Set BIO_TRACE_COMPLETION on new bio during split
  blk-throttle: use queue_is_rq_based
  block: Remove kblockd_schedule_delayed_work{,_on}()
  blk-mq: Avoid that blk_mq_delay_run_hw_queue() introduces unintended delays
  blk-mq: Rename blk_mq_request_direct_issue() into blk_mq_request_issue_directly()
  lib/scatterlist: Fix chaining support in sgl_alloc_order()
  blk-throttle: track read and write request individually
  block: add bdev_read_only() checks to common helpers
  block: fail op_is_write() requests to read-only partitions
  blk-throttle: export io_serviced_recursive, io_service_bytes_recursive
  ...

block/bfq-cgroup.c

@@ -775,10 +775,11 @@ static void bfq_pd_offline(struct blkg_policy_data *pd)
 	unsigned long flags;
 	int i;
 
+	spin_lock_irqsave(&bfqd->lock, flags);
+
 	if (!entity) /* root group */
-		return;
+		goto put_async_queues;
 
-	spin_lock_irqsave(&bfqd->lock, flags);
 	/*
 	 * Empty all service_trees belonging to this group before
 	 * deactivating the group itself.
@@ -809,6 +810,8 @@ static void bfq_pd_offline(struct blkg_policy_data *pd)
 	}
 
 	__bfq_deactivate_entity(entity, false);
+
+put_async_queues:
 	bfq_put_async_queues(bfqd, bfqg);
 
 	spin_unlock_irqrestore(&bfqd->lock, flags);

block/bfq-iosched.c

@@ -166,6 +166,20 @@ static const int bfq_async_charge_factor = 10;
 /* Default timeout values, in jiffies, approximating CFQ defaults. */
 const int bfq_timeout = HZ / 8;
 
+/*
+ * Time limit for merging (see comments in bfq_setup_cooperator). Set
+ * to the slowest value that, in our tests, proved to be effective in
+ * removing false positives, while not causing true positives to miss
+ * queue merging.
+ *
+ * As can be deduced from the low time limit below, queue merging, if
+ * successful, happens at the very beggining of the I/O of the involved
+ * cooperating processes, as a consequence of the arrival of the very
+ * first requests from each cooperator.  After that, there is very
+ * little chance to find cooperators.
+ */
+static const unsigned long bfq_merge_time_limit = HZ/10;
+
 static struct kmem_cache *bfq_pool;
 
 /* Below this threshold (in ns), we consider thinktime immediate. */
@@ -178,7 +192,7 @@ static struct kmem_cache *bfq_pool;
 #define BFQQ_SEEK_THR		(sector_t)(8 * 100)
 #define BFQQ_SECT_THR_NONROT	(sector_t)(2 * 32)
 #define BFQQ_CLOSE_THR		(sector_t)(8 * 1024)
-#define BFQQ_SEEKY(bfqq)	(hweight32(bfqq->seek_history) > 32/8)
+#define BFQQ_SEEKY(bfqq)	(hweight32(bfqq->seek_history) > 19)
 
 /* Min number of samples required to perform peak-rate update */
 #define BFQ_RATE_MIN_SAMPLES	32
@@ -195,15 +209,17 @@ static struct kmem_cache *bfq_pool;
  * interactive applications automatically, using the following formula:
  * duration = (R / r) * T, where r is the peak rate of the device, and
  * R and T are two reference parameters.
- * In particular, R is the peak rate of the reference device (see below),
- * and T is a reference time: given the systems that are likely to be
- * installed on the reference device according to its speed class, T is
- * about the maximum time needed, under BFQ and while reading two files in
- * parallel, to load typical large applications on these systems.
- * In practice, the slower/faster the device at hand is, the more/less it
- * takes to load applications with respect to the reference device.
- * Accordingly, the longer/shorter BFQ grants weight raising to interactive
- * applications.
+ * In particular, R is the peak rate of the reference device (see
+ * below), and T is a reference time: given the systems that are
+ * likely to be installed on the reference device according to its
+ * speed class, T is about the maximum time needed, under BFQ and
+ * while reading two files in parallel, to load typical large
+ * applications on these systems (see the comments on
+ * max_service_from_wr below, for more details on how T is obtained).
+ * In practice, the slower/faster the device at hand is, the more/less
+ * it takes to load applications with respect to the reference device.
+ * Accordingly, the longer/shorter BFQ grants weight raising to
+ * interactive applications.
  *
  * BFQ uses four different reference pairs (R, T), depending on:
  * . whether the device is rotational or non-rotational;
@@ -240,6 +256,60 @@ static int T_slow[2];
 static int T_fast[2];
 static int device_speed_thresh[2];
 
+/*
+ * BFQ uses the above-detailed, time-based weight-raising mechanism to
+ * privilege interactive tasks. This mechanism is vulnerable to the
+ * following false positives: I/O-bound applications that will go on
+ * doing I/O for much longer than the duration of weight
+ * raising. These applications have basically no benefit from being
+ * weight-raised at the beginning of their I/O. On the opposite end,
+ * while being weight-raised, these applications
+ * a) unjustly steal throughput to applications that may actually need
+ * low latency;
+ * b) make BFQ uselessly perform device idling; device idling results
+ * in loss of device throughput with most flash-based storage, and may
+ * increase latencies when used purposelessly.
+ *
+ * BFQ tries to reduce these problems, by adopting the following
+ * countermeasure. To introduce this countermeasure, we need first to
+ * finish explaining how the duration of weight-raising for
+ * interactive tasks is computed.
+ *
+ * For a bfq_queue deemed as interactive, the duration of weight
+ * raising is dynamically adjusted, as a function of the estimated
+ * peak rate of the device, so as to be equal to the time needed to
+ * execute the 'largest' interactive task we benchmarked so far. By
+ * largest task, we mean the task for which each involved process has
+ * to do more I/O than for any of the other tasks we benchmarked. This
+ * reference interactive task is the start-up of LibreOffice Writer,
+ * and in this task each process/bfq_queue needs to have at most ~110K
+ * sectors transferred.
+ *
+ * This last piece of information enables BFQ to reduce the actual
+ * duration of weight-raising for at least one class of I/O-bound
+ * applications: those doing sequential or quasi-sequential I/O. An
+ * example is file copy. In fact, once started, the main I/O-bound
+ * processes of these applications usually consume the above 110K
+ * sectors in much less time than the processes of an application that
+ * is starting, because these I/O-bound processes will greedily devote
+ * almost all their CPU cycles only to their target,
+ * throughput-friendly I/O operations. This is even more true if BFQ
+ * happens to be underestimating the device peak rate, and thus
+ * overestimating the duration of weight raising. But, according to
+ * our measurements, once transferred 110K sectors, these processes
+ * have no right to be weight-raised any longer.
+ *
+ * Basing on the last consideration, BFQ ends weight-raising for a
+ * bfq_queue if the latter happens to have received an amount of
+ * service at least equal to the following constant. The constant is
+ * set to slightly more than 110K, to have a minimum safety margin.
+ *
+ * This early ending of weight-raising reduces the amount of time
+ * during which interactive false positives cause the two problems
+ * described at the beginning of these comments.
+ */
+static const unsigned long max_service_from_wr = 120000;
+
 #define RQ_BIC(rq)		icq_to_bic((rq)->elv.priv[0])
 #define RQ_BFQQ(rq)		((rq)->elv.priv[1])
 
@@ -403,6 +473,82 @@ static struct request *bfq_choose_req(struct bfq_data *bfqd,
 	}
 }
 
+/*
+ * See the comments on bfq_limit_depth for the purpose of
+ * the depths set in the function.
+ */
+static void bfq_update_depths(struct bfq_data *bfqd, struct sbitmap_queue *bt)
+{
+	bfqd->sb_shift = bt->sb.shift;
+
+	/*
+	 * In-word depths if no bfq_queue is being weight-raised:
+	 * leaving 25% of tags only for sync reads.
+	 *
+	 * In next formulas, right-shift the value
+	 * (1U<<bfqd->sb_shift), instead of computing directly
+	 * (1U<<(bfqd->sb_shift - something)), to be robust against
+	 * any possible value of bfqd->sb_shift, without having to
+	 * limit 'something'.
+	 */
+	/* no more than 50% of tags for async I/O */
+	bfqd->word_depths[0][0] = max((1U<<bfqd->sb_shift)>>1, 1U);
+	/*
+	 * no more than 75% of tags for sync writes (25% extra tags
+	 * w.r.t. async I/O, to prevent async I/O from starving sync
+	 * writes)
+	 */
+	bfqd->word_depths[0][1] = max(((1U<<bfqd->sb_shift) * 3)>>2, 1U);
+
+	/*
+	 * In-word depths in case some bfq_queue is being weight-
+	 * raised: leaving ~63% of tags for sync reads. This is the
+	 * highest percentage for which, in our tests, application
+	 * start-up times didn't suffer from any regression due to tag
+	 * shortage.
+	 */
+	/* no more than ~18% of tags for async I/O */
+	bfqd->word_depths[1][0] = max(((1U<<bfqd->sb_shift) * 3)>>4, 1U);
+	/* no more than ~37% of tags for sync writes (~20% extra tags) */
+	bfqd->word_depths[1][1] = max(((1U<<bfqd->sb_shift) * 6)>>4, 1U);
+}
+
+/*
+ * Async I/O can easily starve sync I/O (both sync reads and sync
+ * writes), by consuming all tags. Similarly, storms of sync writes,
+ * such as those that sync(2) may trigger, can starve sync reads.
+ * Limit depths of async I/O and sync writes so as to counter both
+ * problems.
+ */
+static void bfq_limit_depth(unsigned int op, struct blk_mq_alloc_data *data)
+{
+	struct blk_mq_tags *tags = blk_mq_tags_from_data(data);
+	struct bfq_data *bfqd = data->q->elevator->elevator_data;
+	struct sbitmap_queue *bt;
+
+	if (op_is_sync(op) && !op_is_write(op))
+		return;
+
+	if (data->flags & BLK_MQ_REQ_RESERVED) {
+		if (unlikely(!tags->nr_reserved_tags)) {
+			WARN_ON_ONCE(1);
+			return;
+		}
+		bt = &tags->breserved_tags;
+	} else
+		bt = &tags->bitmap_tags;
+
+	if (unlikely(bfqd->sb_shift != bt->sb.shift))
+		bfq_update_depths(bfqd, bt);
+
+	data->shallow_depth =
+		bfqd->word_depths[!!bfqd->wr_busy_queues][op_is_sync(op)];
+
+	bfq_log(bfqd, "[%s] wr_busy %d sync %d depth %u",
+			__func__, bfqd->wr_busy_queues, op_is_sync(op),
+			data->shallow_depth);
+}
+
 static struct bfq_queue *
 bfq_rq_pos_tree_lookup(struct bfq_data *bfqd, struct rb_root *root,
 		     sector_t sector, struct rb_node **ret_parent,
@@ -444,6 +590,13 @@ bfq_rq_pos_tree_lookup(struct bfq_data *bfqd, struct rb_root *root,
 	return bfqq;
 }
 
+static bool bfq_too_late_for_merging(struct bfq_queue *bfqq)
+{
+	return bfqq->service_from_backlogged > 0 &&
+		time_is_before_jiffies(bfqq->first_IO_time +
+				       bfq_merge_time_limit);
+}
+
 void bfq_pos_tree_add_move(struct bfq_data *bfqd, struct bfq_queue *bfqq)
 {
 	struct rb_node **p, *parent;
@@ -454,6 +607,14 @@ void bfq_pos_tree_add_move(struct bfq_data *bfqd, struct bfq_queue *bfqq)
 		bfqq->pos_root = NULL;
 	}
 
+	/*
+	 * bfqq cannot be merged any longer (see comments in
+	 * bfq_setup_cooperator): no point in adding bfqq into the
+	 * position tree.
+	 */
+	if (bfq_too_late_for_merging(bfqq))
+		return;
+
 	if (bfq_class_idle(bfqq))
 		return;
 	if (!bfqq->next_rq)
@@ -1247,6 +1408,7 @@ static void bfq_update_bfqq_wr_on_rq_arrival(struct bfq_data *bfqd,
 	if (old_wr_coeff == 1 && wr_or_deserves_wr) {
 		/* start a weight-raising period */
 		if (interactive) {
+			bfqq->service_from_wr = 0;
 			bfqq->wr_coeff = bfqd->bfq_wr_coeff;
 			bfqq->wr_cur_max_time = bfq_wr_duration(bfqd);
 		} else {
@@ -1627,6 +1789,8 @@ static void bfq_remove_request(struct request_queue *q,
 			rb_erase(&bfqq->pos_node, bfqq->pos_root);
 			bfqq->pos_root = NULL;
 		}
+	} else {
+		bfq_pos_tree_add_move(bfqd, bfqq);
 	}
 
 	if (rq->cmd_flags & REQ_META)
@@ -1933,6 +2097,9 @@ bfq_setup_merge(struct bfq_queue *bfqq, struct bfq_queue *new_bfqq)
 static bool bfq_may_be_close_cooperator(struct bfq_queue *bfqq,
 					struct bfq_queue *new_bfqq)
 {
+	if (bfq_too_late_for_merging(new_bfqq))
+		return false;
+
 	if (bfq_class_idle(bfqq) || bfq_class_idle(new_bfqq) ||
 	    (bfqq->ioprio_class != new_bfqq->ioprio_class))
 		return false;
@@ -1956,20 +2123,6 @@ static bool bfq_may_be_close_cooperator(struct bfq_queue *bfqq,
 	return true;
 }
 
-/*
- * If this function returns true, then bfqq cannot be merged. The idea
- * is that true cooperation happens very early after processes start
- * to do I/O. Usually, late cooperations are just accidental false
- * positives. In case bfqq is weight-raised, such false positives
- * would evidently degrade latency guarantees for bfqq.
- */
-static bool wr_from_too_long(struct bfq_queue *bfqq)
-{
-	return bfqq->wr_coeff > 1 &&
-		time_is_before_jiffies(bfqq->last_wr_start_finish +
-				       msecs_to_jiffies(100));
-}
-
 /*
  * Attempt to schedule a merge of bfqq with the currently in-service
  * queue or with a close queue among the scheduled queues.  Return
@@ -1983,11 +2136,6 @@ static bool wr_from_too_long(struct bfq_queue *bfqq)
  * to maintain. Besides, in such a critical condition as an out of memory,
  * the benefits of queue merging may be little relevant, or even negligible.
  *
- * Weight-raised queues can be merged only if their weight-raising
- * period has just started. In fact cooperating processes are usually
- * started together. Thus, with this filter we avoid false positives
- * that would jeopardize low-latency guarantees.
- *
  * WARNING: queue merging may impair fairness among non-weight raised
  * queues, for at least two reasons: 1) the original weight of a
  * merged queue may change during the merged state, 2) even being the
@@ -2001,12 +2149,24 @@ bfq_setup_cooperator(struct bfq_data *bfqd, struct bfq_queue *bfqq,
 {
 	struct bfq_queue *in_service_bfqq, *new_bfqq;
 
+	/*
+	 * Prevent bfqq from being merged if it has been created too
+	 * long ago. The idea is that true cooperating processes, and
+	 * thus their associated bfq_queues, are supposed to be
+	 * created shortly after each other. This is the case, e.g.,
+	 * for KVM/QEMU and dump I/O threads. Basing on this
+	 * assumption, the following filtering greatly reduces the
+	 * probability that two non-cooperating processes, which just
+	 * happen to do close I/O for some short time interval, have
+	 * their queues merged by mistake.
+	 */
+	if (bfq_too_late_for_merging(bfqq))
+		return NULL;
+
 	if (bfqq->new_bfqq)
 		return bfqq->new_bfqq;
 
-	if (!io_struct ||
-	    wr_from_too_long(bfqq) ||
-	    unlikely(bfqq == &bfqd->oom_bfqq))
+	if (!io_struct || unlikely(bfqq == &bfqd->oom_bfqq))
 		return NULL;
 
 	/* If there is only one backlogged queue, don't search. */
@@ -2015,12 +2175,9 @@ bfq_setup_cooperator(struct bfq_data *bfqd, struct bfq_queue *bfqq,
 
 	in_service_bfqq = bfqd->in_service_queue;
 
-	if (!in_service_bfqq || in_service_bfqq == bfqq
-	    || wr_from_too_long(in_service_bfqq) ||
-	    unlikely(in_service_bfqq == &bfqd->oom_bfqq))
-		goto check_scheduled;
-
-	if (bfq_rq_close_to_sector(io_struct, request, bfqd->last_position) &&
+	if (in_service_bfqq && in_service_bfqq != bfqq &&
+	    likely(in_service_bfqq != &bfqd->oom_bfqq) &&
+	    bfq_rq_close_to_sector(io_struct, request, bfqd->last_position) &&
 	    bfqq->entity.parent == in_service_bfqq->entity.parent &&
 	    bfq_may_be_close_cooperator(bfqq, in_service_bfqq)) {
 		new_bfqq = bfq_setup_merge(bfqq, in_service_bfqq);
@@ -2032,12 +2189,10 @@ bfq_setup_cooperator(struct bfq_data *bfqd, struct bfq_queue *bfqq,
 	 * queues. The only thing we need is that the bio/request is not
 	 * NULL, as we need it to establish whether a cooperator exists.
 	 */
-check_scheduled:
 	new_bfqq = bfq_find_close_cooperator(bfqd, bfqq,
 			bfq_io_struct_pos(io_struct, request));
 
-	if (new_bfqq && !wr_from_too_long(new_bfqq) &&
-	    likely(new_bfqq != &bfqd->oom_bfqq) &&
+	if (new_bfqq && likely(new_bfqq != &bfqd->oom_bfqq) &&
 	    bfq_may_be_close_cooperator(bfqq, new_bfqq))
 		return bfq_setup_merge(bfqq, new_bfqq);
 
@@ -2062,7 +2217,8 @@ static void bfq_bfqq_save_state(struct bfq_queue *bfqq)
 	bic->saved_in_large_burst = bfq_bfqq_in_large_burst(bfqq);
 	bic->was_in_burst_list = !hlist_unhashed(&bfqq->burst_list_node);
 	if (unlikely(bfq_bfqq_just_created(bfqq) &&
-		     !bfq_bfqq_in_large_burst(bfqq))) {
+		     !bfq_bfqq_in_large_burst(bfqq) &&
+		     bfqq->bfqd->low_latency)) {
 		/*
 		 * bfqq being merged right after being created: bfqq
 		 * would have deserved interactive weight raising, but
@@ -2917,45 +3073,87 @@ static bool bfq_bfqq_is_slow(struct bfq_data *bfqd, struct bfq_queue *bfqq,
  * whereas soft_rt_next_start is set to infinity for applications that do
  * not.
  *
- * Unfortunately, even a greedy application may happen to behave in an
- * isochronous way if the CPU load is high. In fact, the application may
- * stop issuing requests while the CPUs are busy serving other processes,
- * then restart, then stop again for a while, and so on. In addition, if
- * the disk achieves a low enough throughput with the request pattern
- * issued by the application (e.g., because the request pattern is random
- * and/or the device is slow), then the application may meet the above
- * bandwidth requirement too. To prevent such a greedy application to be
- * deemed as soft real-time, a further rule is used in the computation of
- * soft_rt_next_start: soft_rt_next_start must be higher than the current
- * time plus the maximum time for which the arrival of a request is waited
- * for when a sync queue becomes idle, namely bfqd->bfq_slice_idle.
- * This filters out greedy applications, as the latter issue instead their
- * next request as soon as possible after the last one has been completed
- * (in contrast, when a batch of requests is completed, a soft real-time
- * application spends some time processing data).
+ * Unfortunately, even a greedy (i.e., I/O-bound) application may
+ * happen to meet, occasionally or systematically, both the above
+ * bandwidth and isochrony requirements. This may happen at least in
+ * the following circumstances. First, if the CPU load is high. The
+ * application may stop issuing requests while the CPUs are busy
+ * serving other processes, then restart, then stop again for a while,
+ * and so on. The other circumstances are related to the storage
+ * device: the storage device is highly loaded or reaches a low-enough
+ * throughput with the I/O of the application (e.g., because the I/O
+ * is random and/or the device is slow). In all these cases, the
+ * I/O of the application may be simply slowed down enough to meet
+ * the bandwidth and isochrony requirements. To reduce the probability
+ * that greedy applications are deemed as soft real-time in these
+ * corner cases, a further rule is used in the computation of
+ * soft_rt_next_start: the return value of this function is forced to
+ * be higher than the maximum between the following two quantities.
+ *
+ * (a) Current time plus: (1) the maximum time for which the arrival
+ *     of a request is waited for when a sync queue becomes idle,
+ *     namely bfqd->bfq_slice_idle, and (2) a few extra jiffies. We
+ *     postpone for a moment the reason for adding a few extra
+ *     jiffies; we get back to it after next item (b).  Lower-bounding
+ *     the return value of this function with the current time plus
+ *     bfqd->bfq_slice_idle tends to filter out greedy applications,
+ *     because the latter issue their next request as soon as possible
+ *     after the last one has been completed. In contrast, a soft
+ *     real-time application spends some time processing data, after a
+ *     batch of its requests has been completed.
  *
- * Unfortunately, the last filter may easily generate false positives if
- * only bfqd->bfq_slice_idle is used as a reference time interval and one
- * or both the following cases occur:
- * 1) HZ is so low that the duration of a jiffy is comparable to or higher
- *    than bfqd->bfq_slice_idle. This happens, e.g., on slow devices with
- *    HZ=100.
+ * (b) Current value of bfqq->soft_rt_next_start. As pointed out
+ *     above, greedy applications may happen to meet both the
+ *     bandwidth and isochrony requirements under heavy CPU or
+ *     storage-device load. In more detail, in these scenarios, these
+ *     applications happen, only for limited time periods, to do I/O
+ *     slowly enough to meet all the requirements described so far,
+ *     including the filtering in above item (a). These slow-speed
+ *     time intervals are usually interspersed between other time
+ *     intervals during which these applications do I/O at a very high
+ *     speed. Fortunately, exactly because of the high speed of the
+ *     I/O in the high-speed intervals, the values returned by this
+ *     function happen to be so high, near the end of any such
+ *     high-speed interval, to be likely to fall *after* the end of
+ *     the low-speed time interval that follows. These high values are
+ *     stored in bfqq->soft_rt_next_start after each invocation of
+ *     this function. As a consequence, if the last value of
+ *     bfqq->soft_rt_next_start is constantly used to lower-bound the
+ *     next value that this function may return, then, from the very
+ *     beginning of a low-speed interval, bfqq->soft_rt_next_start is
+ *     likely to be constantly kept so high that any I/O request
+ *     issued during the low-speed interval is considered as arriving
+ *     to soon for the application to be deemed as soft
+ *     real-time. Then, in the high-speed interval that follows, the
+ *     application will not be deemed as soft real-time, just because
+ *     it will do I/O at a high speed. And so on.
+ *
+ * Getting back to the filtering in item (a), in the following two
+ * cases this filtering might be easily passed by a greedy
+ * application, if the reference quantity was just
+ * bfqd->bfq_slice_idle:
+ * 1) HZ is so low that the duration of a jiffy is comparable to or
+ *    higher than bfqd->bfq_slice_idle. This happens, e.g., on slow
+ *    devices with HZ=100. The time granularity may be so coarse
+ *    that the approximation, in jiffies, of bfqd->bfq_slice_idle
+ *    is rather lower than the exact value.
  * 2) jiffies, instead of increasing at a constant rate, may stop increasing
  *    for a while, then suddenly 'jump' by several units to recover the lost
  *    increments. This seems to happen, e.g., inside virtual machines.
- * To address this issue, we do not use as a reference time interval just
- * bfqd->bfq_slice_idle, but bfqd->bfq_slice_idle plus a few jiffies. In
- * particular we add the minimum number of jiffies for which the filter
- * seems to be quite precise also in embedded systems and KVM/QEMU virtual
- * machines.
+ * To address this issue, in the filtering in (a) we do not use as a
+ * reference time interval just bfqd->bfq_slice_idle, but
+ * bfqd->bfq_slice_idle plus a few jiffies. In particular, we add the
+ * minimum number of jiffies for which the filter seems to be quite
+ * precise also in embedded systems and KVM/QEMU virtual machines.
  */
 static unsigned long bfq_bfqq_softrt_next_start(struct bfq_data *bfqd,
 						struct bfq_queue *bfqq)
 {
-	return max(bfqq->last_idle_bklogged +
-		   HZ * bfqq->service_from_backlogged /
-		   bfqd->bfq_wr_max_softrt_rate,
-		   jiffies + nsecs_to_jiffies(bfqq->bfqd->bfq_slice_idle) + 4);
+	return max3(bfqq->soft_rt_next_start,
+		    bfqq->last_idle_bklogged +
+		    HZ * bfqq->service_from_backlogged /
+		    bfqd->bfq_wr_max_softrt_rate,
+		    jiffies + nsecs_to_jiffies(bfqq->bfqd->bfq_slice_idle) + 4);
 }
 
 /**
@@ -2999,17 +3197,6 @@ void bfq_bfqq_expire(struct bfq_data *bfqd,
 	 */
 	slow = bfq_bfqq_is_slow(bfqd, bfqq, compensate, reason, &delta);
 
-	/*
-	 * Increase service_from_backlogged before next statement,
-	 * because the possible next invocation of
-	 * bfq_bfqq_charge_time would likely inflate
-	 * entity->service. In contrast, service_from_backlogged must
-	 * contain real service, to enable the soft real-time
-	 * heuristic to correctly compute the bandwidth consumed by
-	 * bfqq.
-	 */
-	bfqq->service_from_backlogged += entity->service;
-
 	/*
 	 * As above explained, charge slow (typically seeky) and
 	 * timed-out queues with the time and not the service
@@ -3535,6 +3722,12 @@ static void bfq_update_wr_data(struct bfq_data *bfqd, struct bfq_queue *bfqq)
 				bfqq->entity.prio_changed = 1;
 			}
 		}
+		if (bfqq->wr_coeff > 1 &&
+		    bfqq->wr_cur_max_time != bfqd->bfq_wr_rt_max_time &&
+		    bfqq->service_from_wr > max_service_from_wr) {
+			/* see comments on max_service_from_wr */
+			bfq_bfqq_end_wr(bfqq);
+		}
 	}
 	/*
 	 * To improve latency (for this or other queues), immediately
@@ -3630,8 +3823,8 @@ static struct request *__bfq_dispatch_request(struct blk_mq_hw_ctx *hctx)
 		}
 
 		/*
-		 * We exploit the put_rq_private hook to decrement
-		 * rq_in_driver, but put_rq_private will not be
+		 * We exploit the bfq_finish_request hook to decrement
+		 * rq_in_driver, but bfq_finish_request will not be
 		 * invoked on this request. So, to avoid unbalance,
 		 * just start this request, without incrementing
 		 * rq_in_driver. As a negative consequence,
@@ -3640,14 +3833,14 @@ static struct request *__bfq_dispatch_request(struct blk_mq_hw_ctx *hctx)
 		 * bfq_schedule_dispatch to be invoked uselessly.
 		 *
 		 * As for implementing an exact solution, the
-		 * put_request hook, if defined, is probably invoked
-		 * also on this request. So, by exploiting this hook,
-		 * we could 1) increment rq_in_driver here, and 2)
-		 * decrement it in put_request. Such a solution would
-		 * let the value of the counter be always accurate,
-		 * but it would entail using an extra interface
-		 * function. This cost seems higher than the benefit,
-		 * being the frequency of non-elevator-private
+		 * bfq_finish_request hook, if defined, is probably
+		 * invoked also on this request. So, by exploiting
+		 * this hook, we could 1) increment rq_in_driver here,
+		 * and 2) decrement it in bfq_finish_request. Such a
+		 * solution would let the value of the counter be
+		 * always accurate, but it would entail using an extra
+		 * interface function. This cost seems higher than the
+		 * benefit, being the frequency of non-elevator-private
 		 * requests very low.
 		 */
 		goto start_rq;
@@ -3689,35 +3882,16 @@ exit:
 	return rq;
 }
 
-static struct request *bfq_dispatch_request(struct blk_mq_hw_ctx *hctx)
-{
-	struct bfq_data *bfqd = hctx->queue->elevator->elevator_data;
-	struct request *rq;
 #if defined(CONFIG_BFQ_GROUP_IOSCHED) && defined(CONFIG_DEBUG_BLK_CGROUP)
-	struct bfq_queue *in_serv_queue, *bfqq;
-	bool waiting_rq, idle_timer_disabled;
-#endif
-
-	spin_lock_irq(&bfqd->lock);
-
-#if defined(CONFIG_BFQ_GROUP_IOSCHED) && defined(CONFIG_DEBUG_BLK_CGROUP)
-	in_serv_queue = bfqd->in_service_queue;
-	waiting_rq = in_serv_queue && bfq_bfqq_wait_request(in_serv_queue);
-
-	rq = __bfq_dispatch_request(hctx);
-
-	idle_timer_disabled =
-		waiting_rq && !bfq_bfqq_wait_request(in_serv_queue);
-
-#else
-	rq = __bfq_dispatch_request(hctx);
-#endif
-	spin_unlock_irq(&bfqd->lock);
+static void bfq_update_dispatch_stats(struct request_queue *q,
+				      struct request *rq,
+				      struct bfq_queue *in_serv_queue,
+				      bool idle_timer_disabled)
+{
+	struct bfq_queue *bfqq = rq ? RQ_BFQQ(rq) : NULL;
 
-#if defined(CONFIG_BFQ_GROUP_IOSCHED) && defined(CONFIG_DEBUG_BLK_CGROUP)
-	bfqq = rq ? RQ_BFQQ(rq) : NULL;
 	if (!idle_timer_disabled && !bfqq)
-		return rq;
+		return;
 
 	/*
 	 * rq and bfqq are guaranteed to exist until this function
@@ -3732,7 +3906,7 @@ static struct request *bfq_dispatch_request(struct blk_mq_hw_ctx *hctx)
 	 * In addition, the following queue lock guarantees that
 	 * bfqq_group(bfqq) exists as well.
 	 */
-	spin_lock_irq(hctx->queue->queue_lock);
+	spin_lock_irq(q->queue_lock);
 	if (idle_timer_disabled)
 		/*
 		 * Since the idle timer has been disabled,
@@ -3751,9 +3925,37 @@ static struct request *bfq_dispatch_request(struct blk_mq_hw_ctx *hctx)
 		bfqg_stats_set_start_empty_time(bfqg);
 		bfqg_stats_update_io_remove(bfqg, rq->cmd_flags);
 	}
-	spin_unlock_irq(hctx->queue->queue_lock);
+	spin_unlock_irq(q->queue_lock);
+}
+#else
+static inline void bfq_update_dispatch_stats(struct request_queue *q,
+					     struct request *rq,
+					     struct bfq_queue *in_serv_queue,
+					     bool idle_timer_disabled) {}
 #endif
 
+static struct request *bfq_dispatch_request(struct blk_mq_hw_ctx *hctx)
+{
+	struct bfq_data *bfqd = hctx->queue->elevator->elevator_data;
+	struct request *rq;
+	struct bfq_queue *in_serv_queue;
+	bool waiting_rq, idle_timer_disabled;
+
+	spin_lock_irq(&bfqd->lock);
+
+	in_serv_queue = bfqd->in_service_queue;
+	waiting_rq = in_serv_queue && bfq_bfqq_wait_request(in_serv_queue);
+
+	rq = __bfq_dispatch_request(hctx);
+
+	idle_timer_disabled =
+		waiting_rq && !bfq_bfqq_wait_request(in_serv_queue);
+
+	spin_unlock_irq(&bfqd->lock);
+
+	bfq_update_dispatch_stats(hctx->queue, rq, in_serv_queue,
+				  idle_timer_disabled);
+
 	return rq;
 }
 
@@ -4002,10 +4204,15 @@ static void bfq_init_bfqq(struct bfq_data *bfqd, struct bfq_queue *bfqq,
 	bfqq->split_time = bfq_smallest_from_now();
 
 	/*
-	 * Set to the value for which bfqq will not be deemed as
-	 * soft rt when it becomes backlogged.
+	 * To not forget the possibly high bandwidth consumed by a
+	 * process/queue in the recent past,
+	 * bfq_bfqq_softrt_next_start() returns a value at least equal
+	 * to the current value of bfqq->soft_rt_next_start (see
+	 * comments on bfq_bfqq_softrt_next_start).  Set
+	 * soft_rt_next_start to now, to mean that bfqq has consumed
+	 * no bandwidth so far.
 	 */
-	bfqq->soft_rt_next_start = bfq_greatest_from_now();
+	bfqq->soft_rt_next_start = jiffies;
 
 	/* first request is almost certainly seeky */
 	bfqq->seek_history = 1;
@@ -4276,16 +4483,46 @@ static bool __bfq_insert_request(struct bfq_data *bfqd, struct request *rq)
 	return idle_timer_disabled;
 }
 
+#if defined(CONFIG_BFQ_GROUP_IOSCHED) && defined(CONFIG_DEBUG_BLK_CGROUP)
+static void bfq_update_insert_stats(struct request_queue *q,
+				    struct bfq_queue *bfqq,
+				    bool idle_timer_disabled,
+				    unsigned int cmd_flags)
+{
+	if (!bfqq)
+		return;
+
+	/*
+	 * bfqq still exists, because it can disappear only after
+	 * either it is merged with another queue, or the process it
+	 * is associated with exits. But both actions must be taken by
+	 * the same process currently executing this flow of
+	 * instructions.
+	 *
+	 * In addition, the following queue lock guarantees that
+	 * bfqq_group(bfqq) exists as well.
+	 */
+	spin_lock_irq(q->queue_lock);
+	bfqg_stats_update_io_add(bfqq_group(bfqq), bfqq, cmd_flags);
+	if (idle_timer_disabled)
+		bfqg_stats_update_idle_time(bfqq_group(bfqq));
+	spin_unlock_irq(q->queue_lock);
+}
+#else
+static inline void bfq_update_insert_stats(struct request_queue *q,
+					   struct bfq_queue *bfqq,
+					   bool idle_timer_disabled,
+					   unsigned int cmd_flags) {}
+#endif
+
 static void bfq_insert_request(struct blk_mq_hw_ctx *hctx, struct request *rq,
 			       bool at_head)
 {
 	struct request_queue *q = hctx->queue;
 	struct bfq_data *bfqd = q->elevator->elevator_data;
-#if defined(CONFIG_BFQ_GROUP_IOSCHED) && defined(CONFIG_DEBUG_BLK_CGROUP)
 	struct bfq_queue *bfqq = RQ_BFQQ(rq);
 	bool idle_timer_disabled = false;
 	unsigned int cmd_flags;
-#endif
 
 	spin_lock_irq(&bfqd->lock);
 	if (blk_mq_sched_try_insert_merge(q, rq)) {
@@ -4304,7 +4541,6 @@ static void bfq_insert_request(struct blk_mq_hw_ctx *hctx, struct request *rq,
 		else
 			list_add_tail(&rq->queuelist, &bfqd->dispatch);
 	} else {
-#if defined(CONFIG_BFQ_GROUP_IOSCHED) && defined(CONFIG_DEBUG_BLK_CGROUP)
 		idle_timer_disabled = __bfq_insert_request(bfqd, rq);
 		/*
 		 * Update bfqq, because, if a queue merge has occurred
@@ -4312,9 +4548,6 @@ static void bfq_insert_request(struct blk_mq_hw_ctx *hctx, struct request *rq,
 		 * redirected into a new queue.
 		 */
 		bfqq = RQ_BFQQ(rq);
-#else
-		__bfq_insert_request(bfqd, rq);
-#endif
 
 		if (rq_mergeable(rq)) {
 			elv_rqhash_add(q, rq);
@@ -4323,35 +4556,17 @@ static void bfq_insert_request(struct blk_mq_hw_ctx *hctx, struct request *rq,
 		}
 	}
 
-#if defined(CONFIG_BFQ_GROUP_IOSCHED) && defined(CONFIG_DEBUG_BLK_CGROUP)
 	/*
 	 * Cache cmd_flags before releasing scheduler lock, because rq
 	 * may disappear afterwards (for example, because of a request
 	 * merge).
 	 */
 	cmd_flags = rq->cmd_flags;
-#endif
+
 	spin_unlock_irq(&bfqd->lock);
 
-#if defined(CONFIG_BFQ_GROUP_IOSCHED) && defined(CONFIG_DEBUG_BLK_CGROUP)
-	if (!bfqq)
-		return;
-	/*
-	 * bfqq still exists, because it can disappear only after
-	 * either it is merged with another queue, or the process it
-	 * is associated with exits. But both actions must be taken by
-	 * the same process currently executing this flow of
-	 * instruction.
-	 *
-	 * In addition, the following queue lock guarantees that
-	 * bfqq_group(bfqq) exists as well.
-	 */
-	spin_lock_irq(q->queue_lock);
-	bfqg_stats_update_io_add(bfqq_group(bfqq), bfqq, cmd_flags);
-	if (idle_timer_disabled)
-		bfqg_stats_update_idle_time(bfqq_group(bfqq));
-	spin_unlock_irq(q->queue_lock);
-#endif
+	bfq_update_insert_stats(q, bfqq, idle_timer_disabled,
+				cmd_flags);
 }
 
 static void bfq_insert_requests(struct blk_mq_hw_ctx *hctx,
@@ -4482,7 +4697,7 @@ static void bfq_completed_request(struct bfq_queue *bfqq, struct bfq_data *bfqd)
 		bfq_schedule_dispatch(bfqd);
 }
 
-static void bfq_put_rq_priv_body(struct bfq_queue *bfqq)
+static void bfq_finish_request_body(struct bfq_queue *bfqq)
 {
 	bfqq->allocated--;
 
@@ -4512,7 +4727,7 @@ static void bfq_finish_request(struct request *rq)
 		spin_lock_irqsave(&bfqd->lock, flags);
 
 		bfq_completed_request(bfqq, bfqd);
-		bfq_put_rq_priv_body(bfqq);
+		bfq_finish_request_body(bfqq);
 
 		spin_unlock_irqrestore(&bfqd->lock, flags);
 	} else {
@@ -4533,7 +4748,7 @@ static void bfq_finish_request(struct request *rq)
 			bfqg_stats_update_io_remove(bfqq_group(bfqq),
 						    rq->cmd_flags);
 		}
-		bfq_put_rq_priv_body(bfqq);
+		bfq_finish_request_body(bfqq);
 	}
 
 	rq->elv.priv[0] = NULL;
@@ -4818,6 +5033,9 @@ static void bfq_exit_queue(struct elevator_queue *e)
 	hrtimer_cancel(&bfqd->idle_slice_timer);
 
 #ifdef CONFIG_BFQ_GROUP_IOSCHED
+	/* release oom-queue reference to root group */
+	bfqg_and_blkg_put(bfqd->root_group);
+
 	blkcg_deactivate_policy(bfqd->queue, &blkcg_policy_bfq);
 #else
 	spin_lock_irq(&bfqd->lock);
@@ -5206,6 +5424,7 @@ static struct elv_fs_entry bfq_attrs[] = {
 
 static struct elevator_type iosched_bfq_mq = {
 	.ops.mq = {
+		.limit_depth		= bfq_limit_depth,
 		.prepare_request	= bfq_prepare_request,
 		.finish_request		= bfq_finish_request,
 		.exit_icq		= bfq_exit_icq,

block/bfq-iosched.h

@@ -337,6 +337,11 @@ struct bfq_queue {
 	 * last transition from idle to backlogged.
 	 */
 	unsigned long service_from_backlogged;
+	/*
+	 * Cumulative service received from the @bfq_queue since its
+	 * last transition to weight-raised state.
+	 */
+	unsigned long service_from_wr;
 
 	/*
 	 * Value of wr start time when switching to soft rt
@@ -344,6 +349,8 @@ struct bfq_queue {
 	unsigned long wr_start_at_switch_to_srt;
 
 	unsigned long split_time; /* time of last split */
+
+	unsigned long first_IO_time; /* time of first I/O for this queue */
 };
 
 /**
@@ -627,6 +634,18 @@ struct bfq_data {
 	struct bfq_io_cq *bio_bic;
 	/* bfqq associated with the task issuing current bio for merging */
 	struct bfq_queue *bio_bfqq;
+
+	/*
+	 * Cached sbitmap shift, used to compute depth limits in
+	 * bfq_update_depths.
+	 */
+	unsigned int sb_shift;
+
+	/*
+	 * Depth limits used in bfq_limit_depth (see comments on the
+	 * function)
+	 */
+	unsigned int word_depths[2][2];
 };
 
 enum bfqq_state_flags {

block/bfq-wf2q.c

@@ -835,6 +835,13 @@ void bfq_bfqq_served(struct bfq_queue *bfqq, int served)
 	struct bfq_entity *entity = &bfqq->entity;
 	struct bfq_service_tree *st;
 
+	if (!bfqq->service_from_backlogged)
+		bfqq->first_IO_time = jiffies;
+
+	if (bfqq->wr_coeff > 1)
+		bfqq->service_from_wr += served;
+
+	bfqq->service_from_backlogged += served;
 	for_each_entity(entity) {
 		st = bfq_entity_service_tree(entity);
 

block/bio-integrity.c

@@ -374,7 +374,6 @@ static void bio_integrity_verify_fn(struct work_struct *work)
 /**
  * __bio_integrity_endio - Integrity I/O completion function
  * @bio:	Protected bio
- * @error:	Pointer to errno
  *
  * Description: Completion for integrity I/O
  *

block/bio.c

@@ -970,34 +970,6 @@ void bio_advance(struct bio *bio, unsigned bytes)
 }
 EXPORT_SYMBOL(bio_advance);
 
-/**
- * bio_alloc_pages - allocates a single page for each bvec in a bio
- * @bio: bio to allocate pages for
- * @gfp_mask: flags for allocation
- *
- * Allocates pages up to @bio->bi_vcnt.
- *
- * Returns 0 on success, -ENOMEM on failure. On failure, any allocated pages are
- * freed.
- */
-int bio_alloc_pages(struct bio *bio, gfp_t gfp_mask)
-{
-	int i;
-	struct bio_vec *bv;
-
-	bio_for_each_segment_all(bv, bio, i) {
-		bv->bv_page = alloc_page(gfp_mask);
-		if (!bv->bv_page) {
-			while (--bv >= bio->bi_io_vec)
-				__free_page(bv->bv_page);
-			return -ENOMEM;
-		}
-	}
-
-	return 0;
-}
-EXPORT_SYMBOL(bio_alloc_pages);
-
 /**
  * bio_copy_data - copy contents of data buffers from one chain of bios to
  * another
@@ -1838,7 +1810,7 @@ struct bio *bio_split(struct bio *bio, int sectors,
 	bio_advance(bio, split->bi_iter.bi_size);
 
 	if (bio_flagged(bio, BIO_TRACE_COMPLETION))
-		bio_set_flag(bio, BIO_TRACE_COMPLETION);
+		bio_set_flag(split, BIO_TRACE_COMPLETION);
 
 	return split;
 }

block/blk-core.c

@@ -126,6 +126,8 @@ void blk_rq_init(struct request_queue *q, struct request *rq)
 	rq->start_time = jiffies;
 	set_start_time_ns(rq);
 	rq->part = NULL;
+	seqcount_init(&rq->gstate_seq);
+	u64_stats_init(&rq->aborted_gstate_sync);
 }
 EXPORT_SYMBOL(blk_rq_init);
 
@@ -699,6 +701,15 @@ void blk_cleanup_queue(struct request_queue *q)
 	queue_flag_set(QUEUE_FLAG_DEAD, q);
 	spin_unlock_irq(lock);
 
+	/*
+	 * make sure all in-progress dispatch are completed because
+	 * blk_freeze_queue() can only complete all requests, and
+	 * dispatch may still be in-progress since we dispatch requests
+	 * from more than one contexts
+	 */
+	if (q->mq_ops)
+		blk_mq_quiesce_queue(q);
+
 	/* for synchronous bio-based driver finish in-flight integrity i/o */
 	blk_flush_integrity();
 
@@ -1646,6 +1657,7 @@ void __blk_put_request(struct request_queue *q, struct request *req)
 
 	lockdep_assert_held(q->queue_lock);
 
+	blk_req_zone_write_unlock(req);
 	blk_pm_put_request(req);
 
 	elv_completed_request(q, req);
@@ -2055,6 +2067,21 @@ static inline bool should_fail_request(struct hd_struct *part,
 
 #endif /* CONFIG_FAIL_MAKE_REQUEST */
 
+static inline bool bio_check_ro(struct bio *bio, struct hd_struct *part)
+{
+	if (part->policy && op_is_write(bio_op(bio))) {
+		char b[BDEVNAME_SIZE];
+
+		printk(KERN_ERR
+		       "generic_make_request: Trying to write "
+			"to read-only block-device %s (partno %d)\n",
+			bio_devname(bio, b), part->partno);
+		return true;
+	}
+
+	return false;
+}
+
 /*
  * Remap block n of partition p to block n+start(p) of the disk.
  */
@@ -2063,27 +2090,28 @@ static inline int blk_partition_remap(struct bio *bio)
 	struct hd_struct *p;
 	int ret = 0;
 
+	rcu_read_lock();
+	p = __disk_get_part(bio->bi_disk, bio->bi_partno);
+	if (unlikely(!p || should_fail_request(p, bio->bi_iter.bi_size) ||
+		     bio_check_ro(bio, p))) {
+		ret = -EIO;
+		goto out;
+	}
+
 	/*
 	 * Zone reset does not include bi_size so bio_sectors() is always 0.
 	 * Include a test for the reset op code and perform the remap if needed.
 	 */
-	if (!bio->bi_partno ||
-	    (!bio_sectors(bio) && bio_op(bio) != REQ_OP_ZONE_RESET))
-		return 0;
+	if (!bio_sectors(bio) && bio_op(bio) != REQ_OP_ZONE_RESET)
+		goto out;
 
-	rcu_read_lock();
-	p = __disk_get_part(bio->bi_disk, bio->bi_partno);
-	if (likely(p && !should_fail_request(p, bio->bi_iter.bi_size))) {
-		bio->bi_iter.bi_sector += p->start_sect;
-		bio->bi_partno = 0;
-		trace_block_bio_remap(bio->bi_disk->queue, bio, part_devt(p),
-				bio->bi_iter.bi_sector - p->start_sect);
-	} else {
-		printk("%s: fail for partition %d\n", __func__, bio->bi_partno);
-		ret = -EIO;
-	}
-	rcu_read_unlock();
+	bio->bi_iter.bi_sector += p->start_sect;
+	bio->bi_partno = 0;
+	trace_block_bio_remap(bio->bi_disk->queue, bio, part_devt(p),
+			      bio->bi_iter.bi_sector - p->start_sect);
 
+out:
+	rcu_read_unlock();
 	return ret;
 }
 
@@ -2142,15 +2170,19 @@ generic_make_request_checks(struct bio *bio)
 	 * For a REQ_NOWAIT based request, return -EOPNOTSUPP
 	 * if queue is not a request based queue.
 	 */
-
 	if ((bio->bi_opf & REQ_NOWAIT) && !queue_is_rq_based(q))
 		goto not_supported;
 
 	if (should_fail_request(&bio->bi_disk->part0, bio->bi_iter.bi_size))
 		goto end_io;
 
-	if (blk_partition_remap(bio))
-		goto end_io;
+	if (!bio->bi_partno) {
+		if (unlikely(bio_check_ro(bio, &bio->bi_disk->part0)))
+			goto end_io;
+	} else {
+		if (blk_partition_remap(bio))
+			goto end_io;
+	}
 
 	if (bio_check_eod(bio, nr_sectors))
 		goto end_io;
@@ -2493,8 +2525,7 @@ blk_status_t blk_insert_cloned_request(struct request_queue *q, struct request *
 		 * bypass a potential scheduler on the bottom device for
 		 * insert.
 		 */
-		blk_mq_request_bypass_insert(rq, true);
-		return BLK_STS_OK;
+		return blk_mq_request_issue_directly(rq);
 	}
 
 	spin_lock_irqsave(q->queue_lock, flags);
@@ -2846,7 +2877,7 @@ void blk_start_request(struct request *req)
 		wbt_issue(req->q->rq_wb, &req->issue_stat);
 	}
 
-	BUG_ON(test_bit(REQ_ATOM_COMPLETE, &req->atomic_flags));
+	BUG_ON(blk_rq_is_complete(req));
 	blk_add_timer(req);
 }
 EXPORT_SYMBOL(blk_start_request);
@@ -3415,20 +3446,6 @@ int kblockd_mod_delayed_work_on(int cpu, struct delayed_work *dwork,
 }
 EXPORT_SYMBOL(kblockd_mod_delayed_work_on);
 
-int kblockd_schedule_delayed_work(struct delayed_work *dwork,
-				  unsigned long delay)
-{
-	return queue_delayed_work(kblockd_workqueue, dwork, delay);
-}
-EXPORT_SYMBOL(kblockd_schedule_delayed_work);
-
-int kblockd_schedule_delayed_work_on(int cpu, struct delayed_work *dwork,
-				     unsigned long delay)
-{
-	return queue_delayed_work_on(cpu, kblockd_workqueue, dwork, delay);
-}
-EXPORT_SYMBOL(kblockd_schedule_delayed_work_on);
-
 /**
  * blk_start_plug - initialize blk_plug and track it inside the task_struct
  * @plug:	The &struct blk_plug that needs to be initialized

block/blk-exec.c

@@ -61,7 +61,7 @@ void blk_execute_rq_nowait(struct request_queue *q, struct gendisk *bd_disk,
 	 * be reused after dying flag is set
 	 */
 	if (q->mq_ops) {
-		blk_mq_sched_insert_request(rq, at_head, true, false, false);
+		blk_mq_sched_insert_request(rq, at_head, true, false);
 		return;
 	}
 

block/blk-lib.c

@@ -37,6 +37,9 @@ int __blkdev_issue_discard(struct block_device *bdev, sector_t sector,
 	if (!q)
 		return -ENXIO;
 
+	if (bdev_read_only(bdev))
+		return -EPERM;
+
 	if (flags & BLKDEV_DISCARD_SECURE) {
 		if (!blk_queue_secure_erase(q))
 			return -EOPNOTSUPP;
@@ -156,6 +159,9 @@ static int __blkdev_issue_write_same(struct block_device *bdev, sector_t sector,
 	if (!q)
 		return -ENXIO;
 
+	if (bdev_read_only(bdev))
+		return -EPERM;
+
 	bs_mask = (bdev_logical_block_size(bdev) >> 9) - 1;
 	if ((sector | nr_sects) & bs_mask)
 		return -EINVAL;
@@ -233,6 +239,9 @@ static int __blkdev_issue_write_zeroes(struct block_device *bdev,
 	if (!q)
 		return -ENXIO;
 
+	if (bdev_read_only(bdev))
+		return -EPERM;
+
 	/* Ensure that max_write_zeroes_sectors doesn't overflow bi_size */
 	max_write_zeroes_sectors = bdev_write_zeroes_sectors(bdev);
 
@@ -287,6 +296,9 @@ static int __blkdev_issue_zero_pages(struct block_device *bdev,
 	if (!q)
 		return -ENXIO;
 
+	if (bdev_read_only(bdev))
+		return -EPERM;
+
 	while (nr_sects != 0) {
 		bio = next_bio(bio, __blkdev_sectors_to_bio_pages(nr_sects),
 			       gfp_mask);

block/blk-map.c

@@ -119,7 +119,7 @@ int blk_rq_map_user_iov(struct request_queue *q, struct request *rq,
 	unsigned long align = q->dma_pad_mask | queue_dma_alignment(q);
 	struct bio *bio = NULL;
 	struct iov_iter i;
-	int ret;
+	int ret = -EINVAL;
 
 	if (!iter_is_iovec(iter))
 		goto fail;
@@ -148,7 +148,7 @@ unmap_rq:
 	__blk_rq_unmap_user(bio);
 fail:
 	rq->bio = NULL;
-	return -EINVAL;
+	return ret;
 }
 EXPORT_SYMBOL(blk_rq_map_user_iov);
 

block/blk-merge.c

@@ -128,9 +128,7 @@ static struct bio *blk_bio_segment_split(struct request_queue *q,
 				nsegs++;
 				sectors = max_sectors;
 			}
-			if (sectors)
-				goto split;
-			/* Make this single bvec as the 1st segment */
+			goto split;
 		}
 
 		if (bvprvp && blk_queue_cluster(q)) {
@@ -146,22 +144,21 @@ static struct bio *blk_bio_segment_split(struct request_queue *q,
 			bvprvp = &bvprv;
 			sectors += bv.bv_len >> 9;
 
-			if (nsegs == 1 && seg_size > front_seg_size)
-				front_seg_size = seg_size;
 			continue;
 		}
 new_segment:
 		if (nsegs == queue_max_segments(q))
 			goto split;
 
+		if (nsegs == 1 && seg_size > front_seg_size)
+			front_seg_size = seg_size;
+
 		nsegs++;
 		bvprv = bv;
 		bvprvp = &bvprv;
 		seg_size = bv.bv_len;
 		sectors += bv.bv_len >> 9;
 
-		if (nsegs == 1 && seg_size > front_seg_size)
-			front_seg_size = seg_size;
 	}
 
 	do_split = false;
@@ -174,6 +171,8 @@ split:
 			bio = new;
 	}
 
+	if (nsegs == 1 && seg_size > front_seg_size)
+		front_seg_size = seg_size;
 	bio->bi_seg_front_size = front_seg_size;
 	if (seg_size > bio->bi_seg_back_size)
 		bio->bi_seg_back_size = seg_size;

block/blk-mq-debugfs.c

@@ -289,17 +289,12 @@ static const char *const rqf_name[] = {
 	RQF_NAME(HASHED),
 	RQF_NAME(STATS),
 	RQF_NAME(SPECIAL_PAYLOAD),
+	RQF_NAME(ZONE_WRITE_LOCKED),
+	RQF_NAME(MQ_TIMEOUT_EXPIRED),
+	RQF_NAME(MQ_POLL_SLEPT),
 };
 #undef RQF_NAME
 
-#define RQAF_NAME(name) [REQ_ATOM_##name] = #name
-static const char *const rqaf_name[] = {
-	RQAF_NAME(COMPLETE),
-	RQAF_NAME(STARTED),
-	RQAF_NAME(POLL_SLEPT),
-};
-#undef RQAF_NAME
-
 int __blk_mq_debugfs_rq_show(struct seq_file *m, struct request *rq)
 {
 	const struct blk_mq_ops *const mq_ops = rq->q->mq_ops;
@@ -316,8 +311,7 @@ int __blk_mq_debugfs_rq_show(struct seq_file *m, struct request *rq)
 	seq_puts(m, ", .rq_flags=");
 	blk_flags_show(m, (__force unsigned int)rq->rq_flags, rqf_name,
 		       ARRAY_SIZE(rqf_name));
-	seq_puts(m, ", .atomic_flags=");
-	blk_flags_show(m, rq->atomic_flags, rqaf_name, ARRAY_SIZE(rqaf_name));
+	seq_printf(m, ", complete=%d", blk_rq_is_complete(rq));
 	seq_printf(m, ", .tag=%d, .internal_tag=%d", rq->tag,
 		   rq->internal_tag);
 	if (mq_ops->show_rq)
@@ -409,7 +403,7 @@ static void hctx_show_busy_rq(struct request *rq, void *data, bool reserved)
 	const struct show_busy_params *params = data;
 
 	if (blk_mq_map_queue(rq->q, rq->mq_ctx->cpu) == params->hctx &&
-	    test_bit(REQ_ATOM_STARTED, &rq->atomic_flags))
+	    blk_mq_rq_state(rq) != MQ_RQ_IDLE)
 		__blk_mq_debugfs_rq_show(params->m,
 					 list_entry_rq(&rq->queuelist));
 }
@@ -703,7 +697,11 @@ static ssize_t blk_mq_debugfs_write(struct file *file, const char __user *buf,
 	const struct blk_mq_debugfs_attr *attr = m->private;
 	void *data = d_inode(file->f_path.dentry->d_parent)->i_private;
 
-	if (!attr->write)
+	/*
+	 * Attributes that only implement .seq_ops are read-only and 'attr' is
+	 * the same with 'data' in this case.
+	 */
+	if (attr == data || !attr->write)
 		return -EPERM;
 
 	return attr->write(data, buf, count, ppos);

block/blk-mq-sched.c

@@ -172,7 +172,6 @@ static void blk_mq_do_dispatch_ctx(struct blk_mq_hw_ctx *hctx)
 	WRITE_ONCE(hctx->dispatch_from, ctx);
 }
 
-/* return true if hw queue need to be run again */
 void blk_mq_sched_dispatch_requests(struct blk_mq_hw_ctx *hctx)
 {
 	struct request_queue *q = hctx->queue;
@@ -428,7 +427,7 @@ done:
 }
 
 void blk_mq_sched_insert_request(struct request *rq, bool at_head,
-				 bool run_queue, bool async, bool can_block)
+				 bool run_queue, bool async)
 {
 	struct request_queue *q = rq->q;
 	struct elevator_queue *e = q->elevator;

block/blk-mq-sched.h

@@ -18,7 +18,7 @@ bool blk_mq_sched_try_insert_merge(struct request_queue *q, struct request *rq);
 void blk_mq_sched_restart(struct blk_mq_hw_ctx *hctx);
 
 void blk_mq_sched_insert_request(struct request *rq, bool at_head,
-				 bool run_queue, bool async, bool can_block);
+				 bool run_queue, bool async);
 void blk_mq_sched_insert_requests(struct request_queue *q,
 				  struct blk_mq_ctx *ctx,
 				  struct list_head *list, bool run_queue_async);

block/blk-mq-sysfs.c

@@ -248,7 +248,7 @@ static int blk_mq_register_hctx(struct blk_mq_hw_ctx *hctx)
 	return ret;
 }
 
-static void __blk_mq_unregister_dev(struct device *dev, struct request_queue *q)
+void blk_mq_unregister_dev(struct device *dev, struct request_queue *q)
 {
 	struct blk_mq_hw_ctx *hctx;
 	int i;
@@ -265,13 +265,6 @@ static void __blk_mq_unregister_dev(struct device *dev, struct request_queue *q)
 	q->mq_sysfs_init_done = false;
 }
 
-void blk_mq_unregister_dev(struct device *dev, struct request_queue *q)
-{
-	mutex_lock(&q->sysfs_lock);
-	__blk_mq_unregister_dev(dev, q);
-	mutex_unlock(&q->sysfs_lock);
-}
-
 void blk_mq_hctx_kobj_init(struct blk_mq_hw_ctx *hctx)
 {
 	kobject_init(&hctx->kobj, &blk_mq_hw_ktype);

block/blk-mq-tag.c

@@ -134,12 +134,6 @@ unsigned int blk_mq_get_tag(struct blk_mq_alloc_data *data)
 	ws = bt_wait_ptr(bt, data->hctx);
 	drop_ctx = data->ctx == NULL;
 	do {
-		prepare_to_wait(&ws->wait, &wait, TASK_UNINTERRUPTIBLE);
-
-		tag = __blk_mq_get_tag(data, bt);
-		if (tag != -1)
-			break;
-
 		/*
 		 * We're out of tags on this hardware queue, kick any
 		 * pending IO submits before going to sleep waiting for
@@ -155,6 +149,13 @@ unsigned int blk_mq_get_tag(struct blk_mq_alloc_data *data)
 		if (tag != -1)
 			break;
 
+		prepare_to_wait_exclusive(&ws->wait, &wait,
+						TASK_UNINTERRUPTIBLE);
+
+		tag = __blk_mq_get_tag(data, bt);
+		if (tag != -1)
+			break;
+
 		if (data->ctx)
 			blk_mq_put_ctx(data->ctx);
 

block/blk-mq.c

@@ -95,8 +95,7 @@ static void blk_mq_check_inflight(struct blk_mq_hw_ctx *hctx,
 {
 	struct mq_inflight *mi = priv;
 
-	if (test_bit(REQ_ATOM_STARTED, &rq->atomic_flags) &&
-	    !test_bit(REQ_ATOM_COMPLETE, &rq->atomic_flags)) {
+	if (blk_mq_rq_state(rq) == MQ_RQ_IN_FLIGHT) {
 		/*
 		 * index[0] counts the specific partition that was asked
 		 * for. index[1] counts the ones that are active on the
@@ -222,7 +221,7 @@ void blk_mq_quiesce_queue(struct request_queue *q)
 
 	queue_for_each_hw_ctx(q, hctx, i) {
 		if (hctx->flags & BLK_MQ_F_BLOCKING)
-			synchronize_srcu(hctx->queue_rq_srcu);
+			synchronize_srcu(hctx->srcu);
 		else
 			rcu = true;
 	}
@@ -272,15 +271,14 @@ static struct request *blk_mq_rq_ctx_init(struct blk_mq_alloc_data *data,
 {
 	struct blk_mq_tags *tags = blk_mq_tags_from_data(data);
 	struct request *rq = tags->static_rqs[tag];
-
-	rq->rq_flags = 0;
+	req_flags_t rq_flags = 0;
 
 	if (data->flags & BLK_MQ_REQ_INTERNAL) {
 		rq->tag = -1;
 		rq->internal_tag = tag;
 	} else {
 		if (blk_mq_tag_busy(data->hctx)) {
-			rq->rq_flags = RQF_MQ_INFLIGHT;
+			rq_flags = RQF_MQ_INFLIGHT;
 			atomic_inc(&data->hctx->nr_active);
 		}
 		rq->tag = tag;
@@ -288,27 +286,22 @@ static struct request *blk_mq_rq_ctx_init(struct blk_mq_alloc_data *data,
 		data->hctx->tags->rqs[rq->tag] = rq;
 	}
 
-	INIT_LIST_HEAD(&rq->queuelist);
 	/* csd/requeue_work/fifo_time is initialized before use */
 	rq->q = data->q;
 	rq->mq_ctx = data->ctx;
+	rq->rq_flags = rq_flags;
+	rq->cpu = -1;
 	rq->cmd_flags = op;
 	if (data->flags & BLK_MQ_REQ_PREEMPT)
 		rq->rq_flags |= RQF_PREEMPT;
 	if (blk_queue_io_stat(data->q))
 		rq->rq_flags |= RQF_IO_STAT;
-	/* do not touch atomic flags, it needs atomic ops against the timer */
-	rq->cpu = -1;
+	INIT_LIST_HEAD(&rq->queuelist);
 	INIT_HLIST_NODE(&rq->hash);
 	RB_CLEAR_NODE(&rq->rb_node);
 	rq->rq_disk = NULL;
 	rq->part = NULL;
 	rq->start_time = jiffies;
-#ifdef CONFIG_BLK_CGROUP
-	rq->rl = NULL;
-	set_start_time_ns(rq);
-	rq->io_start_time_ns = 0;
-#endif
 	rq->nr_phys_segments = 0;
 #if defined(CONFIG_BLK_DEV_INTEGRITY)
 	rq->nr_integrity_segments = 0;
@@ -316,6 +309,7 @@ static struct request *blk_mq_rq_ctx_init(struct blk_mq_alloc_data *data,
 	rq->special = NULL;
 	/* tag was already set */
 	rq->extra_len = 0;
+	rq->__deadline = 0;
 
 	INIT_LIST_HEAD(&rq->timeout_list);
 	rq->timeout = 0;
@@ -324,6 +318,12 @@ static struct request *blk_mq_rq_ctx_init(struct blk_mq_alloc_data *data,
 	rq->end_io_data = NULL;
 	rq->next_rq = NULL;
 
+#ifdef CONFIG_BLK_CGROUP
+	rq->rl = NULL;
+	set_start_time_ns(rq);
+	rq->io_start_time_ns = 0;
+#endif
+
 	data->ctx->rq_dispatched[op_is_sync(op)]++;
 	return rq;
 }
@@ -443,7 +443,7 @@ struct request *blk_mq_alloc_request_hctx(struct request_queue *q,
 		blk_queue_exit(q);
 		return ERR_PTR(-EXDEV);
 	}
-	cpu = cpumask_first(alloc_data.hctx->cpumask);
+	cpu = cpumask_first_and(alloc_data.hctx->cpumask, cpu_online_mask);
 	alloc_data.ctx = __blk_mq_get_ctx(q, cpu);
 
 	rq = blk_mq_get_request(q, NULL, op, &alloc_data);
@@ -485,8 +485,7 @@ void blk_mq_free_request(struct request *rq)
 	if (blk_rq_rl(rq))
 		blk_put_rl(blk_rq_rl(rq));
 
-	clear_bit(REQ_ATOM_STARTED, &rq->atomic_flags);
-	clear_bit(REQ_ATOM_POLL_SLEPT, &rq->atomic_flags);
+	blk_mq_rq_update_state(rq, MQ_RQ_IDLE);
 	if (rq->tag != -1)
 		blk_mq_put_tag(hctx, hctx->tags, ctx, rq->tag);
 	if (sched_tag != -1)
@@ -532,6 +531,9 @@ static void __blk_mq_complete_request(struct request *rq)
 	bool shared = false;
 	int cpu;
 
+	WARN_ON_ONCE(blk_mq_rq_state(rq) != MQ_RQ_IN_FLIGHT);
+	blk_mq_rq_update_state(rq, MQ_RQ_COMPLETE);
+
 	if (rq->internal_tag != -1)
 		blk_mq_sched_completed_request(rq);
 	if (rq->rq_flags & RQF_STATS) {
@@ -559,6 +561,56 @@ static void __blk_mq_complete_request(struct request *rq)
 	put_cpu();
 }
 
+static void hctx_unlock(struct blk_mq_hw_ctx *hctx, int srcu_idx)
+	__releases(hctx->srcu)
+{
+	if (!(hctx->flags & BLK_MQ_F_BLOCKING))
+		rcu_read_unlock();
+	else
+		srcu_read_unlock(hctx->srcu, srcu_idx);
+}
+
+static void hctx_lock(struct blk_mq_hw_ctx *hctx, int *srcu_idx)
+	__acquires(hctx->srcu)
+{
+	if (!(hctx->flags & BLK_MQ_F_BLOCKING)) {
+		/* shut up gcc false positive */
+		*srcu_idx = 0;
+		rcu_read_lock();
+	} else
+		*srcu_idx = srcu_read_lock(hctx->srcu);
+}
+
+static void blk_mq_rq_update_aborted_gstate(struct request *rq, u64 gstate)
+{
+	unsigned long flags;
+
+	/*
+	 * blk_mq_rq_aborted_gstate() is used from the completion path and
+	 * can thus be called from irq context.  u64_stats_fetch in the
+	 * middle of update on the same CPU leads to lockup.  Disable irq
+	 * while updating.
+	 */
+	local_irq_save(flags);
+	u64_stats_update_begin(&rq->aborted_gstate_sync);
+	rq->aborted_gstate = gstate;
+	u64_stats_update_end(&rq->aborted_gstate_sync);
+	local_irq_restore(flags);
+}
+
+static u64 blk_mq_rq_aborted_gstate(struct request *rq)
+{
+	unsigned int start;
+	u64 aborted_gstate;
+
+	do {
+		start = u64_stats_fetch_begin(&rq->aborted_gstate_sync);
+		aborted_gstate = rq->aborted_gstate;
+	} while (u64_stats_fetch_retry(&rq->aborted_gstate_sync, start));
+
+	return aborted_gstate;
+}
+
 /**
  * blk_mq_complete_request - end I/O on a request
  * @rq:		the request being processed
@@ -570,17 +622,33 @@ static void __blk_mq_complete_request(struct request *rq)
 void blk_mq_complete_request(struct request *rq)
 {
 	struct request_queue *q = rq->q;
+	struct blk_mq_hw_ctx *hctx = blk_mq_map_queue(q, rq->mq_ctx->cpu);
+	int srcu_idx;
 
 	if (unlikely(blk_should_fake_timeout(q)))
 		return;
-	if (!blk_mark_rq_complete(rq))
+
+	/*
+	 * If @rq->aborted_gstate equals the current instance, timeout is
+	 * claiming @rq and we lost.  This is synchronized through
+	 * hctx_lock().  See blk_mq_timeout_work() for details.
+	 *
+	 * Completion path never blocks and we can directly use RCU here
+	 * instead of hctx_lock() which can be either RCU or SRCU.
+	 * However, that would complicate paths which want to synchronize
+	 * against us.  Let stay in sync with the issue path so that
+	 * hctx_lock() covers both issue and completion paths.
+	 */
+	hctx_lock(hctx, &srcu_idx);
+	if (blk_mq_rq_aborted_gstate(rq) != rq->gstate)
 		__blk_mq_complete_request(rq);
+	hctx_unlock(hctx, srcu_idx);
 }
 EXPORT_SYMBOL(blk_mq_complete_request);
 
 int blk_mq_request_started(struct request *rq)
 {
-	return test_bit(REQ_ATOM_STARTED, &rq->atomic_flags);
+	return blk_mq_rq_state(rq) != MQ_RQ_IDLE;
 }
 EXPORT_SYMBOL_GPL(blk_mq_request_started);
 
@@ -598,34 +666,27 @@ void blk_mq_start_request(struct request *rq)
 		wbt_issue(q->rq_wb, &rq->issue_stat);
 	}
 
-	blk_add_timer(rq);
-
-	WARN_ON_ONCE(test_bit(REQ_ATOM_STARTED, &rq->atomic_flags));
+	WARN_ON_ONCE(blk_mq_rq_state(rq) != MQ_RQ_IDLE);
 
 	/*
-	 * Mark us as started and clear complete. Complete might have been
-	 * set if requeue raced with timeout, which then marked it as
-	 * complete. So be sure to clear complete again when we start
-	 * the request, otherwise we'll ignore the completion event.
+	 * Mark @rq in-flight which also advances the generation number,
+	 * and register for timeout.  Protect with a seqcount to allow the
+	 * timeout path to read both @rq->gstate and @rq->deadline
+	 * coherently.
 	 *
-	 * Ensure that ->deadline is visible before we set STARTED, such that
-	 * blk_mq_check_expired() is guaranteed to observe our ->deadline when
-	 * it observes STARTED.
+	 * This is the only place where a request is marked in-flight.  If
+	 * the timeout path reads an in-flight @rq->gstate, the
+	 * @rq->deadline it reads together under @rq->gstate_seq is
+	 * guaranteed to be the matching one.
 	 */
-	smp_wmb();
-	set_bit(REQ_ATOM_STARTED, &rq->atomic_flags);
-	if (test_bit(REQ_ATOM_COMPLETE, &rq->atomic_flags)) {
-		/*
-		 * Coherence order guarantees these consecutive stores to a
-		 * single variable propagate in the specified order. Thus the
-		 * clear_bit() is ordered _after_ the set bit. See
-		 * blk_mq_check_expired().
-		 *
-		 * (the bits must be part of the same byte for this to be
-		 * true).
-		 */
-		clear_bit(REQ_ATOM_COMPLETE, &rq->atomic_flags);
-	}
+	preempt_disable();
+	write_seqcount_begin(&rq->gstate_seq);
+
+	blk_mq_rq_update_state(rq, MQ_RQ_IN_FLIGHT);
+	blk_add_timer(rq);
+
+	write_seqcount_end(&rq->gstate_seq);
+	preempt_enable();
 
 	if (q->dma_drain_size && blk_rq_bytes(rq)) {
 		/*
@@ -639,13 +700,9 @@ void blk_mq_start_request(struct request *rq)
 EXPORT_SYMBOL(blk_mq_start_request);
 
 /*
- * When we reach here because queue is busy, REQ_ATOM_COMPLETE
- * flag isn't set yet, so there may be race with timeout handler,
- * but given rq->deadline is just set in .queue_rq() under
- * this situation, the race won't be possible in reality because
- * rq->timeout should be set as big enough to cover the window
- * between blk_mq_start_request() called from .queue_rq() and
- * clearing REQ_ATOM_STARTED here.
+ * When we reach here because queue is busy, it's safe to change the state
+ * to IDLE without checking @rq->aborted_gstate because we should still be
+ * holding the RCU read lock and thus protected against timeout.
  */
 static void __blk_mq_requeue_request(struct request *rq)
 {
@@ -657,7 +714,8 @@ static void __blk_mq_requeue_request(struct request *rq)
 	wbt_requeue(q->rq_wb, &rq->issue_stat);
 	blk_mq_sched_requeue_request(rq);
 
-	if (test_and_clear_bit(REQ_ATOM_STARTED, &rq->atomic_flags)) {
+	if (blk_mq_rq_state(rq) != MQ_RQ_IDLE) {
+		blk_mq_rq_update_state(rq, MQ_RQ_IDLE);
 		if (q->dma_drain_size && blk_rq_bytes(rq))
 			rq->nr_phys_segments--;
 	}
@@ -689,13 +747,13 @@ static void blk_mq_requeue_work(struct work_struct *work)
 
 		rq->rq_flags &= ~RQF_SOFTBARRIER;
 		list_del_init(&rq->queuelist);
-		blk_mq_sched_insert_request(rq, true, false, false, true);
+		blk_mq_sched_insert_request(rq, true, false, false);
 	}
 
 	while (!list_empty(&rq_list)) {
 		rq = list_entry(rq_list.next, struct request, queuelist);
 		list_del_init(&rq->queuelist);
-		blk_mq_sched_insert_request(rq, false, false, false, true);
+		blk_mq_sched_insert_request(rq, false, false, false);
 	}
 
 	blk_mq_run_hw_queues(q, false);
@@ -729,7 +787,7 @@ EXPORT_SYMBOL(blk_mq_add_to_requeue_list);
 
 void blk_mq_kick_requeue_list(struct request_queue *q)
 {
-	kblockd_schedule_delayed_work(&q->requeue_work, 0);
+	kblockd_mod_delayed_work_on(WORK_CPU_UNBOUND, &q->requeue_work, 0);
 }
 EXPORT_SYMBOL(blk_mq_kick_requeue_list);
 
@@ -755,24 +813,15 @@ EXPORT_SYMBOL(blk_mq_tag_to_rq);
 struct blk_mq_timeout_data {
 	unsigned long next;
 	unsigned int next_set;
+	unsigned int nr_expired;
 };
 
-void blk_mq_rq_timed_out(struct request *req, bool reserved)
+static void blk_mq_rq_timed_out(struct request *req, bool reserved)
 {
 	const struct blk_mq_ops *ops = req->q->mq_ops;
 	enum blk_eh_timer_return ret = BLK_EH_RESET_TIMER;
 
-	/*
-	 * We know that complete is set at this point. If STARTED isn't set
-	 * anymore, then the request isn't active and the "timeout" should
-	 * just be ignored. This can happen due to the bitflag ordering.
-	 * Timeout first checks if STARTED is set, and if it is, assumes
-	 * the request is active. But if we race with completion, then
-	 * both flags will get cleared. So check here again, and ignore
-	 * a timeout event with a request that isn't active.
-	 */
-	if (!test_bit(REQ_ATOM_STARTED, &req->atomic_flags))
-		return;
+	req->rq_flags |= RQF_MQ_TIMEOUT_EXPIRED;
 
 	if (ops->timeout)
 		ret = ops->timeout(req, reserved);
@@ -782,8 +831,13 @@ void blk_mq_rq_timed_out(struct request *req, bool reserved)
 		__blk_mq_complete_request(req);
 		break;
 	case BLK_EH_RESET_TIMER:
+		/*
+		 * As nothing prevents from completion happening while
+		 * ->aborted_gstate is set, this may lead to ignored
+		 * completions and further spurious timeouts.
+		 */
+		blk_mq_rq_update_aborted_gstate(req, 0);
 		blk_add_timer(req);
-		blk_clear_rq_complete(req);
 		break;
 	case BLK_EH_NOT_HANDLED:
 		break;
@@ -797,50 +851,51 @@ static void blk_mq_check_expired(struct blk_mq_hw_ctx *hctx,
 		struct request *rq, void *priv, bool reserved)
 {
 	struct blk_mq_timeout_data *data = priv;
-	unsigned long deadline;
+	unsigned long gstate, deadline;
+	int start;
 
-	if (!test_bit(REQ_ATOM_STARTED, &rq->atomic_flags))
-		return;
+	might_sleep();
 
-	/*
-	 * Ensures that if we see STARTED we must also see our
-	 * up-to-date deadline, see blk_mq_start_request().
-	 */
-	smp_rmb();
+	if (rq->rq_flags & RQF_MQ_TIMEOUT_EXPIRED)
+		return;
 
-	deadline = READ_ONCE(rq->deadline);
+	/* read coherent snapshots of @rq->state_gen and @rq->deadline */
+	while (true) {
+		start = read_seqcount_begin(&rq->gstate_seq);
+		gstate = READ_ONCE(rq->gstate);
+		deadline = blk_rq_deadline(rq);
+		if (!read_seqcount_retry(&rq->gstate_seq, start))
+			break;
+		cond_resched();
+	}
 
-	/*
-	 * The rq being checked may have been freed and reallocated
-	 * out already here, we avoid this race by checking rq->deadline
-	 * and REQ_ATOM_COMPLETE flag together:
-	 *
-	 * - if rq->deadline is observed as new value because of
-	 *   reusing, the rq won't be timed out because of timing.
-	 * - if rq->deadline is observed as previous value,
-	 *   REQ_ATOM_COMPLETE flag won't be cleared in reuse path
-	 *   because we put a barrier between setting rq->deadline
-	 *   and clearing the flag in blk_mq_start_request(), so
-	 *   this rq won't be timed out too.
-	 */
-	if (time_after_eq(jiffies, deadline)) {
-		if (!blk_mark_rq_complete(rq)) {
-			/*
-			 * Again coherence order ensures that consecutive reads
-			 * from the same variable must be in that order. This
-			 * ensures that if we see COMPLETE clear, we must then
-			 * see STARTED set and we'll ignore this timeout.
-			 *
-			 * (There's also the MB implied by the test_and_clear())
-			 */
-			blk_mq_rq_timed_out(rq, reserved);
-		}
+	/* if in-flight && overdue, mark for abortion */
+	if ((gstate & MQ_RQ_STATE_MASK) == MQ_RQ_IN_FLIGHT &&
+	    time_after_eq(jiffies, deadline)) {
+		blk_mq_rq_update_aborted_gstate(rq, gstate);
+		data->nr_expired++;
+		hctx->nr_expired++;
 	} else if (!data->next_set || time_after(data->next, deadline)) {
 		data->next = deadline;
 		data->next_set = 1;
 	}
 }
 
+static void blk_mq_terminate_expired(struct blk_mq_hw_ctx *hctx,
+		struct request *rq, void *priv, bool reserved)
+{
+	/*
+	 * We marked @rq->aborted_gstate and waited for RCU.  If there were
+	 * completions that we lost to, they would have finished and
+	 * updated @rq->gstate by now; otherwise, the completion path is
+	 * now guaranteed to see @rq->aborted_gstate and yield.  If
+	 * @rq->aborted_gstate still matches @rq->gstate, @rq is ours.
+	 */
+	if (!(rq->rq_flags & RQF_MQ_TIMEOUT_EXPIRED) &&
+	    READ_ONCE(rq->gstate) == rq->aborted_gstate)
+		blk_mq_rq_timed_out(rq, reserved);
+}
+
 static void blk_mq_timeout_work(struct work_struct *work)
 {
 	struct request_queue *q =
@@ -848,7 +903,9 @@ static void blk_mq_timeout_work(struct work_struct *work)
 	struct blk_mq_timeout_data data = {
 		.next		= 0,
 		.next_set	= 0,
+		.nr_expired	= 0,
 	};
+	struct blk_mq_hw_ctx *hctx;
 	int i;
 
 	/* A deadlock might occur if a request is stuck requiring a
@@ -867,14 +924,46 @@ static void blk_mq_timeout_work(struct work_struct *work)
 	if (!percpu_ref_tryget(&q->q_usage_counter))
 		return;
 
+	/* scan for the expired ones and set their ->aborted_gstate */
 	blk_mq_queue_tag_busy_iter(q, blk_mq_check_expired, &data);
 
+	if (data.nr_expired) {
+		bool has_rcu = false;
+
+		/*
+		 * Wait till everyone sees ->aborted_gstate.  The
+		 * sequential waits for SRCUs aren't ideal.  If this ever
+		 * becomes a problem, we can add per-hw_ctx rcu_head and
+		 * wait in parallel.
+		 */
+		queue_for_each_hw_ctx(q, hctx, i) {
+			if (!hctx->nr_expired)
+				continue;
+
+			if (!(hctx->flags & BLK_MQ_F_BLOCKING))
+				has_rcu = true;
+			else
+				synchronize_srcu(hctx->srcu);
+
+			hctx->nr_expired = 0;
+		}
+		if (has_rcu)
+			synchronize_rcu();
+
+		/* terminate the ones we won */
+		blk_mq_queue_tag_busy_iter(q, blk_mq_terminate_expired, NULL);
+	}
+
 	if (data.next_set) {
 		data.next = blk_rq_timeout(round_jiffies_up(data.next));
 		mod_timer(&q->timeout, data.next);
 	} else {
-		struct blk_mq_hw_ctx *hctx;
-
+		/*
+		 * Request timeouts are handled as a forward rolling timer. If
+		 * we end up here it means that no requests are pending and
+		 * also that no request has been pending for a while. Mark
+		 * each hctx as idle.
+		 */
 		queue_for_each_hw_ctx(q, hctx, i) {
 			/* the hctx may be unmapped, so check it here */
 			if (blk_mq_hw_queue_mapped(hctx))
@@ -1010,66 +1099,67 @@ static int blk_mq_dispatch_wake(wait_queue_entry_t *wait, unsigned mode,
 
 /*
  * Mark us waiting for a tag. For shared tags, this involves hooking us into
- * the tag wakeups. For non-shared tags, we can simply mark us nedeing a
- * restart. For both caes, take care to check the condition again after
+ * the tag wakeups. For non-shared tags, we can simply mark us needing a
+ * restart. For both cases, take care to check the condition again after
  * marking us as waiting.
  */
 static bool blk_mq_mark_tag_wait(struct blk_mq_hw_ctx **hctx,
 				 struct request *rq)
 {
 	struct blk_mq_hw_ctx *this_hctx = *hctx;
-	bool shared_tags = (this_hctx->flags & BLK_MQ_F_TAG_SHARED) != 0;
 	struct sbq_wait_state *ws;
 	wait_queue_entry_t *wait;
 	bool ret;
 
-	if (!shared_tags) {
+	if (!(this_hctx->flags & BLK_MQ_F_TAG_SHARED)) {
 		if (!test_bit(BLK_MQ_S_SCHED_RESTART, &this_hctx->state))
 			set_bit(BLK_MQ_S_SCHED_RESTART, &this_hctx->state);
-	} else {
-		wait = &this_hctx->dispatch_wait;
-		if (!list_empty_careful(&wait->entry))
-			return false;
 
-		spin_lock(&this_hctx->lock);
-		if (!list_empty(&wait->entry)) {
-			spin_unlock(&this_hctx->lock);
-			return false;
-		}
+		/*
+		 * It's possible that a tag was freed in the window between the
+		 * allocation failure and adding the hardware queue to the wait
+		 * queue.
+		 *
+		 * Don't clear RESTART here, someone else could have set it.
+		 * At most this will cost an extra queue run.
+		 */
+		return blk_mq_get_driver_tag(rq, hctx, false);
+	}
 
-		ws = bt_wait_ptr(&this_hctx->tags->bitmap_tags, this_hctx);
-		add_wait_queue(&ws->wait, wait);
+	wait = &this_hctx->dispatch_wait;
+	if (!list_empty_careful(&wait->entry))
+		return false;
+
+	spin_lock(&this_hctx->lock);
+	if (!list_empty(&wait->entry)) {
+		spin_unlock(&this_hctx->lock);
+		return false;
 	}
 
+	ws = bt_wait_ptr(&this_hctx->tags->bitmap_tags, this_hctx);
+	add_wait_queue(&ws->wait, wait);
+
 	/*
 	 * It's possible that a tag was freed in the window between the
 	 * allocation failure and adding the hardware queue to the wait
 	 * queue.
 	 */
 	ret = blk_mq_get_driver_tag(rq, hctx, false);
-
-	if (!shared_tags) {
-		/*
-		 * Don't clear RESTART here, someone else could have set it.
-		 * At most this will cost an extra queue run.
-		 */
-		return ret;
-	} else {
-		if (!ret) {
-			spin_unlock(&this_hctx->lock);
-			return false;
-		}
-
-		/*
-		 * We got a tag, remove ourselves from the wait queue to ensure
-		 * someone else gets the wakeup.
-		 */
-		spin_lock_irq(&ws->wait.lock);
-		list_del_init(&wait->entry);
-		spin_unlock_irq(&ws->wait.lock);
+	if (!ret) {
 		spin_unlock(&this_hctx->lock);
-		return true;
+		return false;
 	}
+
+	/*
+	 * We got a tag, remove ourselves from the wait queue to ensure
+	 * someone else gets the wakeup.
+	 */
+	spin_lock_irq(&ws->wait.lock);
+	list_del_init(&wait->entry);
+	spin_unlock_irq(&ws->wait.lock);
+	spin_unlock(&this_hctx->lock);
+
+	return true;
 }
 
 bool blk_mq_dispatch_rq_list(struct request_queue *q, struct list_head *list,
@@ -1206,9 +1296,27 @@ static void __blk_mq_run_hw_queue(struct blk_mq_hw_ctx *hctx)
 	/*
 	 * We should be running this queue from one of the CPUs that
 	 * are mapped to it.
+	 *
+	 * There are at least two related races now between setting
+	 * hctx->next_cpu from blk_mq_hctx_next_cpu() and running
+	 * __blk_mq_run_hw_queue():
+	 *
+	 * - hctx->next_cpu is found offline in blk_mq_hctx_next_cpu(),
+	 *   but later it becomes online, then this warning is harmless
+	 *   at all
+	 *
+	 * - hctx->next_cpu is found online in blk_mq_hctx_next_cpu(),
+	 *   but later it becomes offline, then the warning can't be
+	 *   triggered, and we depend on blk-mq timeout handler to
+	 *   handle dispatched requests to this hctx
 	 */
-	WARN_ON(!cpumask_test_cpu(raw_smp_processor_id(), hctx->cpumask) &&
-		cpu_online(hctx->next_cpu));
+	if (!cpumask_test_cpu(raw_smp_processor_id(), hctx->cpumask) &&
+		cpu_online(hctx->next_cpu)) {
+		printk(KERN_WARNING "run queue from wrong CPU %d, hctx %s\n",
+			raw_smp_processor_id(),
+			cpumask_empty(hctx->cpumask) ? "inactive": "active");
+		dump_stack();
+	}
 
 	/*
 	 * We can't run the queue inline with ints disabled. Ensure that
@@ -1216,17 +1324,11 @@ static void __blk_mq_run_hw_queue(struct blk_mq_hw_ctx *hctx)
 	 */
 	WARN_ON_ONCE(in_interrupt());
 
-	if (!(hctx->flags & BLK_MQ_F_BLOCKING)) {
-		rcu_read_lock();
-		blk_mq_sched_dispatch_requests(hctx);
-		rcu_read_unlock();
-	} else {
-		might_sleep();
+	might_sleep_if(hctx->flags & BLK_MQ_F_BLOCKING);
 
-		srcu_idx = srcu_read_lock(hctx->queue_rq_srcu);
-		blk_mq_sched_dispatch_requests(hctx);
-		srcu_read_unlock(hctx->queue_rq_srcu, srcu_idx);
-	}
+	hctx_lock(hctx, &srcu_idx);
+	blk_mq_sched_dispatch_requests(hctx);
+	hctx_unlock(hctx, srcu_idx);
 }
 
 /*
@@ -1237,20 +1339,47 @@ static void __blk_mq_run_hw_queue(struct blk_mq_hw_ctx *hctx)
  */
 static int blk_mq_hctx_next_cpu(struct blk_mq_hw_ctx *hctx)
 {
+	bool tried = false;
+
 	if (hctx->queue->nr_hw_queues == 1)
 		return WORK_CPU_UNBOUND;
 
 	if (--hctx->next_cpu_batch <= 0) {
 		int next_cpu;
-
-		next_cpu = cpumask_next(hctx->next_cpu, hctx->cpumask);
+select_cpu:
+		next_cpu = cpumask_next_and(hctx->next_cpu, hctx->cpumask,
+				cpu_online_mask);
 		if (next_cpu >= nr_cpu_ids)
-			next_cpu = cpumask_first(hctx->cpumask);
+			next_cpu = cpumask_first_and(hctx->cpumask,cpu_online_mask);
 
-		hctx->next_cpu = next_cpu;
+		/*
+		 * No online CPU is found, so have to make sure hctx->next_cpu
+		 * is set correctly for not breaking workqueue.
+		 */
+		if (next_cpu >= nr_cpu_ids)
+			hctx->next_cpu = cpumask_first(hctx->cpumask);
+		else
+			hctx->next_cpu = next_cpu;
 		hctx->next_cpu_batch = BLK_MQ_CPU_WORK_BATCH;
 	}
 
+	/*
+	 * Do unbound schedule if we can't find a online CPU for this hctx,
+	 * and it should only happen in the path of handling CPU DEAD.
+	 */
+	if (!cpu_online(hctx->next_cpu)) {
+		if (!tried) {
+			tried = true;
+			goto select_cpu;
+		}
+
+		/*
+		 * Make sure to re-select CPU next time once after CPUs
+		 * in hctx->cpumask become online again.
+		 */
+		hctx->next_cpu_batch = 1;
+		return WORK_CPU_UNBOUND;
+	}
 	return hctx->next_cpu;
 }
 
@@ -1274,9 +1403,8 @@ static void __blk_mq_delay_run_hw_queue(struct blk_mq_hw_ctx *hctx, bool async,
 		put_cpu();
 	}
 
-	kblockd_schedule_delayed_work_on(blk_mq_hctx_next_cpu(hctx),
-					 &hctx->run_work,
-					 msecs_to_jiffies(msecs));
+	kblockd_mod_delayed_work_on(blk_mq_hctx_next_cpu(hctx), &hctx->run_work,
+				    msecs_to_jiffies(msecs));
 }
 
 void blk_mq_delay_run_hw_queue(struct blk_mq_hw_ctx *hctx, unsigned long msecs)
@@ -1287,7 +1415,23 @@ EXPORT_SYMBOL(blk_mq_delay_run_hw_queue);
 
 bool blk_mq_run_hw_queue(struct blk_mq_hw_ctx *hctx, bool async)
 {
-	if (blk_mq_hctx_has_pending(hctx)) {
+	int srcu_idx;
+	bool need_run;
+
+	/*
+	 * When queue is quiesced, we may be switching io scheduler, or
+	 * updating nr_hw_queues, or other things, and we can't run queue
+	 * any more, even __blk_mq_hctx_has_pending() can't be called safely.
+	 *
+	 * And queue will be rerun in blk_mq_unquiesce_queue() if it is
+	 * quiesced.
+	 */
+	hctx_lock(hctx, &srcu_idx);
+	need_run = !blk_queue_quiesced(hctx->queue) &&
+		blk_mq_hctx_has_pending(hctx);
+	hctx_unlock(hctx, srcu_idx);
+
+	if (need_run) {
 		__blk_mq_delay_run_hw_queue(hctx, async, 0);
 		return true;
 	}
@@ -1595,9 +1739,9 @@ static blk_qc_t request_to_qc_t(struct blk_mq_hw_ctx *hctx, struct request *rq)
 	return blk_tag_to_qc_t(rq->internal_tag, hctx->queue_num, true);
 }
 
-static void __blk_mq_try_issue_directly(struct blk_mq_hw_ctx *hctx,
-					struct request *rq,
-					blk_qc_t *cookie, bool may_sleep)
+static blk_status_t __blk_mq_issue_directly(struct blk_mq_hw_ctx *hctx,
+					    struct request *rq,
+					    blk_qc_t *cookie)
 {
 	struct request_queue *q = rq->q;
 	struct blk_mq_queue_data bd = {
@@ -1606,15 +1750,52 @@ static void __blk_mq_try_issue_directly(struct blk_mq_hw_ctx *hctx,
 	};
 	blk_qc_t new_cookie;
 	blk_status_t ret;
+
+	new_cookie = request_to_qc_t(hctx, rq);
+
+	/*
+	 * For OK queue, we are done. For error, caller may kill it.
+	 * Any other error (busy), just add it to our list as we
+	 * previously would have done.
+	 */
+	ret = q->mq_ops->queue_rq(hctx, &bd);
+	switch (ret) {
+	case BLK_STS_OK:
+		*cookie = new_cookie;
+		break;
+	case BLK_STS_RESOURCE:
+		__blk_mq_requeue_request(rq);
+		break;
+	default:
+		*cookie = BLK_QC_T_NONE;
+		break;
+	}
+
+	return ret;
+}
+
+static blk_status_t __blk_mq_try_issue_directly(struct blk_mq_hw_ctx *hctx,
+						struct request *rq,
+						blk_qc_t *cookie,
+						bool bypass_insert)
+{
+	struct request_queue *q = rq->q;
 	bool run_queue = true;
 
-	/* RCU or SRCU read lock is needed before checking quiesced flag */
+	/*
+	 * RCU or SRCU read lock is needed before checking quiesced flag.
+	 *
+	 * When queue is stopped or quiesced, ignore 'bypass_insert' from
+	 * blk_mq_request_issue_directly(), and return BLK_STS_OK to caller,
+	 * and avoid driver to try to dispatch again.
+	 */
 	if (blk_mq_hctx_stopped(hctx) || blk_queue_quiesced(q)) {
 		run_queue = false;
+		bypass_insert = false;
 		goto insert;
 	}
 
-	if (q->elevator)
+	if (q->elevator && !bypass_insert)
 		goto insert;
 
 	if (!blk_mq_get_driver_tag(rq, NULL, false))
@@ -1625,47 +1806,47 @@ static void __blk_mq_try_issue_directly(struct blk_mq_hw_ctx *hctx,
 		goto insert;
 	}
 
-	new_cookie = request_to_qc_t(hctx, rq);
-
-	/*
-	 * For OK queue, we are done. For error, kill it. Any other
-	 * error (busy), just add it to our list as we previously
-	 * would have done
-	 */
-	ret = q->mq_ops->queue_rq(hctx, &bd);
-	switch (ret) {
-	case BLK_STS_OK:
-		*cookie = new_cookie;
-		return;
-	case BLK_STS_RESOURCE:
-		__blk_mq_requeue_request(rq);
-		goto insert;
-	default:
-		*cookie = BLK_QC_T_NONE;
-		blk_mq_end_request(rq, ret);
-		return;
-	}
-
+	return __blk_mq_issue_directly(hctx, rq, cookie);
 insert:
-	blk_mq_sched_insert_request(rq, false, run_queue, false, may_sleep);
+	if (bypass_insert)
+		return BLK_STS_RESOURCE;
+
+	blk_mq_sched_insert_request(rq, false, run_queue, false);
+	return BLK_STS_OK;
 }
 
 static void blk_mq_try_issue_directly(struct blk_mq_hw_ctx *hctx,
 		struct request *rq, blk_qc_t *cookie)
 {
-	if (!(hctx->flags & BLK_MQ_F_BLOCKING)) {
-		rcu_read_lock();
-		__blk_mq_try_issue_directly(hctx, rq, cookie, false);
-		rcu_read_unlock();
-	} else {
-		unsigned int srcu_idx;
+	blk_status_t ret;
+	int srcu_idx;
 
-		might_sleep();
+	might_sleep_if(hctx->flags & BLK_MQ_F_BLOCKING);
 
-		srcu_idx = srcu_read_lock(hctx->queue_rq_srcu);
-		__blk_mq_try_issue_directly(hctx, rq, cookie, true);
-		srcu_read_unlock(hctx->queue_rq_srcu, srcu_idx);
-	}
+	hctx_lock(hctx, &srcu_idx);
+
+	ret = __blk_mq_try_issue_directly(hctx, rq, cookie, false);
+	if (ret == BLK_STS_RESOURCE)
+		blk_mq_sched_insert_request(rq, false, true, false);
+	else if (ret != BLK_STS_OK)
+		blk_mq_end_request(rq, ret);
+
+	hctx_unlock(hctx, srcu_idx);
+}
+
+blk_status_t blk_mq_request_issue_directly(struct request *rq)
+{
+	blk_status_t ret;
+	int srcu_idx;
+	blk_qc_t unused_cookie;
+	struct blk_mq_ctx *ctx = rq->mq_ctx;
+	struct blk_mq_hw_ctx *hctx = blk_mq_map_queue(rq->q, ctx->cpu);
+
+	hctx_lock(hctx, &srcu_idx);
+	ret = __blk_mq_try_issue_directly(hctx, rq, &unused_cookie, true);
+	hctx_unlock(hctx, srcu_idx);
+
+	return ret;
 }
 
 static blk_qc_t blk_mq_make_request(struct request_queue *q, struct bio *bio)
@@ -1776,7 +1957,7 @@ static blk_qc_t blk_mq_make_request(struct request_queue *q, struct bio *bio)
 	} else if (q->elevator) {
 		blk_mq_put_ctx(data.ctx);
 		blk_mq_bio_to_request(rq, bio);
-		blk_mq_sched_insert_request(rq, false, true, true, true);
+		blk_mq_sched_insert_request(rq, false, true, true);
 	} else {
 		blk_mq_put_ctx(data.ctx);
 		blk_mq_bio_to_request(rq, bio);
@@ -1869,6 +2050,22 @@ static size_t order_to_size(unsigned int order)
 	return (size_t)PAGE_SIZE << order;
 }
 
+static int blk_mq_init_request(struct blk_mq_tag_set *set, struct request *rq,
+			       unsigned int hctx_idx, int node)
+{
+	int ret;
+
+	if (set->ops->init_request) {
+		ret = set->ops->init_request(set, rq, hctx_idx, node);
+		if (ret)
+			return ret;
+	}
+
+	seqcount_init(&rq->gstate_seq);
+	u64_stats_init(&rq->aborted_gstate_sync);
+	return 0;
+}
+
 int blk_mq_alloc_rqs(struct blk_mq_tag_set *set, struct blk_mq_tags *tags,
 		     unsigned int hctx_idx, unsigned int depth)
 {
@@ -1930,12 +2127,9 @@ int blk_mq_alloc_rqs(struct blk_mq_tag_set *set, struct blk_mq_tags *tags,
 			struct request *rq = p;
 
 			tags->static_rqs[i] = rq;
-			if (set->ops->init_request) {
-				if (set->ops->init_request(set, rq, hctx_idx,
-						node)) {
-					tags->static_rqs[i] = NULL;
-					goto fail;
-				}
+			if (blk_mq_init_request(set, rq, hctx_idx, node)) {
+				tags->static_rqs[i] = NULL;
+				goto fail;
 			}
 
 			p += rq_size;
@@ -1994,7 +2188,8 @@ static void blk_mq_exit_hctx(struct request_queue *q,
 {
 	blk_mq_debugfs_unregister_hctx(hctx);
 
-	blk_mq_tag_idle(hctx);
+	if (blk_mq_hw_queue_mapped(hctx))
+		blk_mq_tag_idle(hctx);
 
 	if (set->ops->exit_request)
 		set->ops->exit_request(set, hctx->fq->flush_rq, hctx_idx);
@@ -2005,7 +2200,7 @@ static void blk_mq_exit_hctx(struct request_queue *q,
 		set->ops->exit_hctx(hctx, hctx_idx);
 
 	if (hctx->flags & BLK_MQ_F_BLOCKING)
-		cleanup_srcu_struct(hctx->queue_rq_srcu);
+		cleanup_srcu_struct(hctx->srcu);
 
 	blk_mq_remove_cpuhp(hctx);
 	blk_free_flush_queue(hctx->fq);
@@ -2074,13 +2269,11 @@ static int blk_mq_init_hctx(struct request_queue *q,
 	if (!hctx->fq)
 		goto sched_exit_hctx;
 
-	if (set->ops->init_request &&
-	    set->ops->init_request(set, hctx->fq->flush_rq, hctx_idx,
-				   node))
+	if (blk_mq_init_request(set, hctx->fq->flush_rq, hctx_idx, node))
 		goto free_fq;
 
 	if (hctx->flags & BLK_MQ_F_BLOCKING)
-		init_srcu_struct(hctx->queue_rq_srcu);
+		init_srcu_struct(hctx->srcu);
 
 	blk_mq_debugfs_register_hctx(q, hctx);
 
@@ -2116,16 +2309,11 @@ static void blk_mq_init_cpu_queues(struct request_queue *q,
 		INIT_LIST_HEAD(&__ctx->rq_list);
 		__ctx->queue = q;
 
-		/* If the cpu isn't present, the cpu is mapped to first hctx */
-		if (!cpu_present(i))
-			continue;
-
-		hctx = blk_mq_map_queue(q, i);
-
 		/*
 		 * Set local node, IFF we have more than one hw queue. If
 		 * not, we remain on the home node of the device
 		 */
+		hctx = blk_mq_map_queue(q, i);
 		if (nr_hw_queues > 1 && hctx->numa_node == NUMA_NO_NODE)
 			hctx->numa_node = local_memory_node(cpu_to_node(i));
 	}
@@ -2182,7 +2370,7 @@ static void blk_mq_map_swqueue(struct request_queue *q)
 	 *
 	 * If the cpu isn't present, the cpu is mapped to first hctx.
 	 */
-	for_each_present_cpu(i) {
+	for_each_possible_cpu(i) {
 		hctx_idx = q->mq_map[i];
 		/* unmapped hw queue can be remapped after CPU topo changed */
 		if (!set->tags[hctx_idx] &&
@@ -2236,7 +2424,8 @@ static void blk_mq_map_swqueue(struct request_queue *q)
 		/*
 		 * Initialize batch roundrobin counts
 		 */
-		hctx->next_cpu = cpumask_first(hctx->cpumask);
+		hctx->next_cpu = cpumask_first_and(hctx->cpumask,
+				cpu_online_mask);
 		hctx->next_cpu_batch = BLK_MQ_CPU_WORK_BATCH;
 	}
 }
@@ -2369,7 +2558,7 @@ static int blk_mq_hw_ctx_size(struct blk_mq_tag_set *tag_set)
 {
 	int hw_ctx_size = sizeof(struct blk_mq_hw_ctx);
 
-	BUILD_BUG_ON(ALIGN(offsetof(struct blk_mq_hw_ctx, queue_rq_srcu),
+	BUILD_BUG_ON(ALIGN(offsetof(struct blk_mq_hw_ctx, srcu),
 			   __alignof__(struct blk_mq_hw_ctx)) !=
 		     sizeof(struct blk_mq_hw_ctx));
 
@@ -2386,6 +2575,9 @@ static void blk_mq_realloc_hw_ctxs(struct blk_mq_tag_set *set,
 	struct blk_mq_hw_ctx **hctxs = q->queue_hw_ctx;
 
 	blk_mq_sysfs_unregister(q);
+
+	/* protect against switching io scheduler  */
+	mutex_lock(&q->sysfs_lock);
 	for (i = 0; i < set->nr_hw_queues; i++) {
 		int node;
 
@@ -2430,6 +2622,7 @@ static void blk_mq_realloc_hw_ctxs(struct blk_mq_tag_set *set,
 		}
 	}
 	q->nr_hw_queues = i;
+	mutex_unlock(&q->sysfs_lock);
 	blk_mq_sysfs_register(q);
 }
 
@@ -2601,9 +2794,27 @@ static int blk_mq_alloc_rq_maps(struct blk_mq_tag_set *set)
 
 static int blk_mq_update_queue_map(struct blk_mq_tag_set *set)
 {
-	if (set->ops->map_queues)
+	if (set->ops->map_queues) {
+		int cpu;
+		/*
+		 * transport .map_queues is usually done in the following
+		 * way:
+		 *
+		 * for (queue = 0; queue < set->nr_hw_queues; queue++) {
+		 * 	mask = get_cpu_mask(queue)
+		 * 	for_each_cpu(cpu, mask)
+		 * 		set->mq_map[cpu] = queue;
+		 * }
+		 *
+		 * When we need to remap, the table has to be cleared for
+		 * killing stale mapping since one CPU may not be mapped
+		 * to any hw queue.
+		 */
+		for_each_possible_cpu(cpu)
+			set->mq_map[cpu] = 0;
+
 		return set->ops->map_queues(set);
-	else
+	} else
 		return blk_mq_map_queues(set);
 }
 
@@ -2712,6 +2923,7 @@ int blk_mq_update_nr_requests(struct request_queue *q, unsigned int nr)
 		return -EINVAL;
 
 	blk_mq_freeze_queue(q);
+	blk_mq_quiesce_queue(q);
 
 	ret = 0;
 	queue_for_each_hw_ctx(q, hctx, i) {
@@ -2735,6 +2947,7 @@ int blk_mq_update_nr_requests(struct request_queue *q, unsigned int nr)
 	if (!ret)
 		q->nr_requests = nr;
 
+	blk_mq_unquiesce_queue(q);
 	blk_mq_unfreeze_queue(q);
 
 	return ret;
@@ -2850,7 +3063,7 @@ static bool blk_mq_poll_hybrid_sleep(struct request_queue *q,
 	unsigned int nsecs;
 	ktime_t kt;
 
-	if (test_bit(REQ_ATOM_POLL_SLEPT, &rq->atomic_flags))
+	if (rq->rq_flags & RQF_MQ_POLL_SLEPT)
 		return false;
 
 	/*
@@ -2870,7 +3083,7 @@ static bool blk_mq_poll_hybrid_sleep(struct request_queue *q,
 	if (!nsecs)
 		return false;
 
-	set_bit(REQ_ATOM_POLL_SLEPT, &rq->atomic_flags);
+	rq->rq_flags |= RQF_MQ_POLL_SLEPT;
 
 	/*
 	 * This will be replaced with the stats tracking code, using
@@ -2884,7 +3097,7 @@ static bool blk_mq_poll_hybrid_sleep(struct request_queue *q,
 
 	hrtimer_init_sleeper(&hs, current);
 	do {
-		if (test_bit(REQ_ATOM_COMPLETE, &rq->atomic_flags))
+		if (blk_mq_rq_state(rq) == MQ_RQ_COMPLETE)
 			break;
 		set_current_state(TASK_UNINTERRUPTIBLE);
 		hrtimer_start_expires(&hs.timer, mode);
@@ -2970,12 +3183,6 @@ static bool blk_mq_poll(struct request_queue *q, blk_qc_t cookie)
 
 static int __init blk_mq_init(void)
 {
-	/*
-	 * See comment in block/blk.h rq_atomic_flags enum
-	 */
-	BUILD_BUG_ON((REQ_ATOM_STARTED / BITS_PER_BYTE) !=
-			(REQ_ATOM_COMPLETE / BITS_PER_BYTE));
-
 	cpuhp_setup_state_multi(CPUHP_BLK_MQ_DEAD, "block/mq:dead", NULL,
 				blk_mq_hctx_notify_dead);
 	return 0;

block/blk-mq.h

@@ -27,6 +27,20 @@ struct blk_mq_ctx {
 	struct kobject		kobj;
 } ____cacheline_aligned_in_smp;
 
+/*
+ * Bits for request->gstate.  The lower two bits carry MQ_RQ_* state value
+ * and the upper bits the generation number.
+ */
+enum mq_rq_state {
+	MQ_RQ_IDLE		= 0,
+	MQ_RQ_IN_FLIGHT		= 1,
+	MQ_RQ_COMPLETE		= 2,
+
+	MQ_RQ_STATE_BITS	= 2,
+	MQ_RQ_STATE_MASK	= (1 << MQ_RQ_STATE_BITS) - 1,
+	MQ_RQ_GEN_INC		= 1 << MQ_RQ_STATE_BITS,
+};
+
 void blk_mq_freeze_queue(struct request_queue *q);
 void blk_mq_free_queue(struct request_queue *q);
 int blk_mq_update_nr_requests(struct request_queue *q, unsigned int nr);
@@ -60,6 +74,9 @@ void blk_mq_request_bypass_insert(struct request *rq, bool run_queue);
 void blk_mq_insert_requests(struct blk_mq_hw_ctx *hctx, struct blk_mq_ctx *ctx,
 				struct list_head *list);
 
+/* Used by blk_insert_cloned_request() to issue request directly */
+blk_status_t blk_mq_request_issue_directly(struct request *rq);
+
 /*
  * CPU -> queue mappings
  */
@@ -81,10 +98,41 @@ extern int blk_mq_sysfs_register(struct request_queue *q);
 extern void blk_mq_sysfs_unregister(struct request_queue *q);
 extern void blk_mq_hctx_kobj_init(struct blk_mq_hw_ctx *hctx);
 
-extern void blk_mq_rq_timed_out(struct request *req, bool reserved);
-
 void blk_mq_release(struct request_queue *q);
 
+/**
+ * blk_mq_rq_state() - read the current MQ_RQ_* state of a request
+ * @rq: target request.
+ */
+static inline int blk_mq_rq_state(struct request *rq)
+{
+	return READ_ONCE(rq->gstate) & MQ_RQ_STATE_MASK;
+}
+
+/**
+ * blk_mq_rq_update_state() - set the current MQ_RQ_* state of a request
+ * @rq: target request.
+ * @state: new state to set.
+ *
+ * Set @rq's state to @state.  The caller is responsible for ensuring that
+ * there are no other updaters.  A request can transition into IN_FLIGHT
+ * only from IDLE and doing so increments the generation number.
+ */
+static inline void blk_mq_rq_update_state(struct request *rq,
+					  enum mq_rq_state state)
+{
+	u64 old_val = READ_ONCE(rq->gstate);
+	u64 new_val = (old_val & ~MQ_RQ_STATE_MASK) | state;
+
+	if (state == MQ_RQ_IN_FLIGHT) {
+		WARN_ON_ONCE((old_val & MQ_RQ_STATE_MASK) != MQ_RQ_IDLE);
+		new_val += MQ_RQ_GEN_INC;
+	}
+
+	/* avoid exposing interim values */
+	WRITE_ONCE(rq->gstate, new_val);
+}
+
 static inline struct blk_mq_ctx *__blk_mq_get_ctx(struct request_queue *q,
 					   unsigned int cpu)
 {

block/blk-sysfs.c

@@ -853,6 +853,10 @@ struct kobj_type blk_queue_ktype = {
 	.release	= blk_release_queue,
 };
 
+/**
+ * blk_register_queue - register a block layer queue with sysfs
+ * @disk: Disk of which the request queue should be registered with sysfs.
+ */
 int blk_register_queue(struct gendisk *disk)
 {
 	int ret;
@@ -909,11 +913,12 @@ int blk_register_queue(struct gendisk *disk)
 	if (q->request_fn || (q->mq_ops && q->elevator)) {
 		ret = elv_register_queue(q);
 		if (ret) {
+			mutex_unlock(&q->sysfs_lock);
 			kobject_uevent(&q->kobj, KOBJ_REMOVE);
 			kobject_del(&q->kobj);
 			blk_trace_remove_sysfs(dev);
 			kobject_put(&dev->kobj);
-			goto unlock;
+			return ret;
 		}
 	}
 	ret = 0;
@@ -921,7 +926,15 @@ unlock:
 	mutex_unlock(&q->sysfs_lock);
 	return ret;
 }
+EXPORT_SYMBOL_GPL(blk_register_queue);
 
+/**
+ * blk_unregister_queue - counterpart of blk_register_queue()
+ * @disk: Disk of which the request queue should be unregistered from sysfs.
+ *
+ * Note: the caller is responsible for guaranteeing that this function is called
+ * after blk_register_queue() has finished.
+ */
 void blk_unregister_queue(struct gendisk *disk)
 {
 	struct request_queue *q = disk->queue;
@@ -929,21 +942,39 @@ void blk_unregister_queue(struct gendisk *disk)
 	if (WARN_ON(!q))
 		return;
 
-	mutex_lock(&q->sysfs_lock);
-	queue_flag_clear_unlocked(QUEUE_FLAG_REGISTERED, q);
-	mutex_unlock(&q->sysfs_lock);
+	/* Return early if disk->queue was never registered. */
+	if (!test_bit(QUEUE_FLAG_REGISTERED, &q->queue_flags))
+		return;
 
-	wbt_exit(q);
+	/*
+	 * Since sysfs_remove_dir() prevents adding new directory entries
+	 * before removal of existing entries starts, protect against
+	 * concurrent elv_iosched_store() calls.
+	 */
+	mutex_lock(&q->sysfs_lock);
 
+	spin_lock_irq(q->queue_lock);
+	queue_flag_clear(QUEUE_FLAG_REGISTERED, q);
+	spin_unlock_irq(q->queue_lock);
 
+	/*
+	 * Remove the sysfs attributes before unregistering the queue data
+	 * structures that can be modified through sysfs.
+	 */
 	if (q->mq_ops)
 		blk_mq_unregister_dev(disk_to_dev(disk), q);
-
-	if (q->request_fn || (q->mq_ops && q->elevator))
-		elv_unregister_queue(q);
+	mutex_unlock(&q->sysfs_lock);
 
 	kobject_uevent(&q->kobj, KOBJ_REMOVE);
 	kobject_del(&q->kobj);
 	blk_trace_remove_sysfs(disk_to_dev(disk));
+
+	wbt_exit(q);
+
+	mutex_lock(&q->sysfs_lock);
+	if (q->request_fn || (q->mq_ops && q->elevator))
+		elv_unregister_queue(q);
+	mutex_unlock(&q->sysfs_lock);
+
 	kobject_put(&disk_to_dev(disk)->kobj);
 }

block/blk-throttle.c

@@ -216,9 +216,9 @@ struct throtl_data
 
 	unsigned int scale;
 
-	struct latency_bucket tmp_buckets[LATENCY_BUCKET_SIZE];
-	struct avg_latency_bucket avg_buckets[LATENCY_BUCKET_SIZE];
-	struct latency_bucket __percpu *latency_buckets;
+	struct latency_bucket tmp_buckets[2][LATENCY_BUCKET_SIZE];
+	struct avg_latency_bucket avg_buckets[2][LATENCY_BUCKET_SIZE];
+	struct latency_bucket __percpu *latency_buckets[2];
 	unsigned long last_calculate_time;
 	unsigned long filtered_latency;
 
@@ -1510,11 +1510,21 @@ static struct cftype throtl_legacy_files[] = {
 		.private = (unsigned long)&blkcg_policy_throtl,
 		.seq_show = blkg_print_stat_bytes,
 	},
+	{
+		.name = "throttle.io_service_bytes_recursive",
+		.private = (unsigned long)&blkcg_policy_throtl,
+		.seq_show = blkg_print_stat_bytes_recursive,
+	},
 	{
 		.name = "throttle.io_serviced",
 		.private = (unsigned long)&blkcg_policy_throtl,
 		.seq_show = blkg_print_stat_ios,
 	},
+	{
+		.name = "throttle.io_serviced_recursive",
+		.private = (unsigned long)&blkcg_policy_throtl,
+		.seq_show = blkg_print_stat_ios_recursive,
+	},
 	{ }	/* terminate */
 };
 
@@ -2040,10 +2050,10 @@ static void blk_throtl_update_idletime(struct throtl_grp *tg)
 #ifdef CONFIG_BLK_DEV_THROTTLING_LOW
 static void throtl_update_latency_buckets(struct throtl_data *td)
 {
-	struct avg_latency_bucket avg_latency[LATENCY_BUCKET_SIZE];
-	int i, cpu;
-	unsigned long last_latency = 0;
-	unsigned long latency;
+	struct avg_latency_bucket avg_latency[2][LATENCY_BUCKET_SIZE];
+	int i, cpu, rw;
+	unsigned long last_latency[2] = { 0 };
+	unsigned long latency[2];
 
 	if (!blk_queue_nonrot(td->queue))
 		return;
@@ -2052,56 +2062,67 @@ static void throtl_update_latency_buckets(struct throtl_data *td)
 	td->last_calculate_time = jiffies;
 
 	memset(avg_latency, 0, sizeof(avg_latency));
-	for (i = 0; i < LATENCY_BUCKET_SIZE; i++) {
-		struct latency_bucket *tmp = &td->tmp_buckets[i];
-
-		for_each_possible_cpu(cpu) {
-			struct latency_bucket *bucket;
-
-			/* this isn't race free, but ok in practice */
-			bucket = per_cpu_ptr(td->latency_buckets, cpu);
-			tmp->total_latency += bucket[i].total_latency;
-			tmp->samples += bucket[i].samples;
-			bucket[i].total_latency = 0;
-			bucket[i].samples = 0;
-		}
+	for (rw = READ; rw <= WRITE; rw++) {
+		for (i = 0; i < LATENCY_BUCKET_SIZE; i++) {
+			struct latency_bucket *tmp = &td->tmp_buckets[rw][i];
+
+			for_each_possible_cpu(cpu) {
+				struct latency_bucket *bucket;
+
+				/* this isn't race free, but ok in practice */
+				bucket = per_cpu_ptr(td->latency_buckets[rw],
+					cpu);
+				tmp->total_latency += bucket[i].total_latency;
+				tmp->samples += bucket[i].samples;
+				bucket[i].total_latency = 0;
+				bucket[i].samples = 0;
+			}
 
-		if (tmp->samples >= 32) {
-			int samples = tmp->samples;
+			if (tmp->samples >= 32) {
+				int samples = tmp->samples;
 
-			latency = tmp->total_latency;
+				latency[rw] = tmp->total_latency;
 
-			tmp->total_latency = 0;
-			tmp->samples = 0;
-			latency /= samples;
-			if (latency == 0)
-				continue;
-			avg_latency[i].latency = latency;
+				tmp->total_latency = 0;
+				tmp->samples = 0;
+				latency[rw] /= samples;
+				if (latency[rw] == 0)
+					continue;
+				avg_latency[rw][i].latency = latency[rw];
+			}
 		}
 	}
 
-	for (i = 0; i < LATENCY_BUCKET_SIZE; i++) {
-		if (!avg_latency[i].latency) {
-			if (td->avg_buckets[i].latency < last_latency)
-				td->avg_buckets[i].latency = last_latency;
-			continue;
-		}
+	for (rw = READ; rw <= WRITE; rw++) {
+		for (i = 0; i < LATENCY_BUCKET_SIZE; i++) {
+			if (!avg_latency[rw][i].latency) {
+				if (td->avg_buckets[rw][i].latency < last_latency[rw])
+					td->avg_buckets[rw][i].latency =
+						last_latency[rw];
+				continue;
+			}
 
-		if (!td->avg_buckets[i].valid)
-			latency = avg_latency[i].latency;
-		else
-			latency = (td->avg_buckets[i].latency * 7 +
-				avg_latency[i].latency) >> 3;
+			if (!td->avg_buckets[rw][i].valid)
+				latency[rw] = avg_latency[rw][i].latency;
+			else
+				latency[rw] = (td->avg_buckets[rw][i].latency * 7 +
+					avg_latency[rw][i].latency) >> 3;
 
-		td->avg_buckets[i].latency = max(latency, last_latency);
-		td->avg_buckets[i].valid = true;
-		last_latency = td->avg_buckets[i].latency;
+			td->avg_buckets[rw][i].latency = max(latency[rw],
+				last_latency[rw]);
+			td->avg_buckets[rw][i].valid = true;
+			last_latency[rw] = td->avg_buckets[rw][i].latency;
+		}
 	}
 
 	for (i = 0; i < LATENCY_BUCKET_SIZE; i++)
 		throtl_log(&td->service_queue,
-			"Latency bucket %d: latency=%ld, valid=%d", i,
-			td->avg_buckets[i].latency, td->avg_buckets[i].valid);
+			"Latency bucket %d: read latency=%ld, read valid=%d, "
+			"write latency=%ld, write valid=%d", i,
+			td->avg_buckets[READ][i].latency,
+			td->avg_buckets[READ][i].valid,
+			td->avg_buckets[WRITE][i].latency,
+			td->avg_buckets[WRITE][i].valid);
 }
 #else
 static inline void throtl_update_latency_buckets(struct throtl_data *td)
@@ -2242,16 +2263,17 @@ static void throtl_track_latency(struct throtl_data *td, sector_t size,
 	struct latency_bucket *latency;
 	int index;
 
-	if (!td || td->limit_index != LIMIT_LOW || op != REQ_OP_READ ||
+	if (!td || td->limit_index != LIMIT_LOW ||
+	    !(op == REQ_OP_READ || op == REQ_OP_WRITE) ||
 	    !blk_queue_nonrot(td->queue))
 		return;
 
 	index = request_bucket_index(size);
 
-	latency = get_cpu_ptr(td->latency_buckets);
+	latency = get_cpu_ptr(td->latency_buckets[op]);
 	latency[index].total_latency += time;
 	latency[index].samples++;
-	put_cpu_ptr(td->latency_buckets);
+	put_cpu_ptr(td->latency_buckets[op]);
 }
 
 void blk_throtl_stat_add(struct request *rq, u64 time_ns)
@@ -2270,6 +2292,7 @@ void blk_throtl_bio_endio(struct bio *bio)
 	unsigned long finish_time;
 	unsigned long start_time;
 	unsigned long lat;
+	int rw = bio_data_dir(bio);
 
 	tg = bio->bi_cg_private;
 	if (!tg)
@@ -2298,7 +2321,7 @@ void blk_throtl_bio_endio(struct bio *bio)
 
 		bucket = request_bucket_index(
 			blk_stat_size(&bio->bi_issue_stat));
-		threshold = tg->td->avg_buckets[bucket].latency +
+		threshold = tg->td->avg_buckets[rw][bucket].latency +
 			tg->latency_target;
 		if (lat > threshold)
 			tg->bad_bio_cnt++;
@@ -2391,9 +2414,16 @@ int blk_throtl_init(struct request_queue *q)
 	td = kzalloc_node(sizeof(*td), GFP_KERNEL, q->node);
 	if (!td)
 		return -ENOMEM;
-	td->latency_buckets = __alloc_percpu(sizeof(struct latency_bucket) *
+	td->latency_buckets[READ] = __alloc_percpu(sizeof(struct latency_bucket) *
 		LATENCY_BUCKET_SIZE, __alignof__(u64));
-	if (!td->latency_buckets) {
+	if (!td->latency_buckets[READ]) {
+		kfree(td);
+		return -ENOMEM;
+	}
+	td->latency_buckets[WRITE] = __alloc_percpu(sizeof(struct latency_bucket) *
+		LATENCY_BUCKET_SIZE, __alignof__(u64));
+	if (!td->latency_buckets[WRITE]) {
+		free_percpu(td->latency_buckets[READ]);
 		kfree(td);
 		return -ENOMEM;
 	}
@@ -2412,7 +2442,8 @@ int blk_throtl_init(struct request_queue *q)
 	/* activate policy */
 	ret = blkcg_activate_policy(q, &blkcg_policy_throtl);
 	if (ret) {
-		free_percpu(td->latency_buckets);
+		free_percpu(td->latency_buckets[READ]);
+		free_percpu(td->latency_buckets[WRITE]);
 		kfree(td);
 	}
 	return ret;
@@ -2423,7 +2454,8 @@ void blk_throtl_exit(struct request_queue *q)
 	BUG_ON(!q->td);
 	throtl_shutdown_wq(q);
 	blkcg_deactivate_policy(q, &blkcg_policy_throtl);
-	free_percpu(q->td->latency_buckets);
+	free_percpu(q->td->latency_buckets[READ]);
+	free_percpu(q->td->latency_buckets[WRITE]);
 	kfree(q->td);
 }
 
@@ -2441,15 +2473,17 @@ void blk_throtl_register_queue(struct request_queue *q)
 	} else {
 		td->throtl_slice = DFL_THROTL_SLICE_HD;
 		td->filtered_latency = LATENCY_FILTERED_HD;
-		for (i = 0; i < LATENCY_BUCKET_SIZE; i++)
-			td->avg_buckets[i].latency = DFL_HD_BASELINE_LATENCY;
+		for (i = 0; i < LATENCY_BUCKET_SIZE; i++) {
+			td->avg_buckets[READ][i].latency = DFL_HD_BASELINE_LATENCY;
+			td->avg_buckets[WRITE][i].latency = DFL_HD_BASELINE_LATENCY;
+		}
 	}
 #ifndef CONFIG_BLK_DEV_THROTTLING_LOW
 	/* if no low limit, use previous default */
 	td->throtl_slice = DFL_THROTL_SLICE_HD;
 #endif
 
-	td->track_bio_latency = !q->mq_ops && !q->request_fn;
+	td->track_bio_latency = !queue_is_rq_based(q);
 	if (!td->track_bio_latency)
 		blk_stat_enable_accounting(q);
 }

block/blk-timeout.c

@@ -112,7 +112,9 @@ static void blk_rq_timed_out(struct request *req)
 static void blk_rq_check_expired(struct request *rq, unsigned long *next_timeout,
 			  unsigned int *next_set)
 {
-	if (time_after_eq(jiffies, rq->deadline)) {
+	const unsigned long deadline = blk_rq_deadline(rq);
+
+	if (time_after_eq(jiffies, deadline)) {
 		list_del_init(&rq->timeout_list);
 
 		/*
@@ -120,8 +122,8 @@ static void blk_rq_check_expired(struct request *rq, unsigned long *next_timeout
 		 */
 		if (!blk_mark_rq_complete(rq))
 			blk_rq_timed_out(rq);
-	} else if (!*next_set || time_after(*next_timeout, rq->deadline)) {
-		*next_timeout = rq->deadline;
+	} else if (!*next_set || time_after(*next_timeout, deadline)) {
+		*next_timeout = deadline;
 		*next_set = 1;
 	}
 }
@@ -156,12 +158,17 @@ void blk_timeout_work(struct work_struct *work)
  */
 void blk_abort_request(struct request *req)
 {
-	if (blk_mark_rq_complete(req))
-		return;
-
 	if (req->q->mq_ops) {
-		blk_mq_rq_timed_out(req, false);
+		/*
+		 * All we need to ensure is that timeout scan takes place
+		 * immediately and that scan sees the new timeout value.
+		 * No need for fancy synchronizations.
+		 */
+		blk_rq_set_deadline(req, jiffies);
+		mod_timer(&req->q->timeout, 0);
 	} else {
+		if (blk_mark_rq_complete(req))
+			return;
 		blk_delete_timer(req);
 		blk_rq_timed_out(req);
 	}
@@ -208,7 +215,8 @@ void blk_add_timer(struct request *req)
 	if (!req->timeout)
 		req->timeout = q->rq_timeout;
 
-	WRITE_ONCE(req->deadline, jiffies + req->timeout);
+	blk_rq_set_deadline(req, jiffies + req->timeout);
+	req->rq_flags &= ~RQF_MQ_TIMEOUT_EXPIRED;
 
 	/*
 	 * Only the non-mq case needs to add the request to a protected list.
@@ -222,7 +230,7 @@ void blk_add_timer(struct request *req)
 	 * than an existing one, modify the timer. Round up to next nearest
 	 * second.
 	 */
-	expiry = blk_rq_timeout(round_jiffies_up(req->deadline));
+	expiry = blk_rq_timeout(round_jiffies_up(blk_rq_deadline(req)));
 
 	if (!timer_pending(&q->timeout) ||
 	    time_before(expiry, q->timeout.expires)) {

block/blk-zoned.c

@@ -21,6 +21,48 @@ static inline sector_t blk_zone_start(struct request_queue *q,
 	return sector & ~zone_mask;
 }
 
+/*
+ * Return true if a request is a write requests that needs zone write locking.
+ */
+bool blk_req_needs_zone_write_lock(struct request *rq)
+{
+	if (!rq->q->seq_zones_wlock)
+		return false;
+
+	if (blk_rq_is_passthrough(rq))
+		return false;
+
+	switch (req_op(rq)) {
+	case REQ_OP_WRITE_ZEROES:
+	case REQ_OP_WRITE_SAME:
+	case REQ_OP_WRITE:
+		return blk_rq_zone_is_seq(rq);
+	default:
+		return false;
+	}
+}
+EXPORT_SYMBOL_GPL(blk_req_needs_zone_write_lock);
+
+void __blk_req_zone_write_lock(struct request *rq)
+{
+	if (WARN_ON_ONCE(test_and_set_bit(blk_rq_zone_no(rq),
+					  rq->q->seq_zones_wlock)))
+		return;
+
+	WARN_ON_ONCE(rq->rq_flags & RQF_ZONE_WRITE_LOCKED);
+	rq->rq_flags |= RQF_ZONE_WRITE_LOCKED;
+}
+EXPORT_SYMBOL_GPL(__blk_req_zone_write_lock);
+
+void __blk_req_zone_write_unlock(struct request *rq)
+{
+	rq->rq_flags &= ~RQF_ZONE_WRITE_LOCKED;
+	if (rq->q->seq_zones_wlock)
+		WARN_ON_ONCE(!test_and_clear_bit(blk_rq_zone_no(rq),
+						 rq->q->seq_zones_wlock));
+}
+EXPORT_SYMBOL_GPL(__blk_req_zone_write_unlock);
+
 /*
  * Check that a zone report belongs to the partition.
  * If yes, fix its start sector and write pointer, copy it in the

block/blk.h

@@ -119,34 +119,24 @@ void blk_account_io_start(struct request *req, bool new_io);
 void blk_account_io_completion(struct request *req, unsigned int bytes);
 void blk_account_io_done(struct request *req);
 
-/*
- * Internal atomic flags for request handling
- */
-enum rq_atomic_flags {
-	/*
-	 * Keep these two bits first - not because we depend on the
-	 * value of them, but we do depend on them being in the same
-	 * byte of storage to ensure ordering on writes. Keeping them
-	 * first will achieve that nicely.
-	 */
-	REQ_ATOM_COMPLETE = 0,
-	REQ_ATOM_STARTED,
-
-	REQ_ATOM_POLL_SLEPT,
-};
-
 /*
  * EH timer and IO completion will both attempt to 'grab' the request, make
- * sure that only one of them succeeds
+ * sure that only one of them succeeds. Steal the bottom bit of the
+ * __deadline field for this.
  */
 static inline int blk_mark_rq_complete(struct request *rq)
 {
-	return test_and_set_bit(REQ_ATOM_COMPLETE, &rq->atomic_flags);
+	return test_and_set_bit(0, &rq->__deadline);
 }
 
 static inline void blk_clear_rq_complete(struct request *rq)
 {
-	clear_bit(REQ_ATOM_COMPLETE, &rq->atomic_flags);
+	clear_bit(0, &rq->__deadline);
+}
+
+static inline bool blk_rq_is_complete(struct request *rq)
+{
+	return test_bit(0, &rq->__deadline);
 }
 
 /*
@@ -172,6 +162,9 @@ static inline void elv_deactivate_rq(struct request_queue *q, struct request *rq
 		e->type->ops.sq.elevator_deactivate_req_fn(q, rq);
 }
 
+int elv_register_queue(struct request_queue *q);
+void elv_unregister_queue(struct request_queue *q);
+
 struct hd_struct *__disk_get_part(struct gendisk *disk, int partno);
 
 #ifdef CONFIG_FAIL_IO_TIMEOUT
@@ -245,6 +238,21 @@ static inline void req_set_nomerge(struct request_queue *q, struct request *req)
 		q->last_merge = NULL;
 }
 
+/*
+ * Steal a bit from this field for legacy IO path atomic IO marking. Note that
+ * setting the deadline clears the bottom bit, potentially clearing the
+ * completed bit. The user has to be OK with this (current ones are fine).
+ */
+static inline void blk_rq_set_deadline(struct request *rq, unsigned long time)
+{
+	rq->__deadline = time & ~0x1UL;
+}
+
+static inline unsigned long blk_rq_deadline(struct request *rq)
+{
+	return rq->__deadline & ~0x1UL;
+}
+
 /*
  * Internal io_context interface
  */

block/bounce.c

@@ -113,45 +113,50 @@ int init_emergency_isa_pool(void)
 static void copy_to_high_bio_irq(struct bio *to, struct bio *from)
 {
 	unsigned char *vfrom;
-	struct bio_vec tovec, *fromvec = from->bi_io_vec;
+	struct bio_vec tovec, fromvec;
 	struct bvec_iter iter;
+	/*
+	 * The bio of @from is created by bounce, so we can iterate
+	 * its bvec from start to end, but the @from->bi_iter can't be
+	 * trusted because it might be changed by splitting.
+	 */
+	struct bvec_iter from_iter = BVEC_ITER_ALL_INIT;
 
 	bio_for_each_segment(tovec, to, iter) {
-		if (tovec.bv_page != fromvec->bv_page) {
+		fromvec = bio_iter_iovec(from, from_iter);
+		if (tovec.bv_page != fromvec.bv_page) {
 			/*
 			 * fromvec->bv_offset and fromvec->bv_len might have
 			 * been modified by the block layer, so use the original
 			 * copy, bounce_copy_vec already uses tovec->bv_len
 			 */
-			vfrom = page_address(fromvec->bv_page) +
+			vfrom = page_address(fromvec.bv_page) +
 				tovec.bv_offset;
 
 			bounce_copy_vec(&tovec, vfrom);
 			flush_dcache_page(tovec.bv_page);
 		}
-
-		fromvec++;
+		bio_advance_iter(from, &from_iter, tovec.bv_len);
 	}
 }
 
 static void bounce_end_io(struct bio *bio, mempool_t *pool)
 {
 	struct bio *bio_orig = bio->bi_private;
-	struct bio_vec *bvec, *org_vec;
+	struct bio_vec *bvec, orig_vec;
 	int i;
-	int start = bio_orig->bi_iter.bi_idx;
+	struct bvec_iter orig_iter = bio_orig->bi_iter;
 
 	/*
 	 * free up bounce indirect pages used
 	 */
 	bio_for_each_segment_all(bvec, bio, i) {
-		org_vec = bio_orig->bi_io_vec + i + start;
-
-		if (bvec->bv_page == org_vec->bv_page)
-			continue;
-
-		dec_zone_page_state(bvec->bv_page, NR_BOUNCE);
-		mempool_free(bvec->bv_page, pool);
+		orig_vec = bio_iter_iovec(bio_orig, orig_iter);
+		if (bvec->bv_page != orig_vec.bv_page) {
+			dec_zone_page_state(bvec->bv_page, NR_BOUNCE);
+			mempool_free(bvec->bv_page, pool);
+		}
+		bio_advance_iter(bio_orig, &orig_iter, orig_vec.bv_len);
 	}
 
 	bio_orig->bi_status = bio->bi_status;

block/bsg-lib.c

@@ -30,7 +30,7 @@
 
 /**
  * bsg_teardown_job - routine to teardown a bsg job
- * @job: bsg_job that is to be torn down
+ * @kref: kref inside bsg_job that is to be torn down
  */
 static void bsg_teardown_job(struct kref *kref)
 {
@@ -251,6 +251,7 @@ static void bsg_exit_rq(struct request_queue *q, struct request *req)
  * @name: device to give bsg device
  * @job_fn: bsg job handler
  * @dd_job_size: size of LLD data needed for each job
+ * @release: @dev release function
  */
 struct request_queue *bsg_setup_queue(struct device *dev, const char *name,
 		bsg_job_fn *job_fn, int dd_job_size,

block/bsg.c

@@ -32,6 +32,9 @@
 #define BSG_DESCRIPTION	"Block layer SCSI generic (bsg) driver"
 #define BSG_VERSION	"0.4"
 
+#define bsg_dbg(bd, fmt, ...) \
+	pr_debug("%s: " fmt, (bd)->name, ##__VA_ARGS__)
+
 struct bsg_device {
 	struct request_queue *queue;
 	spinlock_t lock;
@@ -55,14 +58,6 @@ enum {
 #define BSG_DEFAULT_CMDS	64
 #define BSG_MAX_DEVS		32768
 
-#undef BSG_DEBUG
-
-#ifdef BSG_DEBUG
-#define dprintk(fmt, args...) printk(KERN_ERR "%s: " fmt, __func__, ##args)
-#else
-#define dprintk(fmt, args...)
-#endif
-
 static DEFINE_MUTEX(bsg_mutex);
 static DEFINE_IDR(bsg_minor_idr);
 
@@ -123,7 +118,7 @@ static struct bsg_command *bsg_alloc_command(struct bsg_device *bd)
 
 	bc->bd = bd;
 	INIT_LIST_HEAD(&bc->list);
-	dprintk("%s: returning free cmd %p\n", bd->name, bc);
+	bsg_dbg(bd, "returning free cmd %p\n", bc);
 	return bc;
 out:
 	spin_unlock_irq(&bd->lock);
@@ -222,7 +217,8 @@ bsg_map_hdr(struct bsg_device *bd, struct sg_io_v4 *hdr, fmode_t mode)
 	if (!bcd->class_dev)
 		return ERR_PTR(-ENXIO);
 
-	dprintk("map hdr %llx/%u %llx/%u\n", (unsigned long long) hdr->dout_xferp,
+	bsg_dbg(bd, "map hdr %llx/%u %llx/%u\n",
+		(unsigned long long) hdr->dout_xferp,
 		hdr->dout_xfer_len, (unsigned long long) hdr->din_xferp,
 		hdr->din_xfer_len);
 
@@ -299,8 +295,8 @@ static void bsg_rq_end_io(struct request *rq, blk_status_t status)
 	struct bsg_device *bd = bc->bd;
 	unsigned long flags;
 
-	dprintk("%s: finished rq %p bc %p, bio %p\n",
-		bd->name, rq, bc, bc->bio);
+	bsg_dbg(bd, "finished rq %p bc %p, bio %p\n",
+		rq, bc, bc->bio);
 
 	bc->hdr.duration = jiffies_to_msecs(jiffies - bc->hdr.duration);
 
@@ -333,7 +329,7 @@ static void bsg_add_command(struct bsg_device *bd, struct request_queue *q,
 	list_add_tail(&bc->list, &bd->busy_list);
 	spin_unlock_irq(&bd->lock);
 
-	dprintk("%s: queueing rq %p, bc %p\n", bd->name, rq, bc);
+	bsg_dbg(bd, "queueing rq %p, bc %p\n", rq, bc);
 
 	rq->end_io_data = bc;
 	blk_execute_rq_nowait(q, NULL, rq, at_head, bsg_rq_end_io);
@@ -379,7 +375,7 @@ static struct bsg_command *bsg_get_done_cmd(struct bsg_device *bd)
 		}
 	} while (1);
 
-	dprintk("%s: returning done %p\n", bd->name, bc);
+	bsg_dbg(bd, "returning done %p\n", bc);
 
 	return bc;
 }
@@ -390,7 +386,7 @@ static int blk_complete_sgv4_hdr_rq(struct request *rq, struct sg_io_v4 *hdr,
 	struct scsi_request *req = scsi_req(rq);
 	int ret = 0;
 
-	dprintk("rq %p bio %p 0x%x\n", rq, bio, req->result);
+	pr_debug("rq %p bio %p 0x%x\n", rq, bio, req->result);
 	/*
 	 * fill in all the output members
 	 */
@@ -469,7 +465,7 @@ static int bsg_complete_all_commands(struct bsg_device *bd)
 	struct bsg_command *bc;
 	int ret, tret;
 
-	dprintk("%s: entered\n", bd->name);
+	bsg_dbg(bd, "entered\n");
 
 	/*
 	 * wait for all commands to complete
@@ -572,7 +568,7 @@ bsg_read(struct file *file, char __user *buf, size_t count, loff_t *ppos)
 	int ret;
 	ssize_t bytes_read;
 
-	dprintk("%s: read %zd bytes\n", bd->name, count);
+	bsg_dbg(bd, "read %zd bytes\n", count);
 
 	bsg_set_block(bd, file);
 
@@ -646,7 +642,7 @@ bsg_write(struct file *file, const char __user *buf, size_t count, loff_t *ppos)
 	ssize_t bytes_written;
 	int ret;
 
-	dprintk("%s: write %zd bytes\n", bd->name, count);
+	bsg_dbg(bd, "write %zd bytes\n", count);
 
 	if (unlikely(uaccess_kernel()))
 		return -EINVAL;
@@ -664,7 +660,7 @@ bsg_write(struct file *file, const char __user *buf, size_t count, loff_t *ppos)
 	if (!bytes_written || err_block_err(ret))
 		bytes_written = ret;
 
-	dprintk("%s: returning %zd\n", bd->name, bytes_written);
+	bsg_dbg(bd, "returning %zd\n", bytes_written);
 	return bytes_written;
 }
 
@@ -717,7 +713,7 @@ static int bsg_put_device(struct bsg_device *bd)
 	hlist_del(&bd->dev_list);
 	mutex_unlock(&bsg_mutex);
 
-	dprintk("%s: tearing down\n", bd->name);
+	bsg_dbg(bd, "tearing down\n");
 
 	/*
 	 * close can always block
@@ -744,9 +740,7 @@ static struct bsg_device *bsg_add_device(struct inode *inode,
 					 struct file *file)
 {
 	struct bsg_device *bd;
-#ifdef BSG_DEBUG
 	unsigned char buf[32];
-#endif
 
 	if (!blk_queue_scsi_passthrough(rq)) {
 		WARN_ONCE(true, "Attempt to register a non-SCSI queue\n");
@@ -771,7 +765,7 @@ static struct bsg_device *bsg_add_device(struct inode *inode,
 	hlist_add_head(&bd->dev_list, bsg_dev_idx_hash(iminor(inode)));
 
 	strncpy(bd->name, dev_name(rq->bsg_dev.class_dev), sizeof(bd->name) - 1);
-	dprintk("bound to <%s>, max queue %d\n",
+	bsg_dbg(bd, "bound to <%s>, max queue %d\n",
 		format_dev_t(buf, inode->i_rdev), bd->max_queue);
 
 	mutex_unlock(&bsg_mutex);

block/deadline-iosched.c

@@ -50,8 +50,6 @@ struct deadline_data {
 	int front_merges;
 };
 
-static void deadline_move_request(struct deadline_data *, struct request *);
-
 static inline struct rb_root *
 deadline_rb_root(struct deadline_data *dd, struct request *rq)
 {
@@ -100,6 +98,12 @@ deadline_add_request(struct request_queue *q, struct request *rq)
 	struct deadline_data *dd = q->elevator->elevator_data;
 	const int data_dir = rq_data_dir(rq);
 
+	/*
+	 * This may be a requeue of a write request that has locked its
+	 * target zone. If it is the case, this releases the zone lock.
+	 */
+	blk_req_zone_write_unlock(rq);
+
 	deadline_add_rq_rb(dd, rq);
 
 	/*
@@ -190,6 +194,12 @@ deadline_move_to_dispatch(struct deadline_data *dd, struct request *rq)
 {
 	struct request_queue *q = rq->q;
 
+	/*
+	 * For a zoned block device, write requests must write lock their
+	 * target zone.
+	 */
+	blk_req_zone_write_lock(rq);
+
 	deadline_remove_request(q, rq);
 	elv_dispatch_add_tail(q, rq);
 }
@@ -230,6 +240,69 @@ static inline int deadline_check_fifo(struct deadline_data *dd, int ddir)
 	return 0;
 }
 
+/*
+ * For the specified data direction, return the next request to dispatch using
+ * arrival ordered lists.
+ */
+static struct request *
+deadline_fifo_request(struct deadline_data *dd, int data_dir)
+{
+	struct request *rq;
+
+	if (WARN_ON_ONCE(data_dir != READ && data_dir != WRITE))
+		return NULL;
+
+	if (list_empty(&dd->fifo_list[data_dir]))
+		return NULL;
+
+	rq = rq_entry_fifo(dd->fifo_list[data_dir].next);
+	if (data_dir == READ || !blk_queue_is_zoned(rq->q))
+		return rq;
+
+	/*
+	 * Look for a write request that can be dispatched, that is one with
+	 * an unlocked target zone.
+	 */
+	list_for_each_entry(rq, &dd->fifo_list[WRITE], queuelist) {
+		if (blk_req_can_dispatch_to_zone(rq))
+			return rq;
+	}
+
+	return NULL;
+}
+
+/*
+ * For the specified data direction, return the next request to dispatch using
+ * sector position sorted lists.
+ */
+static struct request *
+deadline_next_request(struct deadline_data *dd, int data_dir)
+{
+	struct request *rq;
+
+	if (WARN_ON_ONCE(data_dir != READ && data_dir != WRITE))
+		return NULL;
+
+	rq = dd->next_rq[data_dir];
+	if (!rq)
+		return NULL;
+
+	if (data_dir == READ || !blk_queue_is_zoned(rq->q))
+		return rq;
+
+	/*
+	 * Look for a write request that can be dispatched, that is one with
+	 * an unlocked target zone.
+	 */
+	while (rq) {
+		if (blk_req_can_dispatch_to_zone(rq))
+			return rq;
+		rq = deadline_latter_request(rq);
+	}
+
+	return NULL;
+}
+
 /*
  * deadline_dispatch_requests selects the best request according to
  * read/write expire, fifo_batch, etc
@@ -239,16 +312,15 @@ static int deadline_dispatch_requests(struct request_queue *q, int force)
 	struct deadline_data *dd = q->elevator->elevator_data;
 	const int reads = !list_empty(&dd->fifo_list[READ]);
 	const int writes = !list_empty(&dd->fifo_list[WRITE]);
-	struct request *rq;
+	struct request *rq, *next_rq;
 	int data_dir;
 
 	/*
 	 * batches are currently reads XOR writes
 	 */
-	if (dd->next_rq[WRITE])
-		rq = dd->next_rq[WRITE];
-	else
-		rq = dd->next_rq[READ];
+	rq = deadline_next_request(dd, WRITE);
+	if (!rq)
+		rq = deadline_next_request(dd, READ);
 
 	if (rq && dd->batching < dd->fifo_batch)
 		/* we have a next request are still entitled to batch */
@@ -262,7 +334,8 @@ static int deadline_dispatch_requests(struct request_queue *q, int force)
 	if (reads) {
 		BUG_ON(RB_EMPTY_ROOT(&dd->sort_list[READ]));
 
-		if (writes && (dd->starved++ >= dd->writes_starved))
+		if (deadline_fifo_request(dd, WRITE) &&
+		    (dd->starved++ >= dd->writes_starved))
 			goto dispatch_writes;
 
 		data_dir = READ;
@@ -291,21 +364,29 @@ dispatch_find_request:
 	/*
 	 * we are not running a batch, find best request for selected data_dir
 	 */
-	if (deadline_check_fifo(dd, data_dir) || !dd->next_rq[data_dir]) {
+	next_rq = deadline_next_request(dd, data_dir);
+	if (deadline_check_fifo(dd, data_dir) || !next_rq) {
 		/*
 		 * A deadline has expired, the last request was in the other
 		 * direction, or we have run out of higher-sectored requests.
 		 * Start again from the request with the earliest expiry time.
 		 */
-		rq = rq_entry_fifo(dd->fifo_list[data_dir].next);
+		rq = deadline_fifo_request(dd, data_dir);
 	} else {
 		/*
 		 * The last req was the same dir and we have a next request in
 		 * sort order. No expired requests so continue on from here.
 		 */
-		rq = dd->next_rq[data_dir];
+		rq = next_rq;
 	}
 
+	/*
+	 * For a zoned block device, if we only have writes queued and none of
+	 * them can be dispatched, rq will be NULL.
+	 */
+	if (!rq)
+		return 0;
+
 	dd->batching = 0;
 
 dispatch_request:
@@ -318,6 +399,16 @@ dispatch_request:
 	return 1;
 }
 
+/*
+ * For zoned block devices, write unlock the target zone of completed
+ * write requests.
+ */
+static void
+deadline_completed_request(struct request_queue *q, struct request *rq)
+{
+	blk_req_zone_write_unlock(rq);
+}
+
 static void deadline_exit_queue(struct elevator_queue *e)
 {
 	struct deadline_data *dd = e->elevator_data;
@@ -439,6 +530,7 @@ static struct elevator_type iosched_deadline = {
 		.elevator_merged_fn =		deadline_merged_request,
 		.elevator_merge_req_fn =	deadline_merged_requests,
 		.elevator_dispatch_fn =		deadline_dispatch_requests,
+		.elevator_completed_req_fn =	deadline_completed_request,
 		.elevator_add_req_fn =		deadline_add_request,
 		.elevator_former_req_fn =	elv_rb_former_request,
 		.elevator_latter_req_fn =	elv_rb_latter_request,

block/elevator.c

@@ -869,6 +869,8 @@ int elv_register_queue(struct request_queue *q)
 	struct elevator_queue *e = q->elevator;
 	int error;
 
+	lockdep_assert_held(&q->sysfs_lock);
+
 	error = kobject_add(&e->kobj, &q->kobj, "%s", "iosched");
 	if (!error) {
 		struct elv_fs_entry *attr = e->type->elevator_attrs;
@@ -886,10 +888,11 @@ int elv_register_queue(struct request_queue *q)
 	}
 	return error;
 }
-EXPORT_SYMBOL(elv_register_queue);
 
 void elv_unregister_queue(struct request_queue *q)
 {
+	lockdep_assert_held(&q->sysfs_lock);
+
 	if (q) {
 		struct elevator_queue *e = q->elevator;
 
@@ -900,7 +903,6 @@ void elv_unregister_queue(struct request_queue *q)
 		wbt_enable_default(q);
 	}
 }
-EXPORT_SYMBOL(elv_unregister_queue);
 
 int elv_register(struct elevator_type *e)
 {
@@ -967,7 +969,10 @@ static int elevator_switch_mq(struct request_queue *q,
 {
 	int ret;
 
+	lockdep_assert_held(&q->sysfs_lock);
+
 	blk_mq_freeze_queue(q);
+	blk_mq_quiesce_queue(q);
 
 	if (q->elevator) {
 		if (q->elevator->registered)
@@ -994,6 +999,7 @@ static int elevator_switch_mq(struct request_queue *q,
 		blk_add_trace_msg(q, "elv switch: none");
 
 out:
+	blk_mq_unquiesce_queue(q);
 	blk_mq_unfreeze_queue(q);
 	return ret;
 }
@@ -1010,6 +1016,8 @@ static int elevator_switch(struct request_queue *q, struct elevator_type *new_e)
 	bool old_registered = false;
 	int err;
 
+	lockdep_assert_held(&q->sysfs_lock);
+
 	if (q->mq_ops)
 		return elevator_switch_mq(q, new_e);
 

block/genhd.c

@@ -629,16 +629,18 @@ exit:
 }
 
 /**
- * device_add_disk - add partitioning information to kernel list
+ * __device_add_disk - add disk information to kernel list
  * @parent: parent device for the disk
  * @disk: per-device partitioning information
+ * @register_queue: register the queue if set to true
  *
  * This function registers the partitioning information in @disk
  * with the kernel.
  *
  * FIXME: error handling
  */
-void device_add_disk(struct device *parent, struct gendisk *disk)
+static void __device_add_disk(struct device *parent, struct gendisk *disk,
+			      bool register_queue)
 {
 	dev_t devt;
 	int retval;
@@ -682,7 +684,8 @@ void device_add_disk(struct device *parent, struct gendisk *disk)
 				    exact_match, exact_lock, disk);
 	}
 	register_disk(parent, disk);
-	blk_register_queue(disk);
+	if (register_queue)
+		blk_register_queue(disk);
 
 	/*
 	 * Take an extra ref on queue which will be put on disk_release()
@@ -693,8 +696,19 @@ void device_add_disk(struct device *parent, struct gendisk *disk)
 	disk_add_events(disk);
 	blk_integrity_add(disk);
 }
+
+void device_add_disk(struct device *parent, struct gendisk *disk)
+{
+	__device_add_disk(parent, disk, true);
+}
 EXPORT_SYMBOL(device_add_disk);
 
+void device_add_disk_no_queue_reg(struct device *parent, struct gendisk *disk)
+{
+	__device_add_disk(parent, disk, false);
+}
+EXPORT_SYMBOL(device_add_disk_no_queue_reg);
+
 void del_gendisk(struct gendisk *disk)
 {
 	struct disk_part_iter piter;
@@ -725,7 +739,8 @@ void del_gendisk(struct gendisk *disk)
 		 * Unregister bdi before releasing device numbers (as they can
 		 * get reused and we'd get clashes in sysfs).
 		 */
-		bdi_unregister(disk->queue->backing_dev_info);
+		if (!(disk->flags & GENHD_FL_HIDDEN))
+			bdi_unregister(disk->queue->backing_dev_info);
 		blk_unregister_queue(disk);
 	} else {
 		WARN_ON(1);

block/mq-deadline.c

@@ -59,6 +59,7 @@ struct deadline_data {
 	int front_merges;
 
 	spinlock_t lock;
+	spinlock_t zone_lock;
 	struct list_head dispatch;
 };
 
@@ -191,14 +192,84 @@ static inline int deadline_check_fifo(struct deadline_data *dd, int ddir)
 	return 0;
 }
 
+/*
+ * For the specified data direction, return the next request to
+ * dispatch using arrival ordered lists.
+ */
+static struct request *
+deadline_fifo_request(struct deadline_data *dd, int data_dir)
+{
+	struct request *rq;
+	unsigned long flags;
+
+	if (WARN_ON_ONCE(data_dir != READ && data_dir != WRITE))
+		return NULL;
+
+	if (list_empty(&dd->fifo_list[data_dir]))
+		return NULL;
+
+	rq = rq_entry_fifo(dd->fifo_list[data_dir].next);
+	if (data_dir == READ || !blk_queue_is_zoned(rq->q))
+		return rq;
+
+	/*
+	 * Look for a write request that can be dispatched, that is one with
+	 * an unlocked target zone.
+	 */
+	spin_lock_irqsave(&dd->zone_lock, flags);
+	list_for_each_entry(rq, &dd->fifo_list[WRITE], queuelist) {
+		if (blk_req_can_dispatch_to_zone(rq))
+			goto out;
+	}
+	rq = NULL;
+out:
+	spin_unlock_irqrestore(&dd->zone_lock, flags);
+
+	return rq;
+}
+
+/*
+ * For the specified data direction, return the next request to
+ * dispatch using sector position sorted lists.
+ */
+static struct request *
+deadline_next_request(struct deadline_data *dd, int data_dir)
+{
+	struct request *rq;
+	unsigned long flags;
+
+	if (WARN_ON_ONCE(data_dir != READ && data_dir != WRITE))
+		return NULL;
+
+	rq = dd->next_rq[data_dir];
+	if (!rq)
+		return NULL;
+
+	if (data_dir == READ || !blk_queue_is_zoned(rq->q))
+		return rq;
+
+	/*
+	 * Look for a write request that can be dispatched, that is one with
+	 * an unlocked target zone.
+	 */
+	spin_lock_irqsave(&dd->zone_lock, flags);
+	while (rq) {
+		if (blk_req_can_dispatch_to_zone(rq))
+			break;
+		rq = deadline_latter_request(rq);
+	}
+	spin_unlock_irqrestore(&dd->zone_lock, flags);
+
+	return rq;
+}
+
 /*
  * deadline_dispatch_requests selects the best request according to
  * read/write expire, fifo_batch, etc
  */
-static struct request *__dd_dispatch_request(struct blk_mq_hw_ctx *hctx)
+static struct request *__dd_dispatch_request(struct deadline_data *dd)
 {
-	struct deadline_data *dd = hctx->queue->elevator->elevator_data;
-	struct request *rq;
+	struct request *rq, *next_rq;
 	bool reads, writes;
 	int data_dir;
 
@@ -214,10 +285,9 @@ static struct request *__dd_dispatch_request(struct blk_mq_hw_ctx *hctx)
 	/*
 	 * batches are currently reads XOR writes
 	 */
-	if (dd->next_rq[WRITE])
-		rq = dd->next_rq[WRITE];
-	else
-		rq = dd->next_rq[READ];
+	rq = deadline_next_request(dd, WRITE);
+	if (!rq)
+		rq = deadline_next_request(dd, READ);
 
 	if (rq && dd->batching < dd->fifo_batch)
 		/* we have a next request are still entitled to batch */
@@ -231,7 +301,8 @@ static struct request *__dd_dispatch_request(struct blk_mq_hw_ctx *hctx)
 	if (reads) {
 		BUG_ON(RB_EMPTY_ROOT(&dd->sort_list[READ]));
 
-		if (writes && (dd->starved++ >= dd->writes_starved))
+		if (deadline_fifo_request(dd, WRITE) &&
+		    (dd->starved++ >= dd->writes_starved))
 			goto dispatch_writes;
 
 		data_dir = READ;
@@ -260,21 +331,29 @@ dispatch_find_request:
 	/*
 	 * we are not running a batch, find best request for selected data_dir
 	 */
-	if (deadline_check_fifo(dd, data_dir) || !dd->next_rq[data_dir]) {
+	next_rq = deadline_next_request(dd, data_dir);
+	if (deadline_check_fifo(dd, data_dir) || !next_rq) {
 		/*
 		 * A deadline has expired, the last request was in the other
 		 * direction, or we have run out of higher-sectored requests.
 		 * Start again from the request with the earliest expiry time.
 		 */
-		rq = rq_entry_fifo(dd->fifo_list[data_dir].next);
+		rq = deadline_fifo_request(dd, data_dir);
 	} else {
 		/*
 		 * The last req was the same dir and we have a next request in
 		 * sort order. No expired requests so continue on from here.
 		 */
-		rq = dd->next_rq[data_dir];
+		rq = next_rq;
 	}
 
+	/*
+	 * For a zoned block device, if we only have writes queued and none of
+	 * them can be dispatched, rq will be NULL.
+	 */
+	if (!rq)
+		return NULL;
+
 	dd->batching = 0;
 
 dispatch_request:
@@ -284,17 +363,27 @@ dispatch_request:
 	dd->batching++;
 	deadline_move_request(dd, rq);
 done:
+	/*
+	 * If the request needs its target zone locked, do it.
+	 */
+	blk_req_zone_write_lock(rq);
 	rq->rq_flags |= RQF_STARTED;
 	return rq;
 }
 
+/*
+ * One confusing aspect here is that we get called for a specific
+ * hardware queue, but we return a request that may not be for a
+ * different hardware queue. This is because mq-deadline has shared
+ * state for all hardware queues, in terms of sorting, FIFOs, etc.
+ */
 static struct request *dd_dispatch_request(struct blk_mq_hw_ctx *hctx)
 {
 	struct deadline_data *dd = hctx->queue->elevator->elevator_data;
 	struct request *rq;
 
 	spin_lock(&dd->lock);
-	rq = __dd_dispatch_request(hctx);
+	rq = __dd_dispatch_request(dd);
 	spin_unlock(&dd->lock);
 
 	return rq;
@@ -339,6 +428,7 @@ static int dd_init_queue(struct request_queue *q, struct elevator_type *e)
 	dd->front_merges = 1;
 	dd->fifo_batch = fifo_batch;
 	spin_lock_init(&dd->lock);
+	spin_lock_init(&dd->zone_lock);
 	INIT_LIST_HEAD(&dd->dispatch);
 
 	q->elevator = eq;
@@ -395,6 +485,12 @@ static void dd_insert_request(struct blk_mq_hw_ctx *hctx, struct request *rq,
 	struct deadline_data *dd = q->elevator->elevator_data;
 	const int data_dir = rq_data_dir(rq);
 
+	/*
+	 * This may be a requeue of a write request that has locked its
+	 * target zone. If it is the case, this releases the zone lock.
+	 */
+	blk_req_zone_write_unlock(rq);
+
 	if (blk_mq_sched_try_insert_merge(q, rq))
 		return;
 
@@ -439,6 +535,26 @@ static void dd_insert_requests(struct blk_mq_hw_ctx *hctx,
 	spin_unlock(&dd->lock);
 }
 
+/*
+ * For zoned block devices, write unlock the target zone of
+ * completed write requests. Do this while holding the zone lock
+ * spinlock so that the zone is never unlocked while deadline_fifo_request()
+ * while deadline_next_request() are executing.
+ */
+static void dd_completed_request(struct request *rq)
+{
+	struct request_queue *q = rq->q;
+
+	if (blk_queue_is_zoned(q)) {
+		struct deadline_data *dd = q->elevator->elevator_data;
+		unsigned long flags;
+
+		spin_lock_irqsave(&dd->zone_lock, flags);
+		blk_req_zone_write_unlock(rq);
+		spin_unlock_irqrestore(&dd->zone_lock, flags);
+	}
+}
+
 static bool dd_has_work(struct blk_mq_hw_ctx *hctx)
 {
 	struct deadline_data *dd = hctx->queue->elevator->elevator_data;
@@ -640,6 +756,7 @@ static struct elevator_type mq_deadline = {
 	.ops.mq = {
 		.insert_requests	= dd_insert_requests,
 		.dispatch_request	= dd_dispatch_request,
+		.completed_request	= dd_completed_request,
 		.next_request		= elv_rb_latter_request,
 		.former_request		= elv_rb_former_request,
 		.bio_merge		= dd_bio_merge,

block/partitions/msdos.c

@@ -301,7 +301,9 @@ static void parse_bsd(struct parsed_partitions *state,
 			continue;
 		bsd_start = le32_to_cpu(p->p_offset);
 		bsd_size = le32_to_cpu(p->p_size);
-		if (memcmp(flavour, "bsd\0", 4) == 0)
+		/* FreeBSD has relative offset if C partition offset is zero */
+		if (memcmp(flavour, "bsd\0", 4) == 0 &&
+		    le32_to_cpu(l->d_partitions[2].p_offset) == 0)
 			bsd_start += offset;
 		if (offset == bsd_start && size == bsd_size)
 			/* full parent partition, we have it already */

block/scsi_ioctl.c

@@ -384,9 +384,10 @@ out_put_request:
 
 /**
  * sg_scsi_ioctl  --  handle deprecated SCSI_IOCTL_SEND_COMMAND ioctl
- * @file:	file this ioctl operates on (optional)
  * @q:		request queue to send scsi commands down
  * @disk:	gendisk to operate on (option)
+ * @mode:	mode used to open the file through which the ioctl has been
+ *		submitted
  * @sic:	userspace structure describing the command to perform
  *
  * Send down the scsi command described by @sic to the device below
@@ -415,10 +416,10 @@ out_put_request:
  *      Positive numbers returned are the compacted SCSI error codes (4
  *      bytes in one int) where the lowest byte is the SCSI status.
  */
-#define OMAX_SB_LEN 16          /* For backward compatibility */
 int sg_scsi_ioctl(struct request_queue *q, struct gendisk *disk, fmode_t mode,
 		struct scsi_ioctl_command __user *sic)
 {
+	enum { OMAX_SB_LEN = 16 };	/* For backward compatibility */
 	struct request *rq;
 	struct scsi_request *req;
 	int err;
@@ -692,38 +693,9 @@ int scsi_verify_blk_ioctl(struct block_device *bd, unsigned int cmd)
 	if (bd && bd == bd->bd_contains)
 		return 0;
 
-	/* Actually none of these is particularly useful on a partition,
-	 * but they are safe.
-	 */
-	switch (cmd) {
-	case SCSI_IOCTL_GET_IDLUN:
-	case SCSI_IOCTL_GET_BUS_NUMBER:
-	case SCSI_IOCTL_GET_PCI:
-	case SCSI_IOCTL_PROBE_HOST:
-	case SG_GET_VERSION_NUM:
-	case SG_SET_TIMEOUT:
-	case SG_GET_TIMEOUT:
-	case SG_GET_RESERVED_SIZE:
-	case SG_SET_RESERVED_SIZE:
-	case SG_EMULATED_HOST:
-		return 0;
-	case CDROM_GET_CAPABILITY:
-		/* Keep this until we remove the printk below.  udev sends it
-		 * and we do not want to spam dmesg about it.   CD-ROMs do
-		 * not have partitions, so we get here only for disks.
-		 */
-		return -ENOIOCTLCMD;
-	default:
-		break;
-	}
-
 	if (capable(CAP_SYS_RAWIO))
 		return 0;
 
-	/* In particular, rule out all resets and host-specific ioctls.  */
-	printk_ratelimited(KERN_WARNING
-			   "%s: sending ioctl %x to a partition!\n", current->comm, cmd);
-
 	return -ENOIOCTLCMD;
 }
 EXPORT_SYMBOL(scsi_verify_blk_ioctl);

crypto/Kconfig

@@ -106,6 +106,7 @@ config CRYPTO_KPP
 config CRYPTO_ACOMP2
 	tristate
 	select CRYPTO_ALGAPI2
+	select SGL_ALLOC
 
 config CRYPTO_ACOMP
 	tristate

crypto/scompress.c

@@ -140,53 +140,6 @@ static int crypto_scomp_init_tfm(struct crypto_tfm *tfm)
 	return ret;
 }
 
-static void crypto_scomp_sg_free(struct scatterlist *sgl)
-{
-	int i, n;
-	struct page *page;
-
-	if (!sgl)
-		return;
-
-	n = sg_nents(sgl);
-	for_each_sg(sgl, sgl, n, i) {
-		page = sg_page(sgl);
-		if (page)
-			__free_page(page);
-	}
-
-	kfree(sgl);
-}
-
-static struct scatterlist *crypto_scomp_sg_alloc(size_t size, gfp_t gfp)
-{
-	struct scatterlist *sgl;
-	struct page *page;
-	int i, n;
-
-	n = ((size - 1) >> PAGE_SHIFT) + 1;
-
-	sgl = kmalloc_array(n, sizeof(struct scatterlist), gfp);
-	if (!sgl)
-		return NULL;
-
-	sg_init_table(sgl, n);
-
-	for (i = 0; i < n; i++) {
-		page = alloc_page(gfp);
-		if (!page)
-			goto err;
-		sg_set_page(sgl + i, page, PAGE_SIZE, 0);
-	}
-
-	return sgl;
-
-err:
-	sg_mark_end(sgl + i);
-	crypto_scomp_sg_free(sgl);
-	return NULL;
-}
-
 static int scomp_acomp_comp_decomp(struct acomp_req *req, int dir)
 {
 	struct crypto_acomp *tfm = crypto_acomp_reqtfm(req);
@@ -220,7 +173,7 @@ static int scomp_acomp_comp_decomp(struct acomp_req *req, int dir)
 					      scratch_dst, &req->dlen, *ctx);
 	if (!ret) {
 		if (!req->dst) {
-			req->dst = crypto_scomp_sg_alloc(req->dlen, GFP_ATOMIC);
+			req->dst = sgl_alloc(req->dlen, GFP_ATOMIC, NULL);
 			if (!req->dst)
 				goto out;
 		}
@@ -274,7 +227,7 @@ int crypto_init_scomp_ops_async(struct crypto_tfm *tfm)
 
 	crt->compress = scomp_acomp_compress;
 	crt->decompress = scomp_acomp_decompress;
-	crt->dst_free = crypto_scomp_sg_free;
+	crt->dst_free = sgl_free;
 	crt->reqsize = sizeof(void *);
 
 	return 0;

drivers/block/DAC960.c

@@ -6619,43 +6619,27 @@ static void DAC960_DestroyProcEntries(DAC960_Controller_T *Controller)
 
 #ifdef DAC960_GAM_MINOR
 
-/*
- * DAC960_gam_ioctl is the ioctl function for performing RAID operations.
-*/
-
-static long DAC960_gam_ioctl(struct file *file, unsigned int Request,
-						unsigned long Argument)
+static long DAC960_gam_get_controller_info(DAC960_ControllerInfo_T __user *UserSpaceControllerInfo)
 {
-  long ErrorCode = 0;
-  if (!capable(CAP_SYS_ADMIN)) return -EACCES;
-
-  mutex_lock(&DAC960_mutex);
-  switch (Request)
-    {
-    case DAC960_IOCTL_GET_CONTROLLER_COUNT:
-      ErrorCode = DAC960_ControllerCount;
-      break;
-    case DAC960_IOCTL_GET_CONTROLLER_INFO:
-      {
-	DAC960_ControllerInfo_T __user *UserSpaceControllerInfo =
-	  (DAC960_ControllerInfo_T __user *) Argument;
 	DAC960_ControllerInfo_T ControllerInfo;
 	DAC960_Controller_T *Controller;
 	int ControllerNumber;
+	long ErrorCode;
+
 	if (UserSpaceControllerInfo == NULL)
 		ErrorCode = -EINVAL;
 	else ErrorCode = get_user(ControllerNumber,
 			     &UserSpaceControllerInfo->ControllerNumber);
 	if (ErrorCode != 0)
-		break;
+		goto out;
 	ErrorCode = -ENXIO;
 	if (ControllerNumber < 0 ||
 	    ControllerNumber > DAC960_ControllerCount - 1) {
-	  break;
+		goto out;
 	}
 	Controller = DAC960_Controllers[ControllerNumber];
 	if (Controller == NULL)
-		break;
+		goto out;
 	memset(&ControllerInfo, 0, sizeof(DAC960_ControllerInfo_T));
 	ControllerInfo.ControllerNumber = ControllerNumber;
 	ControllerInfo.FirmwareType = Controller->FirmwareType;
@@ -6670,12 +6654,12 @@ static long DAC960_gam_ioctl(struct file *file, unsigned int Request,
 	strcpy(ControllerInfo.FirmwareVersion, Controller->FirmwareVersion);
 	ErrorCode = (copy_to_user(UserSpaceControllerInfo, &ControllerInfo,
 			     sizeof(DAC960_ControllerInfo_T)) ? -EFAULT : 0);
-	break;
-      }
-    case DAC960_IOCTL_V1_EXECUTE_COMMAND:
-      {
-	DAC960_V1_UserCommand_T __user *UserSpaceUserCommand =
-	  (DAC960_V1_UserCommand_T __user *) Argument;
+out:
+	return ErrorCode;
+}
+
+static long DAC960_gam_v1_execute_command(DAC960_V1_UserCommand_T __user *UserSpaceUserCommand)
+{
 	DAC960_V1_UserCommand_T UserCommand;
 	DAC960_Controller_T *Controller;
 	DAC960_Command_T *Command = NULL;
@@ -6688,39 +6672,41 @@ static long DAC960_gam_ioctl(struct file *file, unsigned int Request,
 	int ControllerNumber, DataTransferLength;
 	unsigned char *DataTransferBuffer = NULL;
 	dma_addr_t DataTransferBufferDMA;
+        long ErrorCode;
+
 	if (UserSpaceUserCommand == NULL) {
 		ErrorCode = -EINVAL;
-		break;
+		goto out;
 	}
 	if (copy_from_user(&UserCommand, UserSpaceUserCommand,
 				   sizeof(DAC960_V1_UserCommand_T))) {
 		ErrorCode = -EFAULT;
-		break;
+		goto out;
 	}
 	ControllerNumber = UserCommand.ControllerNumber;
     	ErrorCode = -ENXIO;
 	if (ControllerNumber < 0 ||
 	    ControllerNumber > DAC960_ControllerCount - 1)
-	    	break;
+		goto out;
 	Controller = DAC960_Controllers[ControllerNumber];
 	if (Controller == NULL)
-		break;
+		goto out;
 	ErrorCode = -EINVAL;
 	if (Controller->FirmwareType != DAC960_V1_Controller)
-		break;
+		goto out;
 	CommandOpcode = UserCommand.CommandMailbox.Common.CommandOpcode;
 	DataTransferLength = UserCommand.DataTransferLength;
 	if (CommandOpcode & 0x80)
-		break;
+		goto out;
 	if (CommandOpcode == DAC960_V1_DCDB)
 	  {
 	    if (copy_from_user(&DCDB, UserCommand.DCDB,
 			       sizeof(DAC960_V1_DCDB_T))) {
 		ErrorCode = -EFAULT;
-		break;
+		goto out;
 	    }
 	    if (DCDB.Channel >= DAC960_V1_MaxChannels)
-	    		break;
+		goto out;
 	    if (!((DataTransferLength == 0 &&
 		   DCDB.Direction
 		   == DAC960_V1_DCDB_NoDataTransfer) ||
@@ -6730,15 +6716,15 @@ static long DAC960_gam_ioctl(struct file *file, unsigned int Request,
 		  (DataTransferLength < 0 &&
 		   DCDB.Direction
 		   == DAC960_V1_DCDB_DataTransferSystemToDevice)))
-		   	break;
+			goto out;
 	    if (((DCDB.TransferLengthHigh4 << 16) | DCDB.TransferLength)
 		!= abs(DataTransferLength))
-			break;
+			goto out;
 	    DCDB_IOBUF = pci_alloc_consistent(Controller->PCIDevice,
 			sizeof(DAC960_V1_DCDB_T), &DCDB_IOBUFDMA);
 	    if (DCDB_IOBUF == NULL) {
 	    		ErrorCode = -ENOMEM;
-			break;
+			goto out;
 		}
 	  }
 	ErrorCode = -ENOMEM;
@@ -6748,19 +6734,19 @@ static long DAC960_gam_ioctl(struct file *file, unsigned int Request,
                                                        DataTransferLength,
                                                        &DataTransferBufferDMA);
 	    if (DataTransferBuffer == NULL)
-	    	break;
+		goto out;
 	  }
 	else if (DataTransferLength < 0)
 	  {
 	    DataTransferBuffer = pci_alloc_consistent(Controller->PCIDevice,
 				-DataTransferLength, &DataTransferBufferDMA);
 	    if (DataTransferBuffer == NULL)
-	    	break;
+		goto out;
 	    if (copy_from_user(DataTransferBuffer,
 			       UserCommand.DataTransferBuffer,
 			       -DataTransferLength)) {
 		ErrorCode = -EFAULT;
-		break;
+		goto out;
 	    }
 	  }
 	if (CommandOpcode == DAC960_V1_DCDB)
@@ -6837,12 +6823,12 @@ static long DAC960_gam_ioctl(struct file *file, unsigned int Request,
 	if (DCDB_IOBUF != NULL)
 	  pci_free_consistent(Controller->PCIDevice, sizeof(DAC960_V1_DCDB_T),
 			DCDB_IOBUF, DCDB_IOBUFDMA);
-      	break;
-      }
-    case DAC960_IOCTL_V2_EXECUTE_COMMAND:
-      {
-	DAC960_V2_UserCommand_T __user *UserSpaceUserCommand =
-	  (DAC960_V2_UserCommand_T __user *) Argument;
+	out:
+	return ErrorCode;
+}
+
+static long DAC960_gam_v2_execute_command(DAC960_V2_UserCommand_T __user *UserSpaceUserCommand)
+{
 	DAC960_V2_UserCommand_T UserCommand;
 	DAC960_Controller_T *Controller;
 	DAC960_Command_T *Command = NULL;
@@ -6855,26 +6841,26 @@ static long DAC960_gam_ioctl(struct file *file, unsigned int Request,
 	dma_addr_t DataTransferBufferDMA;
 	unsigned char *RequestSenseBuffer = NULL;
 	dma_addr_t RequestSenseBufferDMA;
+	long ErrorCode = -EINVAL;
 
-	ErrorCode = -EINVAL;
 	if (UserSpaceUserCommand == NULL)
-		break;
+		goto out;
 	if (copy_from_user(&UserCommand, UserSpaceUserCommand,
 			   sizeof(DAC960_V2_UserCommand_T))) {
 		ErrorCode = -EFAULT;
-		break;
+		goto out;
 	}
 	ErrorCode = -ENXIO;
 	ControllerNumber = UserCommand.ControllerNumber;
 	if (ControllerNumber < 0 ||
 	    ControllerNumber > DAC960_ControllerCount - 1)
-	    	break;
+		goto out;
 	Controller = DAC960_Controllers[ControllerNumber];
 	if (Controller == NULL)
-		break;
+		goto out;
 	if (Controller->FirmwareType != DAC960_V2_Controller){
 		ErrorCode = -EINVAL;
-		break;
+		goto out;
 	}
 	DataTransferLength = UserCommand.DataTransferLength;
     	ErrorCode = -ENOMEM;
@@ -6884,14 +6870,14 @@ static long DAC960_gam_ioctl(struct file *file, unsigned int Request,
                                                        DataTransferLength,
                                                        &DataTransferBufferDMA);
 	    if (DataTransferBuffer == NULL)
-	    	break;
+		goto out;
 	  }
 	else if (DataTransferLength < 0)
 	  {
 	    DataTransferBuffer = pci_alloc_consistent(Controller->PCIDevice,
 				-DataTransferLength, &DataTransferBufferDMA);
 	    if (DataTransferBuffer == NULL)
-	    	break;
+		goto out;
 	    if (copy_from_user(DataTransferBuffer,
 			       UserCommand.DataTransferBuffer,
 			       -DataTransferLength)) {
@@ -7001,42 +6987,44 @@ static long DAC960_gam_ioctl(struct file *file, unsigned int Request,
 	if (RequestSenseBuffer != NULL)
 	  pci_free_consistent(Controller->PCIDevice, RequestSenseLength,
 		RequestSenseBuffer, RequestSenseBufferDMA);
-        break;
-      }
-    case DAC960_IOCTL_V2_GET_HEALTH_STATUS:
-      {
-	DAC960_V2_GetHealthStatus_T __user *UserSpaceGetHealthStatus =
-	  (DAC960_V2_GetHealthStatus_T __user *) Argument;
+out:
+        return ErrorCode;
+}
+
+static long DAC960_gam_v2_get_health_status(DAC960_V2_GetHealthStatus_T __user *UserSpaceGetHealthStatus)
+{
 	DAC960_V2_GetHealthStatus_T GetHealthStatus;
 	DAC960_V2_HealthStatusBuffer_T HealthStatusBuffer;
 	DAC960_Controller_T *Controller;
 	int ControllerNumber;
+	long ErrorCode;
+
 	if (UserSpaceGetHealthStatus == NULL) {
 		ErrorCode = -EINVAL;
-		break;
+		goto out;
 	}
 	if (copy_from_user(&GetHealthStatus, UserSpaceGetHealthStatus,
 			   sizeof(DAC960_V2_GetHealthStatus_T))) {
 		ErrorCode = -EFAULT;
-		break;
+		goto out;
 	}
 	ErrorCode = -ENXIO;
 	ControllerNumber = GetHealthStatus.ControllerNumber;
 	if (ControllerNumber < 0 ||
 	    ControllerNumber > DAC960_ControllerCount - 1)
-		    break;
+		goto out;
 	Controller = DAC960_Controllers[ControllerNumber];
 	if (Controller == NULL)
-		break;
+		goto out;
 	if (Controller->FirmwareType != DAC960_V2_Controller) {
 		ErrorCode = -EINVAL;
-		break;
+		goto out;
 	}
 	if (copy_from_user(&HealthStatusBuffer,
 			   GetHealthStatus.HealthStatusBuffer,
 			   sizeof(DAC960_V2_HealthStatusBuffer_T))) {
 		ErrorCode = -EFAULT;
-		break;
+		goto out;
 	}
 	ErrorCode = wait_event_interruptible_timeout(Controller->HealthStatusWaitQueue,
 			!(Controller->V2.HealthStatusBuffer->StatusChangeCounter
@@ -7046,7 +7034,7 @@ static long DAC960_gam_ioctl(struct file *file, unsigned int Request,
 			DAC960_MonitoringTimerInterval);
 	if (ErrorCode == -ERESTARTSYS) {
 		ErrorCode = -EINTR;
-		break;
+		goto out;
 	}
 	if (copy_to_user(GetHealthStatus.HealthStatusBuffer,
 			 Controller->V2.HealthStatusBuffer,
@@ -7054,7 +7042,39 @@ static long DAC960_gam_ioctl(struct file *file, unsigned int Request,
 		ErrorCode = -EFAULT;
 	else
 		ErrorCode =  0;
-      }
+
+out:
+	return ErrorCode;
+}
+
+/*
+ * DAC960_gam_ioctl is the ioctl function for performing RAID operations.
+*/
+
+static long DAC960_gam_ioctl(struct file *file, unsigned int Request,
+						unsigned long Argument)
+{
+  long ErrorCode = 0;
+  void __user *argp = (void __user *)Argument;
+  if (!capable(CAP_SYS_ADMIN)) return -EACCES;
+
+  mutex_lock(&DAC960_mutex);
+  switch (Request)
+    {
+    case DAC960_IOCTL_GET_CONTROLLER_COUNT:
+      ErrorCode = DAC960_ControllerCount;
+      break;
+    case DAC960_IOCTL_GET_CONTROLLER_INFO:
+      ErrorCode = DAC960_gam_get_controller_info(argp);
+      break;
+    case DAC960_IOCTL_V1_EXECUTE_COMMAND:
+      ErrorCode = DAC960_gam_v1_execute_command(argp);
+      break;
+    case DAC960_IOCTL_V2_EXECUTE_COMMAND:
+      ErrorCode = DAC960_gam_v2_execute_command(argp);
+      break;
+    case DAC960_IOCTL_V2_GET_HEALTH_STATUS:
+      ErrorCode = DAC960_gam_v2_get_health_status(argp);
       break;
       default:
 	ErrorCode = -ENOTTY;

drivers/block/Kconfig

@@ -20,6 +20,10 @@ config BLK_DEV_NULL_BLK
 	tristate "Null test block driver"
 	select CONFIGFS_FS
 
+config BLK_DEV_NULL_BLK_FAULT_INJECTION
+	bool "Support fault injection for Null test block driver"
+	depends on BLK_DEV_NULL_BLK && FAULT_INJECTION
+
 config BLK_DEV_FD
 	tristate "Normal floppy disk support"
 	depends on ARCH_MAY_HAVE_PC_FDC

drivers/block/aoe/aoe.h

@@ -112,8 +112,7 @@ enum frame_flags {
 struct frame {
 	struct list_head head;
 	u32 tag;
-	struct timeval sent;	/* high-res time packet was sent */
-	u32 sent_jiffs;		/* low-res jiffies-based sent time */
+	ktime_t sent;			/* high-res time packet was sent */
 	ulong waited;
 	ulong waited_total;
 	struct aoetgt *t;		/* parent target I belong to */

drivers/block/aoe/aoecmd.c

@@ -398,8 +398,7 @@ aoecmd_ata_rw(struct aoedev *d)
 
 	skb = skb_clone(f->skb, GFP_ATOMIC);
 	if (skb) {
-		do_gettimeofday(&f->sent);
-		f->sent_jiffs = (u32) jiffies;
+		f->sent = ktime_get();
 		__skb_queue_head_init(&queue);
 		__skb_queue_tail(&queue, skb);
 		aoenet_xmit(&queue);
@@ -489,8 +488,7 @@ resend(struct aoedev *d, struct frame *f)
 	skb = skb_clone(skb, GFP_ATOMIC);
 	if (skb == NULL)
 		return;
-	do_gettimeofday(&f->sent);
-	f->sent_jiffs = (u32) jiffies;
+	f->sent = ktime_get();
 	__skb_queue_head_init(&queue);
 	__skb_queue_tail(&queue, skb);
 	aoenet_xmit(&queue);
@@ -499,33 +497,17 @@ resend(struct aoedev *d, struct frame *f)
 static int
 tsince_hr(struct frame *f)
 {
-	struct timeval now;
-	int n;
+	u64 delta = ktime_to_ns(ktime_sub(ktime_get(), f->sent));
 
-	do_gettimeofday(&now);
-	n = now.tv_usec - f->sent.tv_usec;
-	n += (now.tv_sec - f->sent.tv_sec) * USEC_PER_SEC;
+	/* delta is normally under 4.2 seconds, avoid 64-bit division */
+	if (likely(delta <= UINT_MAX))
+		return (u32)delta / NSEC_PER_USEC;
 
-	if (n < 0)
-		n = -n;
+	/* avoid overflow after 71 minutes */
+	if (delta > ((u64)INT_MAX * NSEC_PER_USEC))
+		return INT_MAX;
 
-	/* For relatively long periods, use jiffies to avoid
-	 * discrepancies caused by updates to the system time.
-	 *
-	 * On system with HZ of 1000, 32-bits is over 49 days
-	 * worth of jiffies, or over 71 minutes worth of usecs.
-	 *
-	 * Jiffies overflow is handled by subtraction of unsigned ints:
-	 * (gdb) print (unsigned) 2 - (unsigned) 0xfffffffe
-	 * $3 = 4
-	 * (gdb)
-	 */
-	if (n > USEC_PER_SEC / 4) {
-		n = ((u32) jiffies) - f->sent_jiffs;
-		n *= USEC_PER_SEC / HZ;
-	}
-
-	return n;
+	return div_u64(delta, NSEC_PER_USEC);
 }
 
 static int
@@ -589,7 +571,6 @@ reassign_frame(struct frame *f)
 	nf->waited = 0;
 	nf->waited_total = f->waited_total;
 	nf->sent = f->sent;
-	nf->sent_jiffs = f->sent_jiffs;
 	f->skb = skb;
 
 	return nf;
@@ -633,8 +614,7 @@ probe(struct aoetgt *t)
 
 	skb = skb_clone(f->skb, GFP_ATOMIC);
 	if (skb) {
-		do_gettimeofday(&f->sent);
-		f->sent_jiffs = (u32) jiffies;
+		f->sent = ktime_get();
 		__skb_queue_head_init(&queue);
 		__skb_queue_tail(&queue, skb);
 		aoenet_xmit(&queue);
@@ -1432,10 +1412,8 @@ aoecmd_ata_id(struct aoedev *d)
 	d->timer.function = rexmit_timer;
 
 	skb = skb_clone(skb, GFP_ATOMIC);
-	if (skb) {
-		do_gettimeofday(&f->sent);
-		f->sent_jiffs = (u32) jiffies;
-	}
+	if (skb)
+		f->sent = ktime_get();
 
 	return skb;
 }

drivers/block/drbd/drbd_bitmap.c

@@ -953,7 +953,7 @@ static void drbd_bm_endio(struct bio *bio)
 	struct drbd_bm_aio_ctx *ctx = bio->bi_private;
 	struct drbd_device *device = ctx->device;
 	struct drbd_bitmap *b = device->bitmap;
-	unsigned int idx = bm_page_to_idx(bio->bi_io_vec[0].bv_page);
+	unsigned int idx = bm_page_to_idx(bio_first_page_all(bio));
 
 	if ((ctx->flags & BM_AIO_COPY_PAGES) == 0 &&
 	    !bm_test_page_unchanged(b->bm_pages[idx]))

drivers/block/null_blk.c

@@ -12,9 +12,9 @@
 #include <linux/slab.h>
 #include <linux/blk-mq.h>
 #include <linux/hrtimer.h>
-#include <linux/lightnvm.h>
 #include <linux/configfs.h>
 #include <linux/badblocks.h>
+#include <linux/fault-inject.h>
 
 #define SECTOR_SHIFT		9
 #define PAGE_SECTORS_SHIFT	(PAGE_SHIFT - SECTOR_SHIFT)
@@ -27,6 +27,10 @@
 #define TICKS_PER_SEC		50ULL
 #define TIMER_INTERVAL		(NSEC_PER_SEC / TICKS_PER_SEC)
 
+#ifdef CONFIG_BLK_DEV_NULL_BLK_FAULT_INJECTION
+static DECLARE_FAULT_ATTR(null_timeout_attr);
+#endif
+
 static inline u64 mb_per_tick(int mbps)
 {
 	return (1 << 20) / TICKS_PER_SEC * ((u64) mbps);
@@ -107,7 +111,6 @@ struct nullb_device {
 	unsigned int hw_queue_depth; /* queue depth */
 	unsigned int index; /* index of the disk, only valid with a disk */
 	unsigned int mbps; /* Bandwidth throttle cap (in MB/s) */
-	bool use_lightnvm; /* register as a LightNVM device */
 	bool blocking; /* blocking blk-mq device */
 	bool use_per_node_hctx; /* use per-node allocation for hardware context */
 	bool power; /* power on/off the device */
@@ -121,7 +124,6 @@ struct nullb {
 	unsigned int index;
 	struct request_queue *q;
 	struct gendisk *disk;
-	struct nvm_dev *ndev;
 	struct blk_mq_tag_set *tag_set;
 	struct blk_mq_tag_set __tag_set;
 	unsigned int queue_depth;
@@ -139,7 +141,6 @@ static LIST_HEAD(nullb_list);
 static struct mutex lock;
 static int null_major;
 static DEFINE_IDA(nullb_indexes);
-static struct kmem_cache *ppa_cache;
 static struct blk_mq_tag_set tag_set;
 
 enum {
@@ -166,6 +167,11 @@ static int g_home_node = NUMA_NO_NODE;
 module_param_named(home_node, g_home_node, int, S_IRUGO);
 MODULE_PARM_DESC(home_node, "Home node for the device");
 
+#ifdef CONFIG_BLK_DEV_NULL_BLK_FAULT_INJECTION
+static char g_timeout_str[80];
+module_param_string(timeout, g_timeout_str, sizeof(g_timeout_str), S_IRUGO);
+#endif
+
 static int g_queue_mode = NULL_Q_MQ;
 
 static int null_param_store_val(const char *str, int *val, int min, int max)
@@ -208,10 +214,6 @@ static int nr_devices = 1;
 module_param(nr_devices, int, S_IRUGO);
 MODULE_PARM_DESC(nr_devices, "Number of devices to register");
 
-static bool g_use_lightnvm;
-module_param_named(use_lightnvm, g_use_lightnvm, bool, S_IRUGO);
-MODULE_PARM_DESC(use_lightnvm, "Register as a LightNVM device");
-
 static bool g_blocking;
 module_param_named(blocking, g_blocking, bool, S_IRUGO);
 MODULE_PARM_DESC(blocking, "Register as a blocking blk-mq driver device");
@@ -345,7 +347,6 @@ NULLB_DEVICE_ATTR(blocksize, uint);
 NULLB_DEVICE_ATTR(irqmode, uint);
 NULLB_DEVICE_ATTR(hw_queue_depth, uint);
 NULLB_DEVICE_ATTR(index, uint);
-NULLB_DEVICE_ATTR(use_lightnvm, bool);
 NULLB_DEVICE_ATTR(blocking, bool);
 NULLB_DEVICE_ATTR(use_per_node_hctx, bool);
 NULLB_DEVICE_ATTR(memory_backed, bool);
@@ -455,7 +456,6 @@ static struct configfs_attribute *nullb_device_attrs[] = {
 	&nullb_device_attr_irqmode,
 	&nullb_device_attr_hw_queue_depth,
 	&nullb_device_attr_index,
-	&nullb_device_attr_use_lightnvm,
 	&nullb_device_attr_blocking,
 	&nullb_device_attr_use_per_node_hctx,
 	&nullb_device_attr_power,
@@ -573,7 +573,6 @@ static struct nullb_device *null_alloc_dev(void)
 	dev->blocksize = g_bs;
 	dev->irqmode = g_irqmode;
 	dev->hw_queue_depth = g_hw_queue_depth;
-	dev->use_lightnvm = g_use_lightnvm;
 	dev->blocking = g_blocking;
 	dev->use_per_node_hctx = g_use_per_node_hctx;
 	return dev;
@@ -1352,6 +1351,12 @@ static blk_qc_t null_queue_bio(struct request_queue *q, struct bio *bio)
 	return BLK_QC_T_NONE;
 }
 
+static enum blk_eh_timer_return null_rq_timed_out_fn(struct request *rq)
+{
+	pr_info("null: rq %p timed out\n", rq);
+	return BLK_EH_HANDLED;
+}
+
 static int null_rq_prep_fn(struct request_queue *q, struct request *req)
 {
 	struct nullb *nullb = q->queuedata;
@@ -1369,6 +1374,16 @@ static int null_rq_prep_fn(struct request_queue *q, struct request *req)
 	return BLKPREP_DEFER;
 }
 
+static bool should_timeout_request(struct request *rq)
+{
+#ifdef CONFIG_BLK_DEV_NULL_BLK_FAULT_INJECTION
+	if (g_timeout_str[0])
+		return should_fail(&null_timeout_attr, 1);
+#endif
+
+	return false;
+}
+
 static void null_request_fn(struct request_queue *q)
 {
 	struct request *rq;
@@ -1376,12 +1391,20 @@ static void null_request_fn(struct request_queue *q)
 	while ((rq = blk_fetch_request(q)) != NULL) {
 		struct nullb_cmd *cmd = rq->special;
 
-		spin_unlock_irq(q->queue_lock);
-		null_handle_cmd(cmd);
-		spin_lock_irq(q->queue_lock);
+		if (!should_timeout_request(rq)) {
+			spin_unlock_irq(q->queue_lock);
+			null_handle_cmd(cmd);
+			spin_lock_irq(q->queue_lock);
+		}
 	}
 }
 
+static enum blk_eh_timer_return null_timeout_rq(struct request *rq, bool res)
+{
+	pr_info("null: rq %p timed out\n", rq);
+	return BLK_EH_HANDLED;
+}
+
 static blk_status_t null_queue_rq(struct blk_mq_hw_ctx *hctx,
 			 const struct blk_mq_queue_data *bd)
 {
@@ -1399,12 +1422,16 @@ static blk_status_t null_queue_rq(struct blk_mq_hw_ctx *hctx,
 
 	blk_mq_start_request(bd->rq);
 
-	return null_handle_cmd(cmd);
+	if (!should_timeout_request(bd->rq))
+		return null_handle_cmd(cmd);
+
+	return BLK_STS_OK;
 }
 
 static const struct blk_mq_ops null_mq_ops = {
 	.queue_rq       = null_queue_rq,
 	.complete	= null_softirq_done_fn,
+	.timeout	= null_timeout_rq,
 };
 
 static void cleanup_queue(struct nullb_queue *nq)
@@ -1423,170 +1450,6 @@ static void cleanup_queues(struct nullb *nullb)
 	kfree(nullb->queues);
 }
 
-#ifdef CONFIG_NVM
-
-static void null_lnvm_end_io(struct request *rq, blk_status_t status)
-{
-	struct nvm_rq *rqd = rq->end_io_data;
-
-	/* XXX: lighnvm core seems to expect NVM_RSP_* values here.. */
-	rqd->error = status ? -EIO : 0;
-	nvm_end_io(rqd);
-
-	blk_put_request(rq);
-}
-
-static int null_lnvm_submit_io(struct nvm_dev *dev, struct nvm_rq *rqd)
-{
-	struct request_queue *q = dev->q;
-	struct request *rq;
-	struct bio *bio = rqd->bio;
-
-	rq = blk_mq_alloc_request(q,
-		op_is_write(bio_op(bio)) ? REQ_OP_DRV_OUT : REQ_OP_DRV_IN, 0);
-	if (IS_ERR(rq))
-		return -ENOMEM;
-
-	blk_init_request_from_bio(rq, bio);
-
-	rq->end_io_data = rqd;
-
-	blk_execute_rq_nowait(q, NULL, rq, 0, null_lnvm_end_io);
-
-	return 0;
-}
-
-static int null_lnvm_id(struct nvm_dev *dev, struct nvm_id *id)
-{
-	struct nullb *nullb = dev->q->queuedata;
-	sector_t size = (sector_t)nullb->dev->size * 1024 * 1024ULL;
-	sector_t blksize;
-	struct nvm_id_group *grp;
-
-	id->ver_id = 0x1;
-	id->vmnt = 0;
-	id->cap = 0x2;
-	id->dom = 0x1;
-
-	id->ppaf.blk_offset = 0;
-	id->ppaf.blk_len = 16;
-	id->ppaf.pg_offset = 16;
-	id->ppaf.pg_len = 16;
-	id->ppaf.sect_offset = 32;
-	id->ppaf.sect_len = 8;
-	id->ppaf.pln_offset = 40;
-	id->ppaf.pln_len = 8;
-	id->ppaf.lun_offset = 48;
-	id->ppaf.lun_len = 8;
-	id->ppaf.ch_offset = 56;
-	id->ppaf.ch_len = 8;
-
-	sector_div(size, nullb->dev->blocksize); /* convert size to pages */
-	size >>= 8; /* concert size to pgs pr blk */
-	grp = &id->grp;
-	grp->mtype = 0;
-	grp->fmtype = 0;
-	grp->num_ch = 1;
-	grp->num_pg = 256;
-	blksize = size;
-	size >>= 16;
-	grp->num_lun = size + 1;
-	sector_div(blksize, grp->num_lun);
-	grp->num_blk = blksize;
-	grp->num_pln = 1;
-
-	grp->fpg_sz = nullb->dev->blocksize;
-	grp->csecs = nullb->dev->blocksize;
-	grp->trdt = 25000;
-	grp->trdm = 25000;
-	grp->tprt = 500000;
-	grp->tprm = 500000;
-	grp->tbet = 1500000;
-	grp->tbem = 1500000;
-	grp->mpos = 0x010101; /* single plane rwe */
-	grp->cpar = nullb->dev->hw_queue_depth;
-
-	return 0;
-}
-
-static void *null_lnvm_create_dma_pool(struct nvm_dev *dev, char *name)
-{
-	mempool_t *virtmem_pool;
-
-	virtmem_pool = mempool_create_slab_pool(64, ppa_cache);
-	if (!virtmem_pool) {
-		pr_err("null_blk: Unable to create virtual memory pool\n");
-		return NULL;
-	}
-
-	return virtmem_pool;
-}
-
-static void null_lnvm_destroy_dma_pool(void *pool)
-{
-	mempool_destroy(pool);
-}
-
-static void *null_lnvm_dev_dma_alloc(struct nvm_dev *dev, void *pool,
-				gfp_t mem_flags, dma_addr_t *dma_handler)
-{
-	return mempool_alloc(pool, mem_flags);
-}
-
-static void null_lnvm_dev_dma_free(void *pool, void *entry,
-							dma_addr_t dma_handler)
-{
-	mempool_free(entry, pool);
-}
-
-static struct nvm_dev_ops null_lnvm_dev_ops = {
-	.identity		= null_lnvm_id,
-	.submit_io		= null_lnvm_submit_io,
-
-	.create_dma_pool	= null_lnvm_create_dma_pool,
-	.destroy_dma_pool	= null_lnvm_destroy_dma_pool,
-	.dev_dma_alloc		= null_lnvm_dev_dma_alloc,
-	.dev_dma_free		= null_lnvm_dev_dma_free,
-
-	/* Simulate nvme protocol restriction */
-	.max_phys_sect		= 64,
-};
-
-static int null_nvm_register(struct nullb *nullb)
-{
-	struct nvm_dev *dev;
-	int rv;
-
-	dev = nvm_alloc_dev(0);
-	if (!dev)
-		return -ENOMEM;
-
-	dev->q = nullb->q;
-	memcpy(dev->name, nullb->disk_name, DISK_NAME_LEN);
-	dev->ops = &null_lnvm_dev_ops;
-
-	rv = nvm_register(dev);
-	if (rv) {
-		kfree(dev);
-		return rv;
-	}
-	nullb->ndev = dev;
-	return 0;
-}
-
-static void null_nvm_unregister(struct nullb *nullb)
-{
-	nvm_unregister(nullb->ndev);
-}
-#else
-static int null_nvm_register(struct nullb *nullb)
-{
-	pr_err("null_blk: CONFIG_NVM needs to be enabled for LightNVM\n");
-	return -EINVAL;
-}
-static void null_nvm_unregister(struct nullb *nullb) {}
-#endif /* CONFIG_NVM */
-
 static void null_del_dev(struct nullb *nullb)
 {
 	struct nullb_device *dev = nullb->dev;
@@ -1595,10 +1458,7 @@ static void null_del_dev(struct nullb *nullb)
 
 	list_del_init(&nullb->list);
 
-	if (dev->use_lightnvm)
-		null_nvm_unregister(nullb);
-	else
-		del_gendisk(nullb->disk);
+	del_gendisk(nullb->disk);
 
 	if (test_bit(NULLB_DEV_FL_THROTTLED, &nullb->dev->flags)) {
 		hrtimer_cancel(&nullb->bw_timer);
@@ -1610,8 +1470,7 @@ static void null_del_dev(struct nullb *nullb)
 	if (dev->queue_mode == NULL_Q_MQ &&
 	    nullb->tag_set == &nullb->__tag_set)
 		blk_mq_free_tag_set(nullb->tag_set);
-	if (!dev->use_lightnvm)
-		put_disk(nullb->disk);
+	put_disk(nullb->disk);
 	cleanup_queues(nullb);
 	if (null_cache_active(nullb))
 		null_free_device_storage(nullb->dev, true);
@@ -1775,11 +1634,6 @@ static void null_validate_conf(struct nullb_device *dev)
 {
 	dev->blocksize = round_down(dev->blocksize, 512);
 	dev->blocksize = clamp_t(unsigned int, dev->blocksize, 512, 4096);
-	if (dev->use_lightnvm && dev->blocksize != 4096)
-		dev->blocksize = 4096;
-
-	if (dev->use_lightnvm && dev->queue_mode != NULL_Q_MQ)
-		dev->queue_mode = NULL_Q_MQ;
 
 	if (dev->queue_mode == NULL_Q_MQ && dev->use_per_node_hctx) {
 		if (dev->submit_queues != nr_online_nodes)
@@ -1805,6 +1659,20 @@ static void null_validate_conf(struct nullb_device *dev)
 		dev->mbps = 0;
 }
 
+static bool null_setup_fault(void)
+{
+#ifdef CONFIG_BLK_DEV_NULL_BLK_FAULT_INJECTION
+	if (!g_timeout_str[0])
+		return true;
+
+	if (!setup_fault_attr(&null_timeout_attr, g_timeout_str))
+		return false;
+
+	null_timeout_attr.verbose = 0;
+#endif
+	return true;
+}
+
 static int null_add_dev(struct nullb_device *dev)
 {
 	struct nullb *nullb;
@@ -1838,6 +1706,10 @@ static int null_add_dev(struct nullb_device *dev)
 		if (rv)
 			goto out_cleanup_queues;
 
+		if (!null_setup_fault())
+			goto out_cleanup_queues;
+
+		nullb->tag_set->timeout = 5 * HZ;
 		nullb->q = blk_mq_init_queue(nullb->tag_set);
 		if (IS_ERR(nullb->q)) {
 			rv = -ENOMEM;
@@ -1861,8 +1733,14 @@ static int null_add_dev(struct nullb_device *dev)
 			rv = -ENOMEM;
 			goto out_cleanup_queues;
 		}
+
+		if (!null_setup_fault())
+			goto out_cleanup_blk_queue;
+
 		blk_queue_prep_rq(nullb->q, null_rq_prep_fn);
 		blk_queue_softirq_done(nullb->q, null_softirq_done_fn);
+		blk_queue_rq_timed_out(nullb->q, null_rq_timed_out_fn);
+		nullb->q->rq_timeout = 5 * HZ;
 		rv = init_driver_queues(nullb);
 		if (rv)
 			goto out_cleanup_blk_queue;
@@ -1895,11 +1773,7 @@ static int null_add_dev(struct nullb_device *dev)
 
 	sprintf(nullb->disk_name, "nullb%d", nullb->index);
 
-	if (dev->use_lightnvm)
-		rv = null_nvm_register(nullb);
-	else
-		rv = null_gendisk_register(nullb);
-
+	rv = null_gendisk_register(nullb);
 	if (rv)
 		goto out_cleanup_blk_queue;
 
@@ -1938,18 +1812,6 @@ static int __init null_init(void)
 		g_bs = PAGE_SIZE;
 	}
 
-	if (g_use_lightnvm && g_bs != 4096) {
-		pr_warn("null_blk: LightNVM only supports 4k block size\n");
-		pr_warn("null_blk: defaults block size to 4k\n");
-		g_bs = 4096;
-	}
-
-	if (g_use_lightnvm && g_queue_mode != NULL_Q_MQ) {
-		pr_warn("null_blk: LightNVM only supported for blk-mq\n");
-		pr_warn("null_blk: defaults queue mode to blk-mq\n");
-		g_queue_mode = NULL_Q_MQ;
-	}
-
 	if (g_queue_mode == NULL_Q_MQ && g_use_per_node_hctx) {
 		if (g_submit_queues != nr_online_nodes) {
 			pr_warn("null_blk: submit_queues param is set to %u.\n",
@@ -1982,16 +1844,6 @@ static int __init null_init(void)
 		goto err_conf;
 	}
 
-	if (g_use_lightnvm) {
-		ppa_cache = kmem_cache_create("ppa_cache", 64 * sizeof(u64),
-								0, 0, NULL);
-		if (!ppa_cache) {
-			pr_err("null_blk: unable to create ppa cache\n");
-			ret = -ENOMEM;
-			goto err_ppa;
-		}
-	}
-
 	for (i = 0; i < nr_devices; i++) {
 		dev = null_alloc_dev();
 		if (!dev) {
@@ -2015,8 +1867,6 @@ err_dev:
 		null_del_dev(nullb);
 		null_free_dev(dev);
 	}
-	kmem_cache_destroy(ppa_cache);
-err_ppa:
 	unregister_blkdev(null_major, "nullb");
 err_conf:
 	configfs_unregister_subsystem(&nullb_subsys);
@@ -2047,8 +1897,6 @@ static void __exit null_exit(void)
 
 	if (g_queue_mode == NULL_Q_MQ && shared_tags)
 		blk_mq_free_tag_set(&tag_set);
-
-	kmem_cache_destroy(ppa_cache);
 }
 
 module_init(null_init);

drivers/block/pktcdvd.c

@@ -2579,14 +2579,14 @@ static int pkt_new_dev(struct pktcdvd_device *pd, dev_t dev)
 	bdev = bdget(dev);
 	if (!bdev)
 		return -ENOMEM;
+	ret = blkdev_get(bdev, FMODE_READ | FMODE_NDELAY, NULL);
+	if (ret)
+		return ret;
 	if (!blk_queue_scsi_passthrough(bdev_get_queue(bdev))) {
 		WARN_ONCE(true, "Attempt to register a non-SCSI queue\n");
-		bdput(bdev);
+		blkdev_put(bdev, FMODE_READ | FMODE_NDELAY);
 		return -EINVAL;
 	}
-	ret = blkdev_get(bdev, FMODE_READ | FMODE_NDELAY, NULL);
-	if (ret)
-		return ret;
 
 	/* This is safe, since we have a reference from open(). */
 	__module_get(THIS_MODULE);
@@ -2745,7 +2745,7 @@ static int pkt_setup_dev(dev_t dev, dev_t* pkt_dev)
 	pd->pkt_dev = MKDEV(pktdev_major, idx);
 	ret = pkt_new_dev(pd, dev);
 	if (ret)
-		goto out_new_dev;
+		goto out_mem2;
 
 	/* inherit events of the host device */
 	disk->events = pd->bdev->bd_disk->events;
@@ -2763,8 +2763,6 @@ static int pkt_setup_dev(dev_t dev, dev_t* pkt_dev)
 	mutex_unlock(&ctl_mutex);
 	return 0;
 
-out_new_dev:
-	blk_cleanup_queue(disk->queue);
 out_mem2:
 	put_disk(disk);
 out_mem:

drivers/block/smart1,2.h

@@ -1,278 +0,0 @@
-/*
- *    Disk Array driver for Compaq SMART2 Controllers
- *    Copyright 1998 Compaq Computer Corporation
- *
- *    This program is free software; you can redistribute it and/or modify
- *    it under the terms of the GNU General Public License as published by
- *    the Free Software Foundation; either version 2 of the License, or
- *    (at your option) any later version.
- *
- *    This program is distributed in the hope that it will be useful,
- *    but WITHOUT ANY WARRANTY; without even the implied warranty of
- *    MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
- *    NON INFRINGEMENT.  See the GNU General Public License for more details.
- *
- *    You should have received a copy of the GNU General Public License
- *    along with this program; if not, write to the Free Software
- *    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
- *
- *    Questions/Comments/Bugfixes to iss_storagedev@hp.com
- *
- *    If you want to make changes, improve or add functionality to this
- *    driver, you'll probably need the Compaq Array Controller Interface
- *    Specificiation (Document number ECG086/1198)
- */
-
-/*
- * This file contains the controller communication implementation for
- * Compaq SMART-1 and SMART-2 controllers.  To the best of my knowledge,
- * this should support:
- *
- *  PCI:
- *  SMART-2/P, SMART-2DH, SMART-2SL, SMART-221, SMART-3100ES, SMART-3200
- *  Integerated SMART Array Controller, SMART-4200, SMART-4250ES
- *
- *  EISA:
- *  SMART-2/E, SMART, IAES, IDA-2, IDA
- */
-
-/*
- * Memory mapped FIFO interface (SMART 42xx cards)
- */
-static void smart4_submit_command(ctlr_info_t *h, cmdlist_t *c)
-{
-        writel(c->busaddr, h->vaddr + S42XX_REQUEST_PORT_OFFSET);
-}
-
-/*  
- *  This card is the opposite of the other cards.  
- *   0 turns interrupts on... 
- *   0x08 turns them off... 
- */
-static void smart4_intr_mask(ctlr_info_t *h, unsigned long val)
-{
-	if (val) 
-	{ /* Turn interrupts on */
-		writel(0, h->vaddr + S42XX_REPLY_INTR_MASK_OFFSET);
-	} else /* Turn them off */
-	{
-        	writel( S42XX_INTR_OFF, 
-			h->vaddr + S42XX_REPLY_INTR_MASK_OFFSET);
-	}
-}
-
-/*
- *  For older cards FIFO Full = 0. 
- *  On this card 0 means there is room, anything else FIFO Full. 
- * 
- */ 
-static unsigned long smart4_fifo_full(ctlr_info_t *h)
-{
-	
-        return (!readl(h->vaddr + S42XX_REQUEST_PORT_OFFSET));
-}
-
-/* This type of controller returns -1 if the fifo is empty, 
- *    Not 0 like the others.
- *    And we need to let it know we read a value out 
- */ 
-static unsigned long smart4_completed(ctlr_info_t *h)
-{
-	long register_value 
-		= readl(h->vaddr + S42XX_REPLY_PORT_OFFSET);
-
-	/* Fifo is empty */
-	if( register_value == 0xffffffff)
-		return 0; 	
-
-	/* Need to let it know we got the reply */
-	/* We do this by writing a 0 to the port we just read from */
-	writel(0, h->vaddr + S42XX_REPLY_PORT_OFFSET);
-
-	return ((unsigned long) register_value); 
-}
-
- /*
- *  This hardware returns interrupt pending at a different place and 
- *  it does not tell us if the fifo is empty, we will have check  
- *  that by getting a 0 back from the command_completed call. 
- */
-static unsigned long smart4_intr_pending(ctlr_info_t *h)
-{
-	unsigned long register_value  = 
-		readl(h->vaddr + S42XX_INTR_STATUS);
-
-	if( register_value &  S42XX_INTR_PENDING) 
-		return  FIFO_NOT_EMPTY;	
-	return 0 ;
-}
-
-static struct access_method smart4_access = {
-	smart4_submit_command,
-	smart4_intr_mask,
-	smart4_fifo_full,
-	smart4_intr_pending,
-	smart4_completed,
-};
-
-/*
- * Memory mapped FIFO interface (PCI SMART2 and SMART 3xxx cards)
- */
-static void smart2_submit_command(ctlr_info_t *h, cmdlist_t *c)
-{
-	writel(c->busaddr, h->vaddr + COMMAND_FIFO);
-}
-
-static void smart2_intr_mask(ctlr_info_t *h, unsigned long val)
-{
-	writel(val, h->vaddr + INTR_MASK);
-}
-
-static unsigned long smart2_fifo_full(ctlr_info_t *h)
-{
-	return readl(h->vaddr + COMMAND_FIFO);
-}
-
-static unsigned long smart2_completed(ctlr_info_t *h)
-{
-	return readl(h->vaddr + COMMAND_COMPLETE_FIFO);
-}
-
-static unsigned long smart2_intr_pending(ctlr_info_t *h)
-{
-	return readl(h->vaddr + INTR_PENDING);
-}
-
-static struct access_method smart2_access = {
-	smart2_submit_command,
-	smart2_intr_mask,
-	smart2_fifo_full,
-	smart2_intr_pending,
-	smart2_completed,
-};
-
-/*
- *  IO access for SMART-2/E cards
- */
-static void smart2e_submit_command(ctlr_info_t *h, cmdlist_t *c)
-{
-	outl(c->busaddr, h->io_mem_addr + COMMAND_FIFO);
-}
-
-static void smart2e_intr_mask(ctlr_info_t *h, unsigned long val)
-{
-	outl(val, h->io_mem_addr + INTR_MASK);
-}
-
-static unsigned long smart2e_fifo_full(ctlr_info_t *h)
-{
-	return inl(h->io_mem_addr + COMMAND_FIFO);
-}
-
-static unsigned long smart2e_completed(ctlr_info_t *h)
-{
-	return inl(h->io_mem_addr + COMMAND_COMPLETE_FIFO);
-}
-
-static unsigned long smart2e_intr_pending(ctlr_info_t *h)
-{
-	return inl(h->io_mem_addr + INTR_PENDING);
-}
-
-static struct access_method smart2e_access = {
-	smart2e_submit_command,
-	smart2e_intr_mask,
-	smart2e_fifo_full,
-	smart2e_intr_pending,
-	smart2e_completed,
-};
-
-/*
- *  IO access for older SMART-1 type cards
- */
-#define SMART1_SYSTEM_MASK		0xC8E
-#define SMART1_SYSTEM_DOORBELL		0xC8F
-#define SMART1_LOCAL_MASK		0xC8C
-#define SMART1_LOCAL_DOORBELL		0xC8D
-#define SMART1_INTR_MASK		0xC89
-#define SMART1_LISTADDR			0xC90
-#define SMART1_LISTLEN			0xC94
-#define SMART1_TAG			0xC97
-#define SMART1_COMPLETE_ADDR		0xC98
-#define SMART1_LISTSTATUS		0xC9E
-
-#define CHANNEL_BUSY			0x01
-#define CHANNEL_CLEAR			0x02
-
-static void smart1_submit_command(ctlr_info_t *h, cmdlist_t *c)
-{
-	/*
-	 * This __u16 is actually a bunch of control flags on SMART
-	 * and below.  We want them all to be zero.
-	 */
-	c->hdr.size = 0;
-
-	outb(CHANNEL_CLEAR, h->io_mem_addr + SMART1_SYSTEM_DOORBELL);
-
-	outl(c->busaddr, h->io_mem_addr + SMART1_LISTADDR);
-	outw(c->size, h->io_mem_addr + SMART1_LISTLEN);
-
-	outb(CHANNEL_BUSY, h->io_mem_addr + SMART1_LOCAL_DOORBELL);
-}
-
-static void smart1_intr_mask(ctlr_info_t *h, unsigned long val)
-{
-	if (val == 1) {
-		outb(0xFD, h->io_mem_addr + SMART1_SYSTEM_DOORBELL);
-		outb(CHANNEL_BUSY, h->io_mem_addr + SMART1_LOCAL_DOORBELL);
-		outb(0x01, h->io_mem_addr + SMART1_INTR_MASK);
-		outb(0x01, h->io_mem_addr + SMART1_SYSTEM_MASK);
-	} else {
-		outb(0, h->io_mem_addr + 0xC8E);
-	}
-}
-
-static unsigned long smart1_fifo_full(ctlr_info_t *h)
-{
-	unsigned char chan;
-	chan = inb(h->io_mem_addr + SMART1_SYSTEM_DOORBELL) & CHANNEL_CLEAR;
-	return chan;
-}
-
-static unsigned long smart1_completed(ctlr_info_t *h)
-{
-	unsigned char status;
-	unsigned long cmd;
-
-	if (inb(h->io_mem_addr + SMART1_SYSTEM_DOORBELL) & CHANNEL_BUSY) {
-		outb(CHANNEL_BUSY, h->io_mem_addr + SMART1_SYSTEM_DOORBELL);
-
-		cmd = inl(h->io_mem_addr + SMART1_COMPLETE_ADDR);
-		status = inb(h->io_mem_addr + SMART1_LISTSTATUS);
-
-		outb(CHANNEL_CLEAR, h->io_mem_addr + SMART1_LOCAL_DOORBELL);
-
-		/*
-		 * this is x86 (actually compaq x86) only, so it's ok
-		 */
-		if (cmd) ((cmdlist_t*)bus_to_virt(cmd))->req.hdr.rcode = status;
-	} else {
-		cmd = 0;
-	}
-	return cmd;
-}
-
-static unsigned long smart1_intr_pending(ctlr_info_t *h)
-{
-	unsigned char chan;
-	chan = inb(h->io_mem_addr + SMART1_SYSTEM_DOORBELL) & CHANNEL_BUSY;
-	return chan;
-}
-
-static struct access_method smart1_access = {
-	smart1_submit_command,
-	smart1_intr_mask,
-	smart1_fifo_full,
-	smart1_intr_pending,
-	smart1_completed,
-};

drivers/block/zram/zram_drv.c

@@ -430,7 +430,7 @@ static void put_entry_bdev(struct zram *zram, unsigned long entry)
 
 static void zram_page_end_io(struct bio *bio)
 {
-	struct page *page = bio->bi_io_vec[0].bv_page;
+	struct page *page = bio_first_page_all(bio);
 
 	page_endio(page, op_is_write(bio_op(bio)),
 			blk_status_to_errno(bio->bi_status));

drivers/lightnvm/Kconfig

@@ -27,13 +27,6 @@ config NVM_DEBUG
 
 	It is required to create/remove targets without IOCTLs.
 
-config NVM_RRPC
-	tristate "Round-robin Hybrid Open-Channel SSD target"
-	---help---
-	Allows an open-channel SSD to be exposed as a block device to the
-	host. The target is implemented using a linear mapping table and
-	cost-based garbage collection. It is optimized for 4K IO sizes.
-
 config NVM_PBLK
 	tristate "Physical Block Device Open-Channel SSD target"
 	---help---

drivers/lightnvm/Makefile

@@ -4,7 +4,6 @@
 #
 
 obj-$(CONFIG_NVM)		:= core.o
-obj-$(CONFIG_NVM_RRPC)		+= rrpc.o
 obj-$(CONFIG_NVM_PBLK)		+= pblk.o
 pblk-y				:= pblk-init.o pblk-core.o pblk-rb.o \
 				   pblk-write.o pblk-cache.o pblk-read.o \

drivers/lightnvm/core.c

@@ -45,12 +45,6 @@ struct nvm_dev_map {
 	int nr_chnls;
 };
 
-struct nvm_area {
-	struct list_head list;
-	sector_t begin;
-	sector_t end;	/* end is excluded */
-};
-
 static struct nvm_target *nvm_find_target(struct nvm_dev *dev, const char *name)
 {
 	struct nvm_target *tgt;
@@ -62,6 +56,30 @@ static struct nvm_target *nvm_find_target(struct nvm_dev *dev, const char *name)
 	return NULL;
 }
 
+static bool nvm_target_exists(const char *name)
+{
+	struct nvm_dev *dev;
+	struct nvm_target *tgt;
+	bool ret = false;
+
+	down_write(&nvm_lock);
+	list_for_each_entry(dev, &nvm_devices, devices) {
+		mutex_lock(&dev->mlock);
+		list_for_each_entry(tgt, &dev->targets, list) {
+			if (!strcmp(name, tgt->disk->disk_name)) {
+				ret = true;
+				mutex_unlock(&dev->mlock);
+				goto out;
+			}
+		}
+		mutex_unlock(&dev->mlock);
+	}
+
+out:
+	up_write(&nvm_lock);
+	return ret;
+}
+
 static int nvm_reserve_luns(struct nvm_dev *dev, int lun_begin, int lun_end)
 {
 	int i;
@@ -104,7 +122,7 @@ static void nvm_remove_tgt_dev(struct nvm_tgt_dev *tgt_dev, int clear)
 		if (clear) {
 			for (j = 0; j < ch_map->nr_luns; j++) {
 				int lun = j + lun_offs[j];
-				int lunid = (ch * dev->geo.luns_per_chnl) + lun;
+				int lunid = (ch * dev->geo.nr_luns) + lun;
 
 				WARN_ON(!test_and_clear_bit(lunid,
 							dev->lun_map));
@@ -122,7 +140,8 @@ static void nvm_remove_tgt_dev(struct nvm_tgt_dev *tgt_dev, int clear)
 }
 
 static struct nvm_tgt_dev *nvm_create_tgt_dev(struct nvm_dev *dev,
-					      int lun_begin, int lun_end)
+					      u16 lun_begin, u16 lun_end,
+					      u16 op)
 {
 	struct nvm_tgt_dev *tgt_dev = NULL;
 	struct nvm_dev_map *dev_rmap = dev->rmap;
@@ -130,10 +149,10 @@ static struct nvm_tgt_dev *nvm_create_tgt_dev(struct nvm_dev *dev,
 	struct ppa_addr *luns;
 	int nr_luns = lun_end - lun_begin + 1;
 	int luns_left = nr_luns;
-	int nr_chnls = nr_luns / dev->geo.luns_per_chnl;
-	int nr_chnls_mod = nr_luns % dev->geo.luns_per_chnl;
-	int bch = lun_begin / dev->geo.luns_per_chnl;
-	int blun = lun_begin % dev->geo.luns_per_chnl;
+	int nr_chnls = nr_luns / dev->geo.nr_luns;
+	int nr_chnls_mod = nr_luns % dev->geo.nr_luns;
+	int bch = lun_begin / dev->geo.nr_luns;
+	int blun = lun_begin % dev->geo.nr_luns;
 	int lunid = 0;
 	int lun_balanced = 1;
 	int prev_nr_luns;
@@ -154,15 +173,15 @@ static struct nvm_tgt_dev *nvm_create_tgt_dev(struct nvm_dev *dev,
 	if (!luns)
 		goto err_luns;
 
-	prev_nr_luns = (luns_left > dev->geo.luns_per_chnl) ?
-					dev->geo.luns_per_chnl : luns_left;
+	prev_nr_luns = (luns_left > dev->geo.nr_luns) ?
+					dev->geo.nr_luns : luns_left;
 	for (i = 0; i < nr_chnls; i++) {
 		struct nvm_ch_map *ch_rmap = &dev_rmap->chnls[i + bch];
 		int *lun_roffs = ch_rmap->lun_offs;
 		struct nvm_ch_map *ch_map = &dev_map->chnls[i];
 		int *lun_offs;
-		int luns_in_chnl = (luns_left > dev->geo.luns_per_chnl) ?
-					dev->geo.luns_per_chnl : luns_left;
+		int luns_in_chnl = (luns_left > dev->geo.nr_luns) ?
+					dev->geo.nr_luns : luns_left;
 
 		if (lun_balanced && prev_nr_luns != luns_in_chnl)
 			lun_balanced = 0;
@@ -199,8 +218,9 @@ static struct nvm_tgt_dev *nvm_create_tgt_dev(struct nvm_dev *dev,
 	memcpy(&tgt_dev->geo, &dev->geo, sizeof(struct nvm_geo));
 	/* Target device only owns a portion of the physical device */
 	tgt_dev->geo.nr_chnls = nr_chnls;
-	tgt_dev->geo.nr_luns = nr_luns;
-	tgt_dev->geo.luns_per_chnl = (lun_balanced) ? prev_nr_luns : -1;
+	tgt_dev->geo.all_luns = nr_luns;
+	tgt_dev->geo.nr_luns = (lun_balanced) ? prev_nr_luns : -1;
+	tgt_dev->geo.op = op;
 	tgt_dev->total_secs = nr_luns * tgt_dev->geo.sec_per_lun;
 	tgt_dev->q = dev->q;
 	tgt_dev->map = dev_map;
@@ -226,27 +246,79 @@ static const struct block_device_operations nvm_fops = {
 	.owner		= THIS_MODULE,
 };
 
-static struct nvm_tgt_type *nvm_find_target_type(const char *name, int lock)
+static struct nvm_tgt_type *__nvm_find_target_type(const char *name)
 {
-	struct nvm_tgt_type *tmp, *tt = NULL;
+	struct nvm_tgt_type *tt;
 
-	if (lock)
-		down_write(&nvm_tgtt_lock);
+	list_for_each_entry(tt, &nvm_tgt_types, list)
+		if (!strcmp(name, tt->name))
+			return tt;
 
-	list_for_each_entry(tmp, &nvm_tgt_types, list)
-		if (!strcmp(name, tmp->name)) {
-			tt = tmp;
-			break;
-		}
+	return NULL;
+}
+
+static struct nvm_tgt_type *nvm_find_target_type(const char *name)
+{
+	struct nvm_tgt_type *tt;
+
+	down_write(&nvm_tgtt_lock);
+	tt = __nvm_find_target_type(name);
+	up_write(&nvm_tgtt_lock);
 
-	if (lock)
-		up_write(&nvm_tgtt_lock);
 	return tt;
 }
 
+static int nvm_config_check_luns(struct nvm_geo *geo, int lun_begin,
+				 int lun_end)
+{
+	if (lun_begin > lun_end || lun_end >= geo->all_luns) {
+		pr_err("nvm: lun out of bound (%u:%u > %u)\n",
+			lun_begin, lun_end, geo->all_luns - 1);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int __nvm_config_simple(struct nvm_dev *dev,
+			       struct nvm_ioctl_create_simple *s)
+{
+	struct nvm_geo *geo = &dev->geo;
+
+	if (s->lun_begin == -1 && s->lun_end == -1) {
+		s->lun_begin = 0;
+		s->lun_end = geo->all_luns - 1;
+	}
+
+	return nvm_config_check_luns(geo, s->lun_begin, s->lun_end);
+}
+
+static int __nvm_config_extended(struct nvm_dev *dev,
+				 struct nvm_ioctl_create_extended *e)
+{
+	struct nvm_geo *geo = &dev->geo;
+
+	if (e->lun_begin == 0xFFFF && e->lun_end == 0xFFFF) {
+		e->lun_begin = 0;
+		e->lun_end = dev->geo.all_luns - 1;
+	}
+
+	/* op not set falls into target's default */
+	if (e->op == 0xFFFF)
+		e->op = NVM_TARGET_DEFAULT_OP;
+
+	if (e->op < NVM_TARGET_MIN_OP ||
+	    e->op > NVM_TARGET_MAX_OP) {
+		pr_err("nvm: invalid over provisioning value\n");
+		return -EINVAL;
+	}
+
+	return nvm_config_check_luns(geo, e->lun_begin, e->lun_end);
+}
+
 static int nvm_create_tgt(struct nvm_dev *dev, struct nvm_ioctl_create *create)
 {
-	struct nvm_ioctl_create_simple *s = &create->conf.s;
+	struct nvm_ioctl_create_extended e;
 	struct request_queue *tqueue;
 	struct gendisk *tdisk;
 	struct nvm_tgt_type *tt;
@@ -255,22 +327,41 @@ static int nvm_create_tgt(struct nvm_dev *dev, struct nvm_ioctl_create *create)
 	void *targetdata;
 	int ret;
 
-	tt = nvm_find_target_type(create->tgttype, 1);
+	switch (create->conf.type) {
+	case NVM_CONFIG_TYPE_SIMPLE:
+		ret = __nvm_config_simple(dev, &create->conf.s);
+		if (ret)
+			return ret;
+
+		e.lun_begin = create->conf.s.lun_begin;
+		e.lun_end = create->conf.s.lun_end;
+		e.op = NVM_TARGET_DEFAULT_OP;
+		break;
+	case NVM_CONFIG_TYPE_EXTENDED:
+		ret = __nvm_config_extended(dev, &create->conf.e);
+		if (ret)
+			return ret;
+
+		e = create->conf.e;
+		break;
+	default:
+		pr_err("nvm: config type not valid\n");
+		return -EINVAL;
+	}
+
+	tt = nvm_find_target_type(create->tgttype);
 	if (!tt) {
 		pr_err("nvm: target type %s not found\n", create->tgttype);
 		return -EINVAL;
 	}
 
-	mutex_lock(&dev->mlock);
-	t = nvm_find_target(dev, create->tgtname);
-	if (t) {
-		pr_err("nvm: target name already exists.\n");
-		mutex_unlock(&dev->mlock);
+	if (nvm_target_exists(create->tgtname)) {
+		pr_err("nvm: target name already exists (%s)\n",
+							create->tgtname);
 		return -EINVAL;
 	}
-	mutex_unlock(&dev->mlock);
 
-	ret = nvm_reserve_luns(dev, s->lun_begin, s->lun_end);
+	ret = nvm_reserve_luns(dev, e.lun_begin, e.lun_end);
 	if (ret)
 		return ret;
 
@@ -280,7 +371,7 @@ static int nvm_create_tgt(struct nvm_dev *dev, struct nvm_ioctl_create *create)
 		goto err_reserve;
 	}
 
-	tgt_dev = nvm_create_tgt_dev(dev, s->lun_begin, s->lun_end);
+	tgt_dev = nvm_create_tgt_dev(dev, e.lun_begin, e.lun_end, e.op);
 	if (!tgt_dev) {
 		pr_err("nvm: could not create target device\n");
 		ret = -ENOMEM;
@@ -350,7 +441,7 @@ err_dev:
 err_t:
 	kfree(t);
 err_reserve:
-	nvm_release_luns_err(dev, s->lun_begin, s->lun_end);
+	nvm_release_luns_err(dev, e.lun_begin, e.lun_end);
 	return ret;
 }
 
@@ -420,7 +511,7 @@ static int nvm_register_map(struct nvm_dev *dev)
 	for (i = 0; i < dev->geo.nr_chnls; i++) {
 		struct nvm_ch_map *ch_rmap;
 		int *lun_roffs;
-		int luns_in_chnl = dev->geo.luns_per_chnl;
+		int luns_in_chnl = dev->geo.nr_luns;
 
 		ch_rmap = &rmap->chnls[i];
 
@@ -524,41 +615,12 @@ static void nvm_rq_dev_to_tgt(struct nvm_tgt_dev *tgt_dev, struct nvm_rq *rqd)
 	nvm_ppa_dev_to_tgt(tgt_dev, rqd->ppa_list, rqd->nr_ppas);
 }
 
-void nvm_part_to_tgt(struct nvm_dev *dev, sector_t *entries,
-		     int len)
-{
-	struct nvm_geo *geo = &dev->geo;
-	struct nvm_dev_map *dev_rmap = dev->rmap;
-	u64 i;
-
-	for (i = 0; i < len; i++) {
-		struct nvm_ch_map *ch_rmap;
-		int *lun_roffs;
-		struct ppa_addr gaddr;
-		u64 pba = le64_to_cpu(entries[i]);
-		u64 diff;
-
-		if (!pba)
-			continue;
-
-		gaddr = linear_to_generic_addr(geo, pba);
-		ch_rmap = &dev_rmap->chnls[gaddr.g.ch];
-		lun_roffs = ch_rmap->lun_offs;
-
-		diff = ((ch_rmap->ch_off * geo->luns_per_chnl) +
-				(lun_roffs[gaddr.g.lun])) * geo->sec_per_lun;
-
-		entries[i] -= cpu_to_le64(diff);
-	}
-}
-EXPORT_SYMBOL(nvm_part_to_tgt);
-
 int nvm_register_tgt_type(struct nvm_tgt_type *tt)
 {
 	int ret = 0;
 
 	down_write(&nvm_tgtt_lock);
-	if (nvm_find_target_type(tt->name, 0))
+	if (__nvm_find_target_type(tt->name))
 		ret = -EEXIST;
 	else
 		list_add(&tt->list, &nvm_tgt_types);
@@ -726,112 +788,6 @@ int nvm_submit_io_sync(struct nvm_tgt_dev *tgt_dev, struct nvm_rq *rqd)
 }
 EXPORT_SYMBOL(nvm_submit_io_sync);
 
-int nvm_erase_sync(struct nvm_tgt_dev *tgt_dev, struct ppa_addr *ppas,
-								int nr_ppas)
-{
-	struct nvm_geo *geo = &tgt_dev->geo;
-	struct nvm_rq rqd;
-	int ret;
-
-	memset(&rqd, 0, sizeof(struct nvm_rq));
-
-	rqd.opcode = NVM_OP_ERASE;
-	rqd.flags = geo->plane_mode >> 1;
-
-	ret = nvm_set_rqd_ppalist(tgt_dev, &rqd, ppas, nr_ppas);
-	if (ret)
-		return ret;
-
-	ret = nvm_submit_io_sync(tgt_dev, &rqd);
-	if (ret) {
-		pr_err("rrpr: erase I/O submission failed: %d\n", ret);
-		goto free_ppa_list;
-	}
-
-free_ppa_list:
-	nvm_free_rqd_ppalist(tgt_dev, &rqd);
-
-	return ret;
-}
-EXPORT_SYMBOL(nvm_erase_sync);
-
-int nvm_get_l2p_tbl(struct nvm_tgt_dev *tgt_dev, u64 slba, u32 nlb,
-		    nvm_l2p_update_fn *update_l2p, void *priv)
-{
-	struct nvm_dev *dev = tgt_dev->parent;
-
-	if (!dev->ops->get_l2p_tbl)
-		return 0;
-
-	return dev->ops->get_l2p_tbl(dev, slba, nlb, update_l2p, priv);
-}
-EXPORT_SYMBOL(nvm_get_l2p_tbl);
-
-int nvm_get_area(struct nvm_tgt_dev *tgt_dev, sector_t *lba, sector_t len)
-{
-	struct nvm_dev *dev = tgt_dev->parent;
-	struct nvm_geo *geo = &dev->geo;
-	struct nvm_area *area, *prev, *next;
-	sector_t begin = 0;
-	sector_t max_sectors = (geo->sec_size * dev->total_secs) >> 9;
-
-	if (len > max_sectors)
-		return -EINVAL;
-
-	area = kmalloc(sizeof(struct nvm_area), GFP_KERNEL);
-	if (!area)
-		return -ENOMEM;
-
-	prev = NULL;
-
-	spin_lock(&dev->lock);
-	list_for_each_entry(next, &dev->area_list, list) {
-		if (begin + len > next->begin) {
-			begin = next->end;
-			prev = next;
-			continue;
-		}
-		break;
-	}
-
-	if ((begin + len) > max_sectors) {
-		spin_unlock(&dev->lock);
-		kfree(area);
-		return -EINVAL;
-	}
-
-	area->begin = *lba = begin;
-	area->end = begin + len;
-
-	if (prev) /* insert into sorted order */
-		list_add(&area->list, &prev->list);
-	else
-		list_add(&area->list, &dev->area_list);
-	spin_unlock(&dev->lock);
-
-	return 0;
-}
-EXPORT_SYMBOL(nvm_get_area);
-
-void nvm_put_area(struct nvm_tgt_dev *tgt_dev, sector_t begin)
-{
-	struct nvm_dev *dev = tgt_dev->parent;
-	struct nvm_area *area;
-
-	spin_lock(&dev->lock);
-	list_for_each_entry(area, &dev->area_list, list) {
-		if (area->begin != begin)
-			continue;
-
-		list_del(&area->list);
-		spin_unlock(&dev->lock);
-		kfree(area);
-		return;
-	}
-	spin_unlock(&dev->lock);
-}
-EXPORT_SYMBOL(nvm_put_area);
-
 void nvm_end_io(struct nvm_rq *rqd)
 {
 	struct nvm_tgt_dev *tgt_dev = rqd->dev;
@@ -858,10 +814,10 @@ int nvm_bb_tbl_fold(struct nvm_dev *dev, u8 *blks, int nr_blks)
 	struct nvm_geo *geo = &dev->geo;
 	int blk, offset, pl, blktype;
 
-	if (nr_blks != geo->blks_per_lun * geo->plane_mode)
+	if (nr_blks != geo->nr_chks * geo->plane_mode)
 		return -EINVAL;
 
-	for (blk = 0; blk < geo->blks_per_lun; blk++) {
+	for (blk = 0; blk < geo->nr_chks; blk++) {
 		offset = blk * geo->plane_mode;
 		blktype = blks[offset];
 
@@ -877,7 +833,7 @@ int nvm_bb_tbl_fold(struct nvm_dev *dev, u8 *blks, int nr_blks)
 		blks[blk] = blktype;
 	}
 
-	return geo->blks_per_lun;
+	return geo->nr_chks;
 }
 EXPORT_SYMBOL(nvm_bb_tbl_fold);
 
@@ -892,53 +848,6 @@ int nvm_get_tgt_bb_tbl(struct nvm_tgt_dev *tgt_dev, struct ppa_addr ppa,
 }
 EXPORT_SYMBOL(nvm_get_tgt_bb_tbl);
 
-static int nvm_init_slc_tbl(struct nvm_dev *dev, struct nvm_id_group *grp)
-{
-	struct nvm_geo *geo = &dev->geo;
-	int i;
-
-	dev->lps_per_blk = geo->pgs_per_blk;
-	dev->lptbl = kcalloc(dev->lps_per_blk, sizeof(int), GFP_KERNEL);
-	if (!dev->lptbl)
-		return -ENOMEM;
-
-	/* Just a linear array */
-	for (i = 0; i < dev->lps_per_blk; i++)
-		dev->lptbl[i] = i;
-
-	return 0;
-}
-
-static int nvm_init_mlc_tbl(struct nvm_dev *dev, struct nvm_id_group *grp)
-{
-	int i, p;
-	struct nvm_id_lp_mlc *mlc = &grp->lptbl.mlc;
-
-	if (!mlc->num_pairs)
-		return 0;
-
-	dev->lps_per_blk = mlc->num_pairs;
-	dev->lptbl = kcalloc(dev->lps_per_blk, sizeof(int), GFP_KERNEL);
-	if (!dev->lptbl)
-		return -ENOMEM;
-
-	/* The lower page table encoding consists of a list of bytes, where each
-	 * has a lower and an upper half. The first half byte maintains the
-	 * increment value and every value after is an offset added to the
-	 * previous incrementation value
-	 */
-	dev->lptbl[0] = mlc->pairs[0] & 0xF;
-	for (i = 1; i < dev->lps_per_blk; i++) {
-		p = mlc->pairs[i >> 1];
-		if (i & 0x1) /* upper */
-			dev->lptbl[i] = dev->lptbl[i - 1] + ((p & 0xF0) >> 4);
-		else /* lower */
-			dev->lptbl[i] = dev->lptbl[i - 1] + (p & 0xF);
-	}
-
-	return 0;
-}
-
 static int nvm_core_init(struct nvm_dev *dev)
 {
 	struct nvm_id *id = &dev->identity;
@@ -946,66 +855,44 @@ static int nvm_core_init(struct nvm_dev *dev)
 	struct nvm_geo *geo = &dev->geo;
 	int ret;
 
+	memcpy(&geo->ppaf, &id->ppaf, sizeof(struct nvm_addr_format));
+
+	if (grp->mtype != 0) {
+		pr_err("nvm: memory type not supported\n");
+		return -EINVAL;
+	}
+
 	/* Whole device values */
 	geo->nr_chnls = grp->num_ch;
-	geo->luns_per_chnl = grp->num_lun;
-
-	/* Generic device values */
-	geo->pgs_per_blk = grp->num_pg;
-	geo->blks_per_lun = grp->num_blk;
-	geo->nr_planes = grp->num_pln;
-	geo->fpg_size = grp->fpg_sz;
-	geo->pfpg_size = grp->fpg_sz * grp->num_pln;
+	geo->nr_luns = grp->num_lun;
+
+	/* Generic device geometry values */
+	geo->ws_min = grp->ws_min;
+	geo->ws_opt = grp->ws_opt;
+	geo->ws_seq = grp->ws_seq;
+	geo->ws_per_chk = grp->ws_per_chk;
+	geo->nr_chks = grp->num_chk;
 	geo->sec_size = grp->csecs;
 	geo->oob_size = grp->sos;
-	geo->sec_per_pg = grp->fpg_sz / grp->csecs;
 	geo->mccap = grp->mccap;
-	memcpy(&geo->ppaf, &id->ppaf, sizeof(struct nvm_addr_format));
-
-	geo->plane_mode = NVM_PLANE_SINGLE;
 	geo->max_rq_size = dev->ops->max_phys_sect * geo->sec_size;
 
-	if (grp->mpos & 0x020202)
-		geo->plane_mode = NVM_PLANE_DOUBLE;
-	if (grp->mpos & 0x040404)
-		geo->plane_mode = NVM_PLANE_QUAD;
+	geo->sec_per_chk = grp->clba;
+	geo->sec_per_lun = geo->sec_per_chk * geo->nr_chks;
+	geo->all_luns = geo->nr_luns * geo->nr_chnls;
 
-	if (grp->mtype != 0) {
-		pr_err("nvm: memory type not supported\n");
-		return -EINVAL;
-	}
-
-	/* calculated values */
+	/* 1.2 spec device geometry values */
+	geo->plane_mode = 1 << geo->ws_seq;
+	geo->nr_planes = geo->ws_opt / geo->ws_min;
+	geo->sec_per_pg = geo->ws_min;
 	geo->sec_per_pl = geo->sec_per_pg * geo->nr_planes;
-	geo->sec_per_blk = geo->sec_per_pl * geo->pgs_per_blk;
-	geo->sec_per_lun = geo->sec_per_blk * geo->blks_per_lun;
-	geo->nr_luns = geo->luns_per_chnl * geo->nr_chnls;
 
-	dev->total_secs = geo->nr_luns * geo->sec_per_lun;
-	dev->lun_map = kcalloc(BITS_TO_LONGS(geo->nr_luns),
+	dev->total_secs = geo->all_luns * geo->sec_per_lun;
+	dev->lun_map = kcalloc(BITS_TO_LONGS(geo->all_luns),
 					sizeof(unsigned long), GFP_KERNEL);
 	if (!dev->lun_map)
 		return -ENOMEM;
 
-	switch (grp->fmtype) {
-	case NVM_ID_FMTYPE_SLC:
-		if (nvm_init_slc_tbl(dev, grp)) {
-			ret = -ENOMEM;
-			goto err_fmtype;
-		}
-		break;
-	case NVM_ID_FMTYPE_MLC:
-		if (nvm_init_mlc_tbl(dev, grp)) {
-			ret = -ENOMEM;
-			goto err_fmtype;
-		}
-		break;
-	default:
-		pr_err("nvm: flash type not supported\n");
-		ret = -EINVAL;
-		goto err_fmtype;
-	}
-
 	INIT_LIST_HEAD(&dev->area_list);
 	INIT_LIST_HEAD(&dev->targets);
 	mutex_init(&dev->mlock);
@@ -1031,7 +918,6 @@ static void nvm_free(struct nvm_dev *dev)
 		dev->ops->destroy_dma_pool(dev->dma_pool);
 
 	nvm_unregister_map(dev);
-	kfree(dev->lptbl);
 	kfree(dev->lun_map);
 	kfree(dev);
 }
@@ -1062,8 +948,8 @@ static int nvm_init(struct nvm_dev *dev)
 
 	pr_info("nvm: registered %s [%u/%u/%u/%u/%u/%u]\n",
 			dev->name, geo->sec_per_pg, geo->nr_planes,
-			geo->pgs_per_blk, geo->blks_per_lun,
-			geo->nr_luns, geo->nr_chnls);
+			geo->ws_per_chk, geo->nr_chks,
+			geo->all_luns, geo->nr_chnls);
 	return 0;
 err:
 	pr_err("nvm: failed to initialize nvm\n");
@@ -1135,7 +1021,6 @@ EXPORT_SYMBOL(nvm_unregister);
 static int __nvm_configure_create(struct nvm_ioctl_create *create)
 {
 	struct nvm_dev *dev;
-	struct nvm_ioctl_create_simple *s;
 
 	down_write(&nvm_lock);
 	dev = nvm_find_nvm_dev(create->dev);
@@ -1146,23 +1031,6 @@ static int __nvm_configure_create(struct nvm_ioctl_create *create)
 		return -EINVAL;
 	}
 
-	if (create->conf.type != NVM_CONFIG_TYPE_SIMPLE) {
-		pr_err("nvm: config type not valid\n");
-		return -EINVAL;
-	}
-	s = &create->conf.s;
-
-	if (s->lun_begin == -1 && s->lun_end == -1) {
-		s->lun_begin = 0;
-		s->lun_end = dev->geo.nr_luns - 1;
-	}
-
-	if (s->lun_begin > s->lun_end || s->lun_end >= dev->geo.nr_luns) {
-		pr_err("nvm: lun out of bound (%u:%u > %u)\n",
-			s->lun_begin, s->lun_end, dev->geo.nr_luns - 1);
-		return -EINVAL;
-	}
-
 	return nvm_create_tgt(dev, create);
 }
 
@@ -1262,6 +1130,12 @@ static long nvm_ioctl_dev_create(struct file *file, void __user *arg)
 	if (copy_from_user(&create, arg, sizeof(struct nvm_ioctl_create)))
 		return -EFAULT;
 
+	if (create.conf.type == NVM_CONFIG_TYPE_EXTENDED &&
+	    create.conf.e.rsv != 0) {
+		pr_err("nvm: reserved config field in use\n");
+		return -EINVAL;
+	}
+
 	create.dev[DISK_NAME_LEN - 1] = '\0';
 	create.tgttype[NVM_TTYPE_NAME_MAX - 1] = '\0';
 	create.tgtname[DISK_NAME_LEN - 1] = '\0';

drivers/lightnvm/pblk-cache.c

@@ -19,12 +19,16 @@
 
 int pblk_write_to_cache(struct pblk *pblk, struct bio *bio, unsigned long flags)
 {
+	struct request_queue *q = pblk->dev->q;
 	struct pblk_w_ctx w_ctx;
 	sector_t lba = pblk_get_lba(bio);
+	unsigned long start_time = jiffies;
 	unsigned int bpos, pos;
 	int nr_entries = pblk_get_secs(bio);
 	int i, ret;
 
+	generic_start_io_acct(q, WRITE, bio_sectors(bio), &pblk->disk->part0);
+
 	/* Update the write buffer head (mem) with the entries that we can
 	 * write. The write in itself cannot fail, so there is no need to
 	 * rollback from here on.
@@ -67,6 +71,7 @@ retry:
 	pblk_rl_inserted(&pblk->rl, nr_entries);
 
 out:
+	generic_end_io_acct(q, WRITE, &pblk->disk->part0, start_time);
 	pblk_write_should_kick(pblk);
 	return ret;
 }

drivers/lightnvm/pblk-core.c

@@ -32,8 +32,8 @@ static void pblk_line_mark_bb(struct work_struct *work)
 		struct pblk_line *line;
 		int pos;
 
-		line = &pblk->lines[pblk_dev_ppa_to_line(*ppa)];
-		pos = pblk_dev_ppa_to_pos(&dev->geo, *ppa);
+		line = &pblk->lines[pblk_ppa_to_line(*ppa)];
+		pos = pblk_ppa_to_pos(&dev->geo, *ppa);
 
 		pr_err("pblk: failed to mark bb, line:%d, pos:%d\n",
 				line->id, pos);
@@ -48,7 +48,7 @@ static void pblk_mark_bb(struct pblk *pblk, struct pblk_line *line,
 {
 	struct nvm_tgt_dev *dev = pblk->dev;
 	struct nvm_geo *geo = &dev->geo;
-	int pos = pblk_dev_ppa_to_pos(geo, *ppa);
+	int pos = pblk_ppa_to_pos(geo, *ppa);
 
 	pr_debug("pblk: erase failed: line:%d, pos:%d\n", line->id, pos);
 	atomic_long_inc(&pblk->erase_failed);
@@ -66,7 +66,7 @@ static void __pblk_end_io_erase(struct pblk *pblk, struct nvm_rq *rqd)
 {
 	struct pblk_line *line;
 
-	line = &pblk->lines[pblk_dev_ppa_to_line(rqd->ppa_addr)];
+	line = &pblk->lines[pblk_ppa_to_line(rqd->ppa_addr)];
 	atomic_dec(&line->left_seblks);
 
 	if (rqd->error) {
@@ -144,7 +144,7 @@ void pblk_map_invalidate(struct pblk *pblk, struct ppa_addr ppa)
 	BUG_ON(pblk_ppa_empty(ppa));
 #endif
 
-	line_id = pblk_tgt_ppa_to_line(ppa);
+	line_id = pblk_ppa_to_line(ppa);
 	line = &pblk->lines[line_id];
 	paddr = pblk_dev_ppa_to_line_addr(pblk, ppa);
 
@@ -650,7 +650,7 @@ next_rq:
 	} else {
 		for (i = 0; i < rqd.nr_ppas; ) {
 			struct ppa_addr ppa = addr_to_gen_ppa(pblk, paddr, id);
-			int pos = pblk_dev_ppa_to_pos(geo, ppa);
+			int pos = pblk_ppa_to_pos(geo, ppa);
 			int read_type = PBLK_READ_RANDOM;
 
 			if (pblk_io_aligned(pblk, rq_ppas))
@@ -668,7 +668,7 @@ next_rq:
 				}
 
 				ppa = addr_to_gen_ppa(pblk, paddr, id);
-				pos = pblk_dev_ppa_to_pos(geo, ppa);
+				pos = pblk_ppa_to_pos(geo, ppa);
 			}
 
 			if (pblk_boundary_paddr_checks(pblk, paddr + min)) {
@@ -742,7 +742,7 @@ static int pblk_line_submit_smeta_io(struct pblk *pblk, struct pblk_line *line,
 		cmd_op = NVM_OP_PWRITE;
 		flags = pblk_set_progr_mode(pblk, PBLK_WRITE);
 		lba_list = emeta_to_lbas(pblk, line->emeta->buf);
-	} else if (dir == PBLK_READ) {
+	} else if (dir == PBLK_READ_RECOV || dir == PBLK_READ) {
 		bio_op = REQ_OP_READ;
 		cmd_op = NVM_OP_PREAD;
 		flags = pblk_set_read_mode(pblk, PBLK_READ_SEQUENTIAL);
@@ -802,7 +802,7 @@ static int pblk_line_submit_smeta_io(struct pblk *pblk, struct pblk_line *line,
 	if (rqd.error) {
 		if (dir == PBLK_WRITE)
 			pblk_log_write_err(pblk, &rqd);
-		else
+		else if (dir == PBLK_READ)
 			pblk_log_read_err(pblk, &rqd);
 	}
 
@@ -816,7 +816,7 @@ int pblk_line_read_smeta(struct pblk *pblk, struct pblk_line *line)
 {
 	u64 bpaddr = pblk_line_smeta_start(pblk, line);
 
-	return pblk_line_submit_smeta_io(pblk, line, bpaddr, PBLK_READ);
+	return pblk_line_submit_smeta_io(pblk, line, bpaddr, PBLK_READ_RECOV);
 }
 
 int pblk_line_read_emeta(struct pblk *pblk, struct pblk_line *line,
@@ -854,8 +854,8 @@ static int pblk_blk_erase_sync(struct pblk *pblk, struct ppa_addr ppa)
 		struct nvm_geo *geo = &dev->geo;
 
 		pr_err("pblk: could not sync erase line:%d,blk:%d\n",
-					pblk_dev_ppa_to_line(ppa),
-					pblk_dev_ppa_to_pos(geo, ppa));
+					pblk_ppa_to_line(ppa),
+					pblk_ppa_to_pos(geo, ppa));
 
 		rqd.error = ret;
 		goto out;
@@ -979,7 +979,7 @@ static int pblk_line_init_metadata(struct pblk *pblk, struct pblk_line *line,
 
 	/* Start metadata */
 	smeta_buf->seq_nr = cpu_to_le64(line->seq_nr);
-	smeta_buf->window_wr_lun = cpu_to_le32(geo->nr_luns);
+	smeta_buf->window_wr_lun = cpu_to_le32(geo->all_luns);
 
 	/* Fill metadata among lines */
 	if (cur) {
@@ -1032,7 +1032,7 @@ static int pblk_line_init_bb(struct pblk *pblk, struct pblk_line *line,
 							lm->sec_per_line);
 		bitmap_or(line->map_bitmap, line->map_bitmap, l_mg->bb_aux,
 							lm->sec_per_line);
-		line->sec_in_line -= geo->sec_per_blk;
+		line->sec_in_line -= geo->sec_per_chk;
 		if (bit >= lm->emeta_bb)
 			nr_bb++;
 	}
@@ -1145,7 +1145,7 @@ int pblk_line_recov_alloc(struct pblk *pblk, struct pblk_line *line)
 	}
 	spin_unlock(&l_mg->free_lock);
 
-	pblk_rl_free_lines_dec(&pblk->rl, line);
+	pblk_rl_free_lines_dec(&pblk->rl, line, true);
 
 	if (!pblk_line_init_bb(pblk, line, 0)) {
 		list_add(&line->list, &l_mg->free_list);
@@ -1233,7 +1233,7 @@ retry:
 	l_mg->data_line = retry_line;
 	spin_unlock(&l_mg->free_lock);
 
-	pblk_rl_free_lines_dec(&pblk->rl, retry_line);
+	pblk_rl_free_lines_dec(&pblk->rl, line, false);
 
 	if (pblk_line_erase(pblk, retry_line))
 		goto retry;
@@ -1252,7 +1252,6 @@ struct pblk_line *pblk_line_get_first_data(struct pblk *pblk)
 {
 	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
 	struct pblk_line *line;
-	int is_next = 0;
 
 	spin_lock(&l_mg->free_lock);
 	line = pblk_line_get(pblk);
@@ -1280,7 +1279,6 @@ struct pblk_line *pblk_line_get_first_data(struct pblk *pblk)
 	} else {
 		l_mg->data_next->seq_nr = l_mg->d_seq_nr++;
 		l_mg->data_next->type = PBLK_LINETYPE_DATA;
-		is_next = 1;
 	}
 	spin_unlock(&l_mg->free_lock);
 
@@ -1290,10 +1288,6 @@ struct pblk_line *pblk_line_get_first_data(struct pblk *pblk)
 			return NULL;
 	}
 
-	pblk_rl_free_lines_dec(&pblk->rl, line);
-	if (is_next)
-		pblk_rl_free_lines_dec(&pblk->rl, l_mg->data_next);
-
 retry_setup:
 	if (!pblk_line_init_metadata(pblk, line, NULL)) {
 		line = pblk_line_retry(pblk, line);
@@ -1311,6 +1305,8 @@ retry_setup:
 		goto retry_setup;
 	}
 
+	pblk_rl_free_lines_dec(&pblk->rl, line, true);
+
 	return line;
 }
 
@@ -1395,7 +1391,6 @@ struct pblk_line *pblk_line_replace_data(struct pblk *pblk)
 	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
 	struct pblk_line *cur, *new = NULL;
 	unsigned int left_seblks;
-	int is_next = 0;
 
 	cur = l_mg->data_line;
 	new = l_mg->data_next;
@@ -1444,6 +1439,8 @@ retry_setup:
 		goto retry_setup;
 	}
 
+	pblk_rl_free_lines_dec(&pblk->rl, new, true);
+
 	/* Allocate next line for preparation */
 	spin_lock(&l_mg->free_lock);
 	l_mg->data_next = pblk_line_get(pblk);
@@ -1457,13 +1454,9 @@ retry_setup:
 	} else {
 		l_mg->data_next->seq_nr = l_mg->d_seq_nr++;
 		l_mg->data_next->type = PBLK_LINETYPE_DATA;
-		is_next = 1;
 	}
 	spin_unlock(&l_mg->free_lock);
 
-	if (is_next)
-		pblk_rl_free_lines_dec(&pblk->rl, l_mg->data_next);
-
 out:
 	return new;
 }
@@ -1561,8 +1554,8 @@ int pblk_blk_erase_async(struct pblk *pblk, struct ppa_addr ppa)
 		struct nvm_geo *geo = &dev->geo;
 
 		pr_err("pblk: could not async erase line:%d,blk:%d\n",
-					pblk_dev_ppa_to_line(ppa),
-					pblk_dev_ppa_to_pos(geo, ppa));
+					pblk_ppa_to_line(ppa),
+					pblk_ppa_to_pos(geo, ppa));
 	}
 
 	return err;
@@ -1746,7 +1739,7 @@ void pblk_up_rq(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 nr_luns = geo->nr_luns;
+	int nr_luns = geo->all_luns;
 	int bit = -1;
 
 	while ((bit = find_next_bit(lun_bitmap, nr_luns, bit + 1)) < nr_luns) {
@@ -1884,7 +1877,7 @@ void pblk_lookup_l2p_seq(struct pblk *pblk, struct ppa_addr *ppas,
 
 		/* If the L2P entry maps to a line, the reference is valid */
 		if (!pblk_ppa_empty(ppa) && !pblk_addr_in_cache(ppa)) {
-			int line_id = pblk_dev_ppa_to_line(ppa);
+			int line_id = pblk_ppa_to_line(ppa);
 			struct pblk_line *line = &pblk->lines[line_id];
 
 			kref_get(&line->ref);

drivers/lightnvm/pblk-gc.c

@@ -169,7 +169,14 @@ static void pblk_gc_line_prepare_ws(struct work_struct *work)
 	 * the line untouched. TODO: Implement a recovery routine that scans and
 	 * moves all sectors on the line.
 	 */
-	lba_list = pblk_recov_get_lba_list(pblk, emeta_buf);
+
+	ret = pblk_recov_check_emeta(pblk, emeta_buf);
+	if (ret) {
+		pr_err("pblk: inconsistent emeta (line %d)\n", line->id);
+		goto fail_free_emeta;
+	}
+
+	lba_list = emeta_to_lbas(pblk, emeta_buf);
 	if (!lba_list) {
 		pr_err("pblk: could not interpret emeta (line %d)\n", line->id);
 		goto fail_free_emeta;
@@ -519,22 +526,12 @@ void pblk_gc_should_start(struct pblk *pblk)
 	}
 }
 
-/*
- * If flush_wq == 1 then no lock should be held by the caller since
- * flush_workqueue can sleep
- */
-static void pblk_gc_stop(struct pblk *pblk, int flush_wq)
-{
-	pblk->gc.gc_active = 0;
-	pr_debug("pblk: gc stop\n");
-}
-
 void pblk_gc_should_stop(struct pblk *pblk)
 {
 	struct pblk_gc *gc = &pblk->gc;
 
 	if (gc->gc_active && !gc->gc_forced)
-		pblk_gc_stop(pblk, 0);
+		gc->gc_active = 0;
 }
 
 void pblk_gc_should_kick(struct pblk *pblk)
@@ -660,7 +657,7 @@ void pblk_gc_exit(struct pblk *pblk)
 
 	gc->gc_enabled = 0;
 	del_timer_sync(&gc->gc_timer);
-	pblk_gc_stop(pblk, 1);
+	gc->gc_active = 0;
 
 	if (gc->gc_ts)
 		kthread_stop(gc->gc_ts);

drivers/lightnvm/pblk-init.c

@@ -169,8 +169,8 @@ static int pblk_set_ppaf(struct pblk *pblk)
 	}
 	ppaf.ch_len = power_len;
 
-	power_len = get_count_order(geo->luns_per_chnl);
-	if (1 << power_len != geo->luns_per_chnl) {
+	power_len = get_count_order(geo->nr_luns);
+	if (1 << power_len != geo->nr_luns) {
 		pr_err("pblk: supports only power-of-two LUN config.\n");
 		return -EINVAL;
 	}
@@ -254,7 +254,7 @@ static int pblk_core_init(struct pblk *pblk)
 	struct nvm_geo *geo = &dev->geo;
 
 	pblk->pgs_in_buffer = NVM_MEM_PAGE_WRITE * geo->sec_per_pg *
-						geo->nr_planes * geo->nr_luns;
+						geo->nr_planes * geo->all_luns;
 
 	if (pblk_init_global_caches(pblk))
 		return -ENOMEM;
@@ -270,21 +270,22 @@ static int pblk_core_init(struct pblk *pblk)
 	if (!pblk->gen_ws_pool)
 		goto free_page_bio_pool;
 
-	pblk->rec_pool = mempool_create_slab_pool(geo->nr_luns, pblk_rec_cache);
+	pblk->rec_pool = mempool_create_slab_pool(geo->all_luns,
+							pblk_rec_cache);
 	if (!pblk->rec_pool)
 		goto free_gen_ws_pool;
 
-	pblk->r_rq_pool = mempool_create_slab_pool(geo->nr_luns,
+	pblk->r_rq_pool = mempool_create_slab_pool(geo->all_luns,
 							pblk_g_rq_cache);
 	if (!pblk->r_rq_pool)
 		goto free_rec_pool;
 
-	pblk->e_rq_pool = mempool_create_slab_pool(geo->nr_luns,
+	pblk->e_rq_pool = mempool_create_slab_pool(geo->all_luns,
 							pblk_g_rq_cache);
 	if (!pblk->e_rq_pool)
 		goto free_r_rq_pool;
 
-	pblk->w_rq_pool = mempool_create_slab_pool(geo->nr_luns,
+	pblk->w_rq_pool = mempool_create_slab_pool(geo->all_luns,
 							pblk_w_rq_cache);
 	if (!pblk->w_rq_pool)
 		goto free_e_rq_pool;
@@ -354,6 +355,8 @@ static void pblk_core_free(struct pblk *pblk)
 	mempool_destroy(pblk->e_rq_pool);
 	mempool_destroy(pblk->w_rq_pool);
 
+	pblk_rwb_free(pblk);
+
 	pblk_free_global_caches(pblk);
 }
 
@@ -409,7 +412,7 @@ static int pblk_bb_discovery(struct nvm_tgt_dev *dev, struct pblk_lun *rlun)
 	u8 *blks;
 	int nr_blks, ret;
 
-	nr_blks = geo->blks_per_lun * geo->plane_mode;
+	nr_blks = geo->nr_chks * geo->plane_mode;
 	blks = kmalloc(nr_blks, GFP_KERNEL);
 	if (!blks)
 		return -ENOMEM;
@@ -482,20 +485,21 @@ static int pblk_luns_init(struct pblk *pblk, struct ppa_addr *luns)
 	int i, ret;
 
 	/* TODO: Implement unbalanced LUN support */
-	if (geo->luns_per_chnl < 0) {
+	if (geo->nr_luns < 0) {
 		pr_err("pblk: unbalanced LUN config.\n");
 		return -EINVAL;
 	}
 
-	pblk->luns = kcalloc(geo->nr_luns, sizeof(struct pblk_lun), GFP_KERNEL);
+	pblk->luns = kcalloc(geo->all_luns, sizeof(struct pblk_lun),
+								GFP_KERNEL);
 	if (!pblk->luns)
 		return -ENOMEM;
 
-	for (i = 0; i < geo->nr_luns; i++) {
+	for (i = 0; i < geo->all_luns; i++) {
 		/* Stripe across channels */
 		int ch = i % geo->nr_chnls;
 		int lun_raw = i / geo->nr_chnls;
-		int lunid = lun_raw + ch * geo->luns_per_chnl;
+		int lunid = lun_raw + ch * geo->nr_luns;
 
 		rlun = &pblk->luns[i];
 		rlun->bppa = luns[lunid];
@@ -577,22 +581,37 @@ static unsigned int calc_emeta_len(struct pblk *pblk)
 static void pblk_set_provision(struct pblk *pblk, long nr_free_blks)
 {
 	struct nvm_tgt_dev *dev = pblk->dev;
+	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
+	struct pblk_line_meta *lm = &pblk->lm;
 	struct nvm_geo *geo = &dev->geo;
 	sector_t provisioned;
+	int sec_meta, blk_meta;
 
-	pblk->over_pct = 20;
+	if (geo->op == NVM_TARGET_DEFAULT_OP)
+		pblk->op = PBLK_DEFAULT_OP;
+	else
+		pblk->op = geo->op;
 
 	provisioned = nr_free_blks;
-	provisioned *= (100 - pblk->over_pct);
+	provisioned *= (100 - pblk->op);
 	sector_div(provisioned, 100);
 
+	pblk->op_blks = nr_free_blks - provisioned;
+
 	/* Internally pblk manages all free blocks, but all calculations based
 	 * on user capacity consider only provisioned blocks
 	 */
 	pblk->rl.total_blocks = nr_free_blks;
-	pblk->rl.nr_secs = nr_free_blks * geo->sec_per_blk;
-	pblk->capacity = provisioned * geo->sec_per_blk;
+	pblk->rl.nr_secs = nr_free_blks * geo->sec_per_chk;
+
+	/* Consider sectors used for metadata */
+	sec_meta = (lm->smeta_sec + lm->emeta_sec[0]) * l_mg->nr_free_lines;
+	blk_meta = DIV_ROUND_UP(sec_meta, geo->sec_per_chk);
+
+	pblk->capacity = (provisioned - blk_meta) * geo->sec_per_chk;
+
 	atomic_set(&pblk->rl.free_blocks, nr_free_blks);
+	atomic_set(&pblk->rl.free_user_blocks, nr_free_blks);
 }
 
 static int pblk_lines_alloc_metadata(struct pblk *pblk)
@@ -683,7 +702,7 @@ static int pblk_lines_init(struct pblk *pblk)
 	int i, ret;
 
 	pblk->min_write_pgs = geo->sec_per_pl * (geo->sec_size / PAGE_SIZE);
-	max_write_ppas = pblk->min_write_pgs * geo->nr_luns;
+	max_write_ppas = pblk->min_write_pgs * geo->all_luns;
 	pblk->max_write_pgs = (max_write_ppas < nvm_max_phys_sects(dev)) ?
 				max_write_ppas : nvm_max_phys_sects(dev);
 	pblk_set_sec_per_write(pblk, pblk->min_write_pgs);
@@ -693,26 +712,26 @@ static int pblk_lines_init(struct pblk *pblk)
 		return -EINVAL;
 	}
 
-	div_u64_rem(geo->sec_per_blk, pblk->min_write_pgs, &mod);
+	div_u64_rem(geo->sec_per_chk, pblk->min_write_pgs, &mod);
 	if (mod) {
 		pr_err("pblk: bad configuration of sectors/pages\n");
 		return -EINVAL;
 	}
 
-	l_mg->nr_lines = geo->blks_per_lun;
+	l_mg->nr_lines = geo->nr_chks;
 	l_mg->log_line = l_mg->data_line = NULL;
 	l_mg->l_seq_nr = l_mg->d_seq_nr = 0;
 	l_mg->nr_free_lines = 0;
 	bitmap_zero(&l_mg->meta_bitmap, PBLK_DATA_LINES);
 
-	lm->sec_per_line = geo->sec_per_blk * geo->nr_luns;
-	lm->blk_per_line = geo->nr_luns;
-	lm->blk_bitmap_len = BITS_TO_LONGS(geo->nr_luns) * sizeof(long);
+	lm->sec_per_line = geo->sec_per_chk * geo->all_luns;
+	lm->blk_per_line = geo->all_luns;
+	lm->blk_bitmap_len = BITS_TO_LONGS(geo->all_luns) * sizeof(long);
 	lm->sec_bitmap_len = BITS_TO_LONGS(lm->sec_per_line) * sizeof(long);
-	lm->lun_bitmap_len = BITS_TO_LONGS(geo->nr_luns) * sizeof(long);
+	lm->lun_bitmap_len = BITS_TO_LONGS(geo->all_luns) * sizeof(long);
 	lm->mid_thrs = lm->sec_per_line / 2;
 	lm->high_thrs = lm->sec_per_line / 4;
-	lm->meta_distance = (geo->nr_luns / 2) * pblk->min_write_pgs;
+	lm->meta_distance = (geo->all_luns / 2) * pblk->min_write_pgs;
 
 	/* Calculate necessary pages for smeta. See comment over struct
 	 * line_smeta definition
@@ -742,12 +761,12 @@ add_emeta_page:
 		goto add_emeta_page;
 	}
 
-	lm->emeta_bb = geo->nr_luns > i ? geo->nr_luns - i : 0;
+	lm->emeta_bb = geo->all_luns > i ? geo->all_luns - i : 0;
 
 	lm->min_blk_line = 1;
-	if (geo->nr_luns > 1)
+	if (geo->all_luns > 1)
 		lm->min_blk_line += DIV_ROUND_UP(lm->smeta_sec +
-					lm->emeta_sec[0], geo->sec_per_blk);
+					lm->emeta_sec[0], geo->sec_per_chk);
 
 	if (lm->min_blk_line > lm->blk_per_line) {
 		pr_err("pblk: config. not supported. Min. LUN in line:%d\n",
@@ -772,7 +791,7 @@ add_emeta_page:
 		goto fail_free_bb_template;
 	}
 
-	bb_distance = (geo->nr_luns) * geo->sec_per_pl;
+	bb_distance = (geo->all_luns) * geo->sec_per_pl;
 	for (i = 0; i < lm->sec_per_line; i += bb_distance)
 		bitmap_set(l_mg->bb_template, i, geo->sec_per_pl);
 
@@ -844,7 +863,7 @@ add_emeta_page:
 	pblk_set_provision(pblk, nr_free_blks);
 
 	/* Cleanup per-LUN bad block lists - managed within lines on run-time */
-	for (i = 0; i < geo->nr_luns; i++)
+	for (i = 0; i < geo->all_luns; i++)
 		kfree(pblk->luns[i].bb_list);
 
 	return 0;
@@ -858,7 +877,7 @@ fail_free_bb_template:
 fail_free_meta:
 	pblk_line_meta_free(pblk);
 fail:
-	for (i = 0; i < geo->nr_luns; i++)
+	for (i = 0; i < geo->all_luns; i++)
 		kfree(pblk->luns[i].bb_list);
 
 	return ret;
@@ -866,15 +885,19 @@ fail:
 
 static int pblk_writer_init(struct pblk *pblk)
 {
-	timer_setup(&pblk->wtimer, pblk_write_timer_fn, 0);
-	mod_timer(&pblk->wtimer, jiffies + msecs_to_jiffies(100));
-
 	pblk->writer_ts = kthread_create(pblk_write_ts, pblk, "pblk-writer-t");
 	if (IS_ERR(pblk->writer_ts)) {
-		pr_err("pblk: could not allocate writer kthread\n");
-		return PTR_ERR(pblk->writer_ts);
+		int err = PTR_ERR(pblk->writer_ts);
+
+		if (err != -EINTR)
+			pr_err("pblk: could not allocate writer kthread (%d)\n",
+					err);
+		return err;
 	}
 
+	timer_setup(&pblk->wtimer, pblk_write_timer_fn, 0);
+	mod_timer(&pblk->wtimer, jiffies + msecs_to_jiffies(100));
+
 	return 0;
 }
 
@@ -910,7 +933,6 @@ static void pblk_tear_down(struct pblk *pblk)
 	pblk_pipeline_stop(pblk);
 	pblk_writer_stop(pblk);
 	pblk_rb_sync_l2p(&pblk->rwb);
-	pblk_rwb_free(pblk);
 	pblk_rl_free(&pblk->rl);
 
 	pr_debug("pblk: consistent tear down\n");
@@ -1025,7 +1047,8 @@ static void *pblk_init(struct nvm_tgt_dev *dev, struct gendisk *tdisk,
 
 	ret = pblk_writer_init(pblk);
 	if (ret) {
-		pr_err("pblk: could not initialize write thread\n");
+		if (ret != -EINTR)
+			pr_err("pblk: could not initialize write thread\n");
 		goto fail_free_lines;
 	}
 
@@ -1041,13 +1064,14 @@ static void *pblk_init(struct nvm_tgt_dev *dev, struct gendisk *tdisk,
 
 	blk_queue_write_cache(tqueue, true, false);
 
-	tqueue->limits.discard_granularity = geo->pgs_per_blk * geo->pfpg_size;
+	tqueue->limits.discard_granularity = geo->sec_per_chk * geo->sec_size;
 	tqueue->limits.discard_alignment = 0;
 	blk_queue_max_discard_sectors(tqueue, UINT_MAX >> 9);
 	queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, tqueue);
 
-	pr_info("pblk init: luns:%u, lines:%d, secs:%llu, buf entries:%u\n",
-			geo->nr_luns, pblk->l_mg.nr_lines,
+	pr_info("pblk(%s): luns:%u, lines:%d, secs:%llu, buf entries:%u\n",
+			tdisk->disk_name,
+			geo->all_luns, pblk->l_mg.nr_lines,
 			(unsigned long long)pblk->rl.nr_secs,
 			pblk->rwb.nr_entries);
 

drivers/lightnvm/pblk-map.c

@@ -146,7 +146,7 @@ void pblk_map_erase_rq(struct pblk *pblk, struct nvm_rq *rqd,
 		return;
 
 	/* Erase blocks that are bad in this line but might not be in next */
-	if (unlikely(ppa_empty(*erase_ppa)) &&
+	if (unlikely(pblk_ppa_empty(*erase_ppa)) &&
 			bitmap_weight(d_line->blk_bitmap, lm->blk_per_line)) {
 		int bit = -1;
 

drivers/lightnvm/pblk-rb.c

@@ -54,7 +54,7 @@ int pblk_rb_init(struct pblk_rb *rb, struct pblk_rb_entry *rb_entry_base,
 	rb->seg_size = (1 << power_seg_sz);
 	rb->nr_entries = (1 << power_size);
 	rb->mem = rb->subm = rb->sync = rb->l2p_update = 0;
-	rb->sync_point = EMPTY_ENTRY;
+	rb->flush_point = EMPTY_ENTRY;
 
 	spin_lock_init(&rb->w_lock);
 	spin_lock_init(&rb->s_lock);
@@ -112,7 +112,7 @@ int pblk_rb_init(struct pblk_rb *rb, struct pblk_rb_entry *rb_entry_base,
 	up_write(&pblk_rb_lock);
 
 #ifdef CONFIG_NVM_DEBUG
-	atomic_set(&rb->inflight_sync_point, 0);
+	atomic_set(&rb->inflight_flush_point, 0);
 #endif
 
 	/*
@@ -226,7 +226,7 @@ static int __pblk_rb_update_l2p(struct pblk_rb *rb, unsigned int to_update)
 		pblk_update_map_dev(pblk, w_ctx->lba, w_ctx->ppa,
 							entry->cacheline);
 
-		line = &pblk->lines[pblk_tgt_ppa_to_line(w_ctx->ppa)];
+		line = &pblk->lines[pblk_ppa_to_line(w_ctx->ppa)];
 		kref_put(&line->ref, pblk_line_put);
 		clean_wctx(w_ctx);
 		rb->l2p_update = (rb->l2p_update + 1) & (rb->nr_entries - 1);
@@ -349,35 +349,35 @@ void pblk_rb_write_entry_gc(struct pblk_rb *rb, void *data,
 	smp_store_release(&entry->w_ctx.flags, flags);
 }
 
-static int pblk_rb_sync_point_set(struct pblk_rb *rb, struct bio *bio,
+static int pblk_rb_flush_point_set(struct pblk_rb *rb, struct bio *bio,
 				  unsigned int pos)
 {
 	struct pblk_rb_entry *entry;
-	unsigned int subm, sync_point;
+	unsigned int sync, flush_point;
 
-	subm = READ_ONCE(rb->subm);
+	sync = READ_ONCE(rb->sync);
+
+	if (pos == sync)
+		return 0;
 
 #ifdef CONFIG_NVM_DEBUG
-	atomic_inc(&rb->inflight_sync_point);
+	atomic_inc(&rb->inflight_flush_point);
 #endif
 
-	if (pos == subm)
-		return 0;
+	flush_point = (pos == 0) ? (rb->nr_entries - 1) : (pos - 1);
+	entry = &rb->entries[flush_point];
 
-	sync_point = (pos == 0) ? (rb->nr_entries - 1) : (pos - 1);
-	entry = &rb->entries[sync_point];
+	pblk_rb_sync_init(rb, NULL);
 
-	/* Protect syncs */
-	smp_store_release(&rb->sync_point, sync_point);
+	/* Protect flush points */
+	smp_store_release(&rb->flush_point, flush_point);
 
-	if (!bio)
-		return 0;
+	if (bio)
+		bio_list_add(&entry->w_ctx.bios, bio);
 
-	spin_lock_irq(&rb->s_lock);
-	bio_list_add(&entry->w_ctx.bios, bio);
-	spin_unlock_irq(&rb->s_lock);
+	pblk_rb_sync_end(rb, NULL);
 
-	return 1;
+	return bio ? 1 : 0;
 }
 
 static int __pblk_rb_may_write(struct pblk_rb *rb, unsigned int nr_entries,
@@ -416,7 +416,7 @@ void pblk_rb_flush(struct pblk_rb *rb)
 	struct pblk *pblk = container_of(rb, struct pblk, rwb);
 	unsigned int mem = READ_ONCE(rb->mem);
 
-	if (pblk_rb_sync_point_set(rb, NULL, mem))
+	if (pblk_rb_flush_point_set(rb, NULL, mem))
 		return;
 
 	pblk_write_should_kick(pblk);
@@ -440,7 +440,7 @@ static int pblk_rb_may_write_flush(struct pblk_rb *rb, unsigned int nr_entries,
 #ifdef CONFIG_NVM_DEBUG
 		atomic_long_inc(&pblk->nr_flush);
 #endif
-		if (pblk_rb_sync_point_set(&pblk->rwb, bio, mem))
+		if (pblk_rb_flush_point_set(&pblk->rwb, bio, mem))
 			*io_ret = NVM_IO_OK;
 	}
 
@@ -606,21 +606,6 @@ try:
 			return NVM_IO_ERR;
 		}
 
-		if (flags & PBLK_FLUSH_ENTRY) {
-			unsigned int sync_point;
-
-			sync_point = READ_ONCE(rb->sync_point);
-			if (sync_point == pos) {
-				/* Protect syncs */
-				smp_store_release(&rb->sync_point, EMPTY_ENTRY);
-			}
-
-			flags &= ~PBLK_FLUSH_ENTRY;
-#ifdef CONFIG_NVM_DEBUG
-			atomic_dec(&rb->inflight_sync_point);
-#endif
-		}
-
 		flags &= ~PBLK_WRITTEN_DATA;
 		flags |= PBLK_SUBMITTED_ENTRY;
 
@@ -730,15 +715,24 @@ void pblk_rb_sync_end(struct pblk_rb *rb, unsigned long *flags)
 
 unsigned int pblk_rb_sync_advance(struct pblk_rb *rb, unsigned int nr_entries)
 {
-	unsigned int sync;
-	unsigned int i;
-
+	unsigned int sync, flush_point;
 	lockdep_assert_held(&rb->s_lock);
 
 	sync = READ_ONCE(rb->sync);
+	flush_point = READ_ONCE(rb->flush_point);
 
-	for (i = 0; i < nr_entries; i++)
-		sync = (sync + 1) & (rb->nr_entries - 1);
+	if (flush_point != EMPTY_ENTRY) {
+		unsigned int secs_to_flush;
+
+		secs_to_flush = pblk_rb_ring_count(flush_point, sync,
+					rb->nr_entries);
+		if (secs_to_flush < nr_entries) {
+			/* Protect flush points */
+			smp_store_release(&rb->flush_point, EMPTY_ENTRY);
+		}
+	}
+
+	sync = (sync + nr_entries) & (rb->nr_entries - 1);
 
 	/* Protect from counts */
 	smp_store_release(&rb->sync, sync);
@@ -746,22 +740,27 @@ unsigned int pblk_rb_sync_advance(struct pblk_rb *rb, unsigned int nr_entries)
 	return sync;
 }
 
-unsigned int pblk_rb_sync_point_count(struct pblk_rb *rb)
+/* Calculate how many sectors to submit up to the current flush point. */
+unsigned int pblk_rb_flush_point_count(struct pblk_rb *rb)
 {
-	unsigned int subm, sync_point;
-	unsigned int count;
+	unsigned int subm, sync, flush_point;
+	unsigned int submitted, to_flush;
 
-	/* Protect syncs */
-	sync_point = smp_load_acquire(&rb->sync_point);
-	if (sync_point == EMPTY_ENTRY)
+	/* Protect flush points */
+	flush_point = smp_load_acquire(&rb->flush_point);
+	if (flush_point == EMPTY_ENTRY)
 		return 0;
 
+	/* Protect syncs */
+	sync = smp_load_acquire(&rb->sync);
+
 	subm = READ_ONCE(rb->subm);
+	submitted = pblk_rb_ring_count(subm, sync, rb->nr_entries);
 
 	/* The sync point itself counts as a sector to sync */
-	count = pblk_rb_ring_count(sync_point, subm, rb->nr_entries) + 1;
+	to_flush = pblk_rb_ring_count(flush_point, sync, rb->nr_entries) + 1;
 
-	return count;
+	return (submitted < to_flush) ? (to_flush - submitted) : 0;
 }
 
 /*
@@ -801,7 +800,7 @@ int pblk_rb_tear_down_check(struct pblk_rb *rb)
 
 	if ((rb->mem == rb->subm) && (rb->subm == rb->sync) &&
 				(rb->sync == rb->l2p_update) &&
-				(rb->sync_point == EMPTY_ENTRY)) {
+				(rb->flush_point == EMPTY_ENTRY)) {
 		goto out;
 	}
 
@@ -848,7 +847,7 @@ ssize_t pblk_rb_sysfs(struct pblk_rb *rb, char *buf)
 		queued_entries++;
 	spin_unlock_irq(&rb->s_lock);
 
-	if (rb->sync_point != EMPTY_ENTRY)
+	if (rb->flush_point != EMPTY_ENTRY)
 		offset = scnprintf(buf, PAGE_SIZE,
 			"%u\t%u\t%u\t%u\t%u\t%u\t%u - %u/%u/%u - %d\n",
 			rb->nr_entries,
@@ -857,14 +856,14 @@ ssize_t pblk_rb_sysfs(struct pblk_rb *rb, char *buf)
 			rb->sync,
 			rb->l2p_update,
 #ifdef CONFIG_NVM_DEBUG
-			atomic_read(&rb->inflight_sync_point),
+			atomic_read(&rb->inflight_flush_point),
 #else
 			0,
 #endif
-			rb->sync_point,
+			rb->flush_point,
 			pblk_rb_read_count(rb),
 			pblk_rb_space(rb),
-			pblk_rb_sync_point_count(rb),
+			pblk_rb_flush_point_count(rb),
 			queued_entries);
 	else
 		offset = scnprintf(buf, PAGE_SIZE,
@@ -875,13 +874,13 @@ ssize_t pblk_rb_sysfs(struct pblk_rb *rb, char *buf)
 			rb->sync,
 			rb->l2p_update,
 #ifdef CONFIG_NVM_DEBUG
-			atomic_read(&rb->inflight_sync_point),
+			atomic_read(&rb->inflight_flush_point),
 #else
 			0,
 #endif
 			pblk_rb_read_count(rb),
 			pblk_rb_space(rb),
-			pblk_rb_sync_point_count(rb),
+			pblk_rb_flush_point_count(rb),
 			queued_entries);
 
 	return offset;

drivers/lightnvm/pblk-read.c

@@ -141,7 +141,7 @@ static void pblk_read_put_rqd_kref(struct pblk *pblk, struct nvm_rq *rqd)
 		struct ppa_addr ppa = ppa_list[i];
 		struct pblk_line *line;
 
-		line = &pblk->lines[pblk_dev_ppa_to_line(ppa)];
+		line = &pblk->lines[pblk_ppa_to_line(ppa)];
 		kref_put(&line->ref, pblk_line_put_wq);
 	}
 }
@@ -158,8 +158,12 @@ static void pblk_end_user_read(struct bio *bio)
 static void __pblk_end_io_read(struct pblk *pblk, struct nvm_rq *rqd,
 			       bool put_line)
 {
+	struct nvm_tgt_dev *dev = pblk->dev;
 	struct pblk_g_ctx *r_ctx = nvm_rq_to_pdu(rqd);
 	struct bio *bio = rqd->bio;
+	unsigned long start_time = r_ctx->start_time;
+
+	generic_end_io_acct(dev->q, READ, &pblk->disk->part0, start_time);
 
 	if (rqd->error)
 		pblk_log_read_err(pblk, rqd);
@@ -193,9 +197,9 @@ static void pblk_end_io_read(struct nvm_rq *rqd)
 	__pblk_end_io_read(pblk, rqd, true);
 }
 
-static int pblk_fill_partial_read_bio(struct pblk *pblk, struct nvm_rq *rqd,
-				      unsigned int bio_init_idx,
-				      unsigned long *read_bitmap)
+static int pblk_partial_read_bio(struct pblk *pblk, struct nvm_rq *rqd,
+				 unsigned int bio_init_idx,
+				 unsigned long *read_bitmap)
 {
 	struct bio *new_bio, *bio = rqd->bio;
 	struct pblk_sec_meta *meta_list = rqd->meta_list;
@@ -270,7 +274,7 @@ static int pblk_fill_partial_read_bio(struct pblk *pblk, struct nvm_rq *rqd,
 	i = 0;
 	hole = find_first_zero_bit(read_bitmap, nr_secs);
 	do {
-		int line_id = pblk_dev_ppa_to_line(rqd->ppa_list[i]);
+		int line_id = pblk_ppa_to_line(rqd->ppa_list[i]);
 		struct pblk_line *line = &pblk->lines[line_id];
 
 		kref_put(&line->ref, pblk_line_put);
@@ -306,6 +310,8 @@ static int pblk_fill_partial_read_bio(struct pblk *pblk, struct nvm_rq *rqd,
 	return NVM_IO_OK;
 
 err:
+	pr_err("pblk: failed to perform partial read\n");
+
 	/* Free allocated pages in new bio */
 	pblk_bio_free_pages(pblk, bio, 0, new_bio->bi_vcnt);
 	__pblk_end_io_read(pblk, rqd, false);
@@ -357,6 +363,7 @@ retry:
 int pblk_submit_read(struct pblk *pblk, struct bio *bio)
 {
 	struct nvm_tgt_dev *dev = pblk->dev;
+	struct request_queue *q = dev->q;
 	sector_t blba = pblk_get_lba(bio);
 	unsigned int nr_secs = pblk_get_secs(bio);
 	struct pblk_g_ctx *r_ctx;
@@ -372,6 +379,8 @@ int pblk_submit_read(struct pblk *pblk, struct bio *bio)
 		return NVM_IO_ERR;
 	}
 
+	generic_start_io_acct(q, READ, bio_sectors(bio), &pblk->disk->part0);
+
 	bitmap_zero(&read_bitmap, nr_secs);
 
 	rqd = pblk_alloc_rqd(pblk, PBLK_READ);
@@ -383,6 +392,7 @@ int pblk_submit_read(struct pblk *pblk, struct bio *bio)
 	rqd->end_io = pblk_end_io_read;
 
 	r_ctx = nvm_rq_to_pdu(rqd);
+	r_ctx->start_time = jiffies;
 	r_ctx->lba = blba;
 
 	/* Save the index for this bio's start. This is needed in case
@@ -422,7 +432,7 @@ int pblk_submit_read(struct pblk *pblk, struct bio *bio)
 		int_bio = bio_clone_fast(bio, GFP_KERNEL, pblk_bio_set);
 		if (!int_bio) {
 			pr_err("pblk: could not clone read bio\n");
-			return NVM_IO_ERR;
+			goto fail_end_io;
 		}
 
 		rqd->bio = int_bio;
@@ -433,7 +443,7 @@ int pblk_submit_read(struct pblk *pblk, struct bio *bio)
 			pr_err("pblk: read IO submission failed\n");
 			if (int_bio)
 				bio_put(int_bio);
-			return ret;
+			goto fail_end_io;
 		}
 
 		return NVM_IO_OK;
@@ -442,17 +452,14 @@ int pblk_submit_read(struct pblk *pblk, struct bio *bio)
 	/* The read bio request could be partially filled by the write buffer,
 	 * but there are some holes that need to be read from the drive.
 	 */
-	ret = pblk_fill_partial_read_bio(pblk, rqd, bio_init_idx, &read_bitmap);
-	if (ret) {
-		pr_err("pblk: failed to perform partial read\n");
-		return ret;
-	}
-
-	return NVM_IO_OK;
+	return pblk_partial_read_bio(pblk, rqd, bio_init_idx, &read_bitmap);
 
 fail_rqd_free:
 	pblk_free_rqd(pblk, rqd, PBLK_READ);
 	return ret;
+fail_end_io:
+	__pblk_end_io_read(pblk, rqd, false);
+	return ret;
 }
 
 static int read_ppalist_rq_gc(struct pblk *pblk, struct nvm_rq *rqd,

drivers/lightnvm/pblk-recovery.c

@@ -111,18 +111,18 @@ int pblk_recov_setup_rq(struct pblk *pblk, struct pblk_c_ctx *c_ctx,
 	return 0;
 }
 
-__le64 *pblk_recov_get_lba_list(struct pblk *pblk, struct line_emeta *emeta_buf)
+int pblk_recov_check_emeta(struct pblk *pblk, struct line_emeta *emeta_buf)
 {
 	u32 crc;
 
 	crc = pblk_calc_emeta_crc(pblk, emeta_buf);
 	if (le32_to_cpu(emeta_buf->crc) != crc)
-		return NULL;
+		return 1;
 
 	if (le32_to_cpu(emeta_buf->header.identifier) != PBLK_MAGIC)
-		return NULL;
+		return 1;
 
-	return emeta_to_lbas(pblk, emeta_buf);
+	return 0;
 }
 
 static int pblk_recov_l2p_from_emeta(struct pblk *pblk, struct pblk_line *line)
@@ -137,7 +137,7 @@ static int pblk_recov_l2p_from_emeta(struct pblk *pblk, struct pblk_line *line)
 	u64 nr_valid_lbas, nr_lbas = 0;
 	u64 i;
 
-	lba_list = pblk_recov_get_lba_list(pblk, emeta_buf);
+	lba_list = emeta_to_lbas(pblk, emeta_buf);
 	if (!lba_list)
 		return 1;
 
@@ -149,7 +149,7 @@ static int pblk_recov_l2p_from_emeta(struct pblk *pblk, struct pblk_line *line)
 		struct ppa_addr ppa;
 		int pos;
 
-		ppa = addr_to_pblk_ppa(pblk, i, line->id);
+		ppa = addr_to_gen_ppa(pblk, i, line->id);
 		pos = pblk_ppa_to_pos(geo, ppa);
 
 		/* Do not update bad blocks */
@@ -188,7 +188,7 @@ static int pblk_calc_sec_in_line(struct pblk *pblk, struct pblk_line *line)
 	int nr_bb = bitmap_weight(line->blk_bitmap, lm->blk_per_line);
 
 	return lm->sec_per_line - lm->smeta_sec - lm->emeta_sec[0] -
-				nr_bb * geo->sec_per_blk;
+				nr_bb * geo->sec_per_chk;
 }
 
 struct pblk_recov_alloc {
@@ -263,12 +263,12 @@ next_read_rq:
 		int pos;
 
 		ppa = addr_to_gen_ppa(pblk, r_ptr_int, line->id);
-		pos = pblk_dev_ppa_to_pos(geo, ppa);
+		pos = pblk_ppa_to_pos(geo, ppa);
 
 		while (test_bit(pos, line->blk_bitmap)) {
 			r_ptr_int += pblk->min_write_pgs;
 			ppa = addr_to_gen_ppa(pblk, r_ptr_int, line->id);
-			pos = pblk_dev_ppa_to_pos(geo, ppa);
+			pos = pblk_ppa_to_pos(geo, ppa);
 		}
 
 		for (j = 0; j < pblk->min_write_pgs; j++, i++, r_ptr_int++)
@@ -288,7 +288,7 @@ next_read_rq:
 	/* At this point, the read should not fail. If it does, it is a problem
 	 * we cannot recover from here. Need FTL log.
 	 */
-	if (rqd->error) {
+	if (rqd->error && rqd->error != NVM_RSP_WARN_HIGHECC) {
 		pr_err("pblk: L2P recovery failed (%d)\n", rqd->error);
 		return -EINTR;
 	}
@@ -411,12 +411,12 @@ next_pad_rq:
 		int pos;
 
 		w_ptr = pblk_alloc_page(pblk, line, pblk->min_write_pgs);
-		ppa = addr_to_pblk_ppa(pblk, w_ptr, line->id);
+		ppa = addr_to_gen_ppa(pblk, w_ptr, line->id);
 		pos = pblk_ppa_to_pos(geo, ppa);
 
 		while (test_bit(pos, line->blk_bitmap)) {
 			w_ptr += pblk->min_write_pgs;
-			ppa = addr_to_pblk_ppa(pblk, w_ptr, line->id);
+			ppa = addr_to_gen_ppa(pblk, w_ptr, line->id);
 			pos = pblk_ppa_to_pos(geo, ppa);
 		}
 
@@ -541,12 +541,12 @@ next_rq:
 
 		w_ptr = pblk_alloc_page(pblk, line, pblk->min_write_pgs);
 		ppa = addr_to_gen_ppa(pblk, w_ptr, line->id);
-		pos = pblk_dev_ppa_to_pos(geo, ppa);
+		pos = pblk_ppa_to_pos(geo, ppa);
 
 		while (test_bit(pos, line->blk_bitmap)) {
 			w_ptr += pblk->min_write_pgs;
 			ppa = addr_to_gen_ppa(pblk, w_ptr, line->id);
-			pos = pblk_dev_ppa_to_pos(geo, ppa);
+			pos = pblk_ppa_to_pos(geo, ppa);
 		}
 
 		for (j = 0; j < pblk->min_write_pgs; j++, i++, w_ptr++)
@@ -672,12 +672,12 @@ next_rq:
 
 		paddr = pblk_alloc_page(pblk, line, pblk->min_write_pgs);
 		ppa = addr_to_gen_ppa(pblk, paddr, line->id);
-		pos = pblk_dev_ppa_to_pos(geo, ppa);
+		pos = pblk_ppa_to_pos(geo, ppa);
 
 		while (test_bit(pos, line->blk_bitmap)) {
 			paddr += pblk->min_write_pgs;
 			ppa = addr_to_gen_ppa(pblk, paddr, line->id);
-			pos = pblk_dev_ppa_to_pos(geo, ppa);
+			pos = pblk_ppa_to_pos(geo, ppa);
 		}
 
 		for (j = 0; j < pblk->min_write_pgs; j++, i++, paddr++)
@@ -817,7 +817,7 @@ static u64 pblk_line_emeta_start(struct pblk *pblk, struct pblk_line *line)
 
 	while (emeta_secs) {
 		emeta_start--;
-		ppa = addr_to_pblk_ppa(pblk, emeta_start, line->id);
+		ppa = addr_to_gen_ppa(pblk, emeta_start, line->id);
 		pos = pblk_ppa_to_pos(geo, ppa);
 		if (!test_bit(pos, line->blk_bitmap))
 			emeta_secs--;
@@ -938,6 +938,11 @@ struct pblk_line *pblk_recov_l2p(struct pblk *pblk)
 			goto next;
 		}
 
+		if (pblk_recov_check_emeta(pblk, line->emeta->buf)) {
+			pblk_recov_l2p_from_oob(pblk, line);
+			goto next;
+		}
+
 		if (pblk_recov_l2p_from_emeta(pblk, line))
 			pblk_recov_l2p_from_oob(pblk, line);
 
@@ -984,10 +989,8 @@ next:
 	}
 	spin_unlock(&l_mg->free_lock);
 
-	if (is_next) {
+	if (is_next)
 		pblk_line_erase(pblk, l_mg->data_next);
-		pblk_rl_free_lines_dec(&pblk->rl, l_mg->data_next);
-	}
 
 out:
 	if (found_lines != recovered_lines)

drivers/lightnvm/pblk-rl.c

@@ -89,17 +89,15 @@ unsigned long pblk_rl_nr_free_blks(struct pblk_rl *rl)
 	return atomic_read(&rl->free_blocks);
 }
 
-/*
- * We check for (i) the number of free blocks in the current LUN and (ii) the
- * total number of free blocks in the pblk instance. This is to even out the
- * number of free blocks on each LUN when GC kicks in.
- *
- * Only the total number of free blocks is used to configure the rate limiter.
- */
-void pblk_rl_update_rates(struct pblk_rl *rl)
+unsigned long pblk_rl_nr_user_free_blks(struct pblk_rl *rl)
+{
+	return atomic_read(&rl->free_user_blocks);
+}
+
+static void __pblk_rl_update_rates(struct pblk_rl *rl,
+				   unsigned long free_blocks)
 {
 	struct pblk *pblk = container_of(rl, struct pblk, rl);
-	unsigned long free_blocks = pblk_rl_nr_free_blks(rl);
 	int max = rl->rb_budget;
 
 	if (free_blocks >= rl->high) {
@@ -132,20 +130,37 @@ void pblk_rl_update_rates(struct pblk_rl *rl)
 		pblk_gc_should_stop(pblk);
 }
 
+void pblk_rl_update_rates(struct pblk_rl *rl)
+{
+	__pblk_rl_update_rates(rl, pblk_rl_nr_user_free_blks(rl));
+}
+
 void pblk_rl_free_lines_inc(struct pblk_rl *rl, struct pblk_line *line)
 {
 	int blk_in_line = atomic_read(&line->blk_in_line);
+	int free_blocks;
 
 	atomic_add(blk_in_line, &rl->free_blocks);
-	pblk_rl_update_rates(rl);
+	free_blocks = atomic_add_return(blk_in_line, &rl->free_user_blocks);
+
+	__pblk_rl_update_rates(rl, free_blocks);
 }
 
-void pblk_rl_free_lines_dec(struct pblk_rl *rl, struct pblk_line *line)
+void pblk_rl_free_lines_dec(struct pblk_rl *rl, struct pblk_line *line,
+			    bool used)
 {
 	int blk_in_line = atomic_read(&line->blk_in_line);
+	int free_blocks;
 
 	atomic_sub(blk_in_line, &rl->free_blocks);
-	pblk_rl_update_rates(rl);
+
+	if (used)
+		free_blocks = atomic_sub_return(blk_in_line,
+							&rl->free_user_blocks);
+	else
+		free_blocks = atomic_read(&rl->free_user_blocks);
+
+	__pblk_rl_update_rates(rl, free_blocks);
 }
 
 int pblk_rl_high_thrs(struct pblk_rl *rl)
@@ -174,16 +189,21 @@ void pblk_rl_free(struct pblk_rl *rl)
 void pblk_rl_init(struct pblk_rl *rl, int budget)
 {
 	struct pblk *pblk = container_of(rl, struct pblk, rl);
+	struct nvm_tgt_dev *dev = pblk->dev;
+	struct nvm_geo *geo = &dev->geo;
+	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
 	struct pblk_line_meta *lm = &pblk->lm;
 	int min_blocks = lm->blk_per_line * PBLK_GC_RSV_LINE;
+	int sec_meta, blk_meta;
+
 	unsigned int rb_windows;
 
-	rl->high = rl->total_blocks / PBLK_USER_HIGH_THRS;
-	rl->high_pw = get_count_order(rl->high);
+	/* Consider sectors used for metadata */
+	sec_meta = (lm->smeta_sec + lm->emeta_sec[0]) * l_mg->nr_free_lines;
+	blk_meta = DIV_ROUND_UP(sec_meta, geo->sec_per_chk);
 
-	rl->low = rl->total_blocks / PBLK_USER_LOW_THRS;
-	if (rl->low < min_blocks)
-		rl->low = min_blocks;
+	rl->high = pblk->op_blks - blk_meta - lm->blk_per_line;
+	rl->high_pw = get_count_order(rl->high);
 
 	rl->rsv_blocks = min_blocks;
 

drivers/lightnvm/pblk-sysfs.c

@@ -28,7 +28,7 @@ static ssize_t pblk_sysfs_luns_show(struct pblk *pblk, char *page)
 	ssize_t sz = 0;
 	int i;
 
-	for (i = 0; i < geo->nr_luns; i++) {
+	for (i = 0; i < geo->all_luns; i++) {
 		int active = 1;
 
 		rlun = &pblk->luns[i];
@@ -49,11 +49,12 @@ static ssize_t pblk_sysfs_luns_show(struct pblk *pblk, char *page)
 
 static ssize_t pblk_sysfs_rate_limiter(struct pblk *pblk, char *page)
 {
-	int free_blocks, total_blocks;
+	int free_blocks, free_user_blocks, total_blocks;
 	int rb_user_max, rb_user_cnt;
 	int rb_gc_max, rb_gc_cnt, rb_budget, rb_state;
 
-	free_blocks = atomic_read(&pblk->rl.free_blocks);
+	free_blocks = pblk_rl_nr_free_blks(&pblk->rl);
+	free_user_blocks = pblk_rl_nr_user_free_blks(&pblk->rl);
 	rb_user_max = pblk->rl.rb_user_max;
 	rb_user_cnt = atomic_read(&pblk->rl.rb_user_cnt);
 	rb_gc_max = pblk->rl.rb_gc_max;
@@ -64,16 +65,16 @@ static ssize_t pblk_sysfs_rate_limiter(struct pblk *pblk, char *page)
 	total_blocks = pblk->rl.total_blocks;
 
 	return snprintf(page, PAGE_SIZE,
-		"u:%u/%u,gc:%u/%u(%u/%u)(stop:<%u,full:>%u,free:%d/%d)-%d\n",
+		"u:%u/%u,gc:%u/%u(%u)(stop:<%u,full:>%u,free:%d/%d/%d)-%d\n",
 				rb_user_cnt,
 				rb_user_max,
 				rb_gc_cnt,
 				rb_gc_max,
 				rb_state,
 				rb_budget,
-				pblk->rl.low,
 				pblk->rl.high,
 				free_blocks,
+				free_user_blocks,
 				total_blocks,
 				READ_ONCE(pblk->rl.rb_user_active));
 }
@@ -238,7 +239,7 @@ static ssize_t pblk_sysfs_lines(struct pblk *pblk, char *page)
 
 	sz = snprintf(page, PAGE_SIZE - sz,
 		"line: nluns:%d, nblks:%d, nsecs:%d\n",
-		geo->nr_luns, lm->blk_per_line, lm->sec_per_line);
+		geo->all_luns, lm->blk_per_line, lm->sec_per_line);
 
 	sz += snprintf(page + sz, PAGE_SIZE - sz,
 		"lines:d:%d,l:%d-f:%d,m:%d/%d,c:%d,b:%d,co:%d(d:%d,l:%d)t:%d\n",
@@ -287,7 +288,7 @@ static ssize_t pblk_sysfs_lines_info(struct pblk *pblk, char *page)
 				"blk_line:%d, sec_line:%d, sec_blk:%d\n",
 					lm->blk_per_line,
 					lm->sec_per_line,
-					geo->sec_per_blk);
+					geo->sec_per_chk);
 
 	return sz;
 }

drivers/lightnvm/pblk-write.c

@@ -21,13 +21,28 @@ static unsigned long pblk_end_w_bio(struct pblk *pblk, struct nvm_rq *rqd,
 				    struct pblk_c_ctx *c_ctx)
 {
 	struct bio *original_bio;
+	struct pblk_rb *rwb = &pblk->rwb;
 	unsigned long ret;
 	int i;
 
 	for (i = 0; i < c_ctx->nr_valid; i++) {
 		struct pblk_w_ctx *w_ctx;
+		int pos = c_ctx->sentry + i;
+		int flags;
+
+		w_ctx = pblk_rb_w_ctx(rwb, pos);
+		flags = READ_ONCE(w_ctx->flags);
+
+		if (flags & PBLK_FLUSH_ENTRY) {
+			flags &= ~PBLK_FLUSH_ENTRY;
+			/* Release flags on context. Protect from writes */
+			smp_store_release(&w_ctx->flags, flags);
+
+#ifdef CONFIG_NVM_DEBUG
+			atomic_dec(&rwb->inflight_flush_point);
+#endif
+		}
 
-		w_ctx = pblk_rb_w_ctx(&pblk->rwb, c_ctx->sentry + i);
 		while ((original_bio = bio_list_pop(&w_ctx->bios)))
 			bio_endio(original_bio);
 	}
@@ -439,7 +454,7 @@ static int pblk_submit_io_set(struct pblk *pblk, struct nvm_rq *rqd)
 	struct pblk_line *meta_line;
 	int err;
 
-	ppa_set_empty(&erase_ppa);
+	pblk_ppa_set_empty(&erase_ppa);
 
 	/* Assign lbas to ppas and populate request structure */
 	err = pblk_setup_w_rq(pblk, rqd, &erase_ppa);
@@ -457,7 +472,7 @@ static int pblk_submit_io_set(struct pblk *pblk, struct nvm_rq *rqd)
 		return NVM_IO_ERR;
 	}
 
-	if (!ppa_empty(erase_ppa)) {
+	if (!pblk_ppa_empty(erase_ppa)) {
 		/* Submit erase for next data line */
 		if (pblk_blk_erase_async(pblk, erase_ppa)) {
 			struct pblk_line *e_line = pblk_line_get_erase(pblk);
@@ -508,7 +523,7 @@ static int pblk_submit_write(struct pblk *pblk)
 	if (!secs_avail)
 		return 1;
 
-	secs_to_flush = pblk_rb_sync_point_count(&pblk->rwb);
+	secs_to_flush = pblk_rb_flush_point_count(&pblk->rwb);
 	if (!secs_to_flush && secs_avail < pblk->min_write_pgs)
 		return 1;
 

drivers/lightnvm/pblk.h

@@ -51,17 +51,16 @@
 
 #define NR_PHY_IN_LOG (PBLK_EXPOSED_PAGE_SIZE / PBLK_SECTOR)
 
-#define pblk_for_each_lun(pblk, rlun, i) \
-		for ((i) = 0, rlun = &(pblk)->luns[0]; \
-			(i) < (pblk)->nr_luns; (i)++, rlun = &(pblk)->luns[(i)])
-
 /* Static pool sizes */
 #define PBLK_GEN_WS_POOL_SIZE (2)
 
+#define PBLK_DEFAULT_OP (11)
+
 enum {
 	PBLK_READ		= READ,
 	PBLK_WRITE		= WRITE,/* Write from write buffer */
 	PBLK_WRITE_INT,			/* Internal write - no write buffer */
+	PBLK_READ_RECOV,		/* Recovery read - errors allowed */
 	PBLK_ERASE,
 };
 
@@ -114,6 +113,7 @@ struct pblk_c_ctx {
 /* read context */
 struct pblk_g_ctx {
 	void *private;
+	unsigned long start_time;
 	u64 lba;
 };
 
@@ -170,7 +170,7 @@ struct pblk_rb {
 					 * the last submitted entry that has
 					 * been successfully persisted to media
 					 */
-	unsigned int sync_point;	/* Sync point - last entry that must be
+	unsigned int flush_point;	/* Sync point - last entry that must be
 					 * flushed to the media. Used with
 					 * REQ_FLUSH and REQ_FUA
 					 */
@@ -193,7 +193,7 @@ struct pblk_rb {
 	spinlock_t s_lock;		/* Sync lock */
 
 #ifdef CONFIG_NVM_DEBUG
-	atomic_t inflight_sync_point;	/* Not served REQ_FLUSH | REQ_FUA */
+	atomic_t inflight_flush_point;	/* Not served REQ_FLUSH | REQ_FUA */
 #endif
 };
 
@@ -256,9 +256,6 @@ struct pblk_rl {
 	unsigned int high;	/* Upper threshold for rate limiter (free run -
 				 * user I/O rate limiter
 				 */
-	unsigned int low;	/* Lower threshold for rate limiter (user I/O
-				 * rate limiter - stall)
-				 */
 	unsigned int high_pw;	/* High rounded up as a power of 2 */
 
 #define PBLK_USER_HIGH_THRS 8	/* Begin write limit at 12% available blks */
@@ -292,7 +289,9 @@ struct pblk_rl {
 
 	unsigned long long nr_secs;
 	unsigned long total_blocks;
-	atomic_t free_blocks;
+
+	atomic_t free_blocks;		/* Total number of free blocks (+ OP) */
+	atomic_t free_user_blocks;	/* Number of user free blocks (no OP) */
 };
 
 #define PBLK_LINE_EMPTY (~0U)
@@ -583,7 +582,9 @@ struct pblk {
 			    */
 
 	sector_t capacity; /* Device capacity when bad blocks are subtracted */
-	int over_pct;      /* Percentage of device used for over-provisioning */
+
+	int op;      /* Percentage of device used for over-provisioning */
+	int op_blks; /* Number of blocks used for over-provisioning */
 
 	/* pblk provisioning values. Used by rate limiter */
 	struct pblk_rl rl;
@@ -691,7 +692,7 @@ unsigned int pblk_rb_sync_advance(struct pblk_rb *rb, unsigned int nr_entries);
 struct pblk_rb_entry *pblk_rb_sync_scan_entry(struct pblk_rb *rb,
 					      struct ppa_addr *ppa);
 void pblk_rb_sync_end(struct pblk_rb *rb, unsigned long *flags);
-unsigned int pblk_rb_sync_point_count(struct pblk_rb *rb);
+unsigned int pblk_rb_flush_point_count(struct pblk_rb *rb);
 
 unsigned int pblk_rb_read_count(struct pblk_rb *rb);
 unsigned int pblk_rb_sync_count(struct pblk_rb *rb);
@@ -812,7 +813,7 @@ int pblk_submit_read_gc(struct pblk *pblk, struct pblk_gc_rq *gc_rq);
 void pblk_submit_rec(struct work_struct *work);
 struct pblk_line *pblk_recov_l2p(struct pblk *pblk);
 int pblk_recov_pad(struct pblk *pblk);
-__le64 *pblk_recov_get_lba_list(struct pblk *pblk, struct line_emeta *emeta);
+int pblk_recov_check_emeta(struct pblk *pblk, struct line_emeta *emeta);
 int pblk_recov_setup_rq(struct pblk *pblk, struct pblk_c_ctx *c_ctx,
 			struct pblk_rec_ctx *recovery, u64 *comp_bits,
 			unsigned int comp);
@@ -843,6 +844,7 @@ void pblk_rl_free(struct pblk_rl *rl);
 void pblk_rl_update_rates(struct pblk_rl *rl);
 int pblk_rl_high_thrs(struct pblk_rl *rl);
 unsigned long pblk_rl_nr_free_blks(struct pblk_rl *rl);
+unsigned long pblk_rl_nr_user_free_blks(struct pblk_rl *rl);
 int pblk_rl_user_may_insert(struct pblk_rl *rl, int nr_entries);
 void pblk_rl_inserted(struct pblk_rl *rl, int nr_entries);
 void pblk_rl_user_in(struct pblk_rl *rl, int nr_entries);
@@ -851,7 +853,8 @@ void pblk_rl_gc_in(struct pblk_rl *rl, int nr_entries);
 void pblk_rl_out(struct pblk_rl *rl, int nr_user, int nr_gc);
 int pblk_rl_max_io(struct pblk_rl *rl);
 void pblk_rl_free_lines_inc(struct pblk_rl *rl, struct pblk_line *line);
-void pblk_rl_free_lines_dec(struct pblk_rl *rl, struct pblk_line *line);
+void pblk_rl_free_lines_dec(struct pblk_rl *rl, struct pblk_line *line,
+			    bool used);
 int pblk_rl_is_limit(struct pblk_rl *rl);
 
 /*
@@ -907,28 +910,47 @@ static inline int pblk_pad_distance(struct pblk *pblk)
 	struct nvm_tgt_dev *dev = pblk->dev;
 	struct nvm_geo *geo = &dev->geo;
 
-	return NVM_MEM_PAGE_WRITE * geo->nr_luns * geo->sec_per_pl;
+	return NVM_MEM_PAGE_WRITE * geo->all_luns * geo->sec_per_pl;
 }
 
-static inline int pblk_dev_ppa_to_line(struct ppa_addr p)
+static inline int pblk_ppa_to_line(struct ppa_addr p)
 {
 	return p.g.blk;
 }
 
-static inline int pblk_tgt_ppa_to_line(struct ppa_addr p)
+static inline int pblk_ppa_to_pos(struct nvm_geo *geo, struct ppa_addr p)
 {
-	return p.g.blk;
+	return p.g.lun * geo->nr_chnls + p.g.ch;
 }
 
-static inline int pblk_ppa_to_pos(struct nvm_geo *geo, struct ppa_addr p)
+static inline struct ppa_addr addr_to_gen_ppa(struct pblk *pblk, u64 paddr,
+					      u64 line_id)
 {
-	return p.g.lun * geo->nr_chnls + p.g.ch;
+	struct ppa_addr ppa;
+
+	ppa.ppa = 0;
+	ppa.g.blk = line_id;
+	ppa.g.pg = (paddr & pblk->ppaf.pg_mask) >> pblk->ppaf.pg_offset;
+	ppa.g.lun = (paddr & pblk->ppaf.lun_mask) >> pblk->ppaf.lun_offset;
+	ppa.g.ch = (paddr & pblk->ppaf.ch_mask) >> pblk->ppaf.ch_offset;
+	ppa.g.pl = (paddr & pblk->ppaf.pln_mask) >> pblk->ppaf.pln_offset;
+	ppa.g.sec = (paddr & pblk->ppaf.sec_mask) >> pblk->ppaf.sec_offset;
+
+	return ppa;
 }
 
-/* A block within a line corresponds to the lun */
-static inline int pblk_dev_ppa_to_pos(struct nvm_geo *geo, struct ppa_addr p)
+static inline u64 pblk_dev_ppa_to_line_addr(struct pblk *pblk,
+							struct ppa_addr p)
 {
-	return p.g.lun * geo->nr_chnls + p.g.ch;
+	u64 paddr;
+
+	paddr = (u64)p.g.pg << pblk->ppaf.pg_offset;
+	paddr |= (u64)p.g.lun << pblk->ppaf.lun_offset;
+	paddr |= (u64)p.g.ch << pblk->ppaf.ch_offset;
+	paddr |= (u64)p.g.pl << pblk->ppaf.pln_offset;
+	paddr |= (u64)p.g.sec << pblk->ppaf.sec_offset;
+
+	return paddr;
 }
 
 static inline struct ppa_addr pblk_ppa32_to_ppa64(struct pblk *pblk, u32 ppa32)
@@ -960,24 +982,6 @@ static inline struct ppa_addr pblk_ppa32_to_ppa64(struct pblk *pblk, u32 ppa32)
 	return ppa64;
 }
 
-static inline struct ppa_addr pblk_trans_map_get(struct pblk *pblk,
-								sector_t lba)
-{
-	struct ppa_addr ppa;
-
-	if (pblk->ppaf_bitsize < 32) {
-		u32 *map = (u32 *)pblk->trans_map;
-
-		ppa = pblk_ppa32_to_ppa64(pblk, map[lba]);
-	} else {
-		struct ppa_addr *map = (struct ppa_addr *)pblk->trans_map;
-
-		ppa = map[lba];
-	}
-
-	return ppa;
-}
-
 static inline u32 pblk_ppa64_to_ppa32(struct pblk *pblk, struct ppa_addr ppa64)
 {
 	u32 ppa32 = 0;
@@ -999,33 +1003,36 @@ static inline u32 pblk_ppa64_to_ppa32(struct pblk *pblk, struct ppa_addr ppa64)
 	return ppa32;
 }
 
-static inline void pblk_trans_map_set(struct pblk *pblk, sector_t lba,
-						struct ppa_addr ppa)
+static inline struct ppa_addr pblk_trans_map_get(struct pblk *pblk,
+								sector_t lba)
 {
+	struct ppa_addr ppa;
+
 	if (pblk->ppaf_bitsize < 32) {
 		u32 *map = (u32 *)pblk->trans_map;
 
-		map[lba] = pblk_ppa64_to_ppa32(pblk, ppa);
+		ppa = pblk_ppa32_to_ppa64(pblk, map[lba]);
 	} else {
-		u64 *map = (u64 *)pblk->trans_map;
+		struct ppa_addr *map = (struct ppa_addr *)pblk->trans_map;
 
-		map[lba] = ppa.ppa;
+		ppa = map[lba];
 	}
+
+	return ppa;
 }
 
-static inline u64 pblk_dev_ppa_to_line_addr(struct pblk *pblk,
-							struct ppa_addr p)
+static inline void pblk_trans_map_set(struct pblk *pblk, sector_t lba,
+						struct ppa_addr ppa)
 {
-	u64 paddr;
+	if (pblk->ppaf_bitsize < 32) {
+		u32 *map = (u32 *)pblk->trans_map;
 
-	paddr = 0;
-	paddr |= (u64)p.g.pg << pblk->ppaf.pg_offset;
-	paddr |= (u64)p.g.lun << pblk->ppaf.lun_offset;
-	paddr |= (u64)p.g.ch << pblk->ppaf.ch_offset;
-	paddr |= (u64)p.g.pl << pblk->ppaf.pln_offset;
-	paddr |= (u64)p.g.sec << pblk->ppaf.sec_offset;
+		map[lba] = pblk_ppa64_to_ppa32(pblk, ppa);
+	} else {
+		u64 *map = (u64 *)pblk->trans_map;
 
-	return paddr;
+		map[lba] = ppa.ppa;
+	}
 }
 
 static inline int pblk_ppa_empty(struct ppa_addr ppa_addr)
@@ -1040,10 +1047,7 @@ static inline void pblk_ppa_set_empty(struct ppa_addr *ppa_addr)
 
 static inline bool pblk_ppa_comp(struct ppa_addr lppa, struct ppa_addr rppa)
 {
-	if (lppa.ppa == rppa.ppa)
-		return true;
-
-	return false;
+	return (lppa.ppa == rppa.ppa);
 }
 
 static inline int pblk_addr_in_cache(struct ppa_addr ppa)
@@ -1066,32 +1070,6 @@ static inline struct ppa_addr pblk_cacheline_to_addr(int addr)
 	return p;
 }
 
-static inline struct ppa_addr addr_to_gen_ppa(struct pblk *pblk, u64 paddr,
-					      u64 line_id)
-{
-	struct ppa_addr ppa;
-
-	ppa.ppa = 0;
-	ppa.g.blk = line_id;
-	ppa.g.pg = (paddr & pblk->ppaf.pg_mask) >> pblk->ppaf.pg_offset;
-	ppa.g.lun = (paddr & pblk->ppaf.lun_mask) >> pblk->ppaf.lun_offset;
-	ppa.g.ch = (paddr & pblk->ppaf.ch_mask) >> pblk->ppaf.ch_offset;
-	ppa.g.pl = (paddr & pblk->ppaf.pln_mask) >> pblk->ppaf.pln_offset;
-	ppa.g.sec = (paddr & pblk->ppaf.sec_mask) >> pblk->ppaf.sec_offset;
-
-	return ppa;
-}
-
-static inline struct ppa_addr addr_to_pblk_ppa(struct pblk *pblk, u64 paddr,
-					 u64 line_id)
-{
-	struct ppa_addr ppa;
-
-	ppa = addr_to_gen_ppa(pblk, paddr, line_id);
-
-	return ppa;
-}
-
 static inline u32 pblk_calc_meta_header_crc(struct pblk *pblk,
 					    struct line_header *header)
 {
@@ -1212,10 +1190,10 @@ static inline int pblk_boundary_ppa_checks(struct nvm_tgt_dev *tgt_dev,
 
 		if (!ppa->c.is_cached &&
 				ppa->g.ch < geo->nr_chnls &&
-				ppa->g.lun < geo->luns_per_chnl &&
+				ppa->g.lun < geo->nr_luns &&
 				ppa->g.pl < geo->nr_planes &&
-				ppa->g.blk < geo->blks_per_lun &&
-				ppa->g.pg < geo->pgs_per_blk &&
+				ppa->g.blk < geo->nr_chks &&
+				ppa->g.pg < geo->ws_per_chk &&
 				ppa->g.sec < geo->sec_per_pg)
 			continue;
 
@@ -1245,7 +1223,7 @@ static inline int pblk_check_io(struct pblk *pblk, struct nvm_rq *rqd)
 
 		for (i = 0; i < rqd->nr_ppas; i++) {
 			ppa = ppa_list[i];
-			line = &pblk->lines[pblk_dev_ppa_to_line(ppa)];
+			line = &pblk->lines[pblk_ppa_to_line(ppa)];
 
 			spin_lock(&line->lock);
 			if (line->state != PBLK_LINESTATE_OPEN) {
@@ -1288,11 +1266,6 @@ static inline unsigned int pblk_get_secs(struct bio *bio)
 	return  bio->bi_iter.bi_size / PBLK_EXPOSED_PAGE_SIZE;
 }
 
-static inline sector_t pblk_get_sector(sector_t lba)
-{
-	return lba * NR_PHY_IN_LOG;
-}
-
 static inline void pblk_setup_uuid(struct pblk *pblk)
 {
 	uuid_le uuid;

drivers/lightnvm/rrpc.c

@@ -1,1625 +0,0 @@
-/*
- * Copyright (C) 2015 IT University of Copenhagen
- * Initial release: Matias Bjorling <m@bjorling.me>
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License version
- * 2 as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
- * General Public License for more details.
- *
- * Implementation of a Round-robin page-based Hybrid FTL for Open-channel SSDs.
- */
-
-#include "rrpc.h"
-
-static struct kmem_cache *rrpc_gcb_cache, *rrpc_rq_cache;
-static DECLARE_RWSEM(rrpc_lock);
-
-static int rrpc_submit_io(struct rrpc *rrpc, struct bio *bio,
-				struct nvm_rq *rqd, unsigned long flags);
-
-#define rrpc_for_each_lun(rrpc, rlun, i) \
-		for ((i) = 0, rlun = &(rrpc)->luns[0]; \
-			(i) < (rrpc)->nr_luns; (i)++, rlun = &(rrpc)->luns[(i)])
-
-static void rrpc_page_invalidate(struct rrpc *rrpc, struct rrpc_addr *a)
-{
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	struct rrpc_block *rblk = a->rblk;
-	unsigned int pg_offset;
-
-	lockdep_assert_held(&rrpc->rev_lock);
-
-	if (a->addr == ADDR_EMPTY || !rblk)
-		return;
-
-	spin_lock(&rblk->lock);
-
-	div_u64_rem(a->addr, dev->geo.sec_per_blk, &pg_offset);
-	WARN_ON(test_and_set_bit(pg_offset, rblk->invalid_pages));
-	rblk->nr_invalid_pages++;
-
-	spin_unlock(&rblk->lock);
-
-	rrpc->rev_trans_map[a->addr].addr = ADDR_EMPTY;
-}
-
-static void rrpc_invalidate_range(struct rrpc *rrpc, sector_t slba,
-							unsigned int len)
-{
-	sector_t i;
-
-	spin_lock(&rrpc->rev_lock);
-	for (i = slba; i < slba + len; i++) {
-		struct rrpc_addr *gp = &rrpc->trans_map[i];
-
-		rrpc_page_invalidate(rrpc, gp);
-		gp->rblk = NULL;
-	}
-	spin_unlock(&rrpc->rev_lock);
-}
-
-static struct nvm_rq *rrpc_inflight_laddr_acquire(struct rrpc *rrpc,
-					sector_t laddr, unsigned int pages)
-{
-	struct nvm_rq *rqd;
-	struct rrpc_inflight_rq *inf;
-
-	rqd = mempool_alloc(rrpc->rq_pool, GFP_ATOMIC);
-	if (!rqd)
-		return ERR_PTR(-ENOMEM);
-
-	inf = rrpc_get_inflight_rq(rqd);
-	if (rrpc_lock_laddr(rrpc, laddr, pages, inf)) {
-		mempool_free(rqd, rrpc->rq_pool);
-		return NULL;
-	}
-
-	return rqd;
-}
-
-static void rrpc_inflight_laddr_release(struct rrpc *rrpc, struct nvm_rq *rqd)
-{
-	struct rrpc_inflight_rq *inf = rrpc_get_inflight_rq(rqd);
-
-	rrpc_unlock_laddr(rrpc, inf);
-
-	mempool_free(rqd, rrpc->rq_pool);
-}
-
-static void rrpc_discard(struct rrpc *rrpc, struct bio *bio)
-{
-	sector_t slba = bio->bi_iter.bi_sector / NR_PHY_IN_LOG;
-	sector_t len = bio->bi_iter.bi_size / RRPC_EXPOSED_PAGE_SIZE;
-	struct nvm_rq *rqd;
-
-	while (1) {
-		rqd = rrpc_inflight_laddr_acquire(rrpc, slba, len);
-		if (rqd)
-			break;
-
-		schedule();
-	}
-
-	if (IS_ERR(rqd)) {
-		pr_err("rrpc: unable to acquire inflight IO\n");
-		bio_io_error(bio);
-		return;
-	}
-
-	rrpc_invalidate_range(rrpc, slba, len);
-	rrpc_inflight_laddr_release(rrpc, rqd);
-}
-
-static int block_is_full(struct rrpc *rrpc, struct rrpc_block *rblk)
-{
-	struct nvm_tgt_dev *dev = rrpc->dev;
-
-	return (rblk->next_page == dev->geo.sec_per_blk);
-}
-
-/* Calculate relative addr for the given block, considering instantiated LUNs */
-static u64 block_to_rel_addr(struct rrpc *rrpc, struct rrpc_block *rblk)
-{
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	struct rrpc_lun *rlun = rblk->rlun;
-
-	return rlun->id * dev->geo.sec_per_blk;
-}
-
-static struct ppa_addr rrpc_ppa_to_gaddr(struct nvm_tgt_dev *dev,
-					 struct rrpc_addr *gp)
-{
-	struct rrpc_block *rblk = gp->rblk;
-	struct rrpc_lun *rlun = rblk->rlun;
-	u64 addr = gp->addr;
-	struct ppa_addr paddr;
-
-	paddr.ppa = addr;
-	paddr = rrpc_linear_to_generic_addr(&dev->geo, paddr);
-	paddr.g.ch = rlun->bppa.g.ch;
-	paddr.g.lun = rlun->bppa.g.lun;
-	paddr.g.blk = rblk->id;
-
-	return paddr;
-}
-
-/* requires lun->lock taken */
-static void rrpc_set_lun_cur(struct rrpc_lun *rlun, struct rrpc_block *new_rblk,
-						struct rrpc_block **cur_rblk)
-{
-	struct rrpc *rrpc = rlun->rrpc;
-
-	if (*cur_rblk) {
-		spin_lock(&(*cur_rblk)->lock);
-		WARN_ON(!block_is_full(rrpc, *cur_rblk));
-		spin_unlock(&(*cur_rblk)->lock);
-	}
-	*cur_rblk = new_rblk;
-}
-
-static struct rrpc_block *__rrpc_get_blk(struct rrpc *rrpc,
-							struct rrpc_lun *rlun)
-{
-	struct rrpc_block *rblk = NULL;
-
-	if (list_empty(&rlun->free_list))
-		goto out;
-
-	rblk = list_first_entry(&rlun->free_list, struct rrpc_block, list);
-
-	list_move_tail(&rblk->list, &rlun->used_list);
-	rblk->state = NVM_BLK_ST_TGT;
-	rlun->nr_free_blocks--;
-
-out:
-	return rblk;
-}
-
-static struct rrpc_block *rrpc_get_blk(struct rrpc *rrpc, struct rrpc_lun *rlun,
-							unsigned long flags)
-{
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	struct rrpc_block *rblk;
-	int is_gc = flags & NVM_IOTYPE_GC;
-
-	spin_lock(&rlun->lock);
-	if (!is_gc && rlun->nr_free_blocks < rlun->reserved_blocks) {
-		pr_err("nvm: rrpc: cannot give block to non GC request\n");
-		spin_unlock(&rlun->lock);
-		return NULL;
-	}
-
-	rblk = __rrpc_get_blk(rrpc, rlun);
-	if (!rblk) {
-		pr_err("nvm: rrpc: cannot get new block\n");
-		spin_unlock(&rlun->lock);
-		return NULL;
-	}
-	spin_unlock(&rlun->lock);
-
-	bitmap_zero(rblk->invalid_pages, dev->geo.sec_per_blk);
-	rblk->next_page = 0;
-	rblk->nr_invalid_pages = 0;
-	atomic_set(&rblk->data_cmnt_size, 0);
-
-	return rblk;
-}
-
-static void rrpc_put_blk(struct rrpc *rrpc, struct rrpc_block *rblk)
-{
-	struct rrpc_lun *rlun = rblk->rlun;
-
-	spin_lock(&rlun->lock);
-	if (rblk->state & NVM_BLK_ST_TGT) {
-		list_move_tail(&rblk->list, &rlun->free_list);
-		rlun->nr_free_blocks++;
-		rblk->state = NVM_BLK_ST_FREE;
-	} else if (rblk->state & NVM_BLK_ST_BAD) {
-		list_move_tail(&rblk->list, &rlun->bb_list);
-		rblk->state = NVM_BLK_ST_BAD;
-	} else {
-		WARN_ON_ONCE(1);
-		pr_err("rrpc: erroneous type (ch:%d,lun:%d,blk%d-> %u)\n",
-					rlun->bppa.g.ch, rlun->bppa.g.lun,
-					rblk->id, rblk->state);
-		list_move_tail(&rblk->list, &rlun->bb_list);
-	}
-	spin_unlock(&rlun->lock);
-}
-
-static void rrpc_put_blks(struct rrpc *rrpc)
-{
-	struct rrpc_lun *rlun;
-	int i;
-
-	for (i = 0; i < rrpc->nr_luns; i++) {
-		rlun = &rrpc->luns[i];
-		if (rlun->cur)
-			rrpc_put_blk(rrpc, rlun->cur);
-		if (rlun->gc_cur)
-			rrpc_put_blk(rrpc, rlun->gc_cur);
-	}
-}
-
-static struct rrpc_lun *get_next_lun(struct rrpc *rrpc)
-{
-	int next = atomic_inc_return(&rrpc->next_lun);
-
-	return &rrpc->luns[next % rrpc->nr_luns];
-}
-
-static void rrpc_gc_kick(struct rrpc *rrpc)
-{
-	struct rrpc_lun *rlun;
-	unsigned int i;
-
-	for (i = 0; i < rrpc->nr_luns; i++) {
-		rlun = &rrpc->luns[i];
-		queue_work(rrpc->krqd_wq, &rlun->ws_gc);
-	}
-}
-
-/*
- * timed GC every interval.
- */
-static void rrpc_gc_timer(struct timer_list *t)
-{
-	struct rrpc *rrpc = from_timer(rrpc, t, gc_timer);
-
-	rrpc_gc_kick(rrpc);
-	mod_timer(&rrpc->gc_timer, jiffies + msecs_to_jiffies(10));
-}
-
-static void rrpc_end_sync_bio(struct bio *bio)
-{
-	struct completion *waiting = bio->bi_private;
-
-	if (bio->bi_status)
-		pr_err("nvm: gc request failed (%u).\n", bio->bi_status);
-
-	complete(waiting);
-}
-
-/*
- * rrpc_move_valid_pages -- migrate live data off the block
- * @rrpc: the 'rrpc' structure
- * @block: the block from which to migrate live pages
- *
- * Description:
- *   GC algorithms may call this function to migrate remaining live
- *   pages off the block prior to erasing it. This function blocks
- *   further execution until the operation is complete.
- */
-static int rrpc_move_valid_pages(struct rrpc *rrpc, struct rrpc_block *rblk)
-{
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	struct request_queue *q = dev->q;
-	struct rrpc_rev_addr *rev;
-	struct nvm_rq *rqd;
-	struct bio *bio;
-	struct page *page;
-	int slot;
-	int nr_sec_per_blk = dev->geo.sec_per_blk;
-	u64 phys_addr;
-	DECLARE_COMPLETION_ONSTACK(wait);
-
-	if (bitmap_full(rblk->invalid_pages, nr_sec_per_blk))
-		return 0;
-
-	bio = bio_alloc(GFP_NOIO, 1);
-	if (!bio) {
-		pr_err("nvm: could not alloc bio to gc\n");
-		return -ENOMEM;
-	}
-
-	page = mempool_alloc(rrpc->page_pool, GFP_NOIO);
-
-	while ((slot = find_first_zero_bit(rblk->invalid_pages,
-					    nr_sec_per_blk)) < nr_sec_per_blk) {
-
-		/* Lock laddr */
-		phys_addr = rrpc_blk_to_ppa(rrpc, rblk) + slot;
-
-try:
-		spin_lock(&rrpc->rev_lock);
-		/* Get logical address from physical to logical table */
-		rev = &rrpc->rev_trans_map[phys_addr];
-		/* already updated by previous regular write */
-		if (rev->addr == ADDR_EMPTY) {
-			spin_unlock(&rrpc->rev_lock);
-			continue;
-		}
-
-		rqd = rrpc_inflight_laddr_acquire(rrpc, rev->addr, 1);
-		if (IS_ERR_OR_NULL(rqd)) {
-			spin_unlock(&rrpc->rev_lock);
-			schedule();
-			goto try;
-		}
-
-		spin_unlock(&rrpc->rev_lock);
-
-		/* Perform read to do GC */
-		bio->bi_iter.bi_sector = rrpc_get_sector(rev->addr);
-		bio_set_op_attrs(bio,  REQ_OP_READ, 0);
-		bio->bi_private = &wait;
-		bio->bi_end_io = rrpc_end_sync_bio;
-
-		/* TODO: may fail when EXP_PG_SIZE > PAGE_SIZE */
-		bio_add_pc_page(q, bio, page, RRPC_EXPOSED_PAGE_SIZE, 0);
-
-		if (rrpc_submit_io(rrpc, bio, rqd, NVM_IOTYPE_GC)) {
-			pr_err("rrpc: gc read failed.\n");
-			rrpc_inflight_laddr_release(rrpc, rqd);
-			goto finished;
-		}
-		wait_for_completion_io(&wait);
-		if (bio->bi_status) {
-			rrpc_inflight_laddr_release(rrpc, rqd);
-			goto finished;
-		}
-
-		bio_reset(bio);
-		reinit_completion(&wait);
-
-		bio->bi_iter.bi_sector = rrpc_get_sector(rev->addr);
-		bio_set_op_attrs(bio, REQ_OP_WRITE, 0);
-		bio->bi_private = &wait;
-		bio->bi_end_io = rrpc_end_sync_bio;
-
-		bio_add_pc_page(q, bio, page, RRPC_EXPOSED_PAGE_SIZE, 0);
-
-		/* turn the command around and write the data back to a new
-		 * address
-		 */
-		if (rrpc_submit_io(rrpc, bio, rqd, NVM_IOTYPE_GC)) {
-			pr_err("rrpc: gc write failed.\n");
-			rrpc_inflight_laddr_release(rrpc, rqd);
-			goto finished;
-		}
-		wait_for_completion_io(&wait);
-
-		rrpc_inflight_laddr_release(rrpc, rqd);
-		if (bio->bi_status)
-			goto finished;
-
-		bio_reset(bio);
-	}
-
-finished:
-	mempool_free(page, rrpc->page_pool);
-	bio_put(bio);
-
-	if (!bitmap_full(rblk->invalid_pages, nr_sec_per_blk)) {
-		pr_err("nvm: failed to garbage collect block\n");
-		return -EIO;
-	}
-
-	return 0;
-}
-
-static void rrpc_block_gc(struct work_struct *work)
-{
-	struct rrpc_block_gc *gcb = container_of(work, struct rrpc_block_gc,
-									ws_gc);
-	struct rrpc *rrpc = gcb->rrpc;
-	struct rrpc_block *rblk = gcb->rblk;
-	struct rrpc_lun *rlun = rblk->rlun;
-	struct ppa_addr ppa;
-
-	mempool_free(gcb, rrpc->gcb_pool);
-	pr_debug("nvm: block 'ch:%d,lun:%d,blk:%d' being reclaimed\n",
-			rlun->bppa.g.ch, rlun->bppa.g.lun,
-			rblk->id);
-
-	if (rrpc_move_valid_pages(rrpc, rblk))
-		goto put_back;
-
-	ppa.ppa = 0;
-	ppa.g.ch = rlun->bppa.g.ch;
-	ppa.g.lun = rlun->bppa.g.lun;
-	ppa.g.blk = rblk->id;
-
-	if (nvm_erase_sync(rrpc->dev, &ppa, 1))
-		goto put_back;
-
-	rrpc_put_blk(rrpc, rblk);
-
-	return;
-
-put_back:
-	spin_lock(&rlun->lock);
-	list_add_tail(&rblk->prio, &rlun->prio_list);
-	spin_unlock(&rlun->lock);
-}
-
-/* the block with highest number of invalid pages, will be in the beginning
- * of the list
- */
-static struct rrpc_block *rblk_max_invalid(struct rrpc_block *ra,
-							struct rrpc_block *rb)
-{
-	if (ra->nr_invalid_pages == rb->nr_invalid_pages)
-		return ra;
-
-	return (ra->nr_invalid_pages < rb->nr_invalid_pages) ? rb : ra;
-}
-
-/* linearly find the block with highest number of invalid pages
- * requires lun->lock
- */
-static struct rrpc_block *block_prio_find_max(struct rrpc_lun *rlun)
-{
-	struct list_head *prio_list = &rlun->prio_list;
-	struct rrpc_block *rblk, *max;
-
-	BUG_ON(list_empty(prio_list));
-
-	max = list_first_entry(prio_list, struct rrpc_block, prio);
-	list_for_each_entry(rblk, prio_list, prio)
-		max = rblk_max_invalid(max, rblk);
-
-	return max;
-}
-
-static void rrpc_lun_gc(struct work_struct *work)
-{
-	struct rrpc_lun *rlun = container_of(work, struct rrpc_lun, ws_gc);
-	struct rrpc *rrpc = rlun->rrpc;
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	struct rrpc_block_gc *gcb;
-	unsigned int nr_blocks_need;
-
-	nr_blocks_need = dev->geo.blks_per_lun / GC_LIMIT_INVERSE;
-
-	if (nr_blocks_need < rrpc->nr_luns)
-		nr_blocks_need = rrpc->nr_luns;
-
-	spin_lock(&rlun->lock);
-	while (nr_blocks_need > rlun->nr_free_blocks &&
-					!list_empty(&rlun->prio_list)) {
-		struct rrpc_block *rblk = block_prio_find_max(rlun);
-
-		if (!rblk->nr_invalid_pages)
-			break;
-
-		gcb = mempool_alloc(rrpc->gcb_pool, GFP_ATOMIC);
-		if (!gcb)
-			break;
-
-		list_del_init(&rblk->prio);
-
-		WARN_ON(!block_is_full(rrpc, rblk));
-
-		pr_debug("rrpc: selected block 'ch:%d,lun:%d,blk:%d' for GC\n",
-					rlun->bppa.g.ch, rlun->bppa.g.lun,
-					rblk->id);
-
-		gcb->rrpc = rrpc;
-		gcb->rblk = rblk;
-		INIT_WORK(&gcb->ws_gc, rrpc_block_gc);
-
-		queue_work(rrpc->kgc_wq, &gcb->ws_gc);
-
-		nr_blocks_need--;
-	}
-	spin_unlock(&rlun->lock);
-
-	/* TODO: Hint that request queue can be started again */
-}
-
-static void rrpc_gc_queue(struct work_struct *work)
-{
-	struct rrpc_block_gc *gcb = container_of(work, struct rrpc_block_gc,
-									ws_gc);
-	struct rrpc *rrpc = gcb->rrpc;
-	struct rrpc_block *rblk = gcb->rblk;
-	struct rrpc_lun *rlun = rblk->rlun;
-
-	spin_lock(&rlun->lock);
-	list_add_tail(&rblk->prio, &rlun->prio_list);
-	spin_unlock(&rlun->lock);
-
-	mempool_free(gcb, rrpc->gcb_pool);
-	pr_debug("nvm: block 'ch:%d,lun:%d,blk:%d' full, allow GC (sched)\n",
-					rlun->bppa.g.ch, rlun->bppa.g.lun,
-					rblk->id);
-}
-
-static const struct block_device_operations rrpc_fops = {
-	.owner		= THIS_MODULE,
-};
-
-static struct rrpc_lun *rrpc_get_lun_rr(struct rrpc *rrpc, int is_gc)
-{
-	unsigned int i;
-	struct rrpc_lun *rlun, *max_free;
-
-	if (!is_gc)
-		return get_next_lun(rrpc);
-
-	/* during GC, we don't care about RR, instead we want to make
-	 * sure that we maintain evenness between the block luns.
-	 */
-	max_free = &rrpc->luns[0];
-	/* prevent GC-ing lun from devouring pages of a lun with
-	 * little free blocks. We don't take the lock as we only need an
-	 * estimate.
-	 */
-	rrpc_for_each_lun(rrpc, rlun, i) {
-		if (rlun->nr_free_blocks > max_free->nr_free_blocks)
-			max_free = rlun;
-	}
-
-	return max_free;
-}
-
-static struct rrpc_addr *rrpc_update_map(struct rrpc *rrpc, sector_t laddr,
-					struct rrpc_block *rblk, u64 paddr)
-{
-	struct rrpc_addr *gp;
-	struct rrpc_rev_addr *rev;
-
-	BUG_ON(laddr >= rrpc->nr_sects);
-
-	gp = &rrpc->trans_map[laddr];
-	spin_lock(&rrpc->rev_lock);
-	if (gp->rblk)
-		rrpc_page_invalidate(rrpc, gp);
-
-	gp->addr = paddr;
-	gp->rblk = rblk;
-
-	rev = &rrpc->rev_trans_map[gp->addr];
-	rev->addr = laddr;
-	spin_unlock(&rrpc->rev_lock);
-
-	return gp;
-}
-
-static u64 rrpc_alloc_addr(struct rrpc *rrpc, struct rrpc_block *rblk)
-{
-	u64 addr = ADDR_EMPTY;
-
-	spin_lock(&rblk->lock);
-	if (block_is_full(rrpc, rblk))
-		goto out;
-
-	addr = rblk->next_page;
-
-	rblk->next_page++;
-out:
-	spin_unlock(&rblk->lock);
-	return addr;
-}
-
-/* Map logical address to a physical page. The mapping implements a round robin
- * approach and allocates a page from the next lun available.
- *
- * Returns rrpc_addr with the physical address and block. Returns NULL if no
- * blocks in the next rlun are available.
- */
-static struct ppa_addr rrpc_map_page(struct rrpc *rrpc, sector_t laddr,
-								int is_gc)
-{
-	struct nvm_tgt_dev *tgt_dev = rrpc->dev;
-	struct rrpc_lun *rlun;
-	struct rrpc_block *rblk, **cur_rblk;
-	struct rrpc_addr *p;
-	struct ppa_addr ppa;
-	u64 paddr;
-	int gc_force = 0;
-
-	ppa.ppa = ADDR_EMPTY;
-	rlun = rrpc_get_lun_rr(rrpc, is_gc);
-
-	if (!is_gc && rlun->nr_free_blocks < rrpc->nr_luns * 4)
-		return ppa;
-
-	/*
-	 * page allocation steps:
-	 * 1. Try to allocate new page from current rblk
-	 * 2a. If succeed, proceed to map it in and return
-	 * 2b. If fail, first try to allocate a new block from media manger,
-	 *     and then retry step 1. Retry until the normal block pool is
-	 *     exhausted.
-	 * 3. If exhausted, and garbage collector is requesting the block,
-	 *    go to the reserved block and retry step 1.
-	 *    In the case that this fails as well, or it is not GC
-	 *    requesting, report not able to retrieve a block and let the
-	 *    caller handle further processing.
-	 */
-
-	spin_lock(&rlun->lock);
-	cur_rblk = &rlun->cur;
-	rblk = rlun->cur;
-retry:
-	paddr = rrpc_alloc_addr(rrpc, rblk);
-
-	if (paddr != ADDR_EMPTY)
-		goto done;
-
-	if (!list_empty(&rlun->wblk_list)) {
-new_blk:
-		rblk = list_first_entry(&rlun->wblk_list, struct rrpc_block,
-									prio);
-		rrpc_set_lun_cur(rlun, rblk, cur_rblk);
-		list_del(&rblk->prio);
-		goto retry;
-	}
-	spin_unlock(&rlun->lock);
-
-	rblk = rrpc_get_blk(rrpc, rlun, gc_force);
-	if (rblk) {
-		spin_lock(&rlun->lock);
-		list_add_tail(&rblk->prio, &rlun->wblk_list);
-		/*
-		 * another thread might already have added a new block,
-		 * Therefore, make sure that one is used, instead of the
-		 * one just added.
-		 */
-		goto new_blk;
-	}
-
-	if (unlikely(is_gc) && !gc_force) {
-		/* retry from emergency gc block */
-		cur_rblk = &rlun->gc_cur;
-		rblk = rlun->gc_cur;
-		gc_force = 1;
-		spin_lock(&rlun->lock);
-		goto retry;
-	}
-
-	pr_err("rrpc: failed to allocate new block\n");
-	return ppa;
-done:
-	spin_unlock(&rlun->lock);
-	p = rrpc_update_map(rrpc, laddr, rblk, paddr);
-	if (!p)
-		return ppa;
-
-	/* return global address */
-	return rrpc_ppa_to_gaddr(tgt_dev, p);
-}
-
-static void rrpc_run_gc(struct rrpc *rrpc, struct rrpc_block *rblk)
-{
-	struct rrpc_block_gc *gcb;
-
-	gcb = mempool_alloc(rrpc->gcb_pool, GFP_ATOMIC);
-	if (!gcb) {
-		pr_err("rrpc: unable to queue block for gc.");
-		return;
-	}
-
-	gcb->rrpc = rrpc;
-	gcb->rblk = rblk;
-
-	INIT_WORK(&gcb->ws_gc, rrpc_gc_queue);
-	queue_work(rrpc->kgc_wq, &gcb->ws_gc);
-}
-
-static struct rrpc_lun *rrpc_ppa_to_lun(struct rrpc *rrpc, struct ppa_addr p)
-{
-	struct rrpc_lun *rlun = NULL;
-	int i;
-
-	for (i = 0; i < rrpc->nr_luns; i++) {
-		if (rrpc->luns[i].bppa.g.ch == p.g.ch &&
-				rrpc->luns[i].bppa.g.lun == p.g.lun) {
-			rlun = &rrpc->luns[i];
-			break;
-		}
-	}
-
-	return rlun;
-}
-
-static void __rrpc_mark_bad_block(struct rrpc *rrpc, struct ppa_addr ppa)
-{
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	struct rrpc_lun *rlun;
-	struct rrpc_block *rblk;
-
-	rlun = rrpc_ppa_to_lun(rrpc, ppa);
-	rblk = &rlun->blocks[ppa.g.blk];
-	rblk->state = NVM_BLK_ST_BAD;
-
-	nvm_set_tgt_bb_tbl(dev, &ppa, 1, NVM_BLK_T_GRWN_BAD);
-}
-
-static void rrpc_mark_bad_block(struct rrpc *rrpc, struct nvm_rq *rqd)
-{
-	void *comp_bits = &rqd->ppa_status;
-	struct ppa_addr ppa, prev_ppa;
-	int nr_ppas = rqd->nr_ppas;
-	int bit;
-
-	if (rqd->nr_ppas == 1)
-		__rrpc_mark_bad_block(rrpc, rqd->ppa_addr);
-
-	ppa_set_empty(&prev_ppa);
-	bit = -1;
-	while ((bit = find_next_bit(comp_bits, nr_ppas, bit + 1)) < nr_ppas) {
-		ppa = rqd->ppa_list[bit];
-		if (ppa_cmp_blk(ppa, prev_ppa))
-			continue;
-
-		__rrpc_mark_bad_block(rrpc, ppa);
-	}
-}
-
-static void rrpc_end_io_write(struct rrpc *rrpc, struct rrpc_rq *rrqd,
-						sector_t laddr, uint8_t npages)
-{
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	struct rrpc_addr *p;
-	struct rrpc_block *rblk;
-	int cmnt_size, i;
-
-	for (i = 0; i < npages; i++) {
-		p = &rrpc->trans_map[laddr + i];
-		rblk = p->rblk;
-
-		cmnt_size = atomic_inc_return(&rblk->data_cmnt_size);
-		if (unlikely(cmnt_size == dev->geo.sec_per_blk))
-			rrpc_run_gc(rrpc, rblk);
-	}
-}
-
-static void rrpc_end_io(struct nvm_rq *rqd)
-{
-	struct rrpc *rrpc = rqd->private;
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	struct rrpc_rq *rrqd = nvm_rq_to_pdu(rqd);
-	uint8_t npages = rqd->nr_ppas;
-	sector_t laddr = rrpc_get_laddr(rqd->bio) - npages;
-
-	if (bio_data_dir(rqd->bio) == WRITE) {
-		if (rqd->error == NVM_RSP_ERR_FAILWRITE)
-			rrpc_mark_bad_block(rrpc, rqd);
-
-		rrpc_end_io_write(rrpc, rrqd, laddr, npages);
-	}
-
-	bio_put(rqd->bio);
-
-	if (rrqd->flags & NVM_IOTYPE_GC)
-		return;
-
-	rrpc_unlock_rq(rrpc, rqd);
-
-	if (npages > 1)
-		nvm_dev_dma_free(dev->parent, rqd->ppa_list, rqd->dma_ppa_list);
-
-	mempool_free(rqd, rrpc->rq_pool);
-}
-
-static int rrpc_read_ppalist_rq(struct rrpc *rrpc, struct bio *bio,
-			struct nvm_rq *rqd, unsigned long flags, int npages)
-{
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	struct rrpc_inflight_rq *r = rrpc_get_inflight_rq(rqd);
-	struct rrpc_addr *gp;
-	sector_t laddr = rrpc_get_laddr(bio);
-	int is_gc = flags & NVM_IOTYPE_GC;
-	int i;
-
-	if (!is_gc && rrpc_lock_rq(rrpc, bio, rqd)) {
-		nvm_dev_dma_free(dev->parent, rqd->ppa_list, rqd->dma_ppa_list);
-		return NVM_IO_REQUEUE;
-	}
-
-	for (i = 0; i < npages; i++) {
-		/* We assume that mapping occurs at 4KB granularity */
-		BUG_ON(!(laddr + i < rrpc->nr_sects));
-		gp = &rrpc->trans_map[laddr + i];
-
-		if (gp->rblk) {
-			rqd->ppa_list[i] = rrpc_ppa_to_gaddr(dev, gp);
-		} else {
-			BUG_ON(is_gc);
-			rrpc_unlock_laddr(rrpc, r);
-			nvm_dev_dma_free(dev->parent, rqd->ppa_list,
-							rqd->dma_ppa_list);
-			return NVM_IO_DONE;
-		}
-	}
-
-	rqd->opcode = NVM_OP_HBREAD;
-
-	return NVM_IO_OK;
-}
-
-static int rrpc_read_rq(struct rrpc *rrpc, struct bio *bio, struct nvm_rq *rqd,
-							unsigned long flags)
-{
-	int is_gc = flags & NVM_IOTYPE_GC;
-	sector_t laddr = rrpc_get_laddr(bio);
-	struct rrpc_addr *gp;
-
-	if (!is_gc && rrpc_lock_rq(rrpc, bio, rqd))
-		return NVM_IO_REQUEUE;
-
-	BUG_ON(!(laddr < rrpc->nr_sects));
-	gp = &rrpc->trans_map[laddr];
-
-	if (gp->rblk) {
-		rqd->ppa_addr = rrpc_ppa_to_gaddr(rrpc->dev, gp);
-	} else {
-		BUG_ON(is_gc);
-		rrpc_unlock_rq(rrpc, rqd);
-		return NVM_IO_DONE;
-	}
-
-	rqd->opcode = NVM_OP_HBREAD;
-
-	return NVM_IO_OK;
-}
-
-static int rrpc_write_ppalist_rq(struct rrpc *rrpc, struct bio *bio,
-			struct nvm_rq *rqd, unsigned long flags, int npages)
-{
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	struct rrpc_inflight_rq *r = rrpc_get_inflight_rq(rqd);
-	struct ppa_addr p;
-	sector_t laddr = rrpc_get_laddr(bio);
-	int is_gc = flags & NVM_IOTYPE_GC;
-	int i;
-
-	if (!is_gc && rrpc_lock_rq(rrpc, bio, rqd)) {
-		nvm_dev_dma_free(dev->parent, rqd->ppa_list, rqd->dma_ppa_list);
-		return NVM_IO_REQUEUE;
-	}
-
-	for (i = 0; i < npages; i++) {
-		/* We assume that mapping occurs at 4KB granularity */
-		p = rrpc_map_page(rrpc, laddr + i, is_gc);
-		if (p.ppa == ADDR_EMPTY) {
-			BUG_ON(is_gc);
-			rrpc_unlock_laddr(rrpc, r);
-			nvm_dev_dma_free(dev->parent, rqd->ppa_list,
-							rqd->dma_ppa_list);
-			rrpc_gc_kick(rrpc);
-			return NVM_IO_REQUEUE;
-		}
-
-		rqd->ppa_list[i] = p;
-	}
-
-	rqd->opcode = NVM_OP_HBWRITE;
-
-	return NVM_IO_OK;
-}
-
-static int rrpc_write_rq(struct rrpc *rrpc, struct bio *bio,
-				struct nvm_rq *rqd, unsigned long flags)
-{
-	struct ppa_addr p;
-	int is_gc = flags & NVM_IOTYPE_GC;
-	sector_t laddr = rrpc_get_laddr(bio);
-
-	if (!is_gc && rrpc_lock_rq(rrpc, bio, rqd))
-		return NVM_IO_REQUEUE;
-
-	p = rrpc_map_page(rrpc, laddr, is_gc);
-	if (p.ppa == ADDR_EMPTY) {
-		BUG_ON(is_gc);
-		rrpc_unlock_rq(rrpc, rqd);
-		rrpc_gc_kick(rrpc);
-		return NVM_IO_REQUEUE;
-	}
-
-	rqd->ppa_addr = p;
-	rqd->opcode = NVM_OP_HBWRITE;
-
-	return NVM_IO_OK;
-}
-
-static int rrpc_setup_rq(struct rrpc *rrpc, struct bio *bio,
-			struct nvm_rq *rqd, unsigned long flags, uint8_t npages)
-{
-	struct nvm_tgt_dev *dev = rrpc->dev;
-
-	if (npages > 1) {
-		rqd->ppa_list = nvm_dev_dma_alloc(dev->parent, GFP_KERNEL,
-							&rqd->dma_ppa_list);
-		if (!rqd->ppa_list) {
-			pr_err("rrpc: not able to allocate ppa list\n");
-			return NVM_IO_ERR;
-		}
-
-		if (bio_op(bio) == REQ_OP_WRITE)
-			return rrpc_write_ppalist_rq(rrpc, bio, rqd, flags,
-									npages);
-
-		return rrpc_read_ppalist_rq(rrpc, bio, rqd, flags, npages);
-	}
-
-	if (bio_op(bio) == REQ_OP_WRITE)
-		return rrpc_write_rq(rrpc, bio, rqd, flags);
-
-	return rrpc_read_rq(rrpc, bio, rqd, flags);
-}
-
-static int rrpc_submit_io(struct rrpc *rrpc, struct bio *bio,
-				struct nvm_rq *rqd, unsigned long flags)
-{
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	struct rrpc_rq *rrq = nvm_rq_to_pdu(rqd);
-	uint8_t nr_pages = rrpc_get_pages(bio);
-	int bio_size = bio_sectors(bio) << 9;
-	int err;
-
-	if (bio_size < dev->geo.sec_size)
-		return NVM_IO_ERR;
-	else if (bio_size > dev->geo.max_rq_size)
-		return NVM_IO_ERR;
-
-	err = rrpc_setup_rq(rrpc, bio, rqd, flags, nr_pages);
-	if (err)
-		return err;
-
-	bio_get(bio);
-	rqd->bio = bio;
-	rqd->private = rrpc;
-	rqd->nr_ppas = nr_pages;
-	rqd->end_io = rrpc_end_io;
-	rrq->flags = flags;
-
-	err = nvm_submit_io(dev, rqd);
-	if (err) {
-		pr_err("rrpc: I/O submission failed: %d\n", err);
-		bio_put(bio);
-		if (!(flags & NVM_IOTYPE_GC)) {
-			rrpc_unlock_rq(rrpc, rqd);
-			if (rqd->nr_ppas > 1)
-				nvm_dev_dma_free(dev->parent, rqd->ppa_list,
-							rqd->dma_ppa_list);
-		}
-		return NVM_IO_ERR;
-	}
-
-	return NVM_IO_OK;
-}
-
-static blk_qc_t rrpc_make_rq(struct request_queue *q, struct bio *bio)
-{
-	struct rrpc *rrpc = q->queuedata;
-	struct nvm_rq *rqd;
-	int err;
-
-	blk_queue_split(q, &bio);
-
-	if (bio_op(bio) == REQ_OP_DISCARD) {
-		rrpc_discard(rrpc, bio);
-		return BLK_QC_T_NONE;
-	}
-
-	rqd = mempool_alloc(rrpc->rq_pool, GFP_KERNEL);
-	memset(rqd, 0, sizeof(struct nvm_rq));
-
-	err = rrpc_submit_io(rrpc, bio, rqd, NVM_IOTYPE_NONE);
-	switch (err) {
-	case NVM_IO_OK:
-		return BLK_QC_T_NONE;
-	case NVM_IO_ERR:
-		bio_io_error(bio);
-		break;
-	case NVM_IO_DONE:
-		bio_endio(bio);
-		break;
-	case NVM_IO_REQUEUE:
-		spin_lock(&rrpc->bio_lock);
-		bio_list_add(&rrpc->requeue_bios, bio);
-		spin_unlock(&rrpc->bio_lock);
-		queue_work(rrpc->kgc_wq, &rrpc->ws_requeue);
-		break;
-	}
-
-	mempool_free(rqd, rrpc->rq_pool);
-	return BLK_QC_T_NONE;
-}
-
-static void rrpc_requeue(struct work_struct *work)
-{
-	struct rrpc *rrpc = container_of(work, struct rrpc, ws_requeue);
-	struct bio_list bios;
-	struct bio *bio;
-
-	bio_list_init(&bios);
-
-	spin_lock(&rrpc->bio_lock);
-	bio_list_merge(&bios, &rrpc->requeue_bios);
-	bio_list_init(&rrpc->requeue_bios);
-	spin_unlock(&rrpc->bio_lock);
-
-	while ((bio = bio_list_pop(&bios)))
-		rrpc_make_rq(rrpc->disk->queue, bio);
-}
-
-static void rrpc_gc_free(struct rrpc *rrpc)
-{
-	if (rrpc->krqd_wq)
-		destroy_workqueue(rrpc->krqd_wq);
-
-	if (rrpc->kgc_wq)
-		destroy_workqueue(rrpc->kgc_wq);
-}
-
-static int rrpc_gc_init(struct rrpc *rrpc)
-{
-	rrpc->krqd_wq = alloc_workqueue("rrpc-lun", WQ_MEM_RECLAIM|WQ_UNBOUND,
-								rrpc->nr_luns);
-	if (!rrpc->krqd_wq)
-		return -ENOMEM;
-
-	rrpc->kgc_wq = alloc_workqueue("rrpc-bg", WQ_MEM_RECLAIM, 1);
-	if (!rrpc->kgc_wq)
-		return -ENOMEM;
-
-	timer_setup(&rrpc->gc_timer, rrpc_gc_timer, 0);
-
-	return 0;
-}
-
-static void rrpc_map_free(struct rrpc *rrpc)
-{
-	vfree(rrpc->rev_trans_map);
-	vfree(rrpc->trans_map);
-}
-
-static int rrpc_l2p_update(u64 slba, u32 nlb, __le64 *entries, void *private)
-{
-	struct rrpc *rrpc = (struct rrpc *)private;
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	struct rrpc_addr *addr = rrpc->trans_map + slba;
-	struct rrpc_rev_addr *raddr = rrpc->rev_trans_map;
-	struct rrpc_lun *rlun;
-	struct rrpc_block *rblk;
-	u64 i;
-
-	for (i = 0; i < nlb; i++) {
-		struct ppa_addr gaddr;
-		u64 pba = le64_to_cpu(entries[i]);
-		unsigned int mod;
-
-		/* LNVM treats address-spaces as silos, LBA and PBA are
-		 * equally large and zero-indexed.
-		 */
-		if (unlikely(pba >= dev->total_secs && pba != U64_MAX)) {
-			pr_err("nvm: L2P data entry is out of bounds!\n");
-			pr_err("nvm: Maybe loaded an old target L2P\n");
-			return -EINVAL;
-		}
-
-		/* Address zero is a special one. The first page on a disk is
-		 * protected. As it often holds internal device boot
-		 * information.
-		 */
-		if (!pba)
-			continue;
-
-		div_u64_rem(pba, rrpc->nr_sects, &mod);
-
-		gaddr = rrpc_recov_addr(dev, pba);
-		rlun = rrpc_ppa_to_lun(rrpc, gaddr);
-		if (!rlun) {
-			pr_err("rrpc: l2p corruption on lba %llu\n",
-							slba + i);
-			return -EINVAL;
-		}
-
-		rblk = &rlun->blocks[gaddr.g.blk];
-		if (!rblk->state) {
-			/* at this point, we don't know anything about the
-			 * block. It's up to the FTL on top to re-etablish the
-			 * block state. The block is assumed to be open.
-			 */
-			list_move_tail(&rblk->list, &rlun->used_list);
-			rblk->state = NVM_BLK_ST_TGT;
-			rlun->nr_free_blocks--;
-		}
-
-		addr[i].addr = pba;
-		addr[i].rblk = rblk;
-		raddr[mod].addr = slba + i;
-	}
-
-	return 0;
-}
-
-static int rrpc_map_init(struct rrpc *rrpc)
-{
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	sector_t i;
-	int ret;
-
-	rrpc->trans_map = vzalloc(sizeof(struct rrpc_addr) * rrpc->nr_sects);
-	if (!rrpc->trans_map)
-		return -ENOMEM;
-
-	rrpc->rev_trans_map = vmalloc(sizeof(struct rrpc_rev_addr)
-							* rrpc->nr_sects);
-	if (!rrpc->rev_trans_map)
-		return -ENOMEM;
-
-	for (i = 0; i < rrpc->nr_sects; i++) {
-		struct rrpc_addr *p = &rrpc->trans_map[i];
-		struct rrpc_rev_addr *r = &rrpc->rev_trans_map[i];
-
-		p->addr = ADDR_EMPTY;
-		r->addr = ADDR_EMPTY;
-	}
-
-	/* Bring up the mapping table from device */
-	ret = nvm_get_l2p_tbl(dev, rrpc->soffset, rrpc->nr_sects,
-							rrpc_l2p_update, rrpc);
-	if (ret) {
-		pr_err("nvm: rrpc: could not read L2P table.\n");
-		return -EINVAL;
-	}
-
-	return 0;
-}
-
-/* Minimum pages needed within a lun */
-#define PAGE_POOL_SIZE 16
-#define ADDR_POOL_SIZE 64
-
-static int rrpc_core_init(struct rrpc *rrpc)
-{
-	down_write(&rrpc_lock);
-	if (!rrpc_gcb_cache) {
-		rrpc_gcb_cache = kmem_cache_create("rrpc_gcb",
-				sizeof(struct rrpc_block_gc), 0, 0, NULL);
-		if (!rrpc_gcb_cache) {
-			up_write(&rrpc_lock);
-			return -ENOMEM;
-		}
-
-		rrpc_rq_cache = kmem_cache_create("rrpc_rq",
-				sizeof(struct nvm_rq) + sizeof(struct rrpc_rq),
-				0, 0, NULL);
-		if (!rrpc_rq_cache) {
-			kmem_cache_destroy(rrpc_gcb_cache);
-			up_write(&rrpc_lock);
-			return -ENOMEM;
-		}
-	}
-	up_write(&rrpc_lock);
-
-	rrpc->page_pool = mempool_create_page_pool(PAGE_POOL_SIZE, 0);
-	if (!rrpc->page_pool)
-		return -ENOMEM;
-
-	rrpc->gcb_pool = mempool_create_slab_pool(rrpc->dev->geo.nr_luns,
-								rrpc_gcb_cache);
-	if (!rrpc->gcb_pool)
-		return -ENOMEM;
-
-	rrpc->rq_pool = mempool_create_slab_pool(64, rrpc_rq_cache);
-	if (!rrpc->rq_pool)
-		return -ENOMEM;
-
-	spin_lock_init(&rrpc->inflights.lock);
-	INIT_LIST_HEAD(&rrpc->inflights.reqs);
-
-	return 0;
-}
-
-static void rrpc_core_free(struct rrpc *rrpc)
-{
-	mempool_destroy(rrpc->page_pool);
-	mempool_destroy(rrpc->gcb_pool);
-	mempool_destroy(rrpc->rq_pool);
-}
-
-static void rrpc_luns_free(struct rrpc *rrpc)
-{
-	struct rrpc_lun *rlun;
-	int i;
-
-	if (!rrpc->luns)
-		return;
-
-	for (i = 0; i < rrpc->nr_luns; i++) {
-		rlun = &rrpc->luns[i];
-		vfree(rlun->blocks);
-	}
-
-	kfree(rrpc->luns);
-}
-
-static int rrpc_bb_discovery(struct nvm_tgt_dev *dev, struct rrpc_lun *rlun)
-{
-	struct nvm_geo *geo = &dev->geo;
-	struct rrpc_block *rblk;
-	struct ppa_addr ppa;
-	u8 *blks;
-	int nr_blks;
-	int i;
-	int ret;
-
-	if (!dev->parent->ops->get_bb_tbl)
-		return 0;
-
-	nr_blks = geo->blks_per_lun * geo->plane_mode;
-	blks = kmalloc(nr_blks, GFP_KERNEL);
-	if (!blks)
-		return -ENOMEM;
-
-	ppa.ppa = 0;
-	ppa.g.ch = rlun->bppa.g.ch;
-	ppa.g.lun = rlun->bppa.g.lun;
-
-	ret = nvm_get_tgt_bb_tbl(dev, ppa, blks);
-	if (ret) {
-		pr_err("rrpc: could not get BB table\n");
-		goto out;
-	}
-
-	nr_blks = nvm_bb_tbl_fold(dev->parent, blks, nr_blks);
-	if (nr_blks < 0) {
-		ret = nr_blks;
-		goto out;
-	}
-
-	for (i = 0; i < nr_blks; i++) {
-		if (blks[i] == NVM_BLK_T_FREE)
-			continue;
-
-		rblk = &rlun->blocks[i];
-		list_move_tail(&rblk->list, &rlun->bb_list);
-		rblk->state = NVM_BLK_ST_BAD;
-		rlun->nr_free_blocks--;
-	}
-
-out:
-	kfree(blks);
-	return ret;
-}
-
-static void rrpc_set_lun_ppa(struct rrpc_lun *rlun, struct ppa_addr ppa)
-{
-	rlun->bppa.ppa = 0;
-	rlun->bppa.g.ch = ppa.g.ch;
-	rlun->bppa.g.lun = ppa.g.lun;
-}
-
-static int rrpc_luns_init(struct rrpc *rrpc, struct ppa_addr *luns)
-{
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	struct nvm_geo *geo = &dev->geo;
-	struct rrpc_lun *rlun;
-	int i, j, ret = -EINVAL;
-
-	if (geo->sec_per_blk > MAX_INVALID_PAGES_STORAGE * BITS_PER_LONG) {
-		pr_err("rrpc: number of pages per block too high.");
-		return -EINVAL;
-	}
-
-	spin_lock_init(&rrpc->rev_lock);
-
-	rrpc->luns = kcalloc(rrpc->nr_luns, sizeof(struct rrpc_lun),
-								GFP_KERNEL);
-	if (!rrpc->luns)
-		return -ENOMEM;
-
-	/* 1:1 mapping */
-	for (i = 0; i < rrpc->nr_luns; i++) {
-		rlun = &rrpc->luns[i];
-		rlun->id = i;
-		rrpc_set_lun_ppa(rlun, luns[i]);
-		rlun->blocks = vzalloc(sizeof(struct rrpc_block) *
-							geo->blks_per_lun);
-		if (!rlun->blocks) {
-			ret = -ENOMEM;
-			goto err;
-		}
-
-		INIT_LIST_HEAD(&rlun->free_list);
-		INIT_LIST_HEAD(&rlun->used_list);
-		INIT_LIST_HEAD(&rlun->bb_list);
-
-		for (j = 0; j < geo->blks_per_lun; j++) {
-			struct rrpc_block *rblk = &rlun->blocks[j];
-
-			rblk->id = j;
-			rblk->rlun = rlun;
-			rblk->state = NVM_BLK_T_FREE;
-			INIT_LIST_HEAD(&rblk->prio);
-			INIT_LIST_HEAD(&rblk->list);
-			spin_lock_init(&rblk->lock);
-
-			list_add_tail(&rblk->list, &rlun->free_list);
-		}
-
-		rlun->rrpc = rrpc;
-		rlun->nr_free_blocks = geo->blks_per_lun;
-		rlun->reserved_blocks = 2; /* for GC only */
-
-		INIT_LIST_HEAD(&rlun->prio_list);
-		INIT_LIST_HEAD(&rlun->wblk_list);
-
-		INIT_WORK(&rlun->ws_gc, rrpc_lun_gc);
-		spin_lock_init(&rlun->lock);
-
-		if (rrpc_bb_discovery(dev, rlun))
-			goto err;
-
-	}
-
-	return 0;
-err:
-	return ret;
-}
-
-/* returns 0 on success and stores the beginning address in *begin */
-static int rrpc_area_init(struct rrpc *rrpc, sector_t *begin)
-{
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	sector_t size = rrpc->nr_sects * dev->geo.sec_size;
-	int ret;
-
-	size >>= 9;
-
-	ret = nvm_get_area(dev, begin, size);
-	if (!ret)
-		*begin >>= (ilog2(dev->geo.sec_size) - 9);
-
-	return ret;
-}
-
-static void rrpc_area_free(struct rrpc *rrpc)
-{
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	sector_t begin = rrpc->soffset << (ilog2(dev->geo.sec_size) - 9);
-
-	nvm_put_area(dev, begin);
-}
-
-static void rrpc_free(struct rrpc *rrpc)
-{
-	rrpc_gc_free(rrpc);
-	rrpc_map_free(rrpc);
-	rrpc_core_free(rrpc);
-	rrpc_luns_free(rrpc);
-	rrpc_area_free(rrpc);
-
-	kfree(rrpc);
-}
-
-static void rrpc_exit(void *private)
-{
-	struct rrpc *rrpc = private;
-
-	del_timer(&rrpc->gc_timer);
-
-	flush_workqueue(rrpc->krqd_wq);
-	flush_workqueue(rrpc->kgc_wq);
-
-	rrpc_free(rrpc);
-}
-
-static sector_t rrpc_capacity(void *private)
-{
-	struct rrpc *rrpc = private;
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	sector_t reserved, provisioned;
-
-	/* cur, gc, and two emergency blocks for each lun */
-	reserved = rrpc->nr_luns * dev->geo.sec_per_blk * 4;
-	provisioned = rrpc->nr_sects - reserved;
-
-	if (reserved > rrpc->nr_sects) {
-		pr_err("rrpc: not enough space available to expose storage.\n");
-		return 0;
-	}
-
-	sector_div(provisioned, 10);
-	return provisioned * 9 * NR_PHY_IN_LOG;
-}
-
-/*
- * Looks up the logical address from reverse trans map and check if its valid by
- * comparing the logical to physical address with the physical address.
- * Returns 0 on free, otherwise 1 if in use
- */
-static void rrpc_block_map_update(struct rrpc *rrpc, struct rrpc_block *rblk)
-{
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	int offset;
-	struct rrpc_addr *laddr;
-	u64 bpaddr, paddr, pladdr;
-
-	bpaddr = block_to_rel_addr(rrpc, rblk);
-	for (offset = 0; offset < dev->geo.sec_per_blk; offset++) {
-		paddr = bpaddr + offset;
-
-		pladdr = rrpc->rev_trans_map[paddr].addr;
-		if (pladdr == ADDR_EMPTY)
-			continue;
-
-		laddr = &rrpc->trans_map[pladdr];
-
-		if (paddr == laddr->addr) {
-			laddr->rblk = rblk;
-		} else {
-			set_bit(offset, rblk->invalid_pages);
-			rblk->nr_invalid_pages++;
-		}
-	}
-}
-
-static int rrpc_blocks_init(struct rrpc *rrpc)
-{
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	struct rrpc_lun *rlun;
-	struct rrpc_block *rblk;
-	int lun_iter, blk_iter;
-
-	for (lun_iter = 0; lun_iter < rrpc->nr_luns; lun_iter++) {
-		rlun = &rrpc->luns[lun_iter];
-
-		for (blk_iter = 0; blk_iter < dev->geo.blks_per_lun;
-								blk_iter++) {
-			rblk = &rlun->blocks[blk_iter];
-			rrpc_block_map_update(rrpc, rblk);
-		}
-	}
-
-	return 0;
-}
-
-static int rrpc_luns_configure(struct rrpc *rrpc)
-{
-	struct rrpc_lun *rlun;
-	struct rrpc_block *rblk;
-	int i;
-
-	for (i = 0; i < rrpc->nr_luns; i++) {
-		rlun = &rrpc->luns[i];
-
-		rblk = rrpc_get_blk(rrpc, rlun, 0);
-		if (!rblk)
-			goto err;
-		rrpc_set_lun_cur(rlun, rblk, &rlun->cur);
-
-		/* Emergency gc block */
-		rblk = rrpc_get_blk(rrpc, rlun, 1);
-		if (!rblk)
-			goto err;
-		rrpc_set_lun_cur(rlun, rblk, &rlun->gc_cur);
-	}
-
-	return 0;
-err:
-	rrpc_put_blks(rrpc);
-	return -EINVAL;
-}
-
-static struct nvm_tgt_type tt_rrpc;
-
-static void *rrpc_init(struct nvm_tgt_dev *dev, struct gendisk *tdisk,
-		       int flags)
-{
-	struct request_queue *bqueue = dev->q;
-	struct request_queue *tqueue = tdisk->queue;
-	struct nvm_geo *geo = &dev->geo;
-	struct rrpc *rrpc;
-	sector_t soffset;
-	int ret;
-
-	if (!(dev->identity.dom & NVM_RSP_L2P)) {
-		pr_err("nvm: rrpc: device does not support l2p (%x)\n",
-							dev->identity.dom);
-		return ERR_PTR(-EINVAL);
-	}
-
-	rrpc = kzalloc(sizeof(struct rrpc), GFP_KERNEL);
-	if (!rrpc)
-		return ERR_PTR(-ENOMEM);
-
-	rrpc->dev = dev;
-	rrpc->disk = tdisk;
-
-	bio_list_init(&rrpc->requeue_bios);
-	spin_lock_init(&rrpc->bio_lock);
-	INIT_WORK(&rrpc->ws_requeue, rrpc_requeue);
-
-	rrpc->nr_luns = geo->nr_luns;
-	rrpc->nr_sects = (unsigned long long)geo->sec_per_lun * rrpc->nr_luns;
-
-	/* simple round-robin strategy */
-	atomic_set(&rrpc->next_lun, -1);
-
-	ret = rrpc_area_init(rrpc, &soffset);
-	if (ret < 0) {
-		pr_err("nvm: rrpc: could not initialize area\n");
-		return ERR_PTR(ret);
-	}
-	rrpc->soffset = soffset;
-
-	ret = rrpc_luns_init(rrpc, dev->luns);
-	if (ret) {
-		pr_err("nvm: rrpc: could not initialize luns\n");
-		goto err;
-	}
-
-	ret = rrpc_core_init(rrpc);
-	if (ret) {
-		pr_err("nvm: rrpc: could not initialize core\n");
-		goto err;
-	}
-
-	ret = rrpc_map_init(rrpc);
-	if (ret) {
-		pr_err("nvm: rrpc: could not initialize maps\n");
-		goto err;
-	}
-
-	ret = rrpc_blocks_init(rrpc);
-	if (ret) {
-		pr_err("nvm: rrpc: could not initialize state for blocks\n");
-		goto err;
-	}
-
-	ret = rrpc_luns_configure(rrpc);
-	if (ret) {
-		pr_err("nvm: rrpc: not enough blocks available in LUNs.\n");
-		goto err;
-	}
-
-	ret = rrpc_gc_init(rrpc);
-	if (ret) {
-		pr_err("nvm: rrpc: could not initialize gc\n");
-		goto err;
-	}
-
-	/* inherit the size from the underlying device */
-	blk_queue_logical_block_size(tqueue, queue_physical_block_size(bqueue));
-	blk_queue_max_hw_sectors(tqueue, queue_max_hw_sectors(bqueue));
-
-	pr_info("nvm: rrpc initialized with %u luns and %llu pages.\n",
-			rrpc->nr_luns, (unsigned long long)rrpc->nr_sects);
-
-	mod_timer(&rrpc->gc_timer, jiffies + msecs_to_jiffies(10));
-
-	return rrpc;
-err:
-	rrpc_free(rrpc);
-	return ERR_PTR(ret);
-}
-
-/* round robin, page-based FTL, and cost-based GC */
-static struct nvm_tgt_type tt_rrpc = {
-	.name		= "rrpc",
-	.version	= {1, 0, 0},
-
-	.make_rq	= rrpc_make_rq,
-	.capacity	= rrpc_capacity,
-
-	.init		= rrpc_init,
-	.exit		= rrpc_exit,
-};
-
-static int __init rrpc_module_init(void)
-{
-	return nvm_register_tgt_type(&tt_rrpc);
-}
-
-static void rrpc_module_exit(void)
-{
-	nvm_unregister_tgt_type(&tt_rrpc);
-}
-
-module_init(rrpc_module_init);
-module_exit(rrpc_module_exit);
-MODULE_LICENSE("GPL v2");
-MODULE_DESCRIPTION("Block-Device Target for Open-Channel SSDs");

drivers/lightnvm/rrpc.h

@@ -1,290 +0,0 @@
-/*
- * Copyright (C) 2015 IT University of Copenhagen
- * Initial release: Matias Bjorling <m@bjorling.me>
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License version
- * 2 as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
- * General Public License for more details.
- *
- * Implementation of a Round-robin page-based Hybrid FTL for Open-channel SSDs.
- */
-
-#ifndef RRPC_H_
-#define RRPC_H_
-
-#include <linux/blkdev.h>
-#include <linux/blk-mq.h>
-#include <linux/bio.h>
-#include <linux/module.h>
-#include <linux/kthread.h>
-#include <linux/vmalloc.h>
-
-#include <linux/lightnvm.h>
-
-/* Run only GC if less than 1/X blocks are free */
-#define GC_LIMIT_INVERSE 10
-#define GC_TIME_SECS 100
-
-#define RRPC_SECTOR (512)
-#define RRPC_EXPOSED_PAGE_SIZE (4096)
-
-#define NR_PHY_IN_LOG (RRPC_EXPOSED_PAGE_SIZE / RRPC_SECTOR)
-
-struct rrpc_inflight {
-	struct list_head reqs;
-	spinlock_t lock;
-};
-
-struct rrpc_inflight_rq {
-	struct list_head list;
-	sector_t l_start;
-	sector_t l_end;
-};
-
-struct rrpc_rq {
-	struct rrpc_inflight_rq inflight_rq;
-	unsigned long flags;
-};
-
-struct rrpc_block {
-	int id;				/* id inside of LUN */
-	struct rrpc_lun *rlun;
-
-	struct list_head prio;		/* LUN CG list */
-	struct list_head list;		/* LUN free, used, bb list */
-
-#define MAX_INVALID_PAGES_STORAGE 8
-	/* Bitmap for invalid page intries */
-	unsigned long invalid_pages[MAX_INVALID_PAGES_STORAGE];
-	/* points to the next writable page within a block */
-	unsigned int next_page;
-	/* number of pages that are invalid, wrt host page size */
-	unsigned int nr_invalid_pages;
-
-	int state;
-
-	spinlock_t lock;
-	atomic_t data_cmnt_size; /* data pages committed to stable storage */
-};
-
-struct rrpc_lun {
-	struct rrpc *rrpc;
-
-	int id;
-	struct ppa_addr bppa;
-
-	struct rrpc_block *cur, *gc_cur;
-	struct rrpc_block *blocks;	/* Reference to block allocation */
-
-	struct list_head prio_list;	/* Blocks that may be GC'ed */
-	struct list_head wblk_list;	/* Queued blocks to be written to */
-
-	/* lun block lists */
-	struct list_head used_list;	/* In-use blocks */
-	struct list_head free_list;	/* Not used blocks i.e. released
-					 * and ready for use
-					 */
-	struct list_head bb_list;	/* Bad blocks. Mutually exclusive with
-					 * free_list and used_list
-					 */
-	unsigned int nr_free_blocks;	/* Number of unused blocks */
-
-	struct work_struct ws_gc;
-
-	int reserved_blocks;
-
-	spinlock_t lock;
-};
-
-struct rrpc {
-	struct nvm_tgt_dev *dev;
-	struct gendisk *disk;
-
-	sector_t soffset; /* logical sector offset */
-
-	int nr_luns;
-	struct rrpc_lun *luns;
-
-	/* calculated values */
-	unsigned long long nr_sects;
-
-	/* Write strategy variables. Move these into each for structure for each
-	 * strategy
-	 */
-	atomic_t next_lun; /* Whenever a page is written, this is updated
-			    * to point to the next write lun
-			    */
-
-	spinlock_t bio_lock;
-	struct bio_list requeue_bios;
-	struct work_struct ws_requeue;
-
-	/* Simple translation map of logical addresses to physical addresses.
-	 * The logical addresses is known by the host system, while the physical
-	 * addresses are used when writing to the disk block device.
-	 */
-	struct rrpc_addr *trans_map;
-	/* also store a reverse map for garbage collection */
-	struct rrpc_rev_addr *rev_trans_map;
-	spinlock_t rev_lock;
-
-	struct rrpc_inflight inflights;
-
-	mempool_t *addr_pool;
-	mempool_t *page_pool;
-	mempool_t *gcb_pool;
-	mempool_t *rq_pool;
-
-	struct timer_list gc_timer;
-	struct workqueue_struct *krqd_wq;
-	struct workqueue_struct *kgc_wq;
-};
-
-struct rrpc_block_gc {
-	struct rrpc *rrpc;
-	struct rrpc_block *rblk;
-	struct work_struct ws_gc;
-};
-
-/* Logical to physical mapping */
-struct rrpc_addr {
-	u64 addr;
-	struct rrpc_block *rblk;
-};
-
-/* Physical to logical mapping */
-struct rrpc_rev_addr {
-	u64 addr;
-};
-
-static inline struct ppa_addr rrpc_linear_to_generic_addr(struct nvm_geo *geo,
-							  struct ppa_addr r)
-{
-	struct ppa_addr l;
-	int secs, pgs;
-	sector_t ppa = r.ppa;
-
-	l.ppa = 0;
-
-	div_u64_rem(ppa, geo->sec_per_pg, &secs);
-	l.g.sec = secs;
-
-	sector_div(ppa, geo->sec_per_pg);
-	div_u64_rem(ppa, geo->pgs_per_blk, &pgs);
-	l.g.pg = pgs;
-
-	return l;
-}
-
-static inline struct ppa_addr rrpc_recov_addr(struct nvm_tgt_dev *dev, u64 pba)
-{
-	return linear_to_generic_addr(&dev->geo, pba);
-}
-
-static inline u64 rrpc_blk_to_ppa(struct rrpc *rrpc, struct rrpc_block *rblk)
-{
-	struct nvm_tgt_dev *dev = rrpc->dev;
-	struct nvm_geo *geo = &dev->geo;
-	struct rrpc_lun *rlun = rblk->rlun;
-
-	return (rlun->id * geo->sec_per_lun) + (rblk->id * geo->sec_per_blk);
-}
-
-static inline sector_t rrpc_get_laddr(struct bio *bio)
-{
-	return bio->bi_iter.bi_sector / NR_PHY_IN_LOG;
-}
-
-static inline unsigned int rrpc_get_pages(struct bio *bio)
-{
-	return  bio->bi_iter.bi_size / RRPC_EXPOSED_PAGE_SIZE;
-}
-
-static inline sector_t rrpc_get_sector(sector_t laddr)
-{
-	return laddr * NR_PHY_IN_LOG;
-}
-
-static inline int request_intersects(struct rrpc_inflight_rq *r,
-				sector_t laddr_start, sector_t laddr_end)
-{
-	return (laddr_end >= r->l_start) && (laddr_start <= r->l_end);
-}
-
-static int __rrpc_lock_laddr(struct rrpc *rrpc, sector_t laddr,
-			     unsigned int pages, struct rrpc_inflight_rq *r)
-{
-	sector_t laddr_end = laddr + pages - 1;
-	struct rrpc_inflight_rq *rtmp;
-
-	WARN_ON(irqs_disabled());
-
-	spin_lock_irq(&rrpc->inflights.lock);
-	list_for_each_entry(rtmp, &rrpc->inflights.reqs, list) {
-		if (unlikely(request_intersects(rtmp, laddr, laddr_end))) {
-			/* existing, overlapping request, come back later */
-			spin_unlock_irq(&rrpc->inflights.lock);
-			return 1;
-		}
-	}
-
-	r->l_start = laddr;
-	r->l_end = laddr_end;
-
-	list_add_tail(&r->list, &rrpc->inflights.reqs);
-	spin_unlock_irq(&rrpc->inflights.lock);
-	return 0;
-}
-
-static inline int rrpc_lock_laddr(struct rrpc *rrpc, sector_t laddr,
-				 unsigned int pages,
-				 struct rrpc_inflight_rq *r)
-{
-	BUG_ON((laddr + pages) > rrpc->nr_sects);
-
-	return __rrpc_lock_laddr(rrpc, laddr, pages, r);
-}
-
-static inline struct rrpc_inflight_rq *rrpc_get_inflight_rq(struct nvm_rq *rqd)
-{
-	struct rrpc_rq *rrqd = nvm_rq_to_pdu(rqd);
-
-	return &rrqd->inflight_rq;
-}
-
-static inline int rrpc_lock_rq(struct rrpc *rrpc, struct bio *bio,
-							struct nvm_rq *rqd)
-{
-	sector_t laddr = rrpc_get_laddr(bio);
-	unsigned int pages = rrpc_get_pages(bio);
-	struct rrpc_inflight_rq *r = rrpc_get_inflight_rq(rqd);
-
-	return rrpc_lock_laddr(rrpc, laddr, pages, r);
-}
-
-static inline void rrpc_unlock_laddr(struct rrpc *rrpc,
-						struct rrpc_inflight_rq *r)
-{
-	unsigned long flags;
-
-	spin_lock_irqsave(&rrpc->inflights.lock, flags);
-	list_del_init(&r->list);
-	spin_unlock_irqrestore(&rrpc->inflights.lock, flags);
-}
-
-static inline void rrpc_unlock_rq(struct rrpc *rrpc, struct nvm_rq *rqd)
-{
-	struct rrpc_inflight_rq *r = rrpc_get_inflight_rq(rqd);
-	uint8_t pages = rqd->nr_ppas;
-
-	BUG_ON((r->l_start + pages) > rrpc->nr_sects);
-
-	rrpc_unlock_laddr(rrpc, r);
-}
-
-#endif /* RRPC_H_ */

drivers/md/bcache/alloc.c

@@ -525,15 +525,21 @@ struct open_bucket {
 
 /*
  * We keep multiple buckets open for writes, and try to segregate different
- * write streams for better cache utilization: first we look for a bucket where
- * the last write to it was sequential with the current write, and failing that
- * we look for a bucket that was last used by the same task.
+ * write streams for better cache utilization: first we try to segregate flash
+ * only volume write streams from cached devices, secondly we look for a bucket
+ * where the last write to it was sequential with the current write, and
+ * failing that we look for a bucket that was last used by the same task.
  *
  * The ideas is if you've got multiple tasks pulling data into the cache at the
  * same time, you'll get better cache utilization if you try to segregate their
  * data and preserve locality.
  *
- * For example, say you've starting Firefox at the same time you're copying a
+ * For example, dirty sectors of flash only volume is not reclaimable, if their
+ * dirty sectors mixed with dirty sectors of cached device, such buckets will
+ * be marked as dirty and won't be reclaimed, though the dirty data of cached
+ * device have been written back to backend device.
+ *
+ * And say you've starting Firefox at the same time you're copying a
  * bunch of files. Firefox will likely end up being fairly hot and stay in the
  * cache awhile, but the data you copied might not be; if you wrote all that
  * data to the same buckets it'd get invalidated at the same time.
@@ -550,7 +556,10 @@ static struct open_bucket *pick_data_bucket(struct cache_set *c,
 	struct open_bucket *ret, *ret_task = NULL;
 
 	list_for_each_entry_reverse(ret, &c->data_buckets, list)
-		if (!bkey_cmp(&ret->key, search))
+		if (UUID_FLASH_ONLY(&c->uuids[KEY_INODE(&ret->key)]) !=
+		    UUID_FLASH_ONLY(&c->uuids[KEY_INODE(search)]))
+			continue;
+		else if (!bkey_cmp(&ret->key, search))
 			goto found;
 		else if (ret->last_write_point == write_point)
 			ret_task = ret;

drivers/md/bcache/bcache.h

@@ -320,14 +320,15 @@ struct cached_dev {
 	 */
 	atomic_t		has_dirty;
 
-	struct bch_ratelimit	writeback_rate;
-	struct delayed_work	writeback_rate_update;
-
 	/*
-	 * Internal to the writeback code, so read_dirty() can keep track of
-	 * where it's at.
+	 * Set to zero by things that touch the backing volume-- except
+	 * writeback.  Incremented by writeback.  Used to determine when to
+	 * accelerate idle writeback.
 	 */
-	sector_t		last_read;
+	atomic_t		backing_idle;
+
+	struct bch_ratelimit	writeback_rate;
+	struct delayed_work	writeback_rate_update;
 
 	/* Limit number of writeback bios in flight */
 	struct semaphore	in_flight;
@@ -336,6 +337,14 @@ struct cached_dev {
 
 	struct keybuf		writeback_keys;
 
+	/*
+	 * Order the write-half of writeback operations strongly in dispatch
+	 * order.  (Maintain LBA order; don't allow reads completing out of
+	 * order to re-order the writes...)
+	 */
+	struct closure_waitlist writeback_ordering_wait;
+	atomic_t		writeback_sequence_next;
+
 	/* For tracking sequential IO */
 #define RECENT_IO_BITS	7
 #define RECENT_IO	(1 << RECENT_IO_BITS)
@@ -488,6 +497,7 @@ struct cache_set {
 	int			caches_loaded;
 
 	struct bcache_device	**devices;
+	unsigned		devices_max_used;
 	struct list_head	cached_devs;
 	uint64_t		cached_dev_sectors;
 	struct closure		caching;
@@ -852,7 +862,7 @@ static inline void wake_up_allocators(struct cache_set *c)
 
 /* Forward declarations */
 
-void bch_count_io_errors(struct cache *, blk_status_t, const char *);
+void bch_count_io_errors(struct cache *, blk_status_t, int, const char *);
 void bch_bbio_count_io_errors(struct cache_set *, struct bio *,
 			      blk_status_t, const char *);
 void bch_bbio_endio(struct cache_set *, struct bio *, blk_status_t,

drivers/md/bcache/btree.c

@@ -419,7 +419,7 @@ static void do_btree_node_write(struct btree *b)
 	SET_PTR_OFFSET(&k.key, 0, PTR_OFFSET(&k.key, 0) +
 		       bset_sector_offset(&b->keys, i));
 
-	if (!bio_alloc_pages(b->bio, __GFP_NOWARN|GFP_NOWAIT)) {
+	if (!bch_bio_alloc_pages(b->bio, __GFP_NOWARN|GFP_NOWAIT)) {
 		int j;
 		struct bio_vec *bv;
 		void *base = (void *) ((unsigned long) i & ~(PAGE_SIZE - 1));
@@ -432,6 +432,7 @@ static void do_btree_node_write(struct btree *b)
 
 		continue_at(cl, btree_node_write_done, NULL);
 	} else {
+		/* No problem for multipage bvec since the bio is just allocated */
 		b->bio->bi_vcnt = 0;
 		bch_bio_map(b->bio, i);
 
@@ -1678,7 +1679,7 @@ static void bch_btree_gc_finish(struct cache_set *c)
 
 	/* don't reclaim buckets to which writeback keys point */
 	rcu_read_lock();
-	for (i = 0; i < c->nr_uuids; i++) {
+	for (i = 0; i < c->devices_max_used; i++) {
 		struct bcache_device *d = c->devices[i];
 		struct cached_dev *dc;
 		struct keybuf_key *w, *n;
@@ -1803,10 +1804,7 @@ static int bch_gc_thread(void *arg)
 int bch_gc_thread_start(struct cache_set *c)
 {
 	c->gc_thread = kthread_run(bch_gc_thread, c, "bcache_gc");
-	if (IS_ERR(c->gc_thread))
-		return PTR_ERR(c->gc_thread);
-
-	return 0;
+	return PTR_ERR_OR_ZERO(c->gc_thread);
 }
 
 /* Initial partial gc */

drivers/md/bcache/closure.c

@@ -8,6 +8,7 @@
 #include <linux/debugfs.h>
 #include <linux/module.h>
 #include <linux/seq_file.h>
+#include <linux/sched/debug.h>
 
 #include "closure.h"
 
@@ -18,10 +19,6 @@ static inline void closure_put_after_sub(struct closure *cl, int flags)
 	BUG_ON(flags & CLOSURE_GUARD_MASK);
 	BUG_ON(!r && (flags & ~CLOSURE_DESTRUCTOR));
 
-	/* Must deliver precisely one wakeup */
-	if (r == 1 && (flags & CLOSURE_SLEEPING))
-		wake_up_process(cl->task);
-
 	if (!r) {
 		if (cl->fn && !(flags & CLOSURE_DESTRUCTOR)) {
 			atomic_set(&cl->remaining,
@@ -100,28 +97,34 @@ bool closure_wait(struct closure_waitlist *waitlist, struct closure *cl)
 }
 EXPORT_SYMBOL(closure_wait);
 
-/**
- * closure_sync - sleep until a closure has nothing left to wait on
- *
- * Sleeps until the refcount hits 1 - the thread that's running the closure owns
- * the last refcount.
- */
-void closure_sync(struct closure *cl)
+struct closure_syncer {
+	struct task_struct	*task;
+	int			done;
+};
+
+static void closure_sync_fn(struct closure *cl)
 {
-	while (1) {
-		__closure_start_sleep(cl);
-		closure_set_ret_ip(cl);
+	cl->s->done = 1;
+	wake_up_process(cl->s->task);
+}
 
-		if ((atomic_read(&cl->remaining) &
-		     CLOSURE_REMAINING_MASK) == 1)
-			break;
+void __sched __closure_sync(struct closure *cl)
+{
+	struct closure_syncer s = { .task = current };
 
+	cl->s = &s;
+	continue_at(cl, closure_sync_fn, NULL);
+
+	while (1) {
+		set_current_state(TASK_UNINTERRUPTIBLE);
+		if (s.done)
+			break;
 		schedule();
 	}
 
-	__closure_end_sleep(cl);
+	__set_current_state(TASK_RUNNING);
 }
-EXPORT_SYMBOL(closure_sync);
+EXPORT_SYMBOL(__closure_sync);
 
 #ifdef CONFIG_BCACHE_CLOSURES_DEBUG
 
@@ -168,12 +171,10 @@ static int debug_seq_show(struct seq_file *f, void *data)
 			   cl, (void *) cl->ip, cl->fn, cl->parent,
 			   r & CLOSURE_REMAINING_MASK);
 
-		seq_printf(f, "%s%s%s%s\n",
+		seq_printf(f, "%s%s\n",
 			   test_bit(WORK_STRUCT_PENDING_BIT,
 				    work_data_bits(&cl->work)) ? "Q" : "",
-			   r & CLOSURE_RUNNING	? "R" : "",
-			   r & CLOSURE_STACK	? "S" : "",
-			   r & CLOSURE_SLEEPING	? "Sl" : "");
+			   r & CLOSURE_RUNNING	? "R" : "");
 
 		if (r & CLOSURE_WAITING)
 			seq_printf(f, " W %pF\n",

drivers/md/bcache/closure.h

@@ -103,6 +103,7 @@
  */
 
 struct closure;
+struct closure_syncer;
 typedef void (closure_fn) (struct closure *);
 
 struct closure_waitlist {
@@ -115,10 +116,6 @@ enum closure_state {
 	 * the thread that owns the closure, and cleared by the thread that's
 	 * waking up the closure.
 	 *
-	 * CLOSURE_SLEEPING: Must be set before a thread uses a closure to sleep
-	 * - indicates that cl->task is valid and closure_put() may wake it up.
-	 * Only set or cleared by the thread that owns the closure.
-	 *
 	 * The rest are for debugging and don't affect behaviour:
 	 *
 	 * CLOSURE_RUNNING: Set when a closure is running (i.e. by
@@ -128,22 +125,16 @@ enum closure_state {
 	 * continue_at() and closure_return() clear it for you, if you're doing
 	 * something unusual you can use closure_set_dead() which also helps
 	 * annotate where references are being transferred.
-	 *
-	 * CLOSURE_STACK: Sanity check - remaining should never hit 0 on a
-	 * closure with this flag set
 	 */
 
-	CLOSURE_BITS_START	= (1 << 23),
-	CLOSURE_DESTRUCTOR	= (1 << 23),
-	CLOSURE_WAITING		= (1 << 25),
-	CLOSURE_SLEEPING	= (1 << 27),
-	CLOSURE_RUNNING		= (1 << 29),
-	CLOSURE_STACK		= (1 << 31),
+	CLOSURE_BITS_START	= (1U << 26),
+	CLOSURE_DESTRUCTOR	= (1U << 26),
+	CLOSURE_WAITING		= (1U << 28),
+	CLOSURE_RUNNING		= (1U << 30),
 };
 
 #define CLOSURE_GUARD_MASK					\
-	((CLOSURE_DESTRUCTOR|CLOSURE_WAITING|CLOSURE_SLEEPING|	\
-	  CLOSURE_RUNNING|CLOSURE_STACK) << 1)
+	((CLOSURE_DESTRUCTOR|CLOSURE_WAITING|CLOSURE_RUNNING) << 1)
 
 #define CLOSURE_REMAINING_MASK		(CLOSURE_BITS_START - 1)
 #define CLOSURE_REMAINING_INITIALIZER	(1|CLOSURE_RUNNING)
@@ -152,7 +143,7 @@ struct closure {
 	union {
 		struct {
 			struct workqueue_struct *wq;
-			struct task_struct	*task;
+			struct closure_syncer	*s;
 			struct llist_node	list;
 			closure_fn		*fn;
 		};
@@ -178,7 +169,19 @@ void closure_sub(struct closure *cl, int v);
 void closure_put(struct closure *cl);
 void __closure_wake_up(struct closure_waitlist *list);
 bool closure_wait(struct closure_waitlist *list, struct closure *cl);
-void closure_sync(struct closure *cl);
+void __closure_sync(struct closure *cl);
+
+/**
+ * closure_sync - sleep until a closure a closure has nothing left to wait on
+ *
+ * Sleeps until the refcount hits 1 - the thread that's running the closure owns
+ * the last refcount.
+ */
+static inline void closure_sync(struct closure *cl)
+{
+	if ((atomic_read(&cl->remaining) & CLOSURE_REMAINING_MASK) != 1)
+		__closure_sync(cl);
+}
 
 #ifdef CONFIG_BCACHE_CLOSURES_DEBUG
 
@@ -215,24 +218,6 @@ static inline void closure_set_waiting(struct closure *cl, unsigned long f)
 #endif
 }
 
-static inline void __closure_end_sleep(struct closure *cl)
-{
-	__set_current_state(TASK_RUNNING);
-
-	if (atomic_read(&cl->remaining) & CLOSURE_SLEEPING)
-		atomic_sub(CLOSURE_SLEEPING, &cl->remaining);
-}
-
-static inline void __closure_start_sleep(struct closure *cl)
-{
-	closure_set_ip(cl);
-	cl->task = current;
-	set_current_state(TASK_UNINTERRUPTIBLE);
-
-	if (!(atomic_read(&cl->remaining) & CLOSURE_SLEEPING))
-		atomic_add(CLOSURE_SLEEPING, &cl->remaining);
-}
-
 static inline void closure_set_stopped(struct closure *cl)
 {
 	atomic_sub(CLOSURE_RUNNING, &cl->remaining);
@@ -241,7 +226,6 @@ static inline void closure_set_stopped(struct closure *cl)
 static inline void set_closure_fn(struct closure *cl, closure_fn *fn,
 				  struct workqueue_struct *wq)
 {
-	BUG_ON(object_is_on_stack(cl));
 	closure_set_ip(cl);
 	cl->fn = fn;
 	cl->wq = wq;
@@ -300,7 +284,7 @@ static inline void closure_init(struct closure *cl, struct closure *parent)
 static inline void closure_init_stack(struct closure *cl)
 {
 	memset(cl, 0, sizeof(struct closure));
-	atomic_set(&cl->remaining, CLOSURE_REMAINING_INITIALIZER|CLOSURE_STACK);
+	atomic_set(&cl->remaining, CLOSURE_REMAINING_INITIALIZER);
 }
 
 /**
@@ -322,6 +306,8 @@ static inline void closure_wake_up(struct closure_waitlist *list)
  * This is because after calling continue_at() you no longer have a ref on @cl,
  * and whatever @cl owns may be freed out from under you - a running closure fn
  * has a ref on its own closure which continue_at() drops.
+ *
+ * Note you are expected to immediately return after using this macro.
  */
 #define continue_at(_cl, _fn, _wq)					\
 do {									\

drivers/md/bcache/debug.c

@@ -116,7 +116,7 @@ void bch_data_verify(struct cached_dev *dc, struct bio *bio)
 		return;
 	check->bi_opf = REQ_OP_READ;
 
-	if (bio_alloc_pages(check, GFP_NOIO))
+	if (bch_bio_alloc_pages(check, GFP_NOIO))
 		goto out_put;
 
 	submit_bio_wait(check);
@@ -251,8 +251,7 @@ void bch_debug_exit(void)
 
 int __init bch_debug_init(struct kobject *kobj)
 {
-	int ret = 0;
-
 	debug = debugfs_create_dir("bcache", NULL);
-	return ret;
+
+	return IS_ERR_OR_NULL(debug);
 }

drivers/md/bcache/io.c

@@ -51,7 +51,10 @@ void bch_submit_bbio(struct bio *bio, struct cache_set *c,
 
 /* IO errors */
 
-void bch_count_io_errors(struct cache *ca, blk_status_t error, const char *m)
+void bch_count_io_errors(struct cache *ca,
+			 blk_status_t error,
+			 int is_read,
+			 const char *m)
 {
 	/*
 	 * The halflife of an error is:
@@ -94,8 +97,9 @@ void bch_count_io_errors(struct cache *ca, blk_status_t error, const char *m)
 		errors >>= IO_ERROR_SHIFT;
 
 		if (errors < ca->set->error_limit)
-			pr_err("%s: IO error on %s, recovering",
-			       bdevname(ca->bdev, buf), m);
+			pr_err("%s: IO error on %s%s",
+			       bdevname(ca->bdev, buf), m,
+			       is_read ? ", recovering." : ".");
 		else
 			bch_cache_set_error(ca->set,
 					    "%s: too many IO errors %s",
@@ -108,6 +112,7 @@ void bch_bbio_count_io_errors(struct cache_set *c, struct bio *bio,
 {
 	struct bbio *b = container_of(bio, struct bbio, bio);
 	struct cache *ca = PTR_CACHE(c, &b->key, 0);
+	int is_read = (bio_data_dir(bio) == READ ? 1 : 0);
 
 	unsigned threshold = op_is_write(bio_op(bio))
 		? c->congested_write_threshold_us
@@ -129,7 +134,7 @@ void bch_bbio_count_io_errors(struct cache_set *c, struct bio *bio,
 			atomic_inc(&c->congested);
 	}
 
-	bch_count_io_errors(ca, error, m);
+	bch_count_io_errors(ca, error, is_read, m);
 }
 
 void bch_bbio_endio(struct cache_set *c, struct bio *bio,

drivers/md/bcache/movinggc.c

@@ -162,7 +162,7 @@ static void read_moving(struct cache_set *c)
 		bio_set_op_attrs(bio, REQ_OP_READ, 0);
 		bio->bi_end_io	= read_moving_endio;
 
-		if (bio_alloc_pages(bio, GFP_KERNEL))
+		if (bch_bio_alloc_pages(bio, GFP_KERNEL))
 			goto err;
 
 		trace_bcache_gc_copy(&w->key);

drivers/md/bcache/request.c

@@ -576,6 +576,7 @@ static void cache_lookup(struct closure *cl)
 {
 	struct search *s = container_of(cl, struct search, iop.cl);
 	struct bio *bio = &s->bio.bio;
+	struct cached_dev *dc;
 	int ret;
 
 	bch_btree_op_init(&s->op, -1);
@@ -588,6 +589,27 @@ static void cache_lookup(struct closure *cl)
 		return;
 	}
 
+	/*
+	 * We might meet err when searching the btree, If that happens, we will
+	 * get negative ret, in this scenario we should not recover data from
+	 * backing device (when cache device is dirty) because we don't know
+	 * whether bkeys the read request covered are all clean.
+	 *
+	 * And after that happened, s->iop.status is still its initial value
+	 * before we submit s->bio.bio
+	 */
+	if (ret < 0) {
+		BUG_ON(ret == -EINTR);
+		if (s->d && s->d->c &&
+				!UUID_FLASH_ONLY(&s->d->c->uuids[s->d->id])) {
+			dc = container_of(s->d, struct cached_dev, disk);
+			if (dc && atomic_read(&dc->has_dirty))
+				s->recoverable = false;
+		}
+		if (!s->iop.status)
+			s->iop.status = BLK_STS_IOERR;
+	}
+
 	closure_return(cl);
 }
 
@@ -611,8 +633,8 @@ static void request_endio(struct bio *bio)
 static void bio_complete(struct search *s)
 {
 	if (s->orig_bio) {
-		struct request_queue *q = s->orig_bio->bi_disk->queue;
-		generic_end_io_acct(q, bio_data_dir(s->orig_bio),
+		generic_end_io_acct(s->d->disk->queue,
+				    bio_data_dir(s->orig_bio),
 				    &s->d->disk->part0, s->start_time);
 
 		trace_bcache_request_end(s->d, s->orig_bio);
@@ -841,7 +863,7 @@ static int cached_dev_cache_miss(struct btree *b, struct search *s,
 	cache_bio->bi_private	= &s->cl;
 
 	bch_bio_map(cache_bio, NULL);
-	if (bio_alloc_pages(cache_bio, __GFP_NOWARN|GFP_NOIO))
+	if (bch_bio_alloc_pages(cache_bio, __GFP_NOWARN|GFP_NOIO))
 		goto out_put;
 
 	if (reada)
@@ -974,6 +996,7 @@ static blk_qc_t cached_dev_make_request(struct request_queue *q,
 	struct cached_dev *dc = container_of(d, struct cached_dev, disk);
 	int rw = bio_data_dir(bio);
 
+	atomic_set(&dc->backing_idle, 0);
 	generic_start_io_acct(q, rw, bio_sectors(bio), &d->disk->part0);
 
 	bio_set_dev(bio, dc->bdev);

drivers/md/bcache/super.c

@@ -211,7 +211,7 @@ static void write_bdev_super_endio(struct bio *bio)
 
 static void __write_super(struct cache_sb *sb, struct bio *bio)
 {
-	struct cache_sb *out = page_address(bio->bi_io_vec[0].bv_page);
+	struct cache_sb *out = page_address(bio_first_page_all(bio));
 	unsigned i;
 
 	bio->bi_iter.bi_sector	= SB_SECTOR;
@@ -274,7 +274,9 @@ static void write_super_endio(struct bio *bio)
 {
 	struct cache *ca = bio->bi_private;
 
-	bch_count_io_errors(ca, bio->bi_status, "writing superblock");
+	/* is_read = 0 */
+	bch_count_io_errors(ca, bio->bi_status, 0,
+			    "writing superblock");
 	closure_put(&ca->set->sb_write);
 }
 
@@ -721,6 +723,9 @@ static void bcache_device_attach(struct bcache_device *d, struct cache_set *c,
 	d->c = c;
 	c->devices[id] = d;
 
+	if (id >= c->devices_max_used)
+		c->devices_max_used = id + 1;
+
 	closure_get(&c->caching);
 }
 
@@ -906,6 +911,12 @@ static void cached_dev_detach_finish(struct work_struct *w)
 
 	mutex_lock(&bch_register_lock);
 
+	cancel_delayed_work_sync(&dc->writeback_rate_update);
+	if (!IS_ERR_OR_NULL(dc->writeback_thread)) {
+		kthread_stop(dc->writeback_thread);
+		dc->writeback_thread = NULL;
+	}
+
 	memset(&dc->sb.set_uuid, 0, 16);
 	SET_BDEV_STATE(&dc->sb, BDEV_STATE_NONE);
 
@@ -1166,7 +1177,7 @@ static void register_bdev(struct cache_sb *sb, struct page *sb_page,
 	dc->bdev->bd_holder = dc;
 
 	bio_init(&dc->sb_bio, dc->sb_bio.bi_inline_vecs, 1);
-	dc->sb_bio.bi_io_vec[0].bv_page = sb_page;
+	bio_first_bvec_all(&dc->sb_bio)->bv_page = sb_page;
 	get_page(sb_page);
 
 	if (cached_dev_init(dc, sb->block_size << 9))
@@ -1261,7 +1272,7 @@ static int flash_devs_run(struct cache_set *c)
 	struct uuid_entry *u;
 
 	for (u = c->uuids;
-	     u < c->uuids + c->nr_uuids && !ret;
+	     u < c->uuids + c->devices_max_used && !ret;
 	     u++)
 		if (UUID_FLASH_ONLY(u))
 			ret = flash_dev_run(c, u);
@@ -1427,7 +1438,7 @@ static void __cache_set_unregister(struct closure *cl)
 
 	mutex_lock(&bch_register_lock);
 
-	for (i = 0; i < c->nr_uuids; i++)
+	for (i = 0; i < c->devices_max_used; i++)
 		if (c->devices[i]) {
 			if (!UUID_FLASH_ONLY(&c->uuids[i]) &&
 			    test_bit(CACHE_SET_UNREGISTERING, &c->flags)) {
@@ -1490,7 +1501,7 @@ struct cache_set *bch_cache_set_alloc(struct cache_sb *sb)
 	c->bucket_bits		= ilog2(sb->bucket_size);
 	c->block_bits		= ilog2(sb->block_size);
 	c->nr_uuids		= bucket_bytes(c) / sizeof(struct uuid_entry);
-
+	c->devices_max_used	= 0;
 	c->btree_pages		= bucket_pages(c);
 	if (c->btree_pages > BTREE_MAX_PAGES)
 		c->btree_pages = max_t(int, c->btree_pages / 4,
@@ -1810,7 +1821,7 @@ void bch_cache_release(struct kobject *kobj)
 		free_fifo(&ca->free[i]);
 
 	if (ca->sb_bio.bi_inline_vecs[0].bv_page)
-		put_page(ca->sb_bio.bi_io_vec[0].bv_page);
+		put_page(bio_first_page_all(&ca->sb_bio));
 
 	if (!IS_ERR_OR_NULL(ca->bdev))
 		blkdev_put(ca->bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
@@ -1864,7 +1875,7 @@ static int register_cache(struct cache_sb *sb, struct page *sb_page,
 	ca->bdev->bd_holder = ca;
 
 	bio_init(&ca->sb_bio, ca->sb_bio.bi_inline_vecs, 1);
-	ca->sb_bio.bi_io_vec[0].bv_page = sb_page;
+	bio_first_bvec_all(&ca->sb_bio)->bv_page = sb_page;
 	get_page(sb_page);
 
 	if (blk_queue_discard(bdev_get_queue(ca->bdev)))

drivers/md/bcache/util.c

@@ -249,6 +249,13 @@ uint64_t bch_next_delay(struct bch_ratelimit *d, uint64_t done)
 		: 0;
 }
 
+/*
+ * Generally it isn't good to access .bi_io_vec and .bi_vcnt directly,
+ * the preferred way is bio_add_page, but in this case, bch_bio_map()
+ * supposes that the bvec table is empty, so it is safe to access
+ * .bi_vcnt & .bi_io_vec in this way even after multipage bvec is
+ * supported.
+ */
 void bch_bio_map(struct bio *bio, void *base)
 {
 	size_t size = bio->bi_iter.bi_size;
@@ -276,6 +283,33 @@ start:		bv->bv_len	= min_t(size_t, PAGE_SIZE - bv->bv_offset,
 	}
 }
 
+/**
+ * bch_bio_alloc_pages - allocates a single page for each bvec in a bio
+ * @bio: bio to allocate pages for
+ * @gfp_mask: flags for allocation
+ *
+ * Allocates pages up to @bio->bi_vcnt.
+ *
+ * Returns 0 on success, -ENOMEM on failure. On failure, any allocated pages are
+ * freed.
+ */
+int bch_bio_alloc_pages(struct bio *bio, gfp_t gfp_mask)
+{
+	int i;
+	struct bio_vec *bv;
+
+	bio_for_each_segment_all(bv, bio, i) {
+		bv->bv_page = alloc_page(gfp_mask);
+		if (!bv->bv_page) {
+			while (--bv >= bio->bi_io_vec)
+				__free_page(bv->bv_page);
+			return -ENOMEM;
+		}
+	}
+
+	return 0;
+}
+
 /*
  * Portions Copyright (c) 1996-2001, PostgreSQL Global Development Group (Any
  * use permitted, subject to terms of PostgreSQL license; see.)

drivers/md/bcache/util.h

@@ -558,6 +558,7 @@ static inline unsigned fract_exp_two(unsigned x, unsigned fract_bits)
 }
 
 void bch_bio_map(struct bio *bio, void *base);
+int bch_bio_alloc_pages(struct bio *bio, gfp_t gfp_mask);
 
 static inline sector_t bdev_sectors(struct block_device *bdev)
 {

drivers/md/bcache/writeback.c

@@ -18,17 +18,39 @@
 #include <trace/events/bcache.h>
 
 /* Rate limiting */
-
-static void __update_writeback_rate(struct cached_dev *dc)
+static uint64_t __calc_target_rate(struct cached_dev *dc)
 {
 	struct cache_set *c = dc->disk.c;
+
+	/*
+	 * This is the size of the cache, minus the amount used for
+	 * flash-only devices
+	 */
 	uint64_t cache_sectors = c->nbuckets * c->sb.bucket_size -
 				bcache_flash_devs_sectors_dirty(c);
+
+	/*
+	 * Unfortunately there is no control of global dirty data.  If the
+	 * user states that they want 10% dirty data in the cache, and has,
+	 * e.g., 5 backing volumes of equal size, we try and ensure each
+	 * backing volume uses about 2% of the cache for dirty data.
+	 */
+	uint32_t bdev_share =
+		div64_u64(bdev_sectors(dc->bdev) << WRITEBACK_SHARE_SHIFT,
+				c->cached_dev_sectors);
+
 	uint64_t cache_dirty_target =
 		div_u64(cache_sectors * dc->writeback_percent, 100);
-	int64_t target = div64_u64(cache_dirty_target * bdev_sectors(dc->bdev),
-				   c->cached_dev_sectors);
 
+	/* Ensure each backing dev gets at least one dirty share */
+	if (bdev_share < 1)
+		bdev_share = 1;
+
+	return (cache_dirty_target * bdev_share) >> WRITEBACK_SHARE_SHIFT;
+}
+
+static void __update_writeback_rate(struct cached_dev *dc)
+{
 	/*
 	 * PI controller:
 	 * Figures out the amount that should be written per second.
@@ -49,6 +71,7 @@ static void __update_writeback_rate(struct cached_dev *dc)
 	 * This acts as a slow, long-term average that is not subject to
 	 * variations in usage like the p term.
 	 */
+	int64_t target = __calc_target_rate(dc);
 	int64_t dirty = bcache_dev_sectors_dirty(&dc->disk);
 	int64_t error = dirty - target;
 	int64_t proportional_scaled =
@@ -116,6 +139,7 @@ static unsigned writeback_delay(struct cached_dev *dc, unsigned sectors)
 struct dirty_io {
 	struct closure		cl;
 	struct cached_dev	*dc;
+	uint16_t		sequence;
 	struct bio		bio;
 };
 
@@ -194,6 +218,27 @@ static void write_dirty(struct closure *cl)
 {
 	struct dirty_io *io = container_of(cl, struct dirty_io, cl);
 	struct keybuf_key *w = io->bio.bi_private;
+	struct cached_dev *dc = io->dc;
+
+	uint16_t next_sequence;
+
+	if (atomic_read(&dc->writeback_sequence_next) != io->sequence) {
+		/* Not our turn to write; wait for a write to complete */
+		closure_wait(&dc->writeback_ordering_wait, cl);
+
+		if (atomic_read(&dc->writeback_sequence_next) == io->sequence) {
+			/*
+			 * Edge case-- it happened in indeterminate order
+			 * relative to when we were added to wait list..
+			 */
+			closure_wake_up(&dc->writeback_ordering_wait);
+		}
+
+		continue_at(cl, write_dirty, io->dc->writeback_write_wq);
+		return;
+	}
+
+	next_sequence = io->sequence + 1;
 
 	/*
 	 * IO errors are signalled using the dirty bit on the key.
@@ -211,6 +256,9 @@ static void write_dirty(struct closure *cl)
 		closure_bio_submit(&io->bio, cl);
 	}
 
+	atomic_set(&dc->writeback_sequence_next, next_sequence);
+	closure_wake_up(&dc->writeback_ordering_wait);
+
 	continue_at(cl, write_dirty_finish, io->dc->writeback_write_wq);
 }
 
@@ -219,8 +267,10 @@ static void read_dirty_endio(struct bio *bio)
 	struct keybuf_key *w = bio->bi_private;
 	struct dirty_io *io = w->private;
 
+	/* is_read = 1 */
 	bch_count_io_errors(PTR_CACHE(io->dc->disk.c, &w->key, 0),
-			    bio->bi_status, "reading dirty data from cache");
+			    bio->bi_status, 1,
+			    "reading dirty data from cache");
 
 	dirty_endio(bio);
 }
@@ -237,10 +287,15 @@ static void read_dirty_submit(struct closure *cl)
 static void read_dirty(struct cached_dev *dc)
 {
 	unsigned delay = 0;
-	struct keybuf_key *w;
+	struct keybuf_key *next, *keys[MAX_WRITEBACKS_IN_PASS], *w;
+	size_t size;
+	int nk, i;
 	struct dirty_io *io;
 	struct closure cl;
+	uint16_t sequence = 0;
 
+	BUG_ON(!llist_empty(&dc->writeback_ordering_wait.list));
+	atomic_set(&dc->writeback_sequence_next, sequence);
 	closure_init_stack(&cl);
 
 	/*
@@ -248,45 +303,109 @@ static void read_dirty(struct cached_dev *dc)
 	 * mempools.
 	 */
 
-	while (!kthread_should_stop()) {
-
-		w = bch_keybuf_next(&dc->writeback_keys);
-		if (!w)
-			break;
-
-		BUG_ON(ptr_stale(dc->disk.c, &w->key, 0));
-
-		if (KEY_START(&w->key) != dc->last_read ||
-		    jiffies_to_msecs(delay) > 50)
-			while (!kthread_should_stop() && delay)
-				delay = schedule_timeout_interruptible(delay);
-
-		dc->last_read	= KEY_OFFSET(&w->key);
-
-		io = kzalloc(sizeof(struct dirty_io) + sizeof(struct bio_vec)
-			     * DIV_ROUND_UP(KEY_SIZE(&w->key), PAGE_SECTORS),
-			     GFP_KERNEL);
-		if (!io)
-			goto err;
-
-		w->private	= io;
-		io->dc		= dc;
-
-		dirty_init(w);
-		bio_set_op_attrs(&io->bio, REQ_OP_READ, 0);
-		io->bio.bi_iter.bi_sector = PTR_OFFSET(&w->key, 0);
-		bio_set_dev(&io->bio, PTR_CACHE(dc->disk.c, &w->key, 0)->bdev);
-		io->bio.bi_end_io	= read_dirty_endio;
-
-		if (bio_alloc_pages(&io->bio, GFP_KERNEL))
-			goto err_free;
+	next = bch_keybuf_next(&dc->writeback_keys);
+
+	while (!kthread_should_stop() && next) {
+		size = 0;
+		nk = 0;
+
+		do {
+			BUG_ON(ptr_stale(dc->disk.c, &next->key, 0));
+
+			/*
+			 * Don't combine too many operations, even if they
+			 * are all small.
+			 */
+			if (nk >= MAX_WRITEBACKS_IN_PASS)
+				break;
+
+			/*
+			 * If the current operation is very large, don't
+			 * further combine operations.
+			 */
+			if (size >= MAX_WRITESIZE_IN_PASS)
+				break;
+
+			/*
+			 * Operations are only eligible to be combined
+			 * if they are contiguous.
+			 *
+			 * TODO: add a heuristic willing to fire a
+			 * certain amount of non-contiguous IO per pass,
+			 * so that we can benefit from backing device
+			 * command queueing.
+			 */
+			if ((nk != 0) && bkey_cmp(&keys[nk-1]->key,
+						&START_KEY(&next->key)))
+				break;
+
+			size += KEY_SIZE(&next->key);
+			keys[nk++] = next;
+		} while ((next = bch_keybuf_next(&dc->writeback_keys)));
+
+		/* Now we have gathered a set of 1..5 keys to write back. */
+		for (i = 0; i < nk; i++) {
+			w = keys[i];
+
+			io = kzalloc(sizeof(struct dirty_io) +
+				     sizeof(struct bio_vec) *
+				     DIV_ROUND_UP(KEY_SIZE(&w->key), PAGE_SECTORS),
+				     GFP_KERNEL);
+			if (!io)
+				goto err;
+
+			w->private	= io;
+			io->dc		= dc;
+			io->sequence    = sequence++;
+
+			dirty_init(w);
+			bio_set_op_attrs(&io->bio, REQ_OP_READ, 0);
+			io->bio.bi_iter.bi_sector = PTR_OFFSET(&w->key, 0);
+			bio_set_dev(&io->bio,
+				    PTR_CACHE(dc->disk.c, &w->key, 0)->bdev);
+			io->bio.bi_end_io	= read_dirty_endio;
+
+			if (bch_bio_alloc_pages(&io->bio, GFP_KERNEL))
+				goto err_free;
+
+			trace_bcache_writeback(&w->key);
+
+			down(&dc->in_flight);
+
+			/* We've acquired a semaphore for the maximum
+			 * simultaneous number of writebacks; from here
+			 * everything happens asynchronously.
+			 */
+			closure_call(&io->cl, read_dirty_submit, NULL, &cl);
+		}
 
-		trace_bcache_writeback(&w->key);
+		delay = writeback_delay(dc, size);
 
-		down(&dc->in_flight);
-		closure_call(&io->cl, read_dirty_submit, NULL, &cl);
+		/* If the control system would wait for at least half a
+		 * second, and there's been no reqs hitting the backing disk
+		 * for awhile: use an alternate mode where we have at most
+		 * one contiguous set of writebacks in flight at a time.  If
+		 * someone wants to do IO it will be quick, as it will only
+		 * have to contend with one operation in flight, and we'll
+		 * be round-tripping data to the backing disk as quickly as
+		 * it can accept it.
+		 */
+		if (delay >= HZ / 2) {
+			/* 3 means at least 1.5 seconds, up to 7.5 if we
+			 * have slowed way down.
+			 */
+			if (atomic_inc_return(&dc->backing_idle) >= 3) {
+				/* Wait for current I/Os to finish */
+				closure_sync(&cl);
+				/* And immediately launch a new set. */
+				delay = 0;
+			}
+		}
 
-		delay = writeback_delay(dc, KEY_SIZE(&w->key));
+		while (!kthread_should_stop() && delay) {
+			schedule_timeout_interruptible(delay);
+			delay = writeback_delay(dc, 0);
+		}
 	}
 
 	if (0) {

drivers/md/bcache/writeback.h

@@ -5,6 +5,16 @@
 #define CUTOFF_WRITEBACK	40
 #define CUTOFF_WRITEBACK_SYNC	70
 
+#define MAX_WRITEBACKS_IN_PASS  5
+#define MAX_WRITESIZE_IN_PASS   5000	/* *512b */
+
+/*
+ * 14 (16384ths) is chosen here as something that each backing device
+ * should be a reasonable fraction of the share, and not to blow up
+ * until individual backing devices are a petabyte.
+ */
+#define WRITEBACK_SHARE_SHIFT   14
+
 static inline uint64_t bcache_dev_sectors_dirty(struct bcache_device *d)
 {
 	uint64_t i, ret = 0;
@@ -21,7 +31,7 @@ static inline uint64_t  bcache_flash_devs_sectors_dirty(struct cache_set *c)
 
 	mutex_lock(&bch_register_lock);
 
-	for (i = 0; i < c->nr_uuids; i++) {
+	for (i = 0; i < c->devices_max_used; i++) {
 		struct bcache_device *d = c->devices[i];
 
 		if (!d || !UUID_FLASH_ONLY(&c->uuids[i]))

drivers/md/dm-crypt.c

@@ -1446,7 +1446,6 @@ static void crypt_free_buffer_pages(struct crypt_config *cc, struct bio *clone)
 	bio_for_each_segment_all(bv, clone, i) {
 		BUG_ON(!bv->bv_page);
 		mempool_free(bv->bv_page, cc->page_pool);
-		bv->bv_page = NULL;
 	}
 }
 

drivers/md/dm-mpath.c

@@ -1475,21 +1475,6 @@ static void activate_path_work(struct work_struct *work)
 	activate_or_offline_path(pgpath);
 }
 
-static int noretry_error(blk_status_t error)
-{
-	switch (error) {
-	case BLK_STS_NOTSUPP:
-	case BLK_STS_NOSPC:
-	case BLK_STS_TARGET:
-	case BLK_STS_NEXUS:
-	case BLK_STS_MEDIUM:
-		return 1;
-	}
-
-	/* Anything else could be a path failure, so should be retried */
-	return 0;
-}
-
 static int multipath_end_io(struct dm_target *ti, struct request *clone,
 			    blk_status_t error, union map_info *map_context)
 {
@@ -1508,7 +1493,7 @@ static int multipath_end_io(struct dm_target *ti, struct request *clone,
 	 * request into dm core, which will remake a clone request and
 	 * clone bios for it and resubmit it later.
 	 */
-	if (error && !noretry_error(error)) {
+	if (error && blk_path_error(error)) {
 		struct multipath *m = ti->private;
 
 		r = DM_ENDIO_REQUEUE;
@@ -1544,7 +1529,7 @@ static int multipath_end_io_bio(struct dm_target *ti, struct bio *clone,
 	unsigned long flags;
 	int r = DM_ENDIO_DONE;
 
-	if (!*error || noretry_error(*error))
+	if (!*error || !blk_path_error(*error))
 		goto done;
 
 	if (pgpath)

drivers/md/dm-rq.c

@@ -395,7 +395,7 @@ static void end_clone_request(struct request *clone, blk_status_t error)
 	dm_complete_request(tio->orig, error);
 }
 
-static void dm_dispatch_clone_request(struct request *clone, struct request *rq)
+static blk_status_t dm_dispatch_clone_request(struct request *clone, struct request *rq)
 {
 	blk_status_t r;
 
@@ -404,9 +404,10 @@ static void dm_dispatch_clone_request(struct request *clone, struct request *rq)
 
 	clone->start_time = jiffies;
 	r = blk_insert_cloned_request(clone->q, clone);
-	if (r)
+	if (r != BLK_STS_OK && r != BLK_STS_RESOURCE)
 		/* must complete clone in terms of original request */
 		dm_complete_request(rq, r);
+	return r;
 }
 
 static int dm_rq_bio_constructor(struct bio *bio, struct bio *bio_orig,
@@ -476,8 +477,10 @@ static int map_request(struct dm_rq_target_io *tio)
 	struct mapped_device *md = tio->md;
 	struct request *rq = tio->orig;
 	struct request *clone = NULL;
+	blk_status_t ret;
 
 	r = ti->type->clone_and_map_rq(ti, rq, &tio->info, &clone);
+check_again:
 	switch (r) {
 	case DM_MAPIO_SUBMITTED:
 		/* The target has taken the I/O to submit by itself later */
@@ -492,7 +495,17 @@ static int map_request(struct dm_rq_target_io *tio)
 		/* The target has remapped the I/O so dispatch it */
 		trace_block_rq_remap(clone->q, clone, disk_devt(dm_disk(md)),
 				     blk_rq_pos(rq));
-		dm_dispatch_clone_request(clone, rq);
+		ret = dm_dispatch_clone_request(clone, rq);
+		if (ret == BLK_STS_RESOURCE) {
+			blk_rq_unprep_clone(clone);
+			tio->ti->type->release_clone_rq(clone);
+			tio->clone = NULL;
+			if (!rq->q->mq_ops)
+				r = DM_MAPIO_DELAY_REQUEUE;
+			else
+				r = DM_MAPIO_REQUEUE;
+			goto check_again;
+		}
 		break;
 	case DM_MAPIO_REQUEUE:
 		/* The target wants to requeue the I/O */
@@ -713,8 +726,6 @@ int dm_old_init_request_queue(struct mapped_device *md, struct dm_table *t)
 		return error;
 	}
 
-	elv_register_queue(md->queue);
-
 	return 0;
 }
 
@@ -812,15 +823,8 @@ int dm_mq_init_request_queue(struct mapped_device *md, struct dm_table *t)
 	}
 	dm_init_md_queue(md);
 
-	/* backfill 'mq' sysfs registration normally done in blk_register_queue */
-	err = blk_mq_register_dev(disk_to_dev(md->disk), q);
-	if (err)
-		goto out_cleanup_queue;
-
 	return 0;
 
-out_cleanup_queue:
-	blk_cleanup_queue(q);
 out_tag_set:
 	blk_mq_free_tag_set(md->tag_set);
 out_kfree_tag_set:

drivers/md/dm.c

@@ -920,7 +920,15 @@ int dm_set_target_max_io_len(struct dm_target *ti, sector_t len)
 		return -EINVAL;
 	}
 
-	ti->max_io_len = (uint32_t) len;
+	/*
+	 * BIO based queue uses its own splitting. When multipage bvecs
+	 * is switched on, size of the incoming bio may be too big to
+	 * be handled in some targets, such as crypt.
+	 *
+	 * When these targets are ready for the big bio, we can remove
+	 * the limit.
+	 */
+	ti->max_io_len = min_t(uint32_t, len, BIO_MAX_PAGES * PAGE_SIZE);
 
 	return 0;
 }
@@ -1753,7 +1761,7 @@ static struct mapped_device *alloc_dev(int minor)
 		goto bad;
 	md->dax_dev = dax_dev;
 
-	add_disk(md->disk);
+	add_disk_no_queue_reg(md->disk);
 	format_dev_t(md->name, MKDEV(_major, minor));
 
 	md->wq = alloc_workqueue("kdmflush", WQ_MEM_RECLAIM, 0);
@@ -2013,6 +2021,7 @@ EXPORT_SYMBOL_GPL(dm_get_queue_limits);
 int dm_setup_md_queue(struct mapped_device *md, struct dm_table *t)
 {
 	int r;
+	struct queue_limits limits;
 	enum dm_queue_mode type = dm_get_md_type(md);
 
 	switch (type) {
@@ -2049,6 +2058,14 @@ int dm_setup_md_queue(struct mapped_device *md, struct dm_table *t)
 		break;
 	}
 
+	r = dm_calculate_queue_limits(t, &limits);
+	if (r) {
+		DMERR("Cannot calculate initial queue limits");
+		return r;
+	}
+	dm_table_set_restrictions(t, md->queue, &limits);
+	blk_register_queue(md->disk);
+
 	return 0;
 }
 

drivers/nvme/host/Makefile

@@ -1,4 +1,7 @@
 # SPDX-License-Identifier: GPL-2.0
+
+ccflags-y				+= -I$(src)
+
 obj-$(CONFIG_NVME_CORE)			+= nvme-core.o
 obj-$(CONFIG_BLK_DEV_NVME)		+= nvme.o
 obj-$(CONFIG_NVME_FABRICS)		+= nvme-fabrics.o
@@ -6,6 +9,7 @@ obj-$(CONFIG_NVME_RDMA)			+= nvme-rdma.o
 obj-$(CONFIG_NVME_FC)			+= nvme-fc.o
 
 nvme-core-y				:= core.o
+nvme-core-$(CONFIG_TRACING)		+= trace.o
 nvme-core-$(CONFIG_NVME_MULTIPATH)	+= multipath.o
 nvme-core-$(CONFIG_NVM)			+= lightnvm.o
 

drivers/nvme/host/core.c

@@ -29,6 +29,9 @@
 #include <linux/pm_qos.h>
 #include <asm/unaligned.h>
 
+#define CREATE_TRACE_POINTS
+#include "trace.h"
+
 #include "nvme.h"
 #include "fabrics.h"
 
@@ -65,9 +68,26 @@ static bool streams;
 module_param(streams, bool, 0644);
 MODULE_PARM_DESC(streams, "turn on support for Streams write directives");
 
+/*
+ * nvme_wq - hosts nvme related works that are not reset or delete
+ * nvme_reset_wq - hosts nvme reset works
+ * nvme_delete_wq - hosts nvme delete works
+ *
+ * nvme_wq will host works such are scan, aen handling, fw activation,
+ * keep-alive error recovery, periodic reconnects etc. nvme_reset_wq
+ * runs reset works which also flush works hosted on nvme_wq for
+ * serialization purposes. nvme_delete_wq host controller deletion
+ * works which flush reset works for serialization.
+ */
 struct workqueue_struct *nvme_wq;
 EXPORT_SYMBOL_GPL(nvme_wq);
 
+struct workqueue_struct *nvme_reset_wq;
+EXPORT_SYMBOL_GPL(nvme_reset_wq);
+
+struct workqueue_struct *nvme_delete_wq;
+EXPORT_SYMBOL_GPL(nvme_delete_wq);
+
 static DEFINE_IDA(nvme_subsystems_ida);
 static LIST_HEAD(nvme_subsystems);
 static DEFINE_MUTEX(nvme_subsystems_lock);
@@ -89,13 +109,13 @@ int nvme_reset_ctrl(struct nvme_ctrl *ctrl)
 {
 	if (!nvme_change_ctrl_state(ctrl, NVME_CTRL_RESETTING))
 		return -EBUSY;
-	if (!queue_work(nvme_wq, &ctrl->reset_work))
+	if (!queue_work(nvme_reset_wq, &ctrl->reset_work))
 		return -EBUSY;
 	return 0;
 }
 EXPORT_SYMBOL_GPL(nvme_reset_ctrl);
 
-static int nvme_reset_ctrl_sync(struct nvme_ctrl *ctrl)
+int nvme_reset_ctrl_sync(struct nvme_ctrl *ctrl)
 {
 	int ret;
 
@@ -104,6 +124,7 @@ static int nvme_reset_ctrl_sync(struct nvme_ctrl *ctrl)
 		flush_work(&ctrl->reset_work);
 	return ret;
 }
+EXPORT_SYMBOL_GPL(nvme_reset_ctrl_sync);
 
 static void nvme_delete_ctrl_work(struct work_struct *work)
 {
@@ -122,7 +143,7 @@ int nvme_delete_ctrl(struct nvme_ctrl *ctrl)
 {
 	if (!nvme_change_ctrl_state(ctrl, NVME_CTRL_DELETING))
 		return -EBUSY;
-	if (!queue_work(nvme_wq, &ctrl->delete_work))
+	if (!queue_work(nvme_delete_wq, &ctrl->delete_work))
 		return -EBUSY;
 	return 0;
 }
@@ -157,13 +178,20 @@ static blk_status_t nvme_error_status(struct request *req)
 		return BLK_STS_OK;
 	case NVME_SC_CAP_EXCEEDED:
 		return BLK_STS_NOSPC;
+	case NVME_SC_LBA_RANGE:
+		return BLK_STS_TARGET;
+	case NVME_SC_BAD_ATTRIBUTES:
 	case NVME_SC_ONCS_NOT_SUPPORTED:
+	case NVME_SC_INVALID_OPCODE:
+	case NVME_SC_INVALID_FIELD:
+	case NVME_SC_INVALID_NS:
 		return BLK_STS_NOTSUPP;
 	case NVME_SC_WRITE_FAULT:
 	case NVME_SC_READ_ERROR:
 	case NVME_SC_UNWRITTEN_BLOCK:
 	case NVME_SC_ACCESS_DENIED:
 	case NVME_SC_READ_ONLY:
+	case NVME_SC_COMPARE_FAILED:
 		return BLK_STS_MEDIUM;
 	case NVME_SC_GUARD_CHECK:
 	case NVME_SC_APPTAG_CHECK:
@@ -190,8 +218,12 @@ static inline bool nvme_req_needs_retry(struct request *req)
 
 void nvme_complete_rq(struct request *req)
 {
-	if (unlikely(nvme_req(req)->status && nvme_req_needs_retry(req))) {
-		if (nvme_req_needs_failover(req)) {
+	blk_status_t status = nvme_error_status(req);
+
+	trace_nvme_complete_rq(req);
+
+	if (unlikely(status != BLK_STS_OK && nvme_req_needs_retry(req))) {
+		if (nvme_req_needs_failover(req, status)) {
 			nvme_failover_req(req);
 			return;
 		}
@@ -202,8 +234,7 @@ void nvme_complete_rq(struct request *req)
 			return;
 		}
 	}
-
-	blk_mq_end_request(req, nvme_error_status(req));
+	blk_mq_end_request(req, status);
 }
 EXPORT_SYMBOL_GPL(nvme_complete_rq);
 
@@ -232,6 +263,15 @@ bool nvme_change_ctrl_state(struct nvme_ctrl *ctrl,
 
 	old_state = ctrl->state;
 	switch (new_state) {
+	case NVME_CTRL_ADMIN_ONLY:
+		switch (old_state) {
+		case NVME_CTRL_RECONNECTING:
+			changed = true;
+			/* FALLTHRU */
+		default:
+			break;
+		}
+		break;
 	case NVME_CTRL_LIVE:
 		switch (old_state) {
 		case NVME_CTRL_NEW:
@@ -247,6 +287,7 @@ bool nvme_change_ctrl_state(struct nvme_ctrl *ctrl,
 		switch (old_state) {
 		case NVME_CTRL_NEW:
 		case NVME_CTRL_LIVE:
+		case NVME_CTRL_ADMIN_ONLY:
 			changed = true;
 			/* FALLTHRU */
 		default:
@@ -266,6 +307,7 @@ bool nvme_change_ctrl_state(struct nvme_ctrl *ctrl,
 	case NVME_CTRL_DELETING:
 		switch (old_state) {
 		case NVME_CTRL_LIVE:
+		case NVME_CTRL_ADMIN_ONLY:
 		case NVME_CTRL_RESETTING:
 		case NVME_CTRL_RECONNECTING:
 			changed = true;
@@ -591,6 +633,10 @@ blk_status_t nvme_setup_cmd(struct nvme_ns *ns, struct request *req,
 	}
 
 	cmd->common.command_id = req->tag;
+	if (ns)
+		trace_nvme_setup_nvm_cmd(req->q->id, cmd);
+	else
+		trace_nvme_setup_admin_cmd(cmd);
 	return ret;
 }
 EXPORT_SYMBOL_GPL(nvme_setup_cmd);
@@ -1217,16 +1263,27 @@ static int nvme_open(struct block_device *bdev, fmode_t mode)
 #ifdef CONFIG_NVME_MULTIPATH
 	/* should never be called due to GENHD_FL_HIDDEN */
 	if (WARN_ON_ONCE(ns->head->disk))
-		return -ENXIO;
+		goto fail;
 #endif
 	if (!kref_get_unless_zero(&ns->kref))
-		return -ENXIO;
+		goto fail;
+	if (!try_module_get(ns->ctrl->ops->module))
+		goto fail_put_ns;
+
 	return 0;
+
+fail_put_ns:
+	nvme_put_ns(ns);
+fail:
+	return -ENXIO;
 }
 
 static void nvme_release(struct gendisk *disk, fmode_t mode)
 {
-	nvme_put_ns(disk->private_data);
+	struct nvme_ns *ns = disk->private_data;
+
+	module_put(ns->ctrl->ops->module);
+	nvme_put_ns(ns);
 }
 
 static int nvme_getgeo(struct block_device *bdev, struct hd_geometry *geo)
@@ -2052,6 +2109,22 @@ static const struct attribute_group *nvme_subsys_attrs_groups[] = {
 	NULL,
 };
 
+static int nvme_active_ctrls(struct nvme_subsystem *subsys)
+{
+	int count = 0;
+	struct nvme_ctrl *ctrl;
+
+	mutex_lock(&subsys->lock);
+	list_for_each_entry(ctrl, &subsys->ctrls, subsys_entry) {
+		if (ctrl->state != NVME_CTRL_DELETING &&
+		    ctrl->state != NVME_CTRL_DEAD)
+			count++;
+	}
+	mutex_unlock(&subsys->lock);
+
+	return count;
+}
+
 static int nvme_init_subsystem(struct nvme_ctrl *ctrl, struct nvme_id_ctrl *id)
 {
 	struct nvme_subsystem *subsys, *found;
@@ -2090,7 +2163,7 @@ static int nvme_init_subsystem(struct nvme_ctrl *ctrl, struct nvme_id_ctrl *id)
 		 * Verify that the subsystem actually supports multiple
 		 * controllers, else bail out.
 		 */
-		if (!(id->cmic & (1 << 1))) {
+		if (nvme_active_ctrls(found) && !(id->cmic & (1 << 1))) {
 			dev_err(ctrl->device,
 				"ignoring ctrl due to duplicate subnqn (%s).\n",
 				found->subnqn);
@@ -2257,7 +2330,7 @@ int nvme_init_identify(struct nvme_ctrl *ctrl)
 						 shutdown_timeout, 60);
 
 		if (ctrl->shutdown_timeout != shutdown_timeout)
-			dev_warn(ctrl->device,
+			dev_info(ctrl->device,
 				 "Shutdown timeout set to %u seconds\n",
 				 ctrl->shutdown_timeout);
 	} else
@@ -2341,8 +2414,14 @@ static int nvme_dev_open(struct inode *inode, struct file *file)
 	struct nvme_ctrl *ctrl =
 		container_of(inode->i_cdev, struct nvme_ctrl, cdev);
 
-	if (ctrl->state != NVME_CTRL_LIVE)
+	switch (ctrl->state) {
+	case NVME_CTRL_LIVE:
+	case NVME_CTRL_ADMIN_ONLY:
+		break;
+	default:
 		return -EWOULDBLOCK;
+	}
+
 	file->private_data = ctrl;
 	return 0;
 }
@@ -2606,6 +2685,7 @@ static ssize_t nvme_sysfs_show_state(struct device *dev,
 	static const char *const state_name[] = {
 		[NVME_CTRL_NEW]		= "new",
 		[NVME_CTRL_LIVE]	= "live",
+		[NVME_CTRL_ADMIN_ONLY]	= "only-admin",
 		[NVME_CTRL_RESETTING]	= "resetting",
 		[NVME_CTRL_RECONNECTING]= "reconnecting",
 		[NVME_CTRL_DELETING]	= "deleting",
@@ -3079,6 +3159,8 @@ static void nvme_scan_work(struct work_struct *work)
 	if (ctrl->state != NVME_CTRL_LIVE)
 		return;
 
+	WARN_ON_ONCE(!ctrl->tagset);
+
 	if (nvme_identify_ctrl(ctrl, &id))
 		return;
 
@@ -3099,8 +3181,7 @@ static void nvme_scan_work(struct work_struct *work)
 void nvme_queue_scan(struct nvme_ctrl *ctrl)
 {
 	/*
-	 * Do not queue new scan work when a controller is reset during
-	 * removal.
+	 * Only new queue scan work when admin and IO queues are both alive
 	 */
 	if (ctrl->state == NVME_CTRL_LIVE)
 		queue_work(nvme_wq, &ctrl->scan_work);
@@ -3477,16 +3558,26 @@ EXPORT_SYMBOL_GPL(nvme_reinit_tagset);
 
 int __init nvme_core_init(void)
 {
-	int result;
+	int result = -ENOMEM;
 
 	nvme_wq = alloc_workqueue("nvme-wq",
 			WQ_UNBOUND | WQ_MEM_RECLAIM | WQ_SYSFS, 0);
 	if (!nvme_wq)
-		return -ENOMEM;
+		goto out;
+
+	nvme_reset_wq = alloc_workqueue("nvme-reset-wq",
+			WQ_UNBOUND | WQ_MEM_RECLAIM | WQ_SYSFS, 0);
+	if (!nvme_reset_wq)
+		goto destroy_wq;
+
+	nvme_delete_wq = alloc_workqueue("nvme-delete-wq",
+			WQ_UNBOUND | WQ_MEM_RECLAIM | WQ_SYSFS, 0);
+	if (!nvme_delete_wq)
+		goto destroy_reset_wq;
 
 	result = alloc_chrdev_region(&nvme_chr_devt, 0, NVME_MINORS, "nvme");
 	if (result < 0)
-		goto destroy_wq;
+		goto destroy_delete_wq;
 
 	nvme_class = class_create(THIS_MODULE, "nvme");
 	if (IS_ERR(nvme_class)) {
@@ -3505,8 +3596,13 @@ destroy_class:
 	class_destroy(nvme_class);
 unregister_chrdev:
 	unregister_chrdev_region(nvme_chr_devt, NVME_MINORS);
+destroy_delete_wq:
+	destroy_workqueue(nvme_delete_wq);
+destroy_reset_wq:
+	destroy_workqueue(nvme_reset_wq);
 destroy_wq:
 	destroy_workqueue(nvme_wq);
+out:
 	return result;
 }
 
@@ -3516,6 +3612,8 @@ void nvme_core_exit(void)
 	class_destroy(nvme_subsys_class);
 	class_destroy(nvme_class);
 	unregister_chrdev_region(nvme_chr_devt, NVME_MINORS);
+	destroy_workqueue(nvme_delete_wq);
+	destroy_workqueue(nvme_reset_wq);
 	destroy_workqueue(nvme_wq);
 }
 

drivers/nvme/host/fabrics.c

@@ -493,7 +493,7 @@ EXPORT_SYMBOL_GPL(nvmf_should_reconnect);
  */
 int nvmf_register_transport(struct nvmf_transport_ops *ops)
 {
-	if (!ops->create_ctrl)
+	if (!ops->create_ctrl || !ops->module)
 		return -EINVAL;
 
 	down_write(&nvmf_transports_rwsem);
@@ -739,11 +739,14 @@ static int nvmf_parse_options(struct nvmf_ctrl_options *opts,
 				ret = -ENOMEM;
 				goto out;
 			}
-			if (uuid_parse(p, &hostid)) {
+			ret = uuid_parse(p, &hostid);
+			if (ret) {
 				pr_err("Invalid hostid %s\n", p);
 				ret = -EINVAL;
+				kfree(p);
 				goto out;
 			}
+			kfree(p);
 			break;
 		case NVMF_OPT_DUP_CONNECT:
 			opts->duplicate_connect = true;
@@ -869,32 +872,41 @@ nvmf_create_ctrl(struct device *dev, const char *buf, size_t count)
 		goto out_unlock;
 	}
 
+	if (!try_module_get(ops->module)) {
+		ret = -EBUSY;
+		goto out_unlock;
+	}
+
 	ret = nvmf_check_required_opts(opts, ops->required_opts);
 	if (ret)
-		goto out_unlock;
+		goto out_module_put;
 	ret = nvmf_check_allowed_opts(opts, NVMF_ALLOWED_OPTS |
 				ops->allowed_opts | ops->required_opts);
 	if (ret)
-		goto out_unlock;
+		goto out_module_put;
 
 	ctrl = ops->create_ctrl(dev, opts);
 	if (IS_ERR(ctrl)) {
 		ret = PTR_ERR(ctrl);
-		goto out_unlock;
+		goto out_module_put;
 	}
 
 	if (strcmp(ctrl->subsys->subnqn, opts->subsysnqn)) {
 		dev_warn(ctrl->device,
 			"controller returned incorrect NQN: \"%s\".\n",
 			ctrl->subsys->subnqn);
+		module_put(ops->module);
 		up_read(&nvmf_transports_rwsem);
 		nvme_delete_ctrl_sync(ctrl);
 		return ERR_PTR(-EINVAL);
 	}
 
+	module_put(ops->module);
 	up_read(&nvmf_transports_rwsem);
 	return ctrl;
 
+out_module_put:
+	module_put(ops->module);
 out_unlock:
 	up_read(&nvmf_transports_rwsem);
 out_free_opts:

drivers/nvme/host/fabrics.h

@@ -108,6 +108,7 @@ struct nvmf_ctrl_options {
  *			       fabric implementation of NVMe fabrics.
  * @entry:		Used by the fabrics library to add the new
  *			registration entry to its linked-list internal tree.
+ * @module:             Transport module reference
  * @name:		Name of the NVMe fabric driver implementation.
  * @required_opts:	sysfs command-line options that must be specified
  *			when adding a new NVMe controller.
@@ -126,6 +127,7 @@ struct nvmf_ctrl_options {
  */
 struct nvmf_transport_ops {
 	struct list_head	entry;
+	struct module		*module;
 	const char		*name;
 	int			required_opts;
 	int			allowed_opts;

drivers/nvme/host/fc.c

@@ -2921,6 +2921,9 @@ nvme_fc_delete_association(struct nvme_fc_ctrl *ctrl)
 	__nvme_fc_delete_hw_queue(ctrl, &ctrl->queues[0], 0);
 	nvme_fc_free_queue(&ctrl->queues[0]);
 
+	/* re-enable the admin_q so anything new can fast fail */
+	blk_mq_unquiesce_queue(ctrl->ctrl.admin_q);
+
 	nvme_fc_ctlr_inactive_on_rport(ctrl);
 }
 
@@ -2935,6 +2938,9 @@ nvme_fc_delete_ctrl(struct nvme_ctrl *nctrl)
 	 * waiting for io to terminate
 	 */
 	nvme_fc_delete_association(ctrl);
+
+	/* resume the io queues so that things will fast fail */
+	nvme_start_queues(nctrl);
 }
 
 static void
@@ -3380,6 +3386,7 @@ nvme_fc_create_ctrl(struct device *dev, struct nvmf_ctrl_options *opts)
 
 static struct nvmf_transport_ops nvme_fc_transport = {
 	.name		= "fc",
+	.module		= THIS_MODULE,
 	.required_opts	= NVMF_OPT_TRADDR | NVMF_OPT_HOST_TRADDR,
 	.allowed_opts	= NVMF_OPT_RECONNECT_DELAY | NVMF_OPT_CTRL_LOSS_TMO,
 	.create_ctrl	= nvme_fc_create_ctrl,

drivers/nvme/host/lightnvm.c

@@ -31,27 +31,10 @@
 
 enum nvme_nvm_admin_opcode {
 	nvme_nvm_admin_identity		= 0xe2,
-	nvme_nvm_admin_get_l2p_tbl	= 0xea,
 	nvme_nvm_admin_get_bb_tbl	= 0xf2,
 	nvme_nvm_admin_set_bb_tbl	= 0xf1,
 };
 
-struct nvme_nvm_hb_rw {
-	__u8			opcode;
-	__u8			flags;
-	__u16			command_id;
-	__le32			nsid;
-	__u64			rsvd2;
-	__le64			metadata;
-	__le64			prp1;
-	__le64			prp2;
-	__le64			spba;
-	__le16			length;
-	__le16			control;
-	__le32			dsmgmt;
-	__le64			slba;
-};
-
 struct nvme_nvm_ph_rw {
 	__u8			opcode;
 	__u8			flags;
@@ -80,19 +63,6 @@ struct nvme_nvm_identity {
 	__u32			rsvd11[5];
 };
 
-struct nvme_nvm_l2ptbl {
-	__u8			opcode;
-	__u8			flags;
-	__u16			command_id;
-	__le32			nsid;
-	__le32			cdw2[4];
-	__le64			prp1;
-	__le64			prp2;
-	__le64			slba;
-	__le32			nlb;
-	__le16			cdw14[6];
-};
-
 struct nvme_nvm_getbbtbl {
 	__u8			opcode;
 	__u8			flags;
@@ -139,9 +109,7 @@ struct nvme_nvm_command {
 	union {
 		struct nvme_common_command common;
 		struct nvme_nvm_identity identity;
-		struct nvme_nvm_hb_rw hb_rw;
 		struct nvme_nvm_ph_rw ph_rw;
-		struct nvme_nvm_l2ptbl l2p;
 		struct nvme_nvm_getbbtbl get_bb;
 		struct nvme_nvm_setbbtbl set_bb;
 		struct nvme_nvm_erase_blk erase;
@@ -167,7 +135,7 @@ struct nvme_nvm_id_group {
 	__u8			num_lun;
 	__u8			num_pln;
 	__u8			rsvd1;
-	__le16			num_blk;
+	__le16			num_chk;
 	__le16			num_pg;
 	__le16			fpg_sz;
 	__le16			csecs;
@@ -234,11 +202,9 @@ struct nvme_nvm_bb_tbl {
 static inline void _nvme_nvm_check_size(void)
 {
 	BUILD_BUG_ON(sizeof(struct nvme_nvm_identity) != 64);
-	BUILD_BUG_ON(sizeof(struct nvme_nvm_hb_rw) != 64);
 	BUILD_BUG_ON(sizeof(struct nvme_nvm_ph_rw) != 64);
 	BUILD_BUG_ON(sizeof(struct nvme_nvm_getbbtbl) != 64);
 	BUILD_BUG_ON(sizeof(struct nvme_nvm_setbbtbl) != 64);
-	BUILD_BUG_ON(sizeof(struct nvme_nvm_l2ptbl) != 64);
 	BUILD_BUG_ON(sizeof(struct nvme_nvm_erase_blk) != 64);
 	BUILD_BUG_ON(sizeof(struct nvme_nvm_id_group) != 960);
 	BUILD_BUG_ON(sizeof(struct nvme_nvm_addr_format) != 16);
@@ -249,51 +215,58 @@ static inline void _nvme_nvm_check_size(void)
 static int init_grps(struct nvm_id *nvm_id, struct nvme_nvm_id *nvme_nvm_id)
 {
 	struct nvme_nvm_id_group *src;
-	struct nvm_id_group *dst;
+	struct nvm_id_group *grp;
+	int sec_per_pg, sec_per_pl, pg_per_blk;
 
 	if (nvme_nvm_id->cgrps != 1)
 		return -EINVAL;
 
 	src = &nvme_nvm_id->groups[0];
-	dst = &nvm_id->grp;
-
-	dst->mtype = src->mtype;
-	dst->fmtype = src->fmtype;
-	dst->num_ch = src->num_ch;
-	dst->num_lun = src->num_lun;
-	dst->num_pln = src->num_pln;
-
-	dst->num_pg = le16_to_cpu(src->num_pg);
-	dst->num_blk = le16_to_cpu(src->num_blk);
-	dst->fpg_sz = le16_to_cpu(src->fpg_sz);
-	dst->csecs = le16_to_cpu(src->csecs);
-	dst->sos = le16_to_cpu(src->sos);
-
-	dst->trdt = le32_to_cpu(src->trdt);
-	dst->trdm = le32_to_cpu(src->trdm);
-	dst->tprt = le32_to_cpu(src->tprt);
-	dst->tprm = le32_to_cpu(src->tprm);
-	dst->tbet = le32_to_cpu(src->tbet);
-	dst->tbem = le32_to_cpu(src->tbem);
-	dst->mpos = le32_to_cpu(src->mpos);
-	dst->mccap = le32_to_cpu(src->mccap);
-
-	dst->cpar = le16_to_cpu(src->cpar);
-
-	if (dst->fmtype == NVM_ID_FMTYPE_MLC) {
-		memcpy(dst->lptbl.id, src->lptbl.id, 8);
-		dst->lptbl.mlc.num_pairs =
-				le16_to_cpu(src->lptbl.mlc.num_pairs);
-
-		if (dst->lptbl.mlc.num_pairs > NVME_NVM_LP_MLC_PAIRS) {
-			pr_err("nvm: number of MLC pairs not supported\n");
-			return -EINVAL;
-		}
+	grp = &nvm_id->grp;
+
+	grp->mtype = src->mtype;
+	grp->fmtype = src->fmtype;
+
+	grp->num_ch = src->num_ch;
+	grp->num_lun = src->num_lun;
+
+	grp->num_chk = le16_to_cpu(src->num_chk);
+	grp->csecs = le16_to_cpu(src->csecs);
+	grp->sos = le16_to_cpu(src->sos);
+
+	pg_per_blk = le16_to_cpu(src->num_pg);
+	sec_per_pg = le16_to_cpu(src->fpg_sz) / grp->csecs;
+	sec_per_pl = sec_per_pg * src->num_pln;
+	grp->clba = sec_per_pl * pg_per_blk;
+	grp->ws_per_chk = pg_per_blk;
 
-		memcpy(dst->lptbl.mlc.pairs, src->lptbl.mlc.pairs,
-					dst->lptbl.mlc.num_pairs);
+	grp->mpos = le32_to_cpu(src->mpos);
+	grp->cpar = le16_to_cpu(src->cpar);
+	grp->mccap = le32_to_cpu(src->mccap);
+
+	grp->ws_opt = grp->ws_min = sec_per_pg;
+	grp->ws_seq = NVM_IO_SNGL_ACCESS;
+
+	if (grp->mpos & 0x020202) {
+		grp->ws_seq = NVM_IO_DUAL_ACCESS;
+		grp->ws_opt <<= 1;
+	} else if (grp->mpos & 0x040404) {
+		grp->ws_seq = NVM_IO_QUAD_ACCESS;
+		grp->ws_opt <<= 2;
 	}
 
+	grp->trdt = le32_to_cpu(src->trdt);
+	grp->trdm = le32_to_cpu(src->trdm);
+	grp->tprt = le32_to_cpu(src->tprt);
+	grp->tprm = le32_to_cpu(src->tprm);
+	grp->tbet = le32_to_cpu(src->tbet);
+	grp->tbem = le32_to_cpu(src->tbem);
+
+	/* 1.2 compatibility */
+	grp->num_pln = src->num_pln;
+	grp->num_pg = le16_to_cpu(src->num_pg);
+	grp->fpg_sz = le16_to_cpu(src->fpg_sz);
+
 	return 0;
 }
 
@@ -332,62 +305,6 @@ out:
 	return ret;
 }
 
-static int nvme_nvm_get_l2p_tbl(struct nvm_dev *nvmdev, u64 slba, u32 nlb,
-				nvm_l2p_update_fn *update_l2p, void *priv)
-{
-	struct nvme_ns *ns = nvmdev->q->queuedata;
-	struct nvme_nvm_command c = {};
-	u32 len = queue_max_hw_sectors(ns->ctrl->admin_q) << 9;
-	u32 nlb_pr_rq = len / sizeof(u64);
-	u64 cmd_slba = slba;
-	void *entries;
-	int ret = 0;
-
-	c.l2p.opcode = nvme_nvm_admin_get_l2p_tbl;
-	c.l2p.nsid = cpu_to_le32(ns->head->ns_id);
-	entries = kmalloc(len, GFP_KERNEL);
-	if (!entries)
-		return -ENOMEM;
-
-	while (nlb) {
-		u32 cmd_nlb = min(nlb_pr_rq, nlb);
-		u64 elba = slba + cmd_nlb;
-
-		c.l2p.slba = cpu_to_le64(cmd_slba);
-		c.l2p.nlb = cpu_to_le32(cmd_nlb);
-
-		ret = nvme_submit_sync_cmd(ns->ctrl->admin_q,
-				(struct nvme_command *)&c, entries, len);
-		if (ret) {
-			dev_err(ns->ctrl->device,
-				"L2P table transfer failed (%d)\n", ret);
-			ret = -EIO;
-			goto out;
-		}
-
-		if (unlikely(elba > nvmdev->total_secs)) {
-			pr_err("nvm: L2P data from device is out of bounds!\n");
-			ret = -EINVAL;
-			goto out;
-		}
-
-		/* Transform physical address to target address space */
-		nvm_part_to_tgt(nvmdev, entries, cmd_nlb);
-
-		if (update_l2p(cmd_slba, cmd_nlb, entries, priv)) {
-			ret = -EINTR;
-			goto out;
-		}
-
-		cmd_slba += cmd_nlb;
-		nlb -= cmd_nlb;
-	}
-
-out:
-	kfree(entries);
-	return ret;
-}
-
 static int nvme_nvm_get_bb_tbl(struct nvm_dev *nvmdev, struct ppa_addr ppa,
 								u8 *blks)
 {
@@ -397,7 +314,7 @@ static int nvme_nvm_get_bb_tbl(struct nvm_dev *nvmdev, struct ppa_addr ppa,
 	struct nvme_ctrl *ctrl = ns->ctrl;
 	struct nvme_nvm_command c = {};
 	struct nvme_nvm_bb_tbl *bb_tbl;
-	int nr_blks = geo->blks_per_lun * geo->plane_mode;
+	int nr_blks = geo->nr_chks * geo->plane_mode;
 	int tblsz = sizeof(struct nvme_nvm_bb_tbl) + nr_blks;
 	int ret = 0;
 
@@ -438,7 +355,7 @@ static int nvme_nvm_get_bb_tbl(struct nvm_dev *nvmdev, struct ppa_addr ppa,
 		goto out;
 	}
 
-	memcpy(blks, bb_tbl->blk, geo->blks_per_lun * geo->plane_mode);
+	memcpy(blks, bb_tbl->blk, geo->nr_chks * geo->plane_mode);
 out:
 	kfree(bb_tbl);
 	return ret;
@@ -474,10 +391,6 @@ static inline void nvme_nvm_rqtocmd(struct nvm_rq *rqd, struct nvme_ns *ns,
 	c->ph_rw.metadata = cpu_to_le64(rqd->dma_meta_list);
 	c->ph_rw.control = cpu_to_le16(rqd->flags);
 	c->ph_rw.length = cpu_to_le16(rqd->nr_ppas - 1);
-
-	if (rqd->opcode == NVM_OP_HBWRITE || rqd->opcode == NVM_OP_HBREAD)
-		c->hb_rw.slba = cpu_to_le64(nvme_block_nr(ns,
-					rqd->bio->bi_iter.bi_sector));
 }
 
 static void nvme_nvm_end_io(struct request *rq, blk_status_t status)
@@ -597,8 +510,6 @@ static void nvme_nvm_dev_dma_free(void *pool, void *addr,
 static struct nvm_dev_ops nvme_nvm_dev_ops = {
 	.identity		= nvme_nvm_identity,
 
-	.get_l2p_tbl		= nvme_nvm_get_l2p_tbl,
-
 	.get_bb_tbl		= nvme_nvm_get_bb_tbl,
 	.set_bb_tbl		= nvme_nvm_set_bb_tbl,
 
@@ -883,7 +794,7 @@ static ssize_t nvm_dev_attr_show(struct device *dev,
 	} else if (strcmp(attr->name, "num_planes") == 0) {
 		return scnprintf(page, PAGE_SIZE, "%u\n", grp->num_pln);
 	} else if (strcmp(attr->name, "num_blocks") == 0) {	/* u16 */
-		return scnprintf(page, PAGE_SIZE, "%u\n", grp->num_blk);
+		return scnprintf(page, PAGE_SIZE, "%u\n", grp->num_chk);
 	} else if (strcmp(attr->name, "num_pages") == 0) {
 		return scnprintf(page, PAGE_SIZE, "%u\n", grp->num_pg);
 	} else if (strcmp(attr->name, "page_size") == 0) {

drivers/nvme/host/multipath.c

@@ -33,51 +33,11 @@ void nvme_failover_req(struct request *req)
 	kblockd_schedule_work(&ns->head->requeue_work);
 }
 
-bool nvme_req_needs_failover(struct request *req)
+bool nvme_req_needs_failover(struct request *req, blk_status_t error)
 {
 	if (!(req->cmd_flags & REQ_NVME_MPATH))
 		return false;
-
-	switch (nvme_req(req)->status & 0x7ff) {
-	/*
-	 * Generic command status:
-	 */
-	case NVME_SC_INVALID_OPCODE:
-	case NVME_SC_INVALID_FIELD:
-	case NVME_SC_INVALID_NS:
-	case NVME_SC_LBA_RANGE:
-	case NVME_SC_CAP_EXCEEDED:
-	case NVME_SC_RESERVATION_CONFLICT:
-		return false;
-
-	/*
-	 * I/O command set specific error.  Unfortunately these values are
-	 * reused for fabrics commands, but those should never get here.
-	 */
-	case NVME_SC_BAD_ATTRIBUTES:
-	case NVME_SC_INVALID_PI:
-	case NVME_SC_READ_ONLY:
-	case NVME_SC_ONCS_NOT_SUPPORTED:
-		WARN_ON_ONCE(nvme_req(req)->cmd->common.opcode ==
-			nvme_fabrics_command);
-		return false;
-
-	/*
-	 * Media and Data Integrity Errors:
-	 */
-	case NVME_SC_WRITE_FAULT:
-	case NVME_SC_READ_ERROR:
-	case NVME_SC_GUARD_CHECK:
-	case NVME_SC_APPTAG_CHECK:
-	case NVME_SC_REFTAG_CHECK:
-	case NVME_SC_COMPARE_FAILED:
-	case NVME_SC_ACCESS_DENIED:
-	case NVME_SC_UNWRITTEN_BLOCK:
-		return false;
-	}
-
-	/* Everything else could be a path failure, so should be retried */
-	return true;
+	return blk_path_error(error);
 }
 
 void nvme_kick_requeue_lists(struct nvme_ctrl *ctrl)

drivers/nvme/host/nvme.h

@@ -32,6 +32,8 @@ extern unsigned int admin_timeout;
 #define NVME_KATO_GRACE		10
 
 extern struct workqueue_struct *nvme_wq;
+extern struct workqueue_struct *nvme_reset_wq;
+extern struct workqueue_struct *nvme_delete_wq;
 
 enum {
 	NVME_NS_LBA		= 0,
@@ -119,6 +121,7 @@ static inline struct nvme_request *nvme_req(struct request *req)
 enum nvme_ctrl_state {
 	NVME_CTRL_NEW,
 	NVME_CTRL_LIVE,
+	NVME_CTRL_ADMIN_ONLY,    /* Only admin queue live */
 	NVME_CTRL_RESETTING,
 	NVME_CTRL_RECONNECTING,
 	NVME_CTRL_DELETING,
@@ -393,6 +396,7 @@ int nvme_set_queue_count(struct nvme_ctrl *ctrl, int *count);
 void nvme_start_keep_alive(struct nvme_ctrl *ctrl);
 void nvme_stop_keep_alive(struct nvme_ctrl *ctrl);
 int nvme_reset_ctrl(struct nvme_ctrl *ctrl);
+int nvme_reset_ctrl_sync(struct nvme_ctrl *ctrl);
 int nvme_delete_ctrl(struct nvme_ctrl *ctrl);
 int nvme_delete_ctrl_sync(struct nvme_ctrl *ctrl);
 
@@ -401,7 +405,7 @@ extern const struct block_device_operations nvme_ns_head_ops;
 
 #ifdef CONFIG_NVME_MULTIPATH
 void nvme_failover_req(struct request *req);
-bool nvme_req_needs_failover(struct request *req);
+bool nvme_req_needs_failover(struct request *req, blk_status_t error);
 void nvme_kick_requeue_lists(struct nvme_ctrl *ctrl);
 int nvme_mpath_alloc_disk(struct nvme_ctrl *ctrl,struct nvme_ns_head *head);
 void nvme_mpath_add_disk(struct nvme_ns_head *head);
@@ -430,7 +434,8 @@ static inline void nvme_mpath_check_last_path(struct nvme_ns *ns)
 static inline void nvme_failover_req(struct request *req)
 {
 }
-static inline bool nvme_req_needs_failover(struct request *req)
+static inline bool nvme_req_needs_failover(struct request *req,
+					   blk_status_t error)
 {
 	return false;
 }

drivers/nvme/host/pci.c

@@ -75,7 +75,7 @@ static void nvme_dev_disable(struct nvme_dev *dev, bool shutdown);
  * Represents an NVM Express device.  Each nvme_dev is a PCI function.
  */
 struct nvme_dev {
-	struct nvme_queue **queues;
+	struct nvme_queue *queues;
 	struct blk_mq_tag_set tagset;
 	struct blk_mq_tag_set admin_tagset;
 	u32 __iomem *dbs;
@@ -365,7 +365,7 @@ static int nvme_admin_init_hctx(struct blk_mq_hw_ctx *hctx, void *data,
 				unsigned int hctx_idx)
 {
 	struct nvme_dev *dev = data;
-	struct nvme_queue *nvmeq = dev->queues[0];
+	struct nvme_queue *nvmeq = &dev->queues[0];
 
 	WARN_ON(hctx_idx != 0);
 	WARN_ON(dev->admin_tagset.tags[0] != hctx->tags);
@@ -387,7 +387,7 @@ static int nvme_init_hctx(struct blk_mq_hw_ctx *hctx, void *data,
 			  unsigned int hctx_idx)
 {
 	struct nvme_dev *dev = data;
-	struct nvme_queue *nvmeq = dev->queues[hctx_idx + 1];
+	struct nvme_queue *nvmeq = &dev->queues[hctx_idx + 1];
 
 	if (!nvmeq->tags)
 		nvmeq->tags = &dev->tagset.tags[hctx_idx];
@@ -403,7 +403,7 @@ static int nvme_init_request(struct blk_mq_tag_set *set, struct request *req,
 	struct nvme_dev *dev = set->driver_data;
 	struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
 	int queue_idx = (set == &dev->tagset) ? hctx_idx + 1 : 0;
-	struct nvme_queue *nvmeq = dev->queues[queue_idx];
+	struct nvme_queue *nvmeq = &dev->queues[queue_idx];
 
 	BUG_ON(!nvmeq);
 	iod->nvmeq = nvmeq;
@@ -1044,7 +1044,7 @@ static int nvme_poll(struct blk_mq_hw_ctx *hctx, unsigned int tag)
 static void nvme_pci_submit_async_event(struct nvme_ctrl *ctrl)
 {
 	struct nvme_dev *dev = to_nvme_dev(ctrl);
-	struct nvme_queue *nvmeq = dev->queues[0];
+	struct nvme_queue *nvmeq = &dev->queues[0];
 	struct nvme_command c;
 
 	memset(&c, 0, sizeof(c));
@@ -1138,9 +1138,14 @@ static bool nvme_should_reset(struct nvme_dev *dev, u32 csts)
 	 */
 	bool nssro = dev->subsystem && (csts & NVME_CSTS_NSSRO);
 
-	/* If there is a reset ongoing, we shouldn't reset again. */
-	if (dev->ctrl.state == NVME_CTRL_RESETTING)
+	/* If there is a reset/reinit ongoing, we shouldn't reset again. */
+	switch (dev->ctrl.state) {
+	case NVME_CTRL_RESETTING:
+	case NVME_CTRL_RECONNECTING:
 		return false;
+	default:
+		break;
+	}
 
 	/* We shouldn't reset unless the controller is on fatal error state
 	 * _or_ if we lost the communication with it.
@@ -1280,7 +1285,6 @@ static void nvme_free_queue(struct nvme_queue *nvmeq)
 	if (nvmeq->sq_cmds)
 		dma_free_coherent(nvmeq->q_dmadev, SQ_SIZE(nvmeq->q_depth),
 					nvmeq->sq_cmds, nvmeq->sq_dma_addr);
-	kfree(nvmeq);
 }
 
 static void nvme_free_queues(struct nvme_dev *dev, int lowest)
@@ -1288,10 +1292,8 @@ static void nvme_free_queues(struct nvme_dev *dev, int lowest)
 	int i;
 
 	for (i = dev->ctrl.queue_count - 1; i >= lowest; i--) {
-		struct nvme_queue *nvmeq = dev->queues[i];
 		dev->ctrl.queue_count--;
-		dev->queues[i] = NULL;
-		nvme_free_queue(nvmeq);
+		nvme_free_queue(&dev->queues[i]);
 	}
 }
 
@@ -1323,12 +1325,7 @@ static int nvme_suspend_queue(struct nvme_queue *nvmeq)
 
 static void nvme_disable_admin_queue(struct nvme_dev *dev, bool shutdown)
 {
-	struct nvme_queue *nvmeq = dev->queues[0];
-
-	if (!nvmeq)
-		return;
-	if (nvme_suspend_queue(nvmeq))
-		return;
+	struct nvme_queue *nvmeq = &dev->queues[0];
 
 	if (shutdown)
 		nvme_shutdown_ctrl(&dev->ctrl);
@@ -1367,7 +1364,7 @@ static int nvme_cmb_qdepth(struct nvme_dev *dev, int nr_io_queues,
 static int nvme_alloc_sq_cmds(struct nvme_dev *dev, struct nvme_queue *nvmeq,
 				int qid, int depth)
 {
-	if (qid && dev->cmb && use_cmb_sqes && NVME_CMB_SQS(dev->cmbsz)) {
+	if (qid && dev->cmb && use_cmb_sqes && (dev->cmbsz & NVME_CMBSZ_SQS)) {
 		unsigned offset = (qid - 1) * roundup(SQ_SIZE(depth),
 						      dev->ctrl.page_size);
 		nvmeq->sq_dma_addr = dev->cmb_bus_addr + offset;
@@ -1382,13 +1379,13 @@ static int nvme_alloc_sq_cmds(struct nvme_dev *dev, struct nvme_queue *nvmeq,
 	return 0;
 }
 
-static struct nvme_queue *nvme_alloc_queue(struct nvme_dev *dev, int qid,
-							int depth, int node)
+static int nvme_alloc_queue(struct nvme_dev *dev, int qid,
+		int depth, int node)
 {
-	struct nvme_queue *nvmeq = kzalloc_node(sizeof(*nvmeq), GFP_KERNEL,
-							node);
-	if (!nvmeq)
-		return NULL;
+	struct nvme_queue *nvmeq = &dev->queues[qid];
+
+	if (dev->ctrl.queue_count > qid)
+		return 0;
 
 	nvmeq->cqes = dma_zalloc_coherent(dev->dev, CQ_SIZE(depth),
 					  &nvmeq->cq_dma_addr, GFP_KERNEL);
@@ -1407,17 +1404,15 @@ static struct nvme_queue *nvme_alloc_queue(struct nvme_dev *dev, int qid,
 	nvmeq->q_depth = depth;
 	nvmeq->qid = qid;
 	nvmeq->cq_vector = -1;
-	dev->queues[qid] = nvmeq;
 	dev->ctrl.queue_count++;
 
-	return nvmeq;
+	return 0;
 
  free_cqdma:
 	dma_free_coherent(dev->dev, CQ_SIZE(depth), (void *)nvmeq->cqes,
 							nvmeq->cq_dma_addr);
  free_nvmeq:
-	kfree(nvmeq);
-	return NULL;
+	return -ENOMEM;
 }
 
 static int queue_request_irq(struct nvme_queue *nvmeq)
@@ -1590,14 +1585,12 @@ static int nvme_pci_configure_admin_queue(struct nvme_dev *dev)
 	if (result < 0)
 		return result;
 
-	nvmeq = dev->queues[0];
-	if (!nvmeq) {
-		nvmeq = nvme_alloc_queue(dev, 0, NVME_AQ_DEPTH,
-					dev_to_node(dev->dev));
-		if (!nvmeq)
-			return -ENOMEM;
-	}
+	result = nvme_alloc_queue(dev, 0, NVME_AQ_DEPTH,
+			dev_to_node(dev->dev));
+	if (result)
+		return result;
 
+	nvmeq = &dev->queues[0];
 	aqa = nvmeq->q_depth - 1;
 	aqa |= aqa << 16;
 
@@ -1627,7 +1620,7 @@ static int nvme_create_io_queues(struct nvme_dev *dev)
 
 	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,
+		if (nvme_alloc_queue(dev, i, dev->q_depth,
 		     pci_irq_get_node(to_pci_dev(dev->dev), i - 1))) {
 			ret = -ENOMEM;
 			break;
@@ -1636,15 +1629,15 @@ static int nvme_create_io_queues(struct nvme_dev *dev)
 
 	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);
+		ret = nvme_create_queue(&dev->queues[i], i);
 		if (ret)
 			break;
 	}
 
 	/*
 	 * Ignore failing Create SQ/CQ commands, we can continue with less
-	 * than the desired aount of queues, and even a controller without
-	 * I/O queues an still be used to issue admin commands.  This might
+	 * than the desired amount of queues, and even a controller without
+	 * I/O queues can still be used to issue admin commands.  This might
 	 * be useful to upgrade a buggy firmware for example.
 	 */
 	return ret >= 0 ? 0 : ret;
@@ -1661,30 +1654,40 @@ static ssize_t nvme_cmb_show(struct device *dev,
 }
 static DEVICE_ATTR(cmb, S_IRUGO, nvme_cmb_show, NULL);
 
-static void __iomem *nvme_map_cmb(struct nvme_dev *dev)
+static u64 nvme_cmb_size_unit(struct nvme_dev *dev)
 {
-	u64 szu, size, offset;
+	u8 szu = (dev->cmbsz >> NVME_CMBSZ_SZU_SHIFT) & NVME_CMBSZ_SZU_MASK;
+
+	return 1ULL << (12 + 4 * szu);
+}
+
+static u32 nvme_cmb_size(struct nvme_dev *dev)
+{
+	return (dev->cmbsz >> NVME_CMBSZ_SZ_SHIFT) & NVME_CMBSZ_SZ_MASK;
+}
+
+static void nvme_map_cmb(struct nvme_dev *dev)
+{
+	u64 size, offset;
 	resource_size_t bar_size;
 	struct pci_dev *pdev = to_pci_dev(dev->dev);
-	void __iomem *cmb;
 	int bar;
 
 	dev->cmbsz = readl(dev->bar + NVME_REG_CMBSZ);
-	if (!(NVME_CMB_SZ(dev->cmbsz)))
-		return NULL;
+	if (!dev->cmbsz)
+		return;
 	dev->cmbloc = readl(dev->bar + NVME_REG_CMBLOC);
 
 	if (!use_cmb_sqes)
-		return NULL;
+		return;
 
-	szu = (u64)1 << (12 + 4 * NVME_CMB_SZU(dev->cmbsz));
-	size = szu * NVME_CMB_SZ(dev->cmbsz);
-	offset = szu * NVME_CMB_OFST(dev->cmbloc);
+	size = nvme_cmb_size_unit(dev) * nvme_cmb_size(dev);
+	offset = nvme_cmb_size_unit(dev) * NVME_CMB_OFST(dev->cmbloc);
 	bar = NVME_CMB_BIR(dev->cmbloc);
 	bar_size = pci_resource_len(pdev, bar);
 
 	if (offset > bar_size)
-		return NULL;
+		return;
 
 	/*
 	 * Controllers may support a CMB size larger than their BAR,
@@ -1694,13 +1697,16 @@ static void __iomem *nvme_map_cmb(struct nvme_dev *dev)
 	if (size > bar_size - offset)
 		size = bar_size - offset;
 
-	cmb = ioremap_wc(pci_resource_start(pdev, bar) + offset, size);
-	if (!cmb)
-		return NULL;
-
+	dev->cmb = ioremap_wc(pci_resource_start(pdev, bar) + offset, size);
+	if (!dev->cmb)
+		return;
 	dev->cmb_bus_addr = pci_bus_address(pdev, bar) + offset;
 	dev->cmb_size = size;
-	return cmb;
+
+	if (sysfs_add_file_to_group(&dev->ctrl.device->kobj,
+				    &dev_attr_cmb.attr, NULL))
+		dev_warn(dev->ctrl.device,
+			 "failed to add sysfs attribute for CMB\n");
 }
 
 static inline void nvme_release_cmb(struct nvme_dev *dev)
@@ -1768,7 +1774,7 @@ static int __nvme_alloc_host_mem(struct nvme_dev *dev, u64 preferred,
 	dma_addr_t descs_dma;
 	int i = 0;
 	void **bufs;
-	u64 size = 0, tmp;
+	u64 size, tmp;
 
 	tmp = (preferred + chunk_size - 1);
 	do_div(tmp, chunk_size);
@@ -1851,7 +1857,7 @@ static int nvme_setup_host_mem(struct nvme_dev *dev)
 	u64 preferred = (u64)dev->ctrl.hmpre * 4096;
 	u64 min = (u64)dev->ctrl.hmmin * 4096;
 	u32 enable_bits = NVME_HOST_MEM_ENABLE;
-	int ret = 0;
+	int ret;
 
 	preferred = min(preferred, max);
 	if (min > max) {
@@ -1892,7 +1898,7 @@ static int nvme_setup_host_mem(struct nvme_dev *dev)
 
 static int nvme_setup_io_queues(struct nvme_dev *dev)
 {
-	struct nvme_queue *adminq = dev->queues[0];
+	struct nvme_queue *adminq = &dev->queues[0];
 	struct pci_dev *pdev = to_pci_dev(dev->dev);
 	int result, nr_io_queues;
 	unsigned long size;
@@ -1905,7 +1911,7 @@ static int nvme_setup_io_queues(struct nvme_dev *dev)
 	if (nr_io_queues == 0)
 		return 0;
 
-	if (dev->cmb && NVME_CMB_SQS(dev->cmbsz)) {
+	if (dev->cmb && (dev->cmbsz & NVME_CMBSZ_SQS)) {
 		result = nvme_cmb_qdepth(dev, nr_io_queues,
 				sizeof(struct nvme_command));
 		if (result > 0)
@@ -2005,9 +2011,9 @@ static int nvme_delete_queue(struct nvme_queue *nvmeq, u8 opcode)
 	return 0;
 }
 
-static void nvme_disable_io_queues(struct nvme_dev *dev, int queues)
+static void nvme_disable_io_queues(struct nvme_dev *dev)
 {
-	int pass;
+	int pass, queues = dev->online_queues - 1;
 	unsigned long timeout;
 	u8 opcode = nvme_admin_delete_sq;
 
@@ -2018,7 +2024,7 @@ static void nvme_disable_io_queues(struct nvme_dev *dev, int queues)
  retry:
 		timeout = ADMIN_TIMEOUT;
 		for (; i > 0; i--, sent++)
-			if (nvme_delete_queue(dev->queues[i], opcode))
+			if (nvme_delete_queue(&dev->queues[i], opcode))
 				break;
 
 		while (sent--) {
@@ -2033,13 +2039,12 @@ static void nvme_disable_io_queues(struct nvme_dev *dev, int queues)
 }
 
 /*
- * Return: error value if an error occurred setting up the queues or calling
- * Identify Device.  0 if these succeeded, even if adding some of the
- * namespaces failed.  At the moment, these failures are silent.  TBD which
- * failures should be reported.
+ * return error value only when tagset allocation failed
  */
 static int nvme_dev_add(struct nvme_dev *dev)
 {
+	int ret;
+
 	if (!dev->ctrl.tagset) {
 		dev->tagset.ops = &nvme_mq_ops;
 		dev->tagset.nr_hw_queues = dev->online_queues - 1;
@@ -2055,8 +2060,12 @@ static int nvme_dev_add(struct nvme_dev *dev)
 		dev->tagset.flags = BLK_MQ_F_SHOULD_MERGE;
 		dev->tagset.driver_data = dev;
 
-		if (blk_mq_alloc_tag_set(&dev->tagset))
-			return 0;
+		ret = blk_mq_alloc_tag_set(&dev->tagset);
+		if (ret) {
+			dev_warn(dev->ctrl.device,
+				"IO queues tagset allocation failed %d\n", ret);
+			return ret;
+		}
 		dev->ctrl.tagset = &dev->tagset;
 
 		nvme_dbbuf_set(dev);
@@ -2122,22 +2131,7 @@ static int nvme_pci_enable(struct nvme_dev *dev)
                         "set queue depth=%u\n", dev->q_depth);
 	}
 
-	/*
-	 * CMBs can currently only exist on >=1.2 PCIe devices. We only
-	 * populate sysfs if a CMB is implemented. Since nvme_dev_attrs_group
-	 * has no name we can pass NULL as final argument to
-	 * sysfs_add_file_to_group.
-	 */
-
-	if (readl(dev->bar + NVME_REG_VS) >= NVME_VS(1, 2, 0)) {
-		dev->cmb = nvme_map_cmb(dev);
-		if (dev->cmb) {
-			if (sysfs_add_file_to_group(&dev->ctrl.device->kobj,
-						    &dev_attr_cmb.attr, NULL))
-				dev_warn(dev->ctrl.device,
-					 "failed to add sysfs attribute for CMB\n");
-		}
-	}
+	nvme_map_cmb(dev);
 
 	pci_enable_pcie_error_reporting(pdev);
 	pci_save_state(pdev);
@@ -2170,7 +2164,7 @@ static void nvme_pci_disable(struct nvme_dev *dev)
 
 static void nvme_dev_disable(struct nvme_dev *dev, bool shutdown)
 {
-	int i, queues;
+	int i;
 	bool dead = true;
 	struct pci_dev *pdev = to_pci_dev(dev->dev);
 
@@ -2205,21 +2199,13 @@ static void nvme_dev_disable(struct nvme_dev *dev, bool shutdown)
 	}
 	nvme_stop_queues(&dev->ctrl);
 
-	queues = dev->online_queues - 1;
-	for (i = dev->ctrl.queue_count - 1; i > 0; i--)
-		nvme_suspend_queue(dev->queues[i]);
-
-	if (dead) {
-		/* A device might become IO incapable very soon during
-		 * probe, before the admin queue is configured. Thus,
-		 * queue_count can be 0 here.
-		 */
-		if (dev->ctrl.queue_count)
-			nvme_suspend_queue(dev->queues[0]);
-	} else {
-		nvme_disable_io_queues(dev, queues);
+	if (!dead) {
+		nvme_disable_io_queues(dev);
 		nvme_disable_admin_queue(dev, shutdown);
 	}
+	for (i = dev->ctrl.queue_count - 1; i >= 0; i--)
+		nvme_suspend_queue(&dev->queues[i]);
+
 	nvme_pci_disable(dev);
 
 	blk_mq_tagset_busy_iter(&dev->tagset, nvme_cancel_request, &dev->ctrl);
@@ -2289,6 +2275,7 @@ static void nvme_reset_work(struct work_struct *work)
 		container_of(work, struct nvme_dev, ctrl.reset_work);
 	bool was_suspend = !!(dev->ctrl.ctrl_config & NVME_CC_SHN_NORMAL);
 	int result = -ENODEV;
+	enum nvme_ctrl_state new_state = NVME_CTRL_LIVE;
 
 	if (WARN_ON(dev->ctrl.state != NVME_CTRL_RESETTING))
 		goto out;
@@ -2300,6 +2287,16 @@ static void nvme_reset_work(struct work_struct *work)
 	if (dev->ctrl.ctrl_config & NVME_CC_ENABLE)
 		nvme_dev_disable(dev, false);
 
+	/*
+	 * Introduce RECONNECTING state from nvme-fc/rdma transports to mark the
+	 * initializing procedure here.
+	 */
+	if (!nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_RECONNECTING)) {
+		dev_warn(dev->ctrl.device,
+			"failed to mark controller RECONNECTING\n");
+		goto out;
+	}
+
 	result = nvme_pci_enable(dev);
 	if (result)
 		goto out;
@@ -2352,15 +2349,23 @@ static void nvme_reset_work(struct work_struct *work)
 		dev_warn(dev->ctrl.device, "IO queues not created\n");
 		nvme_kill_queues(&dev->ctrl);
 		nvme_remove_namespaces(&dev->ctrl);
+		new_state = NVME_CTRL_ADMIN_ONLY;
 	} else {
 		nvme_start_queues(&dev->ctrl);
 		nvme_wait_freeze(&dev->ctrl);
-		nvme_dev_add(dev);
+		/* hit this only when allocate tagset fails */
+		if (nvme_dev_add(dev))
+			new_state = NVME_CTRL_ADMIN_ONLY;
 		nvme_unfreeze(&dev->ctrl);
 	}
 
-	if (!nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_LIVE)) {
-		dev_warn(dev->ctrl.device, "failed to mark controller live\n");
+	/*
+	 * If only admin queue live, keep it to do further investigation or
+	 * recovery.
+	 */
+	if (!nvme_change_ctrl_state(&dev->ctrl, new_state)) {
+		dev_warn(dev->ctrl.device,
+			"failed to mark controller state %d\n", new_state);
 		goto out;
 	}
 
@@ -2468,8 +2473,9 @@ static int nvme_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	dev = kzalloc_node(sizeof(*dev), GFP_KERNEL, node);
 	if (!dev)
 		return -ENOMEM;
-	dev->queues = kzalloc_node((num_possible_cpus() + 1) * sizeof(void *),
-							GFP_KERNEL, node);
+
+	dev->queues = kcalloc_node(num_possible_cpus() + 1,
+			sizeof(struct nvme_queue), GFP_KERNEL, node);
 	if (!dev->queues)
 		goto free;
 
@@ -2496,10 +2502,10 @@ static int nvme_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	if (result)
 		goto release_pools;
 
-	nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_RESETTING);
 	dev_info(dev->ctrl.device, "pci function %s\n", dev_name(&pdev->dev));
 
-	queue_work(nvme_wq, &dev->ctrl.reset_work);
+	nvme_reset_ctrl(&dev->ctrl);
+
 	return 0;
 
  release_pools:
@@ -2523,7 +2529,7 @@ static void nvme_reset_prepare(struct pci_dev *pdev)
 static void nvme_reset_done(struct pci_dev *pdev)
 {
 	struct nvme_dev *dev = pci_get_drvdata(pdev);
-	nvme_reset_ctrl(&dev->ctrl);
+	nvme_reset_ctrl_sync(&dev->ctrl);
 }
 
 static void nvme_shutdown(struct pci_dev *pdev)

drivers/nvme/host/rdma.c

@@ -66,7 +66,6 @@ struct nvme_rdma_request {
 	struct ib_sge		sge[1 + NVME_RDMA_MAX_INLINE_SEGMENTS];
 	u32			num_sge;
 	int			nents;
-	bool			inline_data;
 	struct ib_reg_wr	reg_wr;
 	struct ib_cqe		reg_cqe;
 	struct nvme_rdma_queue  *queue;
@@ -1092,7 +1091,6 @@ static int nvme_rdma_map_sg_inline(struct nvme_rdma_queue *queue,
 	sg->length = cpu_to_le32(sg_dma_len(req->sg_table.sgl));
 	sg->type = (NVME_SGL_FMT_DATA_DESC << 4) | NVME_SGL_FMT_OFFSET;
 
-	req->inline_data = true;
 	req->num_sge++;
 	return 0;
 }
@@ -1164,7 +1162,6 @@ static int nvme_rdma_map_data(struct nvme_rdma_queue *queue,
 	int count, ret;
 
 	req->num_sge = 1;
-	req->inline_data = false;
 	refcount_set(&req->ref, 2); /* send and recv completions */
 
 	c->common.flags |= NVME_CMD_SGL_METABUF;
@@ -2018,6 +2015,7 @@ out_free_ctrl:
 
 static struct nvmf_transport_ops nvme_rdma_transport = {
 	.name		= "rdma",
+	.module		= THIS_MODULE,
 	.required_opts	= NVMF_OPT_TRADDR,
 	.allowed_opts	= NVMF_OPT_TRSVCID | NVMF_OPT_RECONNECT_DELAY |
 			  NVMF_OPT_HOST_TRADDR | NVMF_OPT_CTRL_LOSS_TMO,
@@ -2040,7 +2038,7 @@ static void nvme_rdma_remove_one(struct ib_device *ib_device, void *client_data)
 	}
 	mutex_unlock(&nvme_rdma_ctrl_mutex);
 
-	flush_workqueue(nvme_wq);
+	flush_workqueue(nvme_delete_wq);
 }
 
 static struct ib_client nvme_rdma_ib_client = {

drivers/nvme/host/trace.c

@@ -0,0 +1,130 @@
+/*
+ * NVM Express device driver tracepoints
+ * Copyright (c) 2018 Johannes Thumshirn, SUSE Linux GmbH
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ */
+
+#include <asm/unaligned.h>
+#include "trace.h"
+
+static const char *nvme_trace_create_sq(struct trace_seq *p, u8 *cdw10)
+{
+	const char *ret = trace_seq_buffer_ptr(p);
+	u16 sqid = get_unaligned_le16(cdw10);
+	u16 qsize = get_unaligned_le16(cdw10 + 2);
+	u16 sq_flags = get_unaligned_le16(cdw10 + 4);
+	u16 cqid = get_unaligned_le16(cdw10 + 6);
+
+
+	trace_seq_printf(p, "sqid=%u, qsize=%u, sq_flags=0x%x, cqid=%u",
+			 sqid, qsize, sq_flags, cqid);
+	trace_seq_putc(p, 0);
+
+	return ret;
+}
+
+static const char *nvme_trace_create_cq(struct trace_seq *p, u8 *cdw10)
+{
+	const char *ret = trace_seq_buffer_ptr(p);
+	u16 cqid = get_unaligned_le16(cdw10);
+	u16 qsize = get_unaligned_le16(cdw10 + 2);
+	u16 cq_flags = get_unaligned_le16(cdw10 + 4);
+	u16 irq_vector = get_unaligned_le16(cdw10 + 6);
+
+	trace_seq_printf(p, "cqid=%u, qsize=%u, cq_flags=0x%x, irq_vector=%u",
+			 cqid, qsize, cq_flags, irq_vector);
+	trace_seq_putc(p, 0);
+
+	return ret;
+}
+
+static const char *nvme_trace_admin_identify(struct trace_seq *p, u8 *cdw10)
+{
+	const char *ret = trace_seq_buffer_ptr(p);
+	u8 cns = cdw10[0];
+	u16 ctrlid = get_unaligned_le16(cdw10 + 2);
+
+	trace_seq_printf(p, "cns=%u, ctrlid=%u", cns, ctrlid);
+	trace_seq_putc(p, 0);
+
+	return ret;
+}
+
+
+
+static const char *nvme_trace_read_write(struct trace_seq *p, u8 *cdw10)
+{
+	const char *ret = trace_seq_buffer_ptr(p);
+	u64 slba = get_unaligned_le64(cdw10);
+	u16 length = get_unaligned_le16(cdw10 + 8);
+	u16 control = get_unaligned_le16(cdw10 + 10);
+	u32 dsmgmt = get_unaligned_le32(cdw10 + 12);
+	u32 reftag = get_unaligned_le32(cdw10 +  16);
+
+	trace_seq_printf(p,
+			 "slba=%llu, len=%u, ctrl=0x%x, dsmgmt=%u, reftag=%u",
+			 slba, length, control, dsmgmt, reftag);
+	trace_seq_putc(p, 0);
+
+	return ret;
+}
+
+static const char *nvme_trace_dsm(struct trace_seq *p, u8 *cdw10)
+{
+	const char *ret = trace_seq_buffer_ptr(p);
+
+	trace_seq_printf(p, "nr=%u, attributes=%u",
+			 get_unaligned_le32(cdw10),
+			 get_unaligned_le32(cdw10 + 4));
+	trace_seq_putc(p, 0);
+
+	return ret;
+}
+
+static const char *nvme_trace_common(struct trace_seq *p, u8 *cdw10)
+{
+	const char *ret = trace_seq_buffer_ptr(p);
+
+	trace_seq_printf(p, "cdw10=%*ph", 24, cdw10);
+	trace_seq_putc(p, 0);
+
+	return ret;
+}
+
+const char *nvme_trace_parse_admin_cmd(struct trace_seq *p,
+				       u8 opcode, u8 *cdw10)
+{
+	switch (opcode) {
+	case nvme_admin_create_sq:
+		return nvme_trace_create_sq(p, cdw10);
+	case nvme_admin_create_cq:
+		return nvme_trace_create_cq(p, cdw10);
+	case nvme_admin_identify:
+		return nvme_trace_admin_identify(p, cdw10);
+	default:
+		return nvme_trace_common(p, cdw10);
+	}
+}
+
+const char *nvme_trace_parse_nvm_cmd(struct trace_seq *p,
+				     u8 opcode, u8 *cdw10)
+{
+	switch (opcode) {
+	case nvme_cmd_read:
+	case nvme_cmd_write:
+	case nvme_cmd_write_zeroes:
+		return nvme_trace_read_write(p, cdw10);
+	case nvme_cmd_dsm:
+		return nvme_trace_dsm(p, cdw10);
+	default:
+		return nvme_trace_common(p, cdw10);
+	}
+}

drivers/nvme/host/trace.h

@@ -0,0 +1,165 @@
+/*
+ * NVM Express device driver tracepoints
+ * Copyright (c) 2018 Johannes Thumshirn, SUSE Linux GmbH
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ */
+
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM nvme
+
+#if !defined(_TRACE_NVME_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_NVME_H
+
+#include <linux/nvme.h>
+#include <linux/tracepoint.h>
+#include <linux/trace_seq.h>
+
+#include "nvme.h"
+
+#define nvme_admin_opcode_name(opcode)	{ opcode, #opcode }
+#define show_admin_opcode_name(val)					\
+	__print_symbolic(val,						\
+		nvme_admin_opcode_name(nvme_admin_delete_sq),		\
+		nvme_admin_opcode_name(nvme_admin_create_sq),		\
+		nvme_admin_opcode_name(nvme_admin_get_log_page),	\
+		nvme_admin_opcode_name(nvme_admin_delete_cq),		\
+		nvme_admin_opcode_name(nvme_admin_create_cq),		\
+		nvme_admin_opcode_name(nvme_admin_identify),		\
+		nvme_admin_opcode_name(nvme_admin_abort_cmd),		\
+		nvme_admin_opcode_name(nvme_admin_set_features),	\
+		nvme_admin_opcode_name(nvme_admin_get_features),	\
+		nvme_admin_opcode_name(nvme_admin_async_event),		\
+		nvme_admin_opcode_name(nvme_admin_ns_mgmt),		\
+		nvme_admin_opcode_name(nvme_admin_activate_fw),		\
+		nvme_admin_opcode_name(nvme_admin_download_fw),		\
+		nvme_admin_opcode_name(nvme_admin_ns_attach),		\
+		nvme_admin_opcode_name(nvme_admin_keep_alive),		\
+		nvme_admin_opcode_name(nvme_admin_directive_send),	\
+		nvme_admin_opcode_name(nvme_admin_directive_recv),	\
+		nvme_admin_opcode_name(nvme_admin_dbbuf),		\
+		nvme_admin_opcode_name(nvme_admin_format_nvm),		\
+		nvme_admin_opcode_name(nvme_admin_security_send),	\
+		nvme_admin_opcode_name(nvme_admin_security_recv),	\
+		nvme_admin_opcode_name(nvme_admin_sanitize_nvm))
+
+const char *nvme_trace_parse_admin_cmd(struct trace_seq *p, u8 opcode,
+				       u8 *cdw10);
+#define __parse_nvme_admin_cmd(opcode, cdw10) \
+	nvme_trace_parse_admin_cmd(p, opcode, cdw10)
+
+#define nvme_opcode_name(opcode)	{ opcode, #opcode }
+#define show_opcode_name(val)					\
+	__print_symbolic(val,					\
+		nvme_opcode_name(nvme_cmd_flush),		\
+		nvme_opcode_name(nvme_cmd_write),		\
+		nvme_opcode_name(nvme_cmd_read),		\
+		nvme_opcode_name(nvme_cmd_write_uncor),		\
+		nvme_opcode_name(nvme_cmd_compare),		\
+		nvme_opcode_name(nvme_cmd_write_zeroes),	\
+		nvme_opcode_name(nvme_cmd_dsm),			\
+		nvme_opcode_name(nvme_cmd_resv_register),	\
+		nvme_opcode_name(nvme_cmd_resv_report),		\
+		nvme_opcode_name(nvme_cmd_resv_acquire),	\
+		nvme_opcode_name(nvme_cmd_resv_release))
+
+const char *nvme_trace_parse_nvm_cmd(struct trace_seq *p, u8 opcode,
+				     u8 *cdw10);
+#define __parse_nvme_cmd(opcode, cdw10) \
+	nvme_trace_parse_nvm_cmd(p, opcode, cdw10)
+
+TRACE_EVENT(nvme_setup_admin_cmd,
+	    TP_PROTO(struct nvme_command *cmd),
+	    TP_ARGS(cmd),
+	    TP_STRUCT__entry(
+		    __field(u8, opcode)
+		    __field(u8, flags)
+		    __field(u16, cid)
+		    __field(u64, metadata)
+		    __array(u8, cdw10, 24)
+	    ),
+	    TP_fast_assign(
+		    __entry->opcode = cmd->common.opcode;
+		    __entry->flags = cmd->common.flags;
+		    __entry->cid = cmd->common.command_id;
+		    __entry->metadata = le64_to_cpu(cmd->common.metadata);
+		    memcpy(__entry->cdw10, cmd->common.cdw10,
+			   sizeof(__entry->cdw10));
+	    ),
+	    TP_printk(" cmdid=%u, flags=0x%x, meta=0x%llx, cmd=(%s %s)",
+		      __entry->cid, __entry->flags, __entry->metadata,
+		      show_admin_opcode_name(__entry->opcode),
+		      __parse_nvme_admin_cmd(__entry->opcode, __entry->cdw10))
+);
+
+
+TRACE_EVENT(nvme_setup_nvm_cmd,
+	    TP_PROTO(int qid, struct nvme_command *cmd),
+	    TP_ARGS(qid, cmd),
+	    TP_STRUCT__entry(
+		    __field(int, qid)
+		    __field(u8, opcode)
+		    __field(u8, flags)
+		    __field(u16, cid)
+		    __field(u32, nsid)
+		    __field(u64, metadata)
+		    __array(u8, cdw10, 24)
+	    ),
+	    TP_fast_assign(
+		    __entry->qid = qid;
+		    __entry->opcode = cmd->common.opcode;
+		    __entry->flags = cmd->common.flags;
+		    __entry->cid = cmd->common.command_id;
+		    __entry->nsid = le32_to_cpu(cmd->common.nsid);
+		    __entry->metadata = le64_to_cpu(cmd->common.metadata);
+		    memcpy(__entry->cdw10, cmd->common.cdw10,
+			   sizeof(__entry->cdw10));
+	    ),
+	    TP_printk("qid=%d, nsid=%u, cmdid=%u, flags=0x%x, meta=0x%llx, cmd=(%s %s)",
+		      __entry->qid, __entry->nsid, __entry->cid,
+		      __entry->flags, __entry->metadata,
+		      show_opcode_name(__entry->opcode),
+		      __parse_nvme_cmd(__entry->opcode, __entry->cdw10))
+);
+
+TRACE_EVENT(nvme_complete_rq,
+	    TP_PROTO(struct request *req),
+	    TP_ARGS(req),
+	    TP_STRUCT__entry(
+		    __field(int, qid)
+		    __field(int, cid)
+		    __field(u64, result)
+		    __field(u8, retries)
+		    __field(u8, flags)
+		    __field(u16, status)
+	    ),
+	    TP_fast_assign(
+		    __entry->qid = req->q->id;
+		    __entry->cid = req->tag;
+		    __entry->result = le64_to_cpu(nvme_req(req)->result.u64);
+		    __entry->retries = nvme_req(req)->retries;
+		    __entry->flags = nvme_req(req)->flags;
+		    __entry->status = nvme_req(req)->status;
+	    ),
+	    TP_printk("cmdid=%u, qid=%d, res=%llu, retries=%u, flags=0x%x, status=%u",
+		      __entry->cid, __entry->qid, __entry->result,
+		      __entry->retries, __entry->flags, __entry->status)
+
+);
+
+#endif /* _TRACE_NVME_H */
+
+#undef TRACE_INCLUDE_PATH
+#define TRACE_INCLUDE_PATH .
+#undef TRACE_INCLUDE_FILE
+#define TRACE_INCLUDE_FILE trace
+
+/* This part must be outside protection */
+#include <trace/define_trace.h>

drivers/nvme/target/Kconfig

@@ -29,6 +29,7 @@ config NVME_TARGET_RDMA
 	tristate "NVMe over Fabrics RDMA target support"
 	depends on INFINIBAND
 	depends on NVME_TARGET
+	select SGL_ALLOC
 	help
 	  This enables the NVMe RDMA target support, which allows exporting NVMe
 	  devices over RDMA.
@@ -39,6 +40,7 @@ config NVME_TARGET_FC
 	tristate "NVMe over Fabrics FC target driver"
 	depends on NVME_TARGET
 	depends on HAS_DMA
+	select SGL_ALLOC
 	help
 	  This enables the NVMe FC target support, which allows exporting NVMe
 	  devices over FC.

drivers/nvme/target/core.c

@@ -512,6 +512,7 @@ bool nvmet_req_init(struct nvmet_req *req, struct nvmet_cq *cq,
 	req->sg_cnt = 0;
 	req->transfer_len = 0;
 	req->rsp->status = 0;
+	req->ns = NULL;
 
 	/* no support for fused commands yet */
 	if (unlikely(flags & (NVME_CMD_FUSE_FIRST | NVME_CMD_FUSE_SECOND))) {
@@ -557,6 +558,8 @@ EXPORT_SYMBOL_GPL(nvmet_req_init);
 void nvmet_req_uninit(struct nvmet_req *req)
 {
 	percpu_ref_put(&req->sq->ref);
+	if (req->ns)
+		nvmet_put_namespace(req->ns);
 }
 EXPORT_SYMBOL_GPL(nvmet_req_uninit);
 
@@ -830,7 +833,7 @@ u16 nvmet_alloc_ctrl(const char *subsysnqn, const char *hostnqn,
 		/* Don't accept keep-alive timeout for discovery controllers */
 		if (kato) {
 			status = NVME_SC_INVALID_FIELD | NVME_SC_DNR;
-			goto out_free_sqs;
+			goto out_remove_ida;
 		}
 
 		/*
@@ -860,6 +863,8 @@ u16 nvmet_alloc_ctrl(const char *subsysnqn, const char *hostnqn,
 	*ctrlp = ctrl;
 	return 0;
 
+out_remove_ida:
+	ida_simple_remove(&cntlid_ida, ctrl->cntlid);
 out_free_sqs:
 	kfree(ctrl->sqs);
 out_free_cqs:
@@ -877,21 +882,22 @@ static void nvmet_ctrl_free(struct kref *ref)
 	struct nvmet_ctrl *ctrl = container_of(ref, struct nvmet_ctrl, ref);
 	struct nvmet_subsys *subsys = ctrl->subsys;
 
-	nvmet_stop_keep_alive_timer(ctrl);
-
 	mutex_lock(&subsys->lock);
 	list_del(&ctrl->subsys_entry);
 	mutex_unlock(&subsys->lock);
 
+	nvmet_stop_keep_alive_timer(ctrl);
+
 	flush_work(&ctrl->async_event_work);
 	cancel_work_sync(&ctrl->fatal_err_work);
 
 	ida_simple_remove(&cntlid_ida, ctrl->cntlid);
-	nvmet_subsys_put(subsys);
 
 	kfree(ctrl->sqs);
 	kfree(ctrl->cqs);
 	kfree(ctrl);
+
+	nvmet_subsys_put(subsys);
 }
 
 void nvmet_ctrl_put(struct nvmet_ctrl *ctrl)

drivers/nvme/target/fabrics-cmd.c

@@ -225,7 +225,7 @@ static void nvmet_execute_io_connect(struct nvmet_req *req)
 		goto out_ctrl_put;
 	}
 
-	pr_info("adding queue %d to ctrl %d.\n", qid, ctrl->cntlid);
+	pr_debug("adding queue %d to ctrl %d.\n", qid, ctrl->cntlid);
 
 out:
 	kfree(d);

drivers/nvme/target/fc.c

@@ -1697,31 +1697,12 @@ static int
 nvmet_fc_alloc_tgt_pgs(struct nvmet_fc_fcp_iod *fod)
 {
 	struct scatterlist *sg;
-	struct page *page;
 	unsigned int nent;
-	u32 page_len, length;
-	int i = 0;
 
-	length = fod->req.transfer_len;
-	nent = DIV_ROUND_UP(length, PAGE_SIZE);
-	sg = kmalloc_array(nent, sizeof(struct scatterlist), GFP_KERNEL);
+	sg = sgl_alloc(fod->req.transfer_len, GFP_KERNEL, &nent);
 	if (!sg)
 		goto out;
 
-	sg_init_table(sg, nent);
-
-	while (length) {
-		page_len = min_t(u32, length, PAGE_SIZE);
-
-		page = alloc_page(GFP_KERNEL);
-		if (!page)
-			goto out_free_pages;
-
-		sg_set_page(&sg[i], page, page_len, 0);
-		length -= page_len;
-		i++;
-	}
-
 	fod->data_sg = sg;
 	fod->data_sg_cnt = nent;
 	fod->data_sg_cnt = fc_dma_map_sg(fod->tgtport->dev, sg, nent,
@@ -1731,14 +1712,6 @@ nvmet_fc_alloc_tgt_pgs(struct nvmet_fc_fcp_iod *fod)
 
 	return 0;
 
-out_free_pages:
-	while (i > 0) {
-		i--;
-		__free_page(sg_page(&sg[i]));
-	}
-	kfree(sg);
-	fod->data_sg = NULL;
-	fod->data_sg_cnt = 0;
 out:
 	return NVME_SC_INTERNAL;
 }
@@ -1746,18 +1719,13 @@ out:
 static void
 nvmet_fc_free_tgt_pgs(struct nvmet_fc_fcp_iod *fod)
 {
-	struct scatterlist *sg;
-	int count;
-
 	if (!fod->data_sg || !fod->data_sg_cnt)
 		return;
 
 	fc_dma_unmap_sg(fod->tgtport->dev, fod->data_sg, fod->data_sg_cnt,
 				((fod->io_dir == NVMET_FCP_WRITE) ?
 					DMA_FROM_DEVICE : DMA_TO_DEVICE));
-	for_each_sg(fod->data_sg, sg, fod->data_sg_cnt, count)
-		__free_page(sg_page(sg));
-	kfree(fod->data_sg);
+	sgl_free(fod->data_sg);
 	fod->data_sg = NULL;
 	fod->data_sg_cnt = 0;
 }
@@ -2522,14 +2490,8 @@ nvmet_fc_add_port(struct nvmet_port *port)
 	list_for_each_entry(tgtport, &nvmet_fc_target_list, tgt_list) {
 		if ((tgtport->fc_target_port.node_name == traddr.nn) &&
 		    (tgtport->fc_target_port.port_name == traddr.pn)) {
-			/* a FC port can only be 1 nvmet port id */
-			if (!tgtport->port) {
-				tgtport->port = port;
-				port->priv = tgtport;
-				nvmet_fc_tgtport_get(tgtport);
-				ret = 0;
-			} else
-				ret = -EALREADY;
+			tgtport->port = port;
+			ret = 0;
 			break;
 		}
 	}
@@ -2540,19 +2502,7 @@ nvmet_fc_add_port(struct nvmet_port *port)
 static void
 nvmet_fc_remove_port(struct nvmet_port *port)
 {
-	struct nvmet_fc_tgtport *tgtport = port->priv;
-	unsigned long flags;
-	bool matched = false;
-
-	spin_lock_irqsave(&nvmet_fc_tgtlock, flags);
-	if (tgtport->port == port) {
-		matched = true;
-		tgtport->port = NULL;
-	}
-	spin_unlock_irqrestore(&nvmet_fc_tgtlock, flags);
-
-	if (matched)
-		nvmet_fc_tgtport_put(tgtport);
+	/* nothing to do */
 }
 
 static struct nvmet_fabrics_ops nvmet_fc_tgt_fcp_ops = {

drivers/nvme/target/fcloop.c

@@ -204,6 +204,10 @@ struct fcloop_lport {
 	struct completion unreg_done;
 };
 
+struct fcloop_lport_priv {
+	struct fcloop_lport *lport;
+};
+
 struct fcloop_rport {
 	struct nvme_fc_remote_port *remoteport;
 	struct nvmet_fc_target_port *targetport;
@@ -238,21 +242,32 @@ struct fcloop_lsreq {
 	int				status;
 };
 
+enum {
+	INI_IO_START		= 0,
+	INI_IO_ACTIVE		= 1,
+	INI_IO_ABORTED		= 2,
+	INI_IO_COMPLETED	= 3,
+};
+
 struct fcloop_fcpreq {
 	struct fcloop_tport		*tport;
 	struct nvmefc_fcp_req		*fcpreq;
 	spinlock_t			reqlock;
 	u16				status;
+	u32				inistate;
 	bool				active;
 	bool				aborted;
-	struct work_struct		work;
+	struct kref			ref;
+	struct work_struct		fcp_rcv_work;
+	struct work_struct		abort_rcv_work;
+	struct work_struct		tio_done_work;
 	struct nvmefc_tgt_fcp_req	tgt_fcp_req;
 };
 
 struct fcloop_ini_fcpreq {
 	struct nvmefc_fcp_req		*fcpreq;
 	struct fcloop_fcpreq		*tfcp_req;
-	struct work_struct		iniwork;
+	spinlock_t			inilock;
 };
 
 static inline struct fcloop_lsreq *
@@ -343,17 +358,122 @@ fcloop_xmt_ls_rsp(struct nvmet_fc_target_port *tport,
 	return 0;
 }
 
-/*
- * FCP IO operation done by initiator abort.
- * call back up initiator "done" flows.
- */
 static void
-fcloop_tgt_fcprqst_ini_done_work(struct work_struct *work)
+fcloop_tfcp_req_free(struct kref *ref)
 {
-	struct fcloop_ini_fcpreq *inireq =
-		container_of(work, struct fcloop_ini_fcpreq, iniwork);
+	struct fcloop_fcpreq *tfcp_req =
+		container_of(ref, struct fcloop_fcpreq, ref);
+
+	kfree(tfcp_req);
+}
+
+static void
+fcloop_tfcp_req_put(struct fcloop_fcpreq *tfcp_req)
+{
+	kref_put(&tfcp_req->ref, fcloop_tfcp_req_free);
+}
+
+static int
+fcloop_tfcp_req_get(struct fcloop_fcpreq *tfcp_req)
+{
+	return kref_get_unless_zero(&tfcp_req->ref);
+}
+
+static void
+fcloop_call_host_done(struct nvmefc_fcp_req *fcpreq,
+			struct fcloop_fcpreq *tfcp_req, int status)
+{
+	struct fcloop_ini_fcpreq *inireq = NULL;
+
+	if (fcpreq) {
+		inireq = fcpreq->private;
+		spin_lock(&inireq->inilock);
+		inireq->tfcp_req = NULL;
+		spin_unlock(&inireq->inilock);
+
+		fcpreq->status = status;
+		fcpreq->done(fcpreq);
+	}
+
+	/* release original io reference on tgt struct */
+	fcloop_tfcp_req_put(tfcp_req);
+}
+
+static void
+fcloop_fcp_recv_work(struct work_struct *work)
+{
+	struct fcloop_fcpreq *tfcp_req =
+		container_of(work, struct fcloop_fcpreq, fcp_rcv_work);
+	struct nvmefc_fcp_req *fcpreq = tfcp_req->fcpreq;
+	int ret = 0;
+	bool aborted = false;
+
+	spin_lock(&tfcp_req->reqlock);
+	switch (tfcp_req->inistate) {
+	case INI_IO_START:
+		tfcp_req->inistate = INI_IO_ACTIVE;
+		break;
+	case INI_IO_ABORTED:
+		aborted = true;
+		break;
+	default:
+		spin_unlock(&tfcp_req->reqlock);
+		WARN_ON(1);
+		return;
+	}
+	spin_unlock(&tfcp_req->reqlock);
+
+	if (unlikely(aborted))
+		ret = -ECANCELED;
+	else
+		ret = nvmet_fc_rcv_fcp_req(tfcp_req->tport->targetport,
+				&tfcp_req->tgt_fcp_req,
+				fcpreq->cmdaddr, fcpreq->cmdlen);
+	if (ret)
+		fcloop_call_host_done(fcpreq, tfcp_req, ret);
+
+	return;
+}
+
+static void
+fcloop_fcp_abort_recv_work(struct work_struct *work)
+{
+	struct fcloop_fcpreq *tfcp_req =
+		container_of(work, struct fcloop_fcpreq, abort_rcv_work);
+	struct nvmefc_fcp_req *fcpreq;
+	bool completed = false;
+
+	spin_lock(&tfcp_req->reqlock);
+	fcpreq = tfcp_req->fcpreq;
+	switch (tfcp_req->inistate) {
+	case INI_IO_ABORTED:
+		break;
+	case INI_IO_COMPLETED:
+		completed = true;
+		break;
+	default:
+		spin_unlock(&tfcp_req->reqlock);
+		WARN_ON(1);
+		return;
+	}
+	spin_unlock(&tfcp_req->reqlock);
+
+	if (unlikely(completed)) {
+		/* remove reference taken in original abort downcall */
+		fcloop_tfcp_req_put(tfcp_req);
+		return;
+	}
 
-	inireq->fcpreq->done(inireq->fcpreq);
+	if (tfcp_req->tport->targetport)
+		nvmet_fc_rcv_fcp_abort(tfcp_req->tport->targetport,
+					&tfcp_req->tgt_fcp_req);
+
+	spin_lock(&tfcp_req->reqlock);
+	tfcp_req->fcpreq = NULL;
+	spin_unlock(&tfcp_req->reqlock);
+
+	fcloop_call_host_done(fcpreq, tfcp_req, -ECANCELED);
+	/* call_host_done releases reference for abort downcall */
 }
 
 /*
@@ -364,20 +484,15 @@ static void
 fcloop_tgt_fcprqst_done_work(struct work_struct *work)
 {
 	struct fcloop_fcpreq *tfcp_req =
-		container_of(work, struct fcloop_fcpreq, work);
-	struct fcloop_tport *tport = tfcp_req->tport;
+		container_of(work, struct fcloop_fcpreq, tio_done_work);
 	struct nvmefc_fcp_req *fcpreq;
 
 	spin_lock(&tfcp_req->reqlock);
 	fcpreq = tfcp_req->fcpreq;
+	tfcp_req->inistate = INI_IO_COMPLETED;
 	spin_unlock(&tfcp_req->reqlock);
 
-	if (tport->remoteport && fcpreq) {
-		fcpreq->status = tfcp_req->status;
-		fcpreq->done(fcpreq);
-	}
-
-	kfree(tfcp_req);
+	fcloop_call_host_done(fcpreq, tfcp_req, tfcp_req->status);
 }
 
 
@@ -390,7 +505,6 @@ fcloop_fcp_req(struct nvme_fc_local_port *localport,
 	struct fcloop_rport *rport = remoteport->private;
 	struct fcloop_ini_fcpreq *inireq = fcpreq->private;
 	struct fcloop_fcpreq *tfcp_req;
-	int ret = 0;
 
 	if (!rport->targetport)
 		return -ECONNREFUSED;
@@ -401,16 +515,20 @@ fcloop_fcp_req(struct nvme_fc_local_port *localport,
 
 	inireq->fcpreq = fcpreq;
 	inireq->tfcp_req = tfcp_req;
-	INIT_WORK(&inireq->iniwork, fcloop_tgt_fcprqst_ini_done_work);
+	spin_lock_init(&inireq->inilock);
+
 	tfcp_req->fcpreq = fcpreq;
 	tfcp_req->tport = rport->targetport->private;
+	tfcp_req->inistate = INI_IO_START;
 	spin_lock_init(&tfcp_req->reqlock);
-	INIT_WORK(&tfcp_req->work, fcloop_tgt_fcprqst_done_work);
+	INIT_WORK(&tfcp_req->fcp_rcv_work, fcloop_fcp_recv_work);
+	INIT_WORK(&tfcp_req->abort_rcv_work, fcloop_fcp_abort_recv_work);
+	INIT_WORK(&tfcp_req->tio_done_work, fcloop_tgt_fcprqst_done_work);
+	kref_init(&tfcp_req->ref);
 
-	ret = nvmet_fc_rcv_fcp_req(rport->targetport, &tfcp_req->tgt_fcp_req,
-				 fcpreq->cmdaddr, fcpreq->cmdlen);
+	schedule_work(&tfcp_req->fcp_rcv_work);
 
-	return ret;
+	return 0;
 }
 
 static void
@@ -589,7 +707,7 @@ fcloop_fcp_req_release(struct nvmet_fc_target_port *tgtport,
 {
 	struct fcloop_fcpreq *tfcp_req = tgt_fcp_req_to_fcpreq(tgt_fcpreq);
 
-	schedule_work(&tfcp_req->work);
+	schedule_work(&tfcp_req->tio_done_work);
 }
 
 static void
@@ -605,27 +723,47 @@ fcloop_fcp_abort(struct nvme_fc_local_port *localport,
 			void *hw_queue_handle,
 			struct nvmefc_fcp_req *fcpreq)
 {
-	struct fcloop_rport *rport = remoteport->private;
 	struct fcloop_ini_fcpreq *inireq = fcpreq->private;
-	struct fcloop_fcpreq *tfcp_req = inireq->tfcp_req;
+	struct fcloop_fcpreq *tfcp_req;
+	bool abortio = true;
+
+	spin_lock(&inireq->inilock);
+	tfcp_req = inireq->tfcp_req;
+	if (tfcp_req)
+		fcloop_tfcp_req_get(tfcp_req);
+	spin_unlock(&inireq->inilock);
 
 	if (!tfcp_req)
 		/* abort has already been called */
 		return;
 
-	if (rport->targetport)
-		nvmet_fc_rcv_fcp_abort(rport->targetport,
-					&tfcp_req->tgt_fcp_req);
-
 	/* break initiator/target relationship for io */
 	spin_lock(&tfcp_req->reqlock);
-	inireq->tfcp_req = NULL;
-	tfcp_req->fcpreq = NULL;
+	switch (tfcp_req->inistate) {
+	case INI_IO_START:
+	case INI_IO_ACTIVE:
+		tfcp_req->inistate = INI_IO_ABORTED;
+		break;
+	case INI_IO_COMPLETED:
+		abortio = false;
+		break;
+	default:
+		spin_unlock(&tfcp_req->reqlock);
+		WARN_ON(1);
+		return;
+	}
 	spin_unlock(&tfcp_req->reqlock);
 
-	/* post the aborted io completion */
-	fcpreq->status = -ECANCELED;
-	schedule_work(&inireq->iniwork);
+	if (abortio)
+		/* leave the reference while the work item is scheduled */
+		WARN_ON(!schedule_work(&tfcp_req->abort_rcv_work));
+	else  {
+		/*
+		 * as the io has already had the done callback made,
+		 * nothing more to do. So release the reference taken above
+		 */
+		fcloop_tfcp_req_put(tfcp_req);
+	}
 }
 
 static void
@@ -657,7 +795,8 @@ fcloop_nport_get(struct fcloop_nport *nport)
 static void
 fcloop_localport_delete(struct nvme_fc_local_port *localport)
 {
-	struct fcloop_lport *lport = localport->private;
+	struct fcloop_lport_priv *lport_priv = localport->private;
+	struct fcloop_lport *lport = lport_priv->lport;
 
 	/* release any threads waiting for the unreg to complete */
 	complete(&lport->unreg_done);
@@ -697,7 +836,7 @@ static struct nvme_fc_port_template fctemplate = {
 	.max_dif_sgl_segments	= FCLOOP_SGL_SEGS,
 	.dma_boundary		= FCLOOP_DMABOUND_4G,
 	/* sizes of additional private data for data structures */
-	.local_priv_sz		= sizeof(struct fcloop_lport),
+	.local_priv_sz		= sizeof(struct fcloop_lport_priv),
 	.remote_priv_sz		= sizeof(struct fcloop_rport),
 	.lsrqst_priv_sz		= sizeof(struct fcloop_lsreq),
 	.fcprqst_priv_sz	= sizeof(struct fcloop_ini_fcpreq),
@@ -714,8 +853,7 @@ static struct nvmet_fc_target_template tgttemplate = {
 	.max_dif_sgl_segments	= FCLOOP_SGL_SEGS,
 	.dma_boundary		= FCLOOP_DMABOUND_4G,
 	/* optional features */
-	.target_features	= NVMET_FCTGTFEAT_CMD_IN_ISR |
-				  NVMET_FCTGTFEAT_OPDONE_IN_ISR,
+	.target_features	= 0,
 	/* sizes of additional private data for data structures */
 	.target_priv_sz		= sizeof(struct fcloop_tport),
 };
@@ -728,11 +866,17 @@ fcloop_create_local_port(struct device *dev, struct device_attribute *attr,
 	struct fcloop_ctrl_options *opts;
 	struct nvme_fc_local_port *localport;
 	struct fcloop_lport *lport;
-	int ret;
+	struct fcloop_lport_priv *lport_priv;
+	unsigned long flags;
+	int ret = -ENOMEM;
+
+	lport = kzalloc(sizeof(*lport), GFP_KERNEL);
+	if (!lport)
+		return -ENOMEM;
 
 	opts = kzalloc(sizeof(*opts), GFP_KERNEL);
 	if (!opts)
-		return -ENOMEM;
+		goto out_free_lport;
 
 	ret = fcloop_parse_options(opts, buf);
 	if (ret)
@@ -752,23 +896,25 @@ fcloop_create_local_port(struct device *dev, struct device_attribute *attr,
 
 	ret = nvme_fc_register_localport(&pinfo, &fctemplate, NULL, &localport);
 	if (!ret) {
-		unsigned long flags;
-
 		/* success */
-		lport = localport->private;
+		lport_priv = localport->private;
+		lport_priv->lport = lport;
+
 		lport->localport = localport;
 		INIT_LIST_HEAD(&lport->lport_list);
 
 		spin_lock_irqsave(&fcloop_lock, flags);
 		list_add_tail(&lport->lport_list, &fcloop_lports);
 		spin_unlock_irqrestore(&fcloop_lock, flags);
-
-		/* mark all of the input buffer consumed */
-		ret = count;
 	}
 
 out_free_opts:
 	kfree(opts);
+out_free_lport:
+	/* free only if we're going to fail */
+	if (ret)
+		kfree(lport);
+
 	return ret ? ret : count;
 }
 
@@ -790,6 +936,8 @@ __wait_localport_unreg(struct fcloop_lport *lport)
 
 	wait_for_completion(&lport->unreg_done);
 
+	kfree(lport);
+
 	return ret;
 }
 

drivers/nvme/target/loop.c

@@ -686,6 +686,7 @@ static struct nvmet_fabrics_ops nvme_loop_ops = {
 
 static struct nvmf_transport_ops nvme_loop_transport = {
 	.name		= "loop",
+	.module		= THIS_MODULE,
 	.create_ctrl	= nvme_loop_create_ctrl,
 };
 
@@ -716,7 +717,7 @@ static void __exit nvme_loop_cleanup_module(void)
 		nvme_delete_ctrl(&ctrl->ctrl);
 	mutex_unlock(&nvme_loop_ctrl_mutex);
 
-	flush_workqueue(nvme_wq);
+	flush_workqueue(nvme_delete_wq);
 }
 
 module_init(nvme_loop_init_module);

drivers/nvme/target/rdma.c

@@ -185,59 +185,6 @@ nvmet_rdma_put_rsp(struct nvmet_rdma_rsp *rsp)
 	spin_unlock_irqrestore(&rsp->queue->rsps_lock, flags);
 }
 
-static void nvmet_rdma_free_sgl(struct scatterlist *sgl, unsigned int nents)
-{
-	struct scatterlist *sg;
-	int count;
-
-	if (!sgl || !nents)
-		return;
-
-	for_each_sg(sgl, sg, nents, count)
-		__free_page(sg_page(sg));
-	kfree(sgl);
-}
-
-static int nvmet_rdma_alloc_sgl(struct scatterlist **sgl, unsigned int *nents,
-		u32 length)
-{
-	struct scatterlist *sg;
-	struct page *page;
-	unsigned int nent;
-	int i = 0;
-
-	nent = DIV_ROUND_UP(length, PAGE_SIZE);
-	sg = kmalloc_array(nent, sizeof(struct scatterlist), GFP_KERNEL);
-	if (!sg)
-		goto out;
-
-	sg_init_table(sg, nent);
-
-	while (length) {
-		u32 page_len = min_t(u32, length, PAGE_SIZE);
-
-		page = alloc_page(GFP_KERNEL);
-		if (!page)
-			goto out_free_pages;
-
-		sg_set_page(&sg[i], page, page_len, 0);
-		length -= page_len;
-		i++;
-	}
-	*sgl = sg;
-	*nents = nent;
-	return 0;
-
-out_free_pages:
-	while (i > 0) {
-		i--;
-		__free_page(sg_page(&sg[i]));
-	}
-	kfree(sg);
-out:
-	return NVME_SC_INTERNAL;
-}
-
 static int nvmet_rdma_alloc_cmd(struct nvmet_rdma_device *ndev,
 			struct nvmet_rdma_cmd *c, bool admin)
 {
@@ -484,7 +431,7 @@ static void nvmet_rdma_release_rsp(struct nvmet_rdma_rsp *rsp)
 	}
 
 	if (rsp->req.sg != &rsp->cmd->inline_sg)
-		nvmet_rdma_free_sgl(rsp->req.sg, rsp->req.sg_cnt);
+		sgl_free(rsp->req.sg);
 
 	if (unlikely(!list_empty_careful(&queue->rsp_wr_wait_list)))
 		nvmet_rdma_process_wr_wait_list(queue);
@@ -621,16 +568,14 @@ static u16 nvmet_rdma_map_sgl_keyed(struct nvmet_rdma_rsp *rsp,
 	u32 len = get_unaligned_le24(sgl->length);
 	u32 key = get_unaligned_le32(sgl->key);
 	int ret;
-	u16 status;
 
 	/* no data command? */
 	if (!len)
 		return 0;
 
-	status = nvmet_rdma_alloc_sgl(&rsp->req.sg, &rsp->req.sg_cnt,
-			len);
-	if (status)
-		return status;
+	rsp->req.sg = sgl_alloc(len, GFP_KERNEL, &rsp->req.sg_cnt);
+	if (!rsp->req.sg)
+		return NVME_SC_INTERNAL;
 
 	ret = rdma_rw_ctx_init(&rsp->rw, cm_id->qp, cm_id->port_num,
 			rsp->req.sg, rsp->req.sg_cnt, 0, addr, key,
@@ -976,7 +921,7 @@ static void nvmet_rdma_destroy_queue_ib(struct nvmet_rdma_queue *queue)
 
 static void nvmet_rdma_free_queue(struct nvmet_rdma_queue *queue)
 {
-	pr_info("freeing queue %d\n", queue->idx);
+	pr_debug("freeing queue %d\n", queue->idx);
 
 	nvmet_sq_destroy(&queue->nvme_sq);
 
@@ -1558,25 +1503,9 @@ err_ib_client:
 
 static void __exit nvmet_rdma_exit(void)
 {
-	struct nvmet_rdma_queue *queue;
-
 	nvmet_unregister_transport(&nvmet_rdma_ops);
-
-	flush_scheduled_work();
-
-	mutex_lock(&nvmet_rdma_queue_mutex);
-	while ((queue = list_first_entry_or_null(&nvmet_rdma_queue_list,
-			struct nvmet_rdma_queue, queue_list))) {
-		list_del_init(&queue->queue_list);
-
-		mutex_unlock(&nvmet_rdma_queue_mutex);
-		__nvmet_rdma_queue_disconnect(queue);
-		mutex_lock(&nvmet_rdma_queue_mutex);
-	}
-	mutex_unlock(&nvmet_rdma_queue_mutex);
-
-	flush_scheduled_work();
 	ib_unregister_client(&nvmet_rdma_ib_client);
+	WARN_ON_ONCE(!list_empty(&nvmet_rdma_queue_list));
 	ida_destroy(&nvmet_rdma_queue_ida);
 }
 

drivers/target/Kconfig

@@ -5,6 +5,7 @@ menuconfig TARGET_CORE
 	select CONFIGFS_FS
 	select CRC_T10DIF
 	select BLK_SCSI_REQUEST # only for scsi_command_size_tbl..
+	select SGL_ALLOC
 	default n
 	help
 	Say Y or M here to enable the TCM Storage Engine and ConfigFS enabled

drivers/target/target_core_transport.c

@@ -2300,13 +2300,7 @@ queue_full:
 
 void target_free_sgl(struct scatterlist *sgl, int nents)
 {
-	struct scatterlist *sg;
-	int count;
-
-	for_each_sg(sgl, sg, nents, count)
-		__free_page(sg_page(sg));
-
-	kfree(sgl);
+	sgl_free_n_order(sgl, nents, 0);
 }
 EXPORT_SYMBOL(target_free_sgl);
 
@@ -2414,42 +2408,10 @@ int
 target_alloc_sgl(struct scatterlist **sgl, unsigned int *nents, u32 length,
 		 bool zero_page, bool chainable)
 {
-	struct scatterlist *sg;
-	struct page *page;
-	gfp_t zero_flag = (zero_page) ? __GFP_ZERO : 0;
-	unsigned int nalloc, nent;
-	int i = 0;
-
-	nalloc = nent = DIV_ROUND_UP(length, PAGE_SIZE);
-	if (chainable)
-		nalloc++;
-	sg = kmalloc_array(nalloc, sizeof(struct scatterlist), GFP_KERNEL);
-	if (!sg)
-		return -ENOMEM;
+	gfp_t gfp = GFP_KERNEL | (zero_page ? __GFP_ZERO : 0);
 
-	sg_init_table(sg, nalloc);
-
-	while (length) {
-		u32 page_len = min_t(u32, length, PAGE_SIZE);
-		page = alloc_page(GFP_KERNEL | zero_flag);
-		if (!page)
-			goto out;
-
-		sg_set_page(&sg[i], page, page_len, 0);
-		length -= page_len;
-		i++;
-	}
-	*sgl = sg;
-	*nents = nent;
-	return 0;
-
-out:
-	while (i > 0) {
-		i--;
-		__free_page(sg_page(&sg[i]));
-	}
-	kfree(sg);
-	return -ENOMEM;
+	*sgl = sgl_alloc_order(length, 0, chainable, gfp, nents);
+	return *sgl ? 0 : -ENOMEM;
 }
 EXPORT_SYMBOL(target_alloc_sgl);
 

fs/btrfs/compression.c

@@ -411,7 +411,7 @@ blk_status_t btrfs_submit_compressed_write(struct inode *inode, u64 start,
 
 static u64 bio_end_offset(struct bio *bio)
 {
-	struct bio_vec *last = &bio->bi_io_vec[bio->bi_vcnt - 1];
+	struct bio_vec *last = bio_last_bvec_all(bio);
 
 	return page_offset(last->bv_page) + last->bv_len + last->bv_offset;
 }
@@ -563,7 +563,7 @@ blk_status_t btrfs_submit_compressed_read(struct inode *inode, struct bio *bio,
 	/* we need the actual starting offset of this extent in the file */
 	read_lock(&em_tree->lock);
 	em = lookup_extent_mapping(em_tree,
-				   page_offset(bio->bi_io_vec->bv_page),
+				   page_offset(bio_first_page_all(bio)),
 				   PAGE_SIZE);
 	read_unlock(&em_tree->lock);
 	if (!em)