[PATCH v5 00/28] Optimize the hot path in the UFS driver

[PATCH v5 00/28] Optimize the hot path in the UFS driver

[Bart Van Assche]

Hi Martin,

This patch series optimizes the hot path of the UFS driver by making
struct scsi_cmnd and struct ufshcd_lrb adjacent. Making these two data
structures adjacent is realized as follows:

@@ -9040,6 +9046,7 @@ static const struct scsi_host_template ufshcd_driver_template = {
     .name           = UFSHCD,
     .proc_name      = UFSHCD,
     .map_queues     = ufshcd_map_queues,
+    .cmd_size       = sizeof(struct ufshcd_lrb),
     .init_cmd_priv  = ufshcd_init_cmd_priv,
     .queuecommand   = ufshcd_queuecommand,
     .mq_poll        = ufshcd_poll,

The following changes had to be made prior to making these two data
structures adjacent:
* Add support for driver-internal and reserved commands in the SCSI core.
* Instead of making the reserved command slot (hba->reserved_slot)
  invisible to the SCSI core, let the SCSI core allocate a reserved command.
* Remove all UFS data structure members that are no longer needed
  because struct scsi_cmnd and struct ufshcd_lrb are now adjacent
* Call ufshcd_init_lrb() from inside ufshcd_queuecommand() instead of
  calling this function before I/O starts. This is necessary because
  ufshcd_memory_alloc() allocates fewer instances than the block layer
  allocates requests. See also the following code in the block layer
  core:

    if (blk_mq_init_request(set, hctx->fq->flush_rq, hctx_idx,
                hctx->numa_node))

  Although the UFS driver could be modified such that ufshcd_init_lrb()
  is called from ufshcd_init_cmd_priv(), realizing this would require
  moving the memory allocations that happen from inside
  ufshcd_memory_alloc() into ufshcd_init_cmd_priv(). That would make
  this patch series even larger. Although ufshcd_init_lrb() is called for each
  command, the benefits of reduced indirection and better cache efficiency
  outweigh the small overhead of per-command lrb initialization.
* ufshcd_add_scsi_host() happens now before any device management
  commands are submitted. This change is necessary because this patch
  makes device management command allocation happen when the SCSI host
  is allocated.
* Allocate as many command slots as the host controller supports. Decrease
  host->cmds_per_lun if necessary once it is clear whether or not the UFS
  device supports less command slots than the host controller.

Changes compared to v4:
 - Dropped the scsi_execute_cmd() changes.
 - Restored patch "scsi: core: Add scsi_{get,put}_internal_cmd() helpers".
 - Switched back from scsi_execute_cmd() to blk_execute_rq() for submitting
   device management commands in the UFS driver.
 - As suggested by John Garry, modified the scsi_debug patch such that aborting
   a SCSI command happens by submitting a reserved command.

Changes compared to v3:
 - Fixed a spelling error in patch 1 and left out a superfluous if-statement.
 - Left out scsi_host_template.alloc_pseudo_sdev and allocate a pseudo SCSI
   device if either nr_reserved_cmds > 0 or .queue_reserved_commands has been
   set.
 - Left out the 'pseudo_sdev' local variable from scsi_forget_host().
 - Removed a backwards jump from scsi_get_pseudo_dev().
 - Included a bug fix for synchronous scanning.
 - Skip scsi_track_queue_full() and scsi_handle_queue_ramp_up() for pseudo SCSI
   devices.
 - Extended the scsi_execute_rq() functionality.
 - Use scsi_execute_rq() for submitting reserved commands instead of
   blk_execute_rq().
 - Dropped the patch that introduces scsi_get_internal_cmd() and
   scsi_put_internal_cmd().

Changes compared to v2:
 - Removed scsi_host_update_can_queue() and also the UFS driver refactoring
   patches that were introduced to support this call.
 - Added .queue_reserved_command(). Added ufshcd_queue_reserved_command().
 - Removed a BUG_ON() statement from ufshcd_get_dev_mgmt_cmd().
 - Modified and renamed ufshcd_mcq_decide_queue_depth().

Changes compared to v1:
 - Left out the kernel patches related to support for const SCSI command
   arguments.
 - Added SCSI core patches for allocating a pseudo SCSI device and reserved
   command support.
 - Added several kernel patches to switch the UFS driver from a hardcoded
   reserved slot to calling scsi_get_internal_cmd().
 - Enable .alloc_pseudo_sdev in the scsi_debug driver.

Bart Van Assche (24):
  scsi: core: Move two statements
  scsi: core: Make the budget map optional
  scsi_debug: Abort SCSI commands via .queue_reserved_command()
  ufs: core: Move an assignment in ufshcd_mcq_process_cqe()
  ufs: core: Change the type of one ufshcd_add_cmd_upiu_trace() argument
  ufs: core: Only call ufshcd_add_command_trace() for SCSI commands
  ufs: core: Change the type of one ufshcd_add_command_trace() argument
  ufs: core: Change the type of one ufshcd_send_command() argument
  ufs: core: Only call ufshcd_should_inform_monitor() for SCSI commands
  ufs: core: Change the monitor function argument types
  ufs: core: Rework ufshcd_mcq_compl_pending_transfer()
  ufs: core: Rework ufshcd_eh_device_reset_handler()
  ufs: core: Rework the SCSI host queue depth calculation code
  ufs: core: Allocate the SCSI host earlier
  ufs: core: Call ufshcd_init_lrb() later
  ufs: core: Use hba->reserved_slot
  ufs: core: Make the reserved slot a reserved request
  ufs: core: Do not clear driver-private command data
  ufs: core: Optimize the hot path
  ufs: core: Pass a SCSI pointer instead of an LRB pointer
  ufs: core: Remove the ufshcd_lrb task_tag member
  ufs: core: Make blk_mq_tagset_busy_iter() skip reserved requests
  ufs: core: Move code out of ufshcd_wait_for_dev_cmd()
  ufs: core: Switch to scsi_get_internal_cmd()

Hannes Reinecke (3):
  scsi: core: Support allocating reserved commands
  scsi: core: Support allocating a pseudo SCSI device
  scsi: core: Add scsi_{get,put}_internal_cmd() helpers

John Garry (1):
  scsi: core: Introduce .queue_reserved_command()

 drivers/scsi/hosts.c             |  15 +
 drivers/scsi/scsi.c              |  12 +-
 drivers/scsi/scsi_debug.c        | 106 +++-
 drivers/scsi/scsi_error.c        |   3 +
 drivers/scsi/scsi_lib.c          | 145 +++++-
 drivers/scsi/scsi_priv.h         |   2 +
 drivers/scsi/scsi_scan.c         |  77 ++-
 drivers/scsi/scsi_sysfs.c        |   5 +-
 drivers/ufs/core/ufs-mcq.c       |  51 +-
 drivers/ufs/core/ufshcd-crypto.h |  18 +-
 drivers/ufs/core/ufshcd-priv.h   |  20 +-
 drivers/ufs/core/ufshcd.c        | 808 ++++++++++++++++---------------
 include/scsi/scsi_device.h       |  23 +
 include/scsi/scsi_host.h         |  33 +-
 include/ufs/ufshcd.h             |  12 -
 15 files changed, 861 insertions(+), 469 deletions(-)

[PATCH v5 01/28] scsi: core: Support allocating reserved commands

[Bart Van Assche]

From: Hannes Reinecke <hare@suse.de>

Quite some drivers are using management commands internally. These
commands typically use the same tag pool as regular SCSI commands. Tags
for these management commands are set aside before allocating the
block-mq tag bitmap for regular SCSI commands. The block layer already
supports this via the reserved tag mechanism. Add a new field
'nr_reserved_cmds' to the SCSI host template to instruct the block layer
to set aside a tag space for these management commands by using reserved
tags. Exclude reserved commands from .can_queue because .can_queue is
visible in sysfs.

Signed-off-by: Hannes Reinecke <hare@suse.de>
[ bvanassche: modified patch title and patch description. Left out the
  following statements: "if (sht->nr_reserved_cmds)" and also
  "if (sdev->host->nr_reserved_cmds) flags |= BLK_MQ_REQ_RESERVED;". Moved
  nr_reserved_cmds declarations and statements close to the
  corresponding can_queue declarations and statements. See also
  https://lore.kernel.org/linux-scsi/20210503150333.130310-11-hare@suse.de/ ]
Reviewed-by: John Garry <john.g.garry@oracle.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/scsi/hosts.c     |  1 +
 drivers/scsi/scsi_lib.c  |  3 ++-
 include/scsi/scsi_host.h | 21 ++++++++++++++++++++-
 3 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/drivers/scsi/hosts.c b/drivers/scsi/hosts.c
index cc5d05dc395c..9bb7f0114763 100644
--- a/drivers/scsi/hosts.c
+++ b/drivers/scsi/hosts.c
@@ -436,6 +436,7 @@ struct Scsi_Host *scsi_host_alloc(const struct scsi_host_template *sht, int priv
 	shost->hostt = sht;
 	shost->this_id = sht->this_id;
 	shost->can_queue = sht->can_queue;
+	shost->nr_reserved_cmds = sht->nr_reserved_cmds;
 	shost->sg_tablesize = sht->sg_tablesize;
 	shost->sg_prot_tablesize = sht->sg_prot_tablesize;
 	shost->cmd_per_lun = sht->cmd_per_lun;
diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c
index 0c65ecfedfbd..9c67e04265ce 100644
--- a/drivers/scsi/scsi_lib.c
+++ b/drivers/scsi/scsi_lib.c
@@ -2083,7 +2083,8 @@ int scsi_mq_setup_tags(struct Scsi_Host *shost)
 		tag_set->ops = &scsi_mq_ops_no_commit;
 	tag_set->nr_hw_queues = shost->nr_hw_queues ? : 1;
 	tag_set->nr_maps = shost->nr_maps ? : 1;
-	tag_set->queue_depth = shost->can_queue;
+	tag_set->queue_depth = shost->can_queue + shost->nr_reserved_cmds;
+	tag_set->reserved_tags = shost->nr_reserved_cmds;
 	tag_set->cmd_size = cmd_size;
 	tag_set->numa_node = dev_to_node(shost->dma_dev);
 	if (shost->hostt->tag_alloc_policy_rr)
diff --git a/include/scsi/scsi_host.h b/include/scsi/scsi_host.h
index c53812b9026f..91eb3f52b3d0 100644
--- a/include/scsi/scsi_host.h
+++ b/include/scsi/scsi_host.h
@@ -375,10 +375,19 @@ struct scsi_host_template {
 	/*
 	 * This determines if we will use a non-interrupt driven
 	 * or an interrupt driven scheme.  It is set to the maximum number
-	 * of simultaneous commands a single hw queue in HBA will accept.
+	 * of simultaneous commands a single hw queue in HBA will accept
+	 * excluding internal commands.
 	 */
 	int can_queue;
 
+	/*
+	 * This determines how many commands the HBA will set aside
+	 * for internal commands. This number will be added to
+	 * @can_queue to calculate the maximum number of simultaneous
+	 * commands sent to the host.
+	 */
+	int nr_reserved_cmds;
+
 	/*
 	 * In many instances, especially where disconnect / reconnect are
 	 * supported, our host also has an ID on the SCSI bus.  If this is
@@ -611,7 +620,17 @@ struct Scsi_Host {
 	unsigned short max_cmd_len;
 
 	int this_id;
+
+	/*
+	 * Number of commands this host can handle at the same time.
+	 * This excludes reserved commands as specified by nr_reserved_cmds.
+	 */
 	int can_queue;
+	/*
+	 * Number of reserved commands to allocate, if any.
+	 */
+	unsigned int nr_reserved_cmds;
+
 	short cmd_per_lun;
 	short unsigned int sg_tablesize;
 	short unsigned int sg_prot_tablesize;

[PATCH v5 02/28] scsi: core: Move two statements

[Bart Van Assche]

Move two statements that will be needed for pseudo SCSI devices in front
of code that won't be needed for pseudo SCSI devices. No functionality
has been changed.

Reviewed-by: John Garry <john.g.garry@oracle.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/scsi/scsi_scan.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/drivers/scsi/scsi_scan.c b/drivers/scsi/scsi_scan.c
index 3c6e089e80c3..de039efef290 100644
--- a/drivers/scsi/scsi_scan.c
+++ b/drivers/scsi/scsi_scan.c
@@ -347,6 +347,8 @@ static struct scsi_device *scsi_alloc_sdev(struct scsi_target *starget,
 	kref_get(&sdev->host->tagset_refcnt);
 	sdev->request_queue = q;
 
+	scsi_sysfs_device_initialize(sdev);
+
 	depth = sdev->host->cmd_per_lun ?: 1;
 
 	/*
@@ -363,8 +365,6 @@ static struct scsi_device *scsi_alloc_sdev(struct scsi_target *starget,
 
 	scsi_change_queue_depth(sdev, depth);
 
-	scsi_sysfs_device_initialize(sdev);
-
 	if (shost->hostt->sdev_init) {
 		ret = shost->hostt->sdev_init(sdev);
 		if (ret) {
@@ -1068,6 +1068,8 @@ static int scsi_add_lun(struct scsi_device *sdev, unsigned char *inq_result,
 
 	transport_configure_device(&sdev->sdev_gendev);
 
+	sdev->sdev_bflags = *bflags;
+
 	/*
 	 * No need to freeze the queue as it isn't reachable to anyone else yet.
 	 */
@@ -1113,7 +1115,6 @@ static int scsi_add_lun(struct scsi_device *sdev, unsigned char *inq_result,
 
 	sdev->max_queue_depth = sdev->queue_depth;
 	WARN_ON_ONCE(sdev->max_queue_depth > sdev->budget_map.depth);
-	sdev->sdev_bflags = *bflags;
 
 	/*
 	 * Ok, the device is now all set up, we can

[PATCH v5 03/28] scsi: core: Make the budget map optional

[Bart Van Assche]

Prepare for not allocating a budget map for pseudo SCSI devices by
checking whether a budget map has been allocated before using it.

Cc: Hannes Reinecke <hare@suse.de>
Cc: John Garry <john.g.garry@oracle.com>
Cc: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/scsi/scsi.c       | 5 +++++
 drivers/scsi/scsi_error.c | 3 +++
 drivers/scsi/scsi_lib.c   | 9 +++++++--
 drivers/scsi/scsi_scan.c  | 3 +++
 4 files changed, 18 insertions(+), 2 deletions(-)

diff --git a/drivers/scsi/scsi.c b/drivers/scsi/scsi.c
index 9a0f467264b3..a9d3e0242555 100644
--- a/drivers/scsi/scsi.c
+++ b/drivers/scsi/scsi.c
@@ -216,6 +216,9 @@ int scsi_device_max_queue_depth(struct scsi_device *sdev)
  */
 int scsi_change_queue_depth(struct scsi_device *sdev, int depth)
 {
+	if (WARN_ON_ONCE(!sdev->budget_map.map))
+		return -EINVAL;
+
 	depth = min_t(int, depth, scsi_device_max_queue_depth(sdev));
 
 	if (depth > 0) {
@@ -255,6 +258,8 @@ EXPORT_SYMBOL(scsi_change_queue_depth);
  */
 int scsi_track_queue_full(struct scsi_device *sdev, int depth)
 {
+	if (!sdev->budget_map.map)
+		return 0;
 
 	/*
 	 * Don't let QUEUE_FULLs on the same
diff --git a/drivers/scsi/scsi_error.c b/drivers/scsi/scsi_error.c
index 746ff6a1f309..87636068cd37 100644
--- a/drivers/scsi/scsi_error.c
+++ b/drivers/scsi/scsi_error.c
@@ -749,6 +749,9 @@ static void scsi_handle_queue_ramp_up(struct scsi_device *sdev)
 	const struct scsi_host_template *sht = sdev->host->hostt;
 	struct scsi_device *tmp_sdev;
 
+	if (!sdev->budget_map.map)
+		return;
+
 	if (!sht->track_queue_depth ||
 	    sdev->queue_depth >= sdev->max_queue_depth)
 		return;
diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c
index 9c67e04265ce..8d711f6d8959 100644
--- a/drivers/scsi/scsi_lib.c
+++ b/drivers/scsi/scsi_lib.c
@@ -396,7 +396,8 @@ void scsi_device_unbusy(struct scsi_device *sdev, struct scsi_cmnd *cmd)
 	if (starget->can_queue > 0)
 		atomic_dec(&starget->target_busy);
 
-	sbitmap_put(&sdev->budget_map, cmd->budget_token);
+	if (sdev->budget_map.map)
+		sbitmap_put(&sdev->budget_map, cmd->budget_token);
 	cmd->budget_token = -1;
 }
 
@@ -1360,6 +1361,9 @@ static inline int scsi_dev_queue_ready(struct request_queue *q,
 {
 	int token;
 
+	if (!sdev->budget_map.map)
+		return INT_MAX;
+
 	token = sbitmap_get(&sdev->budget_map);
 	if (token < 0)
 		return -1;
@@ -1749,7 +1753,8 @@ static void scsi_mq_put_budget(struct request_queue *q, int budget_token)
 {
 	struct scsi_device *sdev = q->queuedata;
 
-	sbitmap_put(&sdev->budget_map, budget_token);
+	if (sdev->budget_map.map)
+		sbitmap_put(&sdev->budget_map, budget_token);
 }
 
 /*
diff --git a/drivers/scsi/scsi_scan.c b/drivers/scsi/scsi_scan.c
index de039efef290..3267b36a6059 100644
--- a/drivers/scsi/scsi_scan.c
+++ b/drivers/scsi/scsi_scan.c
@@ -225,6 +225,9 @@ static int scsi_realloc_sdev_budget_map(struct scsi_device *sdev,
 	int ret;
 	struct sbitmap sb_backup;
 
+	if (WARN_ON_ONCE(!sdev->budget_map.map))
+		return -EINVAL;
+
 	depth = min_t(unsigned int, depth, scsi_device_max_queue_depth(sdev));
 
 	/*

[PATCH v5 04/28] scsi: core: Support allocating a pseudo SCSI device

[Bart Van Assche]

From: Hannes Reinecke <hare@suse.de>

Allocate a pseudo SCSI device if 'nr_reserved_cmds' has been set. Pseudo
SCSI devices have the SCSI ID <max_id>:U64_MAX so they won't clash with
any devices the LLD might create. Pseudo SCSI devices are excluded from
scanning and will not show up in sysfs. Additionally, pseudo SCSI
devices are skipped by shost_for_each_device(). This prevents that the
SCSI error handler tries to submit a reset to a non-existent logical unit.

Do not allocate a budget map for pseudo SCSI devices since the
cmd_per_lun limit does not apply to pseudo SCSI devices.

Do not perform queue depth ramp up / ramp down for pseudo SCSI devices.

Pseudo SCSI devices will be used to send internal commands to a storage
device.

Cc: John Garry <john.g.garry@oracle.com>
Signed-off-by: Hannes Reinecke <hare@suse.de>
[ bvanassche: edited patch description / renamed host_sdev into
  pseudo_sdev / unexported scsi_get_host_dev() / modified error path in
  scsi_get_pseudo_dev() / skip pseudo devices in __scsi_iterate_devices()
  and also when calling sdev_init(), sdev_configure() and sdev_destroy().
  See also
  https://lore.kernel.org/linux-scsi/20211125151048.103910-2-hare@suse.de/ ]
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/scsi/hosts.c       |  8 +++++
 drivers/scsi/scsi.c        |  7 ++--
 drivers/scsi/scsi_priv.h   |  2 ++
 drivers/scsi/scsi_scan.c   | 67 +++++++++++++++++++++++++++++++++++++-
 drivers/scsi/scsi_sysfs.c  |  5 ++-
 include/scsi/scsi_device.h | 16 +++++++++
 include/scsi/scsi_host.h   |  6 ++++
 7 files changed, 107 insertions(+), 4 deletions(-)

diff --git a/drivers/scsi/hosts.c b/drivers/scsi/hosts.c
index 9bb7f0114763..986586bf67dc 100644
--- a/drivers/scsi/hosts.c
+++ b/drivers/scsi/hosts.c
@@ -307,6 +307,14 @@ int scsi_add_host_with_dma(struct Scsi_Host *shost, struct device *dev,
 	if (error)
 		goto out_del_dev;
 
+	if (sht->nr_reserved_cmds) {
+		shost->pseudo_sdev = scsi_get_pseudo_dev(shost);
+		if (!shost->pseudo_sdev) {
+			error = -ENOMEM;
+			goto out_del_dev;
+		}
+	}
+
 	scsi_proc_host_add(shost);
 	scsi_autopm_put_host(shost);
 	return error;
diff --git a/drivers/scsi/scsi.c b/drivers/scsi/scsi.c
index a9d3e0242555..b74693763af3 100644
--- a/drivers/scsi/scsi.c
+++ b/drivers/scsi/scsi.c
@@ -831,8 +831,11 @@ struct scsi_device *__scsi_iterate_devices(struct Scsi_Host *shost,
 	spin_lock_irqsave(shost->host_lock, flags);
 	while (list->next != &shost->__devices) {
 		next = list_entry(list->next, struct scsi_device, siblings);
-		/* skip devices that we can't get a reference to */
-		if (!scsi_device_get(next))
+		/*
+		 * Skip pseudo devices and also devices for which
+		 * scsi_device_get() fails.
+		 */
+		if (!scsi_device_is_pseudo_dev(next) && !scsi_device_get(next))
 			break;
 		next = NULL;
 		list = list->next;
diff --git a/drivers/scsi/scsi_priv.h b/drivers/scsi/scsi_priv.h
index 5b2b19f5e8ec..da3bc87ac5a6 100644
--- a/drivers/scsi/scsi_priv.h
+++ b/drivers/scsi/scsi_priv.h
@@ -135,6 +135,8 @@ extern int scsi_complete_async_scans(void);
 extern int scsi_scan_host_selected(struct Scsi_Host *, unsigned int,
 				   unsigned int, u64, enum scsi_scan_mode);
 extern void scsi_forget_host(struct Scsi_Host *);
+struct scsi_device *scsi_get_pseudo_dev(struct Scsi_Host *);
+bool scsi_device_is_pseudo_dev(struct scsi_device *sdev);
 
 /* scsi_sysctl.c */
 #ifdef CONFIG_SYSCTL
diff --git a/drivers/scsi/scsi_scan.c b/drivers/scsi/scsi_scan.c
index 3267b36a6059..511cf84c2f89 100644
--- a/drivers/scsi/scsi_scan.c
+++ b/drivers/scsi/scsi_scan.c
@@ -352,6 +352,9 @@ static struct scsi_device *scsi_alloc_sdev(struct scsi_target *starget,
 
 	scsi_sysfs_device_initialize(sdev);
 
+	if (scsi_device_is_pseudo_dev(sdev))
+		return sdev;
+
 	depth = sdev->host->cmd_per_lun ?: 1;
 
 	/*
@@ -1073,6 +1076,9 @@ static int scsi_add_lun(struct scsi_device *sdev, unsigned char *inq_result,
 
 	sdev->sdev_bflags = *bflags;
 
+	if (scsi_device_is_pseudo_dev(sdev))
+		return SCSI_SCAN_LUN_PRESENT;
+
 	/*
 	 * No need to freeze the queue as it isn't reachable to anyone else yet.
 	 */
@@ -1216,6 +1222,12 @@ static int scsi_probe_and_add_lun(struct scsi_target *starget,
 	if (!sdev)
 		goto out;
 
+	if (scsi_device_is_pseudo_dev(sdev)) {
+		if (bflagsp)
+			*bflagsp = BLIST_NOLUN;
+		return SCSI_SCAN_LUN_PRESENT;
+	}
+
 	result = kmalloc(result_len, GFP_KERNEL);
 	if (!result)
 		goto out_free_sdev;
@@ -2087,12 +2099,65 @@ void scsi_forget_host(struct Scsi_Host *shost)
  restart:
 	spin_lock_irqsave(shost->host_lock, flags);
 	list_for_each_entry(sdev, &shost->__devices, siblings) {
-		if (sdev->sdev_state == SDEV_DEL)
+		if (scsi_device_is_pseudo_dev(sdev) ||
+		    sdev->sdev_state == SDEV_DEL)
 			continue;
 		spin_unlock_irqrestore(shost->host_lock, flags);
 		__scsi_remove_device(sdev);
 		goto restart;
 	}
 	spin_unlock_irqrestore(shost->host_lock, flags);
+
+	/*
+	 * Remove the pseudo device last since it may be needed during removal
+	 * of other SCSI devices.
+	 */
+	if (shost->pseudo_sdev)
+		__scsi_remove_device(shost->pseudo_sdev);
 }
 
+/**
+ * scsi_get_pseudo_dev() - Attach a pseudo SCSI device to a SCSI host
+ * @shost: Host that needs a pseudo SCSI device
+ *
+ * Lock status: None assumed.
+ *
+ * Returns:     The scsi_device or NULL
+ *
+ * Notes:
+ *	Attach a single scsi_device to the Scsi_Host. The primary aim for this
+ *	device is to serve as a container from which SCSI commands can be
+ *	allocated. Each SCSI command will carry a command tag allocated by the
+ *	block layer. These SCSI commands can be used by the LLDD to send
+ *	internal or passthrough commands without having to manage tag allocation
+ *	inside the LLDD.
+ */
+struct scsi_device *scsi_get_pseudo_dev(struct Scsi_Host *shost)
+{
+	struct scsi_device *sdev = NULL;
+	struct scsi_target *starget;
+
+	guard(mutex)(&shost->scan_mutex);
+
+	if (!scsi_host_scan_allowed(shost))
+		goto out;
+
+	starget = scsi_alloc_target(&shost->shost_gendev, 0, shost->max_id);
+	if (!starget)
+		goto out;
+
+	sdev = scsi_alloc_sdev(starget, U64_MAX, NULL);
+	if (!sdev) {
+		scsi_target_reap(starget);
+		goto put_target;
+	}
+
+	sdev->borken = 0;
+
+put_target:
+	/* See also the get_device(dev) call in scsi_alloc_target(). */
+	put_device(&starget->dev);
+
+out:
+	return sdev;
+}
diff --git a/drivers/scsi/scsi_sysfs.c b/drivers/scsi/scsi_sysfs.c
index 169af7d47ce7..22f76a1ca23b 100644
--- a/drivers/scsi/scsi_sysfs.c
+++ b/drivers/scsi/scsi_sysfs.c
@@ -1406,6 +1406,9 @@ int scsi_sysfs_add_sdev(struct scsi_device *sdev)
 	int error;
 	struct scsi_target *starget = sdev->sdev_target;
 
+	if (WARN_ON_ONCE(scsi_device_is_pseudo_dev(sdev)))
+		return -EINVAL;
+
 	error = scsi_target_add(starget);
 	if (error)
 		return error;
@@ -1513,7 +1516,7 @@ void __scsi_remove_device(struct scsi_device *sdev)
 	kref_put(&sdev->host->tagset_refcnt, scsi_mq_free_tags);
 	cancel_work_sync(&sdev->requeue_work);
 
-	if (sdev->host->hostt->sdev_destroy)
+	if (!scsi_device_is_pseudo_dev(sdev) && sdev->host->hostt->sdev_destroy)
 		sdev->host->hostt->sdev_destroy(sdev);
 	transport_destroy_device(dev);
 
diff --git a/include/scsi/scsi_device.h b/include/scsi/scsi_device.h
index 6d6500148c4b..3846f5dfc51c 100644
--- a/include/scsi/scsi_device.h
+++ b/include/scsi/scsi_device.h
@@ -589,6 +589,22 @@ static inline unsigned int sdev_id(struct scsi_device *sdev)
 #define scmd_id(scmd) sdev_id((scmd)->device)
 #define scmd_channel(scmd) sdev_channel((scmd)->device)
 
+/**
+ * scsi_device_is_pseudo_dev() - Whether a device is a pseudo SCSI device.
+ * @sdev: SCSI device to examine
+ *
+ * A pseudo SCSI device can be used to allocate SCSI commands but does not show
+ * up in sysfs. Additionally, the logical unit information in *@sdev is made up.
+ *
+ * This function tests the LUN number instead of comparing @sdev with
+ * @sdev->host->pseudo_sdev because this function may be called before
+ * @sdev->host->pseudo_sdev has been initialized.
+ */
+static inline bool scsi_device_is_pseudo_dev(struct scsi_device *sdev)
+{
+	return sdev->lun == U64_MAX;
+}
+
 /*
  * checks for positions of the SCSI state machine
  */
diff --git a/include/scsi/scsi_host.h b/include/scsi/scsi_host.h
index 91eb3f52b3d0..3bfb53cf5dfc 100644
--- a/include/scsi/scsi_host.h
+++ b/include/scsi/scsi_host.h
@@ -721,6 +721,12 @@ struct Scsi_Host {
 	/* ldm bits */
 	struct device		shost_gendev, shost_dev;
 
+	/*
+	 * A SCSI device structure used for sending internal commands to the
+	 * HBA. There is no corresponding logical unit inside the SCSI device.
+	 */
+	struct scsi_device *pseudo_sdev;
+
 	/*
 	 * Points to the transport data (if any) which is allocated
 	 * separately

Re: [PATCH v5 04/28] scsi: core: Support allocating a pseudo SCSI device

[John Garry]

>   		next = NULL;
>   		list = list->next;
> diff --git a/drivers/scsi/scsi_priv.h b/drivers/scsi/scsi_priv.h
> index 5b2b19f5e8ec..da3bc87ac5a6 100644
> --- a/drivers/scsi/scsi_priv.h
> +++ b/drivers/scsi/scsi_priv.h
> @@ -135,6 +135,8 @@ extern int scsi_complete_async_scans(void);
>   extern int scsi_scan_host_selected(struct Scsi_Host *, unsigned int,
>   				   unsigned int, u64, enum scsi_scan_mode);
>   extern void scsi_forget_host(struct Scsi_Host *);
> +struct scsi_device *scsi_get_pseudo_dev(struct Scsi_Host *);
> +bool scsi_device_is_pseudo_dev(struct scsi_device *sdev);


this is defined in scsi_device.h - why also have a prototype here?

generally looks ok apart from that

Re: [PATCH v5 04/28] scsi: core: Support allocating a pseudo SCSI device

[Bart Van Assche]

On 9/26/25 12:14 AM, John Garry wrote:
>> +bool scsi_device_is_pseudo_dev(struct scsi_device *sdev);
> 
> this is defined in scsi_device.h - why also have a prototype here?

When I changed scsi_device_is_pseudo_dev() from an exported function
into an inline function I overlooked that the above line should be
removed. I will remove the above declaration before I repost this patch
series.

Thanks,

Bart.

Re: [PATCH v5 04/28] scsi: core: Support allocating a pseudo SCSI device

[Bart Van Assche]

On 9/24/25 1:30 PM, Bart Van Assche wrote:
> diff --git a/drivers/scsi/hosts.c b/drivers/scsi/hosts.c
> index 9bb7f0114763..986586bf67dc 100644
> --- a/drivers/scsi/hosts.c
> +++ b/drivers/scsi/hosts.c
> @@ -307,6 +307,14 @@ int scsi_add_host_with_dma(struct Scsi_Host *shost, struct device *dev,
>   	if (error)
>   		goto out_del_dev;
>   
> +	if (sht->nr_reserved_cmds) {
> +		shost->pseudo_sdev = scsi_get_pseudo_dev(shost);
> +		if (!shost->pseudo_sdev) {
> +			error = -ENOMEM;
> +			goto out_del_dev;
> +		}
> +	}

(replying to my own email)

Unless anyone objects I will change sht->nr_reserved_cmds into
shost->nr_reserved_cmds in the above code.

Bart.

[PATCH v5 05/28] scsi: core: Introduce .queue_reserved_command()

[Bart Van Assche]

From: John Garry <john.garry@huawei.com>

Reserved commands will be used by SCSI LLDs for submitting internal
commands. Since the SCSI host, target and device limits do not apply to
the reserved command use cases, bypass the SCSI host limit checks for
reserved commands. Introduce the .queue_reserved_command() callback for
reserved commands. Additionally, do not activate the SCSI error handler
if a reserved command fails such that reserved commands can be submitted
from inside the SCSI error handler.

Signed-off-by: John Garry <john.garry@huawei.com>
[ bvanassche: modified patch title and patch description. Renamed
  .reserved_queuecommand() into .queue_reserved_command(). Changed
  the second argument of __blk_mq_end_request() from 0 into error
  code in the completion path if cmd->result != 0. Rewrote the
  scsi_queue_rq() changes. See also
  https://lore.kernel.org/linux-scsi/1666693096-180008-5-git-send-email-john.garry@huawei.com/ ]
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/scsi/hosts.c     |  8 +++++-
 drivers/scsi/scsi_lib.c  | 54 ++++++++++++++++++++++++++++------------
 include/scsi/scsi_host.h |  6 +++++
 3 files changed, 51 insertions(+), 17 deletions(-)

diff --git a/drivers/scsi/hosts.c b/drivers/scsi/hosts.c
index 986586bf67dc..3a62c51379ef 100644
--- a/drivers/scsi/hosts.c
+++ b/drivers/scsi/hosts.c
@@ -231,6 +231,12 @@ int scsi_add_host_with_dma(struct Scsi_Host *shost, struct device *dev,
 		goto fail;
 	}
 
+	if (shost->nr_reserved_cmds && !sht->queue_reserved_command) {
+		shost_printk(KERN_ERR, shost,
+			     "nr_reserved_cmds set but no method to queue\n");
+		goto fail;
+	}
+
 	/* Use min_t(int, ...) in case shost->can_queue exceeds SHRT_MAX */
 	shost->cmd_per_lun = min_t(int, shost->cmd_per_lun,
 				   shost->can_queue);
@@ -307,7 +313,7 @@ int scsi_add_host_with_dma(struct Scsi_Host *shost, struct device *dev,
 	if (error)
 		goto out_del_dev;
 
-	if (sht->nr_reserved_cmds) {
+	if (sht->nr_reserved_cmds || sht->queue_reserved_command) {
 		shost->pseudo_sdev = scsi_get_pseudo_dev(shost);
 		if (!shost->pseudo_sdev) {
 			error = -ENOMEM;
diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c
index 8d711f6d8959..868013cb1f46 100644
--- a/drivers/scsi/scsi_lib.c
+++ b/drivers/scsi/scsi_lib.c
@@ -1534,6 +1534,14 @@ static void scsi_complete(struct request *rq)
 	struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(rq);
 	enum scsi_disposition disposition;
 
+	if (blk_mq_is_reserved_rq(rq)) {
+		/* Only pass-through requests are supported in this code path. */
+		WARN_ON_ONCE(!blk_rq_is_passthrough(scsi_cmd_to_rq(cmd)));
+		scsi_mq_uninit_cmd(cmd);
+		__blk_mq_end_request(rq, scsi_result_to_blk_status(cmd->result));
+		return;
+	}
+
 	INIT_LIST_HEAD(&cmd->eh_entry);
 
 	atomic_inc(&cmd->device->iodone_cnt);
@@ -1823,25 +1831,31 @@ static blk_status_t scsi_queue_rq(struct blk_mq_hw_ctx *hctx,
 	WARN_ON_ONCE(cmd->budget_token < 0);
 
 	/*
-	 * If the device is not in running state we will reject some or all
-	 * commands.
+	 * Bypass the SCSI device, SCSI target and SCSI host checks for
+	 * reserved commands.
 	 */
-	if (unlikely(sdev->sdev_state != SDEV_RUNNING)) {
-		ret = scsi_device_state_check(sdev, req);
-		if (ret != BLK_STS_OK)
-			goto out_put_budget;
-	}
+	if (!blk_mq_is_reserved_rq(req)) {
+		/*
+		 * If the device is not in running state we will reject some or
+		 * all commands.
+		 */
+		if (unlikely(sdev->sdev_state != SDEV_RUNNING)) {
+			ret = scsi_device_state_check(sdev, req);
+			if (ret != BLK_STS_OK)
+				goto out_put_budget;
+		}
 
-	ret = BLK_STS_RESOURCE;
-	if (!scsi_target_queue_ready(shost, sdev))
-		goto out_put_budget;
-	if (unlikely(scsi_host_in_recovery(shost))) {
-		if (cmd->flags & SCMD_FAIL_IF_RECOVERING)
-			ret = BLK_STS_OFFLINE;
-		goto out_dec_target_busy;
+		ret = BLK_STS_RESOURCE;
+		if (!scsi_target_queue_ready(shost, sdev))
+			goto out_put_budget;
+		if (unlikely(scsi_host_in_recovery(shost))) {
+			if (cmd->flags & SCMD_FAIL_IF_RECOVERING)
+				ret = BLK_STS_OFFLINE;
+			goto out_dec_target_busy;
+		}
+		if (!scsi_host_queue_ready(q, shost, sdev, cmd))
+			goto out_dec_target_busy;
 	}
-	if (!scsi_host_queue_ready(q, shost, sdev, cmd))
-		goto out_dec_target_busy;
 
 	/*
 	 * Only clear the driver-private command data if the LLD does not supply
@@ -1870,6 +1884,14 @@ static blk_status_t scsi_queue_rq(struct blk_mq_hw_ctx *hctx,
 	cmd->submitter = SUBMITTED_BY_BLOCK_LAYER;
 
 	blk_mq_start_request(req);
+	if (blk_mq_is_reserved_rq(req)) {
+		reason = shost->hostt->queue_reserved_command(shost, cmd);
+		if (reason) {
+			ret = BLK_STS_RESOURCE;
+			goto out_put_budget;
+		}
+		return BLK_STS_OK;
+	}
 	reason = scsi_dispatch_cmd(cmd);
 	if (reason) {
 		scsi_set_blocked(cmd, reason);
diff --git a/include/scsi/scsi_host.h b/include/scsi/scsi_host.h
index 3bfb53cf5dfc..9a0b07bfd559 100644
--- a/include/scsi/scsi_host.h
+++ b/include/scsi/scsi_host.h
@@ -86,6 +86,12 @@ struct scsi_host_template {
 	 */
 	int (* queuecommand)(struct Scsi_Host *, struct scsi_cmnd *);
 
+	/*
+	 * Queue a reserved command (BLK_MQ_REQ_RESERVED). The .queuecommand()
+	 * documentation also applies to the .queue_reserved_command() callback.
+	 */
+	int (*queue_reserved_command)(struct Scsi_Host *, struct scsi_cmnd *);
+
 	/*
 	 * The commit_rqs function is used to trigger a hardware
 	 * doorbell after some requests have been queued with

Re: [PATCH v5 05/28] scsi: core: Introduce .queue_reserved_command()

[John Garry]

On 24/09/2025 21:30, Bart Van Assche wrote:
> From: John Garry <john.garry@huawei.com>
> 
> Reserved commands will be used by SCSI LLDs for submitting internal
> commands. Since the SCSI host, target and device limits do not apply to
> the reserved command use cases, bypass the SCSI host limit checks for
> reserved commands. Introduce the .queue_reserved_command() callback for
> reserved commands. Additionally, do not activate the SCSI error handler
> if a reserved command fails such that reserved commands can be submitted
> from inside the SCSI error handler.
> 
> Signed-off-by: John Garry <john.garry@huawei.com>
> [ bvanassche: modified patch title and patch description. Renamed
>    .reserved_queuecommand() into .queue_reserved_command(). Changed
>    the second argument of __blk_mq_end_request() from 0 into error
>    code in the completion path if cmd->result != 0. Rewrote the
>    scsi_queue_rq() changes. See also
>    https://urldefense.com/v3/__https://lore.kernel.org/linux-scsi/1666693096-180008-5-git-send-email-john.garry@huawei.com/__;!!ACWV5N9M2RV99hQ!Ku4G8LUl3kt-hX9_3TXoSAPJKn9Wz4_3Bnnz_ahYQYT0tAKkJTwnWZv4av5MhFclGpk6b3N7n06tpzYzyj5X0A$  ]
> Signed-off-by: Bart Van Assche <bvanassche@acm.org>
> ---
>   drivers/scsi/hosts.c     |  8 +++++-
>   drivers/scsi/scsi_lib.c  | 54 ++++++++++++++++++++++++++++------------
>   include/scsi/scsi_host.h |  6 +++++
>   3 files changed, 51 insertions(+), 17 deletions(-)
> 
> diff --git a/drivers/scsi/hosts.c b/drivers/scsi/hosts.c
> index 986586bf67dc..3a62c51379ef 100644
> --- a/drivers/scsi/hosts.c
> +++ b/drivers/scsi/hosts.c
> @@ -231,6 +231,12 @@ int scsi_add_host_with_dma(struct Scsi_Host *shost, struct device *dev,
>   		goto fail;
>   	}
>   
> +	if (shost->nr_reserved_cmds && !sht->queue_reserved_command) {

But having sht->queue_reserved_command and !shost->nr_reserved_cmds does 
not make sense. However, I suppose that we should not change the sht 
when we have a scenario with shost->nr_reserved_cmds == 0, so I suppose 
that it is ok.

> +		shost_printk(KERN_ERR, shost,
> +			     "nr_reserved_cmds set but no method to queue\n");
> +		goto fail;
> +	}
> +
>   	/* Use min_t(int, ...) in case shost->can_queue exceeds SHRT_MAX */
>   	shost->cmd_per_lun = min_t(int, shost->cmd_per_lun,
>   				   shost->can_queue);
> @@ -307,7 +313,7 @@ int scsi_add_host_with_dma(struct Scsi_Host *shost, struct device *dev,
>   	if (error)
>   		goto out_del_dev;
>   
> -	if (sht->nr_reserved_cmds) {
> +	if (sht->nr_reserved_cmds || sht->queue_reserved_command) {

But this really should not change, as we cannot have 
sht->nr_reserved_cmds && !sht->queue_reserved_command (from above)

>   		shost->pseudo_sdev = scsi_get_pseudo_dev(shost);
>   		if (!shost->pseudo_sdev) {
>   			error = -ENOMEM;

generally the rest looks ok

Re: [PATCH v5 05/28] scsi: core: Introduce .queue_reserved_command()

[Bart Van Assche]

On 9/26/25 12:24 AM, John Garry wrote:
> On 24/09/2025 21:30, Bart Van Assche wrote:
>> @@ -307,7 +313,7 @@ int scsi_add_host_with_dma(struct Scsi_Host 
>> *shost, struct device *dev,
>>       if (error)
>>           goto out_del_dev;
>> -    if (sht->nr_reserved_cmds) {
>> +    if (sht->nr_reserved_cmds || sht->queue_reserved_command) {
> 
> But this really should not change, as we cannot have sht- 
>  >nr_reserved_cmds && !sht->queue_reserved_command (from above)

I will drop the above change.

Thanks,

Bart.

[PATCH v5 06/28] scsi: core: Add scsi_{get,put}_internal_cmd() helpers

[Bart Van Assche]

From: Hannes Reinecke <hare@suse.de>

Add helper functions to allow LLDDs to allocate and free internal commands.

Cc: John Garry <john.g.garry@oracle.com>
Signed-off-by: Hannes Reinecke <hare@suse.de>
[ bvanassche: changed the 'nowait' argument into a 'flags' argument /
  added scsi_get_internal_cmd_hctx(). See also
  https://lore.kernel.org/linux-scsi/20211125151048.103910-3-hare@suse.de/ ]
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/scsi/scsi_lib.c    | 79 ++++++++++++++++++++++++++++++++++++++
 include/scsi/scsi_device.h |  7 ++++
 2 files changed, 86 insertions(+)

diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c
index 868013cb1f46..b25677e967d4 100644
--- a/drivers/scsi/scsi_lib.c
+++ b/drivers/scsi/scsi_lib.c
@@ -1247,6 +1247,17 @@ struct request *scsi_alloc_request(struct request_queue *q, blk_opf_t opf,
 }
 EXPORT_SYMBOL_GPL(scsi_alloc_request);
 
+static struct request *scsi_alloc_request_hctx(struct request_queue *q,
+		blk_opf_t opf, blk_mq_req_flags_t flags, unsigned int hctx_idx)
+{
+	struct request *rq;
+
+	rq = blk_mq_alloc_request_hctx(q, opf, flags, hctx_idx);
+	if (!IS_ERR(rq))
+		scsi_initialize_rq(rq);
+	return rq;
+}
+
 /*
  * Only called when the request isn't completed by SCSI, and not freed by
  * SCSI
@@ -2134,6 +2145,74 @@ void scsi_mq_free_tags(struct kref *kref)
 	complete(&shost->tagset_freed);
 }
 
+/**
+ * scsi_get_internal_cmd() - Allocate an internal SCSI command.
+ * @sdev: SCSI device from which to allocate the command
+ * @data_direction: Data direction for the allocated command
+ * @flags: request allocation flags, e.g. BLK_MQ_REQ_RESERVED or
+ *	BLK_MQ_REQ_NOWAIT.
+ *
+ * Allocates a SCSI command for internal LLDD use.
+ */
+struct scsi_cmnd *scsi_get_internal_cmd(struct scsi_device *sdev,
+					enum dma_data_direction data_direction,
+					blk_mq_req_flags_t flags)
+{
+	enum req_op op = data_direction == DMA_TO_DEVICE ? REQ_OP_DRV_OUT :
+							   REQ_OP_DRV_IN;
+	struct scsi_cmnd *scmd;
+	struct request *rq;
+
+	rq = scsi_alloc_request(sdev->request_queue, op, flags);
+	if (IS_ERR(rq))
+		return NULL;
+	scmd = blk_mq_rq_to_pdu(rq);
+	scmd->device = sdev;
+
+	return scmd;
+}
+EXPORT_SYMBOL_GPL(scsi_get_internal_cmd);
+
+/**
+ * scsi_get_internal_cmd_hctx() - Allocate an internal SCSI command from a given
+ *	hardware queue.
+ * @sdev: SCSI device from which to allocate the command
+ * @data_direction: Data direction for the allocated command
+ * @flags: request allocation flags, e.g. BLK_MQ_REQ_RESERVED or
+ *	BLK_MQ_REQ_NOWAIT.
+ * @hctx_idx: Hardware queue index.
+ *
+ * Allocates a SCSI command for internal LLDD use.
+ */
+struct scsi_cmnd *scsi_get_internal_cmd_hctx(struct scsi_device *sdev,
+			enum dma_data_direction data_direction,
+			blk_mq_req_flags_t flags, unsigned int hctx_idx)
+{
+	enum req_op op = data_direction == DMA_TO_DEVICE ? REQ_OP_DRV_OUT :
+							   REQ_OP_DRV_IN;
+	struct scsi_cmnd *scmd;
+	struct request *rq;
+
+	rq = scsi_alloc_request_hctx(sdev->request_queue, op, flags, hctx_idx);
+	if (IS_ERR(rq))
+		return NULL;
+	scmd = blk_mq_rq_to_pdu(rq);
+	scmd->device = sdev;
+
+	return scmd;
+}
+EXPORT_SYMBOL_GPL(scsi_get_internal_cmd_hctx);
+
+/**
+ * scsi_put_internal_cmd() - Free an internal SCSI command.
+ * @scmd: SCSI command to be freed
+ */
+void scsi_put_internal_cmd(struct scsi_cmnd *scmd)
+{
+	blk_mq_free_request(blk_mq_rq_from_pdu(scmd));
+}
+EXPORT_SYMBOL_GPL(scsi_put_internal_cmd);
+
 /**
  * scsi_device_from_queue - return sdev associated with a request_queue
  * @q: The request queue to return the sdev from
diff --git a/include/scsi/scsi_device.h b/include/scsi/scsi_device.h
index 3846f5dfc51c..1d83b17d95e6 100644
--- a/include/scsi/scsi_device.h
+++ b/include/scsi/scsi_device.h
@@ -558,6 +558,13 @@ int scsi_execute_cmd(struct scsi_device *sdev, const unsigned char *cmd,
 		     const struct scsi_exec_args *args);
 void scsi_failures_reset_retries(struct scsi_failures *failures);
 
+struct scsi_cmnd *scsi_get_internal_cmd(struct scsi_device *sdev,
+					enum dma_data_direction data_direction,
+					blk_mq_req_flags_t flags);
+struct scsi_cmnd *scsi_get_internal_cmd_hctx(struct scsi_device *sdev,
+			enum dma_data_direction data_direction,
+			blk_mq_req_flags_t flags, unsigned int hctx_idx);
+void scsi_put_internal_cmd(struct scsi_cmnd *scmd);
 extern void sdev_disable_disk_events(struct scsi_device *sdev);
 extern void sdev_enable_disk_events(struct scsi_device *sdev);
 extern int scsi_vpd_lun_id(struct scsi_device *, char *, size_t);

Re: [PATCH v5 06/28] scsi: core: Add scsi_{get,put}_internal_cmd() helpers

[John Garry]

On 24/09/2025 21:30, Bart Van Assche wrote:
> From: Hannes Reinecke<hare@suse.de>
> 
> Add helper functions to allow LLDDs to allocate and free internal commands.
> 
> Cc: John Garry<john.g.garry@oracle.com>
> Signed-off-by: Hannes Reinecke<hare@suse.de>
> [ bvanassche: changed the 'nowait' argument into a 'flags' argument /
>    added scsi_get_internal_cmd_hctx(). See also
>    https://lore.kernel.org/linux-scsi/20211125151048.103910-3-hare@suse.de/ ]
> Signed-off-by: Bart Van Assche<bvanassche@acm.org>

Reviewed-by: John Garry <john.g.garry@oracle.com>

[PATCH v5 07/28] scsi_debug: Abort SCSI commands via .queue_reserved_command()

[Bart Van Assche]

Add a .queue_reserved_command() implementation and call it from the code
path that aborts SCSI commands. This ensures that the code for
allocating a pseudo SCSI device and also the code for processing
reserved commands gets triggered while running blktests.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/scsi/scsi_debug.c | 106 ++++++++++++++++++++++++++++++++++----
 1 file changed, 97 insertions(+), 9 deletions(-)

diff --git a/drivers/scsi/scsi_debug.c b/drivers/scsi/scsi_debug.c
index 2a8638937d23..b376331c4cce 100644
--- a/drivers/scsi/scsi_debug.c
+++ b/drivers/scsi/scsi_debug.c
@@ -451,6 +451,23 @@ struct sdeb_store_info {
 #define shost_to_sdebug_host(shost)	\
 	dev_to_sdebug_host(shost->dma_dev)
 
+struct scsi_debug_abort_cmd {
+	u16 tag;
+	u16 hwq;
+};
+
+enum scsi_debug_internal_cmd_type {
+	SCSI_DEBUG_ABORT_CMD,
+};
+
+struct scsi_debug_internal_cmd {
+	enum scsi_debug_internal_cmd_type type;
+
+	union {
+		struct scsi_debug_abort_cmd abort_cmd;
+	};
+};
+
 enum sdeb_defer_type {SDEB_DEFER_NONE = 0, SDEB_DEFER_HRT = 1,
 		      SDEB_DEFER_WQ = 2, SDEB_DEFER_POLL = 3};
 
@@ -466,6 +483,8 @@ struct sdebug_defer {
 struct sdebug_scsi_cmd {
 	spinlock_t   lock;
 	struct sdebug_defer sd_dp;
+
+	struct scsi_debug_internal_cmd internal_cmd;
 };
 
 static atomic_t sdebug_cmnd_count;   /* number of incoming commands */
@@ -6729,20 +6748,48 @@ static bool scsi_debug_stop_cmnd(struct scsi_cmnd *cmnd)
 	return false;
 }
 
+static int scsi_debug_setup_abort_cmd(struct scsi_cmnd *cmd,
+			const struct scsi_debug_internal_cmd *internal_cmd)
+{
+	struct sdebug_scsi_cmd *sdsc = scsi_cmd_priv(cmd);
+
+	sdsc->internal_cmd = *internal_cmd;
+
+	return 0;
+}
+
 /*
- * Called from scsi_debug_abort() only, which is for timed-out cmd.
+ * Abort a pending SCSI command. Only called from scsi_debug_abort(). Although
+ * it would be possible to call scsi_debug_stop_cmnd() directly, an internal
+ * command is allocated and submitted to use the reserved command
+ * infrastructure.
  */
 static bool scsi_debug_abort_cmnd(struct scsi_cmnd *cmnd)
 {
-	struct sdebug_scsi_cmd *sdsc = scsi_cmd_priv(cmnd);
-	unsigned long flags;
-	bool res;
-
-	spin_lock_irqsave(&sdsc->lock, flags);
-	res = scsi_debug_stop_cmnd(cmnd);
-	spin_unlock_irqrestore(&sdsc->lock, flags);
+	struct request *rq = scsi_cmd_to_rq(cmnd);
+	u32 unique_tag = blk_mq_unique_tag(rq);
+	u16 hwq = blk_mq_unique_tag_to_hwq(unique_tag);
+	u16 tag = blk_mq_unique_tag_to_tag(unique_tag);
+	struct scsi_device *sdev = cmnd->device;
+	struct Scsi_Host *shost = sdev->host;
+	const struct scsi_debug_internal_cmd ic = {
+		.type = SCSI_DEBUG_ABORT_CMD,
+		.abort_cmd = {
+			.tag = tag,
+			.hwq = hwq,
+		},
+	};
+	struct scsi_cmnd *abort_cmd;
+	struct request *abort_rq;
 
-	return res;
+	abort_cmd = scsi_get_internal_cmd(shost->pseudo_sdev, DMA_TO_DEVICE,
+					  BLK_MQ_REQ_RESERVED);
+	if (WARN_ON_ONCE(!abort_cmd))
+		return false;
+	scsi_debug_setup_abort_cmd(abort_cmd, &ic);
+	abort_rq = scsi_cmd_to_rq(abort_cmd);
+	abort_rq->timeout = secs_to_jiffies(3);
+	return blk_execute_rq(abort_rq, true) == BLK_STS_OK;
 }
 
 /*
@@ -9197,6 +9244,45 @@ static int sdebug_fail_cmd(struct scsi_cmnd *cmnd, int *retval,
 	return ret;
 }
 
+static void scsi_debug_abort_cmd(struct Scsi_Host *shost, struct scsi_cmnd *scp)
+{
+	struct sdebug_scsi_cmd *sdsc = scsi_cmd_priv(scp);
+	struct scsi_debug_abort_cmd *abort_cmd =
+		&sdsc->internal_cmd.abort_cmd;
+	struct blk_mq_tag_set *tag_set = &shost->tag_set;
+	unsigned int tag = abort_cmd->tag;
+	unsigned int hwq = abort_cmd->hwq;
+	struct blk_mq_tags *tags = tag_set->tags[hwq];
+	struct request *abort_rq = blk_mq_tag_to_rq(tags, tag);
+	struct scsi_cmnd *abort_scmd = blk_mq_rq_to_pdu(abort_rq);
+	struct sdebug_scsi_cmd *abort_sdsc = scsi_cmd_priv(abort_scmd);
+	bool res;
+
+	scoped_guard(spinlock_irqsave, &abort_sdsc->lock)
+		res = scsi_debug_stop_cmnd(abort_scmd);
+
+	scp->result = (res ? DID_OK : DID_ERROR) << 16;
+}
+
+static int scsi_debug_queue_reserved_command(struct Scsi_Host *shost,
+					     struct scsi_cmnd *scp)
+{
+	struct sdebug_scsi_cmd *sdsc = scsi_cmd_priv(scp);
+
+	switch (sdsc->internal_cmd.type) {
+	case SCSI_DEBUG_ABORT_CMD:
+		scsi_debug_abort_cmd(shost, scp);
+		break;
+	default:
+		WARN_ON_ONCE(true);
+		scp->result = DID_ERROR << 16;
+		break;
+	}
+
+	scsi_done(scp);
+	return 0;
+}
+
 static int scsi_debug_queuecommand(struct Scsi_Host *shost,
 				   struct scsi_cmnd *scp)
 {
@@ -9416,6 +9502,7 @@ static const struct scsi_host_template sdebug_driver_template = {
 	.sdev_destroy =		scsi_debug_sdev_destroy,
 	.ioctl =		scsi_debug_ioctl,
 	.queuecommand =		scsi_debug_queuecommand,
+	.queue_reserved_command = scsi_debug_queue_reserved_command,
 	.change_queue_depth =	sdebug_change_qdepth,
 	.map_queues =		sdebug_map_queues,
 	.mq_poll =		sdebug_blk_mq_poll,
@@ -9425,6 +9512,7 @@ static const struct scsi_host_template sdebug_driver_template = {
 	.eh_bus_reset_handler = scsi_debug_bus_reset,
 	.eh_host_reset_handler = scsi_debug_host_reset,
 	.can_queue =		SDEBUG_CANQUEUE,
+	.nr_reserved_cmds =	1,
 	.this_id =		7,
 	.sg_tablesize =		SG_MAX_SEGMENTS,
 	.cmd_per_lun =		DEF_CMD_PER_LUN,

Re: [PATCH v5 07/28] scsi_debug: Abort SCSI commands via .queue_reserved_command()

[John Garry]

On 24/09/2025 21:30, Bart Van Assche wrote:
> Add a .queue_reserved_command() implementation and call it from the code
> path that aborts SCSI commands. This ensures that the code for
> allocating a pseudo SCSI device and also the code for processing
> reserved commands gets triggered while running blktests.
> 
> Signed-off-by: Bart Van Assche <bvanassche@acm.org>

At least Suggested-by would be good, if you don't mind

> ---
>   drivers/scsi/scsi_debug.c | 106 ++++++++++++++++++++++++++++++++++----
>   1 file changed, 97 insertions(+), 9 deletions(-)
> 
> diff --git a/drivers/scsi/scsi_debug.c b/drivers/scsi/scsi_debug.c
> index 2a8638937d23..b376331c4cce 100644
> --- a/drivers/scsi/scsi_debug.c
> +++ b/drivers/scsi/scsi_debug.c
> @@ -451,6 +451,23 @@ struct sdeb_store_info {
>   #define shost_to_sdebug_host(shost)	\
>   	dev_to_sdebug_host(shost->dma_dev)
>   
> +struct scsi_debug_abort_cmd {
> +	u16 tag;
> +	u16 hwq;
> +};
> +
> +enum scsi_debug_internal_cmd_type {
> +	SCSI_DEBUG_ABORT_CMD,
> +};
> +
> +struct scsi_debug_internal_cmd {
> +	enum scsi_debug_internal_cmd_type type;
> +
> +	union {
> +		struct scsi_debug_abort_cmd abort_cmd;
> +	};
> +};
> +
>   enum sdeb_defer_type {SDEB_DEFER_NONE = 0, SDEB_DEFER_HRT = 1,
>   		      SDEB_DEFER_WQ = 2, SDEB_DEFER_POLL = 3};
>   
> @@ -466,6 +483,8 @@ struct sdebug_defer {
>   struct sdebug_scsi_cmd {
>   	spinlock_t   lock;
>   	struct sdebug_defer sd_dp;
> +
> +	struct scsi_debug_internal_cmd internal_cmd;

you could prob make this a union with sd_dp - I mean, could they ever 
both be simultaneously used?

>   };
>   
>   static atomic_t sdebug_cmnd_count;   /* number of incoming commands */
> @@ -6729,20 +6748,48 @@ static bool scsi_debug_stop_cmnd(struct scsi_cmnd *cmnd)
>   	return false;
>   }
>   
> +static int scsi_debug_setup_abort_cmd(struct scsi_cmnd *cmd,

nobody checks the return value

> +			const struct scsi_debug_internal_cmd *internal_cmd)
> +{
> +	struct sdebug_scsi_cmd *sdsc = scsi_cmd_priv(cmd);
> +
> +	sdsc->internal_cmd = *internal_cmd;
> +
> +	return 0;
> +}
> +
>   /*
> - * Called from scsi_debug_abort() only, which is for timed-out cmd.
> + * Abort a pending SCSI command. Only called from scsi_debug_abort(). Although
> + * it would be possible to call scsi_debug_stop_cmnd() directly, an internal
> + * command is allocated and submitted to use the reserved command
> + * infrastructure.
>    */
>   static bool scsi_debug_abort_cmnd(struct scsi_cmnd *cmnd)
>   {
> -	struct sdebug_scsi_cmd *sdsc = scsi_cmd_priv(cmnd);
> -	unsigned long flags;
> -	bool res;
> -
> -	spin_lock_irqsave(&sdsc->lock, flags);
> -	res = scsi_debug_stop_cmnd(cmnd);
> -	spin_unlock_irqrestore(&sdsc->lock, flags);
> +	struct request *rq = scsi_cmd_to_rq(cmnd);
> +	u32 unique_tag = blk_mq_unique_tag(rq);
> +	u16 hwq = blk_mq_unique_tag_to_hwq(unique_tag);
> +	u16 tag = blk_mq_unique_tag_to_tag(unique_tag);
> +	struct scsi_device *sdev = cmnd->device;
> +	struct Scsi_Host *shost = sdev->host;
> +	const struct scsi_debug_internal_cmd ic = {
> +		.type = SCSI_DEBUG_ABORT_CMD,
> +		.abort_cmd = {
> +			.tag = tag,
> +			.hwq = hwq,
> +		},
> +	};
> +	struct scsi_cmnd *abort_cmd;
> +	struct request *abort_rq;
>   
> -	return res;
> +	abort_cmd = scsi_get_internal_cmd(shost->pseudo_sdev, DMA_TO_DEVICE,

DMA_NONE?

> +					  BLK_MQ_REQ_RESERVED);
> +	if (WARN_ON_ONCE(!abort_cmd))
> +		return false;

I don't really think that this deserves a WARN_ON_ONCE

> +	scsi_debug_setup_abort_cmd(abort_cmd, &ic);
> +	abort_rq = scsi_cmd_to_rq(abort_cmd);
> +	abort_rq->timeout = secs_to_jiffies(3);
> +	return blk_execute_rq(abort_rq, true) == BLK_STS_OK;
>   }
>   
>   /*
> @@ -9197,6 +9244,45 @@ static int sdebug_fail_cmd(struct scsi_cmnd *cmnd, int *retval,
>   	return ret;
>   }
>   
> +static void scsi_debug_abort_cmd(struct Scsi_Host *shost, struct scsi_cmnd *scp)
> +{
> +	struct sdebug_scsi_cmd *sdsc = scsi_cmd_priv(scp);
> +	struct scsi_debug_abort_cmd *abort_cmd =
> +		&sdsc->internal_cmd.abort_cmd;
> +	struct blk_mq_tag_set *tag_set = &shost->tag_set;
> +	unsigned int tag = abort_cmd->tag;
> +	unsigned int hwq = abort_cmd->hwq;
> +	struct blk_mq_tags *tags = tag_set->tags[hwq];
> +	struct request *abort_rq = blk_mq_tag_to_rq(tags, tag);
> +	struct scsi_cmnd *abort_scmd = blk_mq_rq_to_pdu(abort_rq);
> +	struct sdebug_scsi_cmd *abort_sdsc = scsi_cmd_priv(abort_scmd);
> +	bool res;
> +
> +	scoped_guard(spinlock_irqsave, &abort_sdsc->lock)
> +		res = scsi_debug_stop_cmnd(abort_scmd);
> +
> +	scp->result = (res ? DID_OK : DID_ERROR) << 16;

personally I think that if-else is nicer, but that's just me

And please consider using set_host_byte()

> +}
> +
> +static int scsi_debug_queue_reserved_command(struct Scsi_Host *shost,
> +					     struct scsi_cmnd *scp)
> +{
> +	struct sdebug_scsi_cmd *sdsc = scsi_cmd_priv(scp);
> +
> +	switch (sdsc->internal_cmd.type) {
> +	case SCSI_DEBUG_ABORT_CMD:
> +		scsi_debug_abort_cmd(shost, scp);
> +		break;
> +	default:
> +		WARN_ON_ONCE(true);
> +		scp->result = DID_ERROR << 16;
> +		break;
> +	}
> +
> +	scsi_done(scp);
> +	return 0;
> +}
> +
>   static int scsi_debug_queuecommand(struct Scsi_Host *shost,
>   				   struct scsi_cmnd *scp)
>   {
> @@ -9416,6 +9502,7 @@ static const struct scsi_host_template sdebug_driver_template = {
>   	.sdev_destroy =		scsi_debug_sdev_destroy,
>   	.ioctl =		scsi_debug_ioctl,
>   	.queuecommand =		scsi_debug_queuecommand,
> +	.queue_reserved_command = scsi_debug_queue_reserved_command,
>   	.change_queue_depth =	sdebug_change_qdepth,
>   	.map_queues =		sdebug_map_queues,
>   	.mq_poll =		sdebug_blk_mq_poll,
> @@ -9425,6 +9512,7 @@ static const struct scsi_host_template sdebug_driver_template = {
>   	.eh_bus_reset_handler = scsi_debug_bus_reset,
>   	.eh_host_reset_handler = scsi_debug_host_reset,
>   	.can_queue =		SDEBUG_CANQUEUE,
> +	.nr_reserved_cmds =	1,
>   	.this_id =		7,
>   	.sg_tablesize =		SG_MAX_SEGMENTS,
>   	.cmd_per_lun =		DEF_CMD_PER_LUN,
> 

Re: [PATCH v5 07/28] scsi_debug: Abort SCSI commands via .queue_reserved_command()

[Bart Van Assche]

On 9/26/25 12:32 AM, John Garry wrote:
> On 24/09/2025 21:30, Bart Van Assche wrote:
>> Add a .queue_reserved_command() implementation and call it from the code
>> path that aborts SCSI commands. This ensures that the code for
>> allocating a pseudo SCSI device and also the code for processing
>> reserved commands gets triggered while running blktests.
>>
>> Signed-off-by: Bart Van Assche <bvanassche@acm.org>
> 
> At least Suggested-by would be good, if you don't mind

If nobody objects I will add the following text:

----------------------------------------------------------------------
Most of the code in this patch is a modified version of code from John
Garry. See also
https://lore.kernel.org/linux-scsi/75018e17-4dea-4e1b-8c92-7a224a1e13b9@oracle.com/

Suggested-by: John Garry <john.g.garry@oracle.com>
----------------------------------------------------------------------

>>   enum sdeb_defer_type {SDEB_DEFER_NONE = 0, SDEB_DEFER_HRT = 1,
>>                 SDEB_DEFER_WQ = 2, SDEB_DEFER_POLL = 3};
>> @@ -466,6 +483,8 @@ struct sdebug_defer {
>>   struct sdebug_scsi_cmd {
>>       spinlock_t   lock;
>>       struct sdebug_defer sd_dp;
>> +
>> +    struct scsi_debug_internal_cmd internal_cmd;
> 
> you could prob make this a union with sd_dp - I mean, could they ever 
> both be simultaneously used?

Introducing a union for sd_dp and internal_cmd only would suggest to
readers of the scsi_debug code that 'lock' protects both 'sd_dp' and
'internal_cmd', which is not the case. I propose to include the
following change in this patch:

diff --git a/drivers/scsi/scsi_debug.c b/drivers/scsi/scsi_debug.c
index 0d4bdd38597a..12b885a2f719 100644
--- a/drivers/scsi/scsi_debug.c
+++ b/drivers/scsi/scsi_debug.c
@@ -483,8 +483,6 @@ struct sdebug_defer {
  struct sdebug_scsi_cmd {
  	spinlock_t   lock;
  	struct sdebug_defer sd_dp;
-
-	struct scsi_debug_internal_cmd internal_cmd;
  };

  static atomic_t sdebug_cmnd_count;   /* number of incoming commands */
@@ -9518,7 +9516,8 @@ static const struct scsi_host_template 
sdebug_driver_template = {
  	.module =		THIS_MODULE,
  	.skip_settle_delay =	1,
  	.track_queue_depth =	1,
-	.cmd_size = sizeof(struct sdebug_scsi_cmd),
+	.cmd_size = sizeof(union { struct sdebug_scsi_cmd c;
+				   struct scsi_debug_internal_cmd ic; }),
  	.init_cmd_priv = sdebug_init_cmd_priv,
  	.target_alloc =		sdebug_target_alloc,
  	.target_destroy =	sdebug_target_destroy,

>> +static int scsi_debug_setup_abort_cmd(struct scsi_cmnd *cmd,
> 
> nobody checks the return value

Agreed. I will change the return type from 'int' into 'void'.

>> +    abort_cmd = scsi_get_internal_cmd(shost->pseudo_sdev, DMA_TO_DEVICE,
> 
> DMA_NONE?

DMA_TO_DEVICE is converted into REQ_OP_DRV_OUT by 
scsi_get_internal_cmd(). Isn't that more appropriate for an operation
that has side effects (aborting a SCSI command) rather than
REQ_OP_DRV_IN?

>> +                      BLK_MQ_REQ_RESERVED);
>> +    if (WARN_ON_ONCE(!abort_cmd))
>> +        return false;
> 
> I don't really think that this deserves a WARN_ON_ONCE

OK, I will remove the WARN_ON_ONCE().

>> +    scp->result = (res ? DID_OK : DID_ERROR) << 16;
> 
> personally I think that if-else is nicer, but that's just me
> 
> And please consider using set_host_byte()

I will use set_host_byte().

Thanks,

Bart.

Re: [PATCH v5 07/28] scsi_debug: Abort SCSI commands via .queue_reserved_command()

[John Garry]

On 26/09/2025 21:44, Bart Van Assche wrote:
> On 9/26/25 12:32 AM, John Garry wrote:
>> On 24/09/2025 21:30, Bart Van Assche wrote:
>>> Add a .queue_reserved_command() implementation and call it from the code
>>> path that aborts SCSI commands. This ensures that the code for
>>> allocating a pseudo SCSI device and also the code for processing
>>> reserved commands gets triggered while running blktests.
>>>
>>> Signed-off-by: Bart Van Assche <bvanassche@acm.org>
>>
>> At least Suggested-by would be good, if you don't mind
> 
> If nobody objects I will add the following text:
> 
> ----------------------------------------------------------------------
> Most of the code in this patch is a modified version of code from John
> Garry. See also
> https://lore.kernel.org/linux- 
> scsi/75018e17-4dea-4e1b-8c92-7a224a1e13b9@oracle.com/
> 
> Suggested-by: John Garry <john.g.garry@oracle.com>
> ----------------------------------------------------------------------

fine, thanks

> 
>>>   enum sdeb_defer_type {SDEB_DEFER_NONE = 0, SDEB_DEFER_HRT = 1,
>>>                 SDEB_DEFER_WQ = 2, SDEB_DEFER_POLL = 3};
>>> @@ -466,6 +483,8 @@ struct sdebug_defer {
>>>   struct sdebug_scsi_cmd {
>>>       spinlock_t   lock;
>>>       struct sdebug_defer sd_dp;
>>> +
>>> +    struct scsi_debug_internal_cmd internal_cmd;
>>
>> you could prob make this a union with sd_dp - I mean, could they ever 
>> both be simultaneously used?
> 
> Introducing a union for sd_dp and internal_cmd only would suggest to
> readers of the scsi_debug code that 'lock' protects both 'sd_dp' and
> 'internal_cmd', which is not the case. I propose to include the
> following change in this patch:
> 
> diff --git a/drivers/scsi/scsi_debug.c b/drivers/scsi/scsi_debug.c
> index 0d4bdd38597a..12b885a2f719 100644
> --- a/drivers/scsi/scsi_debug.c
> +++ b/drivers/scsi/scsi_debug.c
> @@ -483,8 +483,6 @@ struct sdebug_defer {
>   struct sdebug_scsi_cmd {
>       spinlock_t   lock;
>       struct sdebug_defer sd_dp;
> -
> -    struct scsi_debug_internal_cmd internal_cmd;
>   };
> 
>   static atomic_t sdebug_cmnd_count;   /* number of incoming commands */
> @@ -9518,7 +9516,8 @@ static const struct scsi_host_template 
> sdebug_driver_template = {
>       .module =        THIS_MODULE,
>       .skip_settle_delay =    1,
>       .track_queue_depth =    1,
> -    .cmd_size = sizeof(struct sdebug_scsi_cmd),
> +    .cmd_size = sizeof(union { struct sdebug_scsi_cmd c;
> +                   struct scsi_debug_internal_cmd ic; }),

ok, but you could also introduce a new structure which defines this 
union. That's looks a little bit better.

>       .init_cmd_priv = sdebug_init_cmd_priv,
>       .target_alloc =        sdebug_target_alloc,
>       .target_destroy =    sdebug_target_destroy,
> 
>>> +static int scsi_debug_setup_abort_cmd(struct scsi_cmnd *cmd,
>>
>> nobody checks the return value
> 
> Agreed. I will change the return type from 'int' into 'void'.
> 
>>> +    abort_cmd = scsi_get_internal_cmd(shost->pseudo_sdev, 
>>> DMA_TO_DEVICE,
>>
>> DMA_NONE?
> 
> DMA_TO_DEVICE is converted into REQ_OP_DRV_OUT by 
> scsi_get_internal_cmd(). Isn't that more appropriate for an operation
> that has side effects (aborting a SCSI command) rather than
> REQ_OP_DRV_IN?

but it it make a difference if we use REQ_OP_DRV_OUT or REQ_OP_DRV_IN? 
If not, then I think DMA_NONE makes more sense.

Thanks,
John

Re: [PATCH v5 07/28] scsi_debug: Abort SCSI commands via .queue_reserved_command()

[Bart Van Assche]

On 9/30/25 12:03 AM, John Garry wrote:
> On 26/09/2025 21:44, Bart Van Assche wrote:
>>   static atomic_t sdebug_cmnd_count;   /* number of incoming commands */
>> @@ -9518,7 +9516,8 @@ static const struct scsi_host_template 
>> sdebug_driver_template = {
>>       .module =        THIS_MODULE,
>>       .skip_settle_delay =    1,
>>       .track_queue_depth =    1,
>> -    .cmd_size = sizeof(struct sdebug_scsi_cmd),
>> +    .cmd_size = sizeof(union { struct sdebug_scsi_cmd c;
>> +                   struct scsi_debug_internal_cmd ic; }),
> 
> ok, but you could also introduce a new structure which defines this 
> union. That's looks a little bit better.
> 
>>>> +    abort_cmd = scsi_get_internal_cmd(shost->pseudo_sdev, 
>>>> DMA_TO_DEVICE,
>>>
>>> DMA_NONE?
>>
>> DMA_TO_DEVICE is converted into REQ_OP_DRV_OUT by 
>> scsi_get_internal_cmd(). Isn't that more appropriate for an operation
>> that has side effects (aborting a SCSI command) rather than
>> REQ_OP_DRV_IN?
> 
> but it it make a difference if we use REQ_OP_DRV_OUT or REQ_OP_DRV_IN? 
> If not, then I think DMA_NONE makes more sense.

The above feedback has been integrated in the patch below. The following
additional changes have been integrated in this patch:
- Use scsi_host_find_tag() to make the implementation more compact.
- Pass a 32-bit tag from scsi_debug_abort_cmnd() to the code that
   implements the abort instead of splitting this 32-bit tag into a
   16-bit hwq index and a 16-bit tag.

This new version passes the blktests test that I wrote to trigger the
new code.

Thanks,

Bart.


Subject: [PATCH] scsi_debug: Abort SCSI commands via an internal command

Add a .queue_reserved_command() implementation and call it from the code
path that aborts SCSI commands. This ensures that the code for
allocating a pseudo SCSI device and also the code for allocating and
processing reserved commands gets triggered while running blktests.

Most of the code in this patch is a modified version of code from John
Garry. See also
https://lore.kernel.org/linux-scsi/75018e17-4dea-4e1b-8c92-7a224a1e13b9@oracle.com/

Suggested-by: John Garry <john.g.garry@oracle.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
  drivers/scsi/scsi_debug.c | 113 ++++++++++++++++++++++++++++++++++----
  1 file changed, 102 insertions(+), 11 deletions(-)

diff --git a/drivers/scsi/scsi_debug.c b/drivers/scsi/scsi_debug.c
index 2a8638937d23..6e68ed131726 100644
--- a/drivers/scsi/scsi_debug.c
+++ b/drivers/scsi/scsi_debug.c
@@ -6729,20 +6729,59 @@ static bool scsi_debug_stop_cmnd(struct 
scsi_cmnd *cmnd)
  	return false;
  }

+struct sdebug_abort_cmd {
+	u32 unique_tag;
+};
+
+enum sdebug_internal_cmd_type {
+	SCSI_DEBUG_ABORT_CMD,
+};
+
+struct sdebug_internal_cmd {
+	enum sdebug_internal_cmd_type type;
+
+	union {
+		struct sdebug_abort_cmd abort_cmd;
+	};
+};
+
+union sdebug_priv {
+	struct sdebug_scsi_cmd c;
+	struct sdebug_internal_cmd ic;
+};
+
  /*
- * Called from scsi_debug_abort() only, which is for timed-out cmd.
+ * Abort a pending SCSI command. Only called from scsi_debug_abort(). 
Although
+ * it would be possible to call scsi_debug_stop_cmnd() directly, an 
internal
+ * command is allocated and submitted to trigger the reserved command
+ * infrastructure.
   */
  static bool scsi_debug_abort_cmnd(struct scsi_cmnd *cmnd)
  {
-	struct sdebug_scsi_cmd *sdsc = scsi_cmd_priv(cmnd);
-	unsigned long flags;
-	bool res;
-
-	spin_lock_irqsave(&sdsc->lock, flags);
-	res = scsi_debug_stop_cmnd(cmnd);
-	spin_unlock_irqrestore(&sdsc->lock, flags);
-
-	return res;
+	struct Scsi_Host *shost = cmnd->device->host;
+	struct request *rq = scsi_cmd_to_rq(cmnd);
+	u32 unique_tag = blk_mq_unique_tag(rq);
+	struct sdebug_internal_cmd *sdic;
+	struct scsi_cmnd *abort_cmd;
+	struct request *abort_rq;
+	blk_status_t res;
+
+	abort_cmd = scsi_get_internal_cmd(shost->pseudo_sdev, DMA_NONE,
+					  BLK_MQ_REQ_RESERVED);
+	if (!abort_cmd)
+		return false;
+	sdic = scsi_cmd_priv(abort_cmd);
+	*sdic = (struct sdebug_internal_cmd) {
+		.type = SCSI_DEBUG_ABORT_CMD,
+		.abort_cmd = {
+			.unique_tag = unique_tag,
+		},
+	};
+	abort_rq = scsi_cmd_to_rq(abort_cmd);
+	abort_rq->timeout = secs_to_jiffies(3);
+	res = blk_execute_rq(abort_rq, true);
+	scsi_put_internal_cmd(abort_cmd);
+	return res == BLK_STS_OK;
  }

  /*
@@ -9197,6 +9236,53 @@ static int sdebug_fail_cmd(struct scsi_cmnd 
*cmnd, int *retval,
  	return ret;
  }

+static void scsi_debug_abort_cmd(struct Scsi_Host *shost, struct 
scsi_cmnd *scp)
+{
+	struct sdebug_internal_cmd *sdic = scsi_cmd_priv(scp);
+	struct sdebug_abort_cmd *abort_cmd = &sdic->abort_cmd;
+	const u32 unique_tag = abort_cmd->unique_tag;
+	struct scsi_cmnd *to_be_aborted_scmd =
+		scsi_host_find_tag(shost, unique_tag);
+	struct sdebug_scsi_cmd *to_be_aborted_sdsc =
+		scsi_cmd_priv(to_be_aborted_scmd);
+	bool res = false;
+
+	if (to_be_aborted_scmd) {
+		scoped_guard(spinlock_irqsave, &to_be_aborted_sdsc->lock)
+			res = scsi_debug_stop_cmnd(to_be_aborted_scmd);
+		if (res)
+			pr_info("%s: aborted command with tag %#x\n",
+				__func__, unique_tag);
+		else
+			pr_err("%s: failed to abort command with tag %#x\n",
+			       __func__, unique_tag);
+	} else {
+		pr_err("%s: command with tag %#x not found\n", __func__,
+		       unique_tag);
+	}
+
+	set_host_byte(scp, res ? DID_OK : DID_ERROR);
+}
+
+static int scsi_debug_process_reserved_command(struct Scsi_Host *shost,
+					       struct scsi_cmnd *scp)
+{
+	struct sdebug_internal_cmd *sdic = scsi_cmd_priv(scp);
+
+	switch (sdic->type) {
+	case SCSI_DEBUG_ABORT_CMD:
+		scsi_debug_abort_cmd(shost, scp);
+		break;
+	default:
+		WARN_ON_ONCE(true);
+		set_host_byte(scp, DID_ERROR);
+		break;
+	}
+
+	scsi_done(scp);
+	return 0;
+}
+
  static int scsi_debug_queuecommand(struct Scsi_Host *shost,
  				   struct scsi_cmnd *scp)
  {
@@ -9397,6 +9483,9 @@ static int sdebug_init_cmd_priv(struct Scsi_Host 
*shost, struct scsi_cmnd *cmd)
  	struct sdebug_scsi_cmd *sdsc = scsi_cmd_priv(cmd);
  	struct sdebug_defer *sd_dp = &sdsc->sd_dp;

+	if (blk_mq_is_reserved_rq(scsi_cmd_to_rq(cmd)))
+		return 0;
+
  	spin_lock_init(&sdsc->lock);
  	hrtimer_setup(&sd_dp->hrt, sdebug_q_cmd_hrt_complete, CLOCK_MONOTONIC,
  		      HRTIMER_MODE_REL_PINNED);
@@ -9416,6 +9505,7 @@ static const struct scsi_host_template 
sdebug_driver_template = {
  	.sdev_destroy =		scsi_debug_sdev_destroy,
  	.ioctl =		scsi_debug_ioctl,
  	.queuecommand =		scsi_debug_queuecommand,
+	.queue_reserved_command = scsi_debug_process_reserved_command,
  	.change_queue_depth =	sdebug_change_qdepth,
  	.map_queues =		sdebug_map_queues,
  	.mq_poll =		sdebug_blk_mq_poll,
@@ -9425,6 +9515,7 @@ static const struct scsi_host_template 
sdebug_driver_template = {
  	.eh_bus_reset_handler = scsi_debug_bus_reset,
  	.eh_host_reset_handler = scsi_debug_host_reset,
  	.can_queue =		SDEBUG_CANQUEUE,
+	.nr_reserved_cmds =	1,
  	.this_id =		7,
  	.sg_tablesize =		SG_MAX_SEGMENTS,
  	.cmd_per_lun =		DEF_CMD_PER_LUN,
@@ -9433,7 +9524,7 @@ static const struct scsi_host_template 
sdebug_driver_template = {
  	.module =		THIS_MODULE,
  	.skip_settle_delay =	1,
  	.track_queue_depth =	1,
-	.cmd_size = sizeof(struct sdebug_scsi_cmd),
+	.cmd_size = sizeof(union sdebug_priv),
  	.init_cmd_priv = sdebug_init_cmd_priv,
  	.target_alloc =		sdebug_target_alloc,
  	.target_destroy =	sdebug_target_destroy,

[PATCH v5 08/28] ufs: core: Move an assignment in ufshcd_mcq_process_cqe()

[Bart Van Assche]

Prepare for introducing a WARN_ON_ONCE() call in ufshcd_mcq_get_tag() if
the tag lookup fails.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufs-mcq.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/ufs/core/ufs-mcq.c b/drivers/ufs/core/ufs-mcq.c
index 1e50675772fe..4eb4f1af7800 100644
--- a/drivers/ufs/core/ufs-mcq.c
+++ b/drivers/ufs/core/ufs-mcq.c
@@ -303,9 +303,10 @@ static void ufshcd_mcq_process_cqe(struct ufs_hba *hba,
 				   struct ufs_hw_queue *hwq)
 {
 	struct cq_entry *cqe = ufshcd_mcq_cur_cqe(hwq);
-	int tag = ufshcd_mcq_get_tag(hba, cqe);
 
 	if (cqe->command_desc_base_addr) {
+		int tag = ufshcd_mcq_get_tag(hba, cqe);
+
 		ufshcd_compl_one_cqe(hba, tag, cqe);
 		/* After processed the cqe, mark it empty (invalid) entry */
 		cqe->command_desc_base_addr = 0;

RE: [PATCH v5 08/28] ufs: core: Move an assignment in ufshcd_mcq_process_cqe()

[Avri Altman]

> Prepare for introducing a WARN_ON_ONCE() call in ufshcd_mcq_get_tag() if
> the tag lookup fails.
> 
> Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Avri Altman <avri.altman@sandisk.com>

> ---
>  drivers/ufs/core/ufs-mcq.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/ufs/core/ufs-mcq.c b/drivers/ufs/core/ufs-mcq.c index
> 1e50675772fe..4eb4f1af7800 100644
> --- a/drivers/ufs/core/ufs-mcq.c
> +++ b/drivers/ufs/core/ufs-mcq.c
> @@ -303,9 +303,10 @@ static void ufshcd_mcq_process_cqe(struct ufs_hba
> *hba,
>                                    struct ufs_hw_queue *hwq)  {
>         struct cq_entry *cqe = ufshcd_mcq_cur_cqe(hwq);
> -       int tag = ufshcd_mcq_get_tag(hba, cqe);
> 
>         if (cqe->command_desc_base_addr) {
> +               int tag = ufshcd_mcq_get_tag(hba, cqe);
> +
>                 ufshcd_compl_one_cqe(hba, tag, cqe);
>                 /* After processed the cqe, mark it empty (invalid) entry */
>                 cqe->command_desc_base_addr = 0;

[PATCH v5 09/28] ufs: core: Change the type of one ufshcd_add_cmd_upiu_trace() argument

[Bart Van Assche]

Change the 'tag' argument into an LRB pointer. This patch prepares for the
removal of the hba->lrb[] array.

Reviewed-by: Avri Altman <avri.altman@sandisk.com>
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufshcd.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index cfc149f8238e..fa3d5a0a7db7 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -408,10 +408,11 @@ static void ufshcd_configure_wb(struct ufs_hba *hba)
 		ufshcd_wb_toggle_buf_flush(hba, true);
 }
 
-static void ufshcd_add_cmd_upiu_trace(struct ufs_hba *hba, unsigned int tag,
+static void ufshcd_add_cmd_upiu_trace(struct ufs_hba *hba,
+				      struct ufshcd_lrb *lrb,
 				      enum ufs_trace_str_t str_t)
 {
-	struct utp_upiu_req *rq = hba->lrb[tag].ucd_req_ptr;
+	struct utp_upiu_req *rq = lrb->ucd_req_ptr;
 	struct utp_upiu_header *header;
 
 	if (!trace_ufshcd_upiu_enabled())
@@ -420,7 +421,7 @@ static void ufshcd_add_cmd_upiu_trace(struct ufs_hba *hba, unsigned int tag,
 	if (str_t == UFS_CMD_SEND)
 		header = &rq->header;
 	else
-		header = &hba->lrb[tag].ucd_rsp_ptr->header;
+		header = &lrb->ucd_rsp_ptr->header;
 
 	trace_ufshcd_upiu(hba, str_t, header, &rq->sc.cdb,
 			  UFS_TSF_CDB);
@@ -494,7 +495,7 @@ static void ufshcd_add_command_trace(struct ufs_hba *hba, unsigned int tag,
 		return;
 
 	/* trace UPIU also */
-	ufshcd_add_cmd_upiu_trace(hba, tag, str_t);
+	ufshcd_add_cmd_upiu_trace(hba, lrbp, str_t);
 	if (!trace_ufshcd_command_enabled())
 		return;
 

[PATCH v5 10/28] ufs: core: Only call ufshcd_add_command_trace() for SCSI commands

[Bart Van Assche]

Instead of checking inside ufshcd_add_command_trace() whether 'cmd' points
at a SCSI command, let the caller perform that check. This patch prepares
for removing the lrbp->cmd pointer.

Reviewed-by: Avri Altman <avri.altman@sandisk.com>
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufshcd.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index fa3d5a0a7db7..394cc1b5719b 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -491,9 +491,6 @@ static void ufshcd_add_command_trace(struct ufs_hba *hba, unsigned int tag,
 	struct request *rq = scsi_cmd_to_rq(cmd);
 	int transfer_len = -1;
 
-	if (!cmd)
-		return;
-
 	/* trace UPIU also */
 	ufshcd_add_cmd_upiu_trace(hba, lrbp, str_t);
 	if (!trace_ufshcd_command_enabled())
@@ -2368,9 +2365,10 @@ void ufshcd_send_command(struct ufs_hba *hba, unsigned int task_tag,
 		lrbp->compl_time_stamp = ktime_set(0, 0);
 		lrbp->compl_time_stamp_local_clock = 0;
 	}
-	ufshcd_add_command_trace(hba, task_tag, UFS_CMD_SEND);
-	if (lrbp->cmd)
+	if (lrbp->cmd) {
+		ufshcd_add_command_trace(hba, task_tag, UFS_CMD_SEND);
 		ufshcd_clk_scaling_start_busy(hba);
+	}
 	if (unlikely(ufshcd_should_inform_monitor(hba, lrbp)))
 		ufshcd_start_monitor(hba, lrbp);
 

[PATCH v5 11/28] ufs: core: Change the type of one ufshcd_add_command_trace() argument

[Bart Van Assche]

Change the 'tag' argument into a SCSI command pointer. This patch prepares
for the removal of the hba->lrb[] array.

Reviewed-by: Avri Altman <avri.altman@sandisk.com>
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufshcd.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index 394cc1b5719b..de04ee74cad6 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -478,7 +478,7 @@ static void ufshcd_add_uic_command_trace(struct ufs_hba *hba,
 				 ufshcd_readl(hba, REG_UIC_COMMAND_ARG_3));
 }
 
-static void ufshcd_add_command_trace(struct ufs_hba *hba, unsigned int tag,
+static void ufshcd_add_command_trace(struct ufs_hba *hba, struct scsi_cmnd *cmd,
 				     enum ufs_trace_str_t str_t)
 {
 	u64 lba = 0;
@@ -486,9 +486,9 @@ static void ufshcd_add_command_trace(struct ufs_hba *hba, unsigned int tag,
 	u32 doorbell = 0;
 	u32 intr;
 	u32 hwq_id = 0;
-	struct ufshcd_lrb *lrbp = &hba->lrb[tag];
-	struct scsi_cmnd *cmd = lrbp->cmd;
 	struct request *rq = scsi_cmd_to_rq(cmd);
+	unsigned int tag = rq->tag;
+	struct ufshcd_lrb *lrbp = &hba->lrb[tag];
 	int transfer_len = -1;
 
 	/* trace UPIU also */
@@ -506,7 +506,7 @@ static void ufshcd_add_command_trace(struct ufs_hba *hba, unsigned int tag,
 		       be32_to_cpu(lrbp->ucd_req_ptr->sc.exp_data_transfer_len);
 		lba = scsi_get_lba(cmd);
 		if (opcode == WRITE_10)
-			group_id = lrbp->cmd->cmnd[6];
+			group_id = cmd->cmnd[6];
 	} else if (opcode == UNMAP) {
 		/*
 		 * The number of Bytes to be unmapped beginning with the lba.
@@ -2366,7 +2366,7 @@ void ufshcd_send_command(struct ufs_hba *hba, unsigned int task_tag,
 		lrbp->compl_time_stamp_local_clock = 0;
 	}
 	if (lrbp->cmd) {
-		ufshcd_add_command_trace(hba, task_tag, UFS_CMD_SEND);
+		ufshcd_add_command_trace(hba, lrbp->cmd, UFS_CMD_SEND);
 		ufshcd_clk_scaling_start_busy(hba);
 	}
 	if (unlikely(ufshcd_should_inform_monitor(hba, lrbp)))
@@ -5633,7 +5633,7 @@ void ufshcd_compl_one_cqe(struct ufs_hba *hba, int task_tag,
 	if (cmd) {
 		if (unlikely(ufshcd_should_inform_monitor(hba, lrbp)))
 			ufshcd_update_monitor(hba, lrbp);
-		ufshcd_add_command_trace(hba, task_tag, UFS_CMD_COMP);
+		ufshcd_add_command_trace(hba, cmd, UFS_CMD_COMP);
 		cmd->result = ufshcd_transfer_rsp_status(hba, lrbp, cqe);
 		ufshcd_release_scsi_cmd(hba, lrbp);
 		/* Do not touch lrbp after scsi done */

[PATCH v5 12/28] ufs: core: Change the type of one ufshcd_send_command() argument

[Bart Van Assche]

Change the 'task_tag' argument into an LRB pointer. This patch prepares
for the removal of the hba->lrb[] array.

Reviewed-by: Avri Altman <avri.altman@sandisk.com>
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufshcd.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index de04ee74cad6..abc843fdbbc4 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -2349,14 +2349,13 @@ static void ufshcd_update_monitor(struct ufs_hba *hba, const struct ufshcd_lrb *
 /**
  * ufshcd_send_command - Send SCSI or device management commands
  * @hba: per adapter instance
- * @task_tag: Task tag of the command
+ * @lrbp: Local reference block of SCSI command
  * @hwq: pointer to hardware queue instance
  */
-static inline
-void ufshcd_send_command(struct ufs_hba *hba, unsigned int task_tag,
-			 struct ufs_hw_queue *hwq)
+static inline void ufshcd_send_command(struct ufs_hba *hba,
+				       struct ufshcd_lrb *lrbp,
+				       struct ufs_hw_queue *hwq)
 {
-	struct ufshcd_lrb *lrbp = &hba->lrb[task_tag];
 	unsigned long flags;
 
 	if (hba->monitor.enabled) {
@@ -3065,7 +3064,7 @@ static int ufshcd_queuecommand(struct Scsi_Host *host, struct scsi_cmnd *cmd)
 	if (hba->mcq_enabled)
 		hwq = ufshcd_mcq_req_to_hwq(hba, scsi_cmd_to_rq(cmd));
 
-	ufshcd_send_command(hba, tag, hwq);
+	ufshcd_send_command(hba, lrbp, hwq);
 
 out:
 	if (ufs_trigger_eh(hba)) {
@@ -3310,7 +3309,7 @@ static int ufshcd_issue_dev_cmd(struct ufs_hba *hba, struct ufshcd_lrb *lrbp,
 	int err;
 
 	ufshcd_add_query_upiu_trace(hba, UFS_QUERY_SEND, lrbp->ucd_req_ptr);
-	ufshcd_send_command(hba, tag, hba->dev_cmd_queue);
+	ufshcd_send_command(hba, lrbp, hba->dev_cmd_queue);
 	err = ufshcd_wait_for_dev_cmd(hba, lrbp, timeout);
 
 	ufshcd_add_query_upiu_trace(hba, err ? UFS_QUERY_ERR : UFS_QUERY_COMP,

[PATCH v5 13/28] ufs: core: Only call ufshcd_should_inform_monitor() for SCSI commands

[Bart Van Assche]

ufshcd_should_inform_monitor() only returns 'true' for SCSI commands.
Instead of checking inside ufshcd_should_inform_monitor() whether its
second argument represents a SCSI command, only call this function for
SCSI commands. This patch prepares for removing the lrbp->cmd member.

Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufshcd.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index abc843fdbbc4..be8fd3b15d5f 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -2292,12 +2292,13 @@ static inline int ufshcd_monitor_opcode2dir(u8 opcode)
 		return -EINVAL;
 }
 
+/* Must only be called for SCSI commands. */
 static inline bool ufshcd_should_inform_monitor(struct ufs_hba *hba,
 						struct ufshcd_lrb *lrbp)
 {
 	const struct ufs_hba_monitor *m = &hba->monitor;
 
-	return (m->enabled && lrbp && lrbp->cmd &&
+	return (m->enabled &&
 		(!m->chunk_size || m->chunk_size == lrbp->cmd->sdb.length) &&
 		ktime_before(hba->monitor.enabled_ts, lrbp->issue_time_stamp));
 }
@@ -2367,9 +2368,9 @@ static inline void ufshcd_send_command(struct ufs_hba *hba,
 	if (lrbp->cmd) {
 		ufshcd_add_command_trace(hba, lrbp->cmd, UFS_CMD_SEND);
 		ufshcd_clk_scaling_start_busy(hba);
+		if (unlikely(ufshcd_should_inform_monitor(hba, lrbp)))
+			ufshcd_start_monitor(hba, lrbp);
 	}
-	if (unlikely(ufshcd_should_inform_monitor(hba, lrbp)))
-		ufshcd_start_monitor(hba, lrbp);
 
 	if (hba->mcq_enabled) {
 		int utrd_size = sizeof(struct utp_transfer_req_desc);

[PATCH v5 14/28] ufs: core: Change the monitor function argument types

[Bart Van Assche]

Pass a SCSI command pointer instead of a struct ufshcd_lrb pointer. This
patch prepares for combining the SCSI command and ufshcd_lrb data
structures into a single data structure.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufshcd.c | 44 +++++++++++++++++++--------------------
 1 file changed, 22 insertions(+), 22 deletions(-)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index be8fd3b15d5f..ec1c8bdf07e6 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -2294,19 +2294,20 @@ static inline int ufshcd_monitor_opcode2dir(u8 opcode)
 
 /* Must only be called for SCSI commands. */
 static inline bool ufshcd_should_inform_monitor(struct ufs_hba *hba,
-						struct ufshcd_lrb *lrbp)
+						struct scsi_cmnd *cmd)
 {
 	const struct ufs_hba_monitor *m = &hba->monitor;
+	struct request *rq = scsi_cmd_to_rq(cmd);
+	struct ufshcd_lrb *lrbp = &hba->lrb[rq->tag];
 
-	return (m->enabled &&
-		(!m->chunk_size || m->chunk_size == lrbp->cmd->sdb.length) &&
-		ktime_before(hba->monitor.enabled_ts, lrbp->issue_time_stamp));
+	return m->enabled &&
+	       (!m->chunk_size || m->chunk_size == cmd->sdb.length) &&
+	       ktime_before(hba->monitor.enabled_ts, lrbp->issue_time_stamp);
 }
 
-static void ufshcd_start_monitor(struct ufs_hba *hba,
-				 const struct ufshcd_lrb *lrbp)
+static void ufshcd_start_monitor(struct ufs_hba *hba, struct scsi_cmnd *cmd)
 {
-	int dir = ufshcd_monitor_opcode2dir(*lrbp->cmd->cmnd);
+	int dir = ufshcd_monitor_opcode2dir(cmd->cmnd[0]);
 	unsigned long flags;
 
 	spin_lock_irqsave(hba->host->host_lock, flags);
@@ -2315,14 +2316,15 @@ static void ufshcd_start_monitor(struct ufs_hba *hba,
 	spin_unlock_irqrestore(hba->host->host_lock, flags);
 }
 
-static void ufshcd_update_monitor(struct ufs_hba *hba, const struct ufshcd_lrb *lrbp)
+static void ufshcd_update_monitor(struct ufs_hba *hba, struct scsi_cmnd *cmd)
 {
-	int dir = ufshcd_monitor_opcode2dir(*lrbp->cmd->cmnd);
+	struct request *req = scsi_cmd_to_rq(cmd);
+	struct ufshcd_lrb *lrbp = &hba->lrb[req->tag];
+	int dir = ufshcd_monitor_opcode2dir(cmd->cmnd[0]);
 	unsigned long flags;
 
 	spin_lock_irqsave(hba->host->host_lock, flags);
 	if (dir >= 0 && hba->monitor.nr_queued[dir] > 0) {
-		const struct request *req = scsi_cmd_to_rq(lrbp->cmd);
 		struct ufs_hba_monitor *m = &hba->monitor;
 		ktime_t now, inc, lat;
 
@@ -2357,6 +2359,7 @@ static inline void ufshcd_send_command(struct ufs_hba *hba,
 				       struct ufshcd_lrb *lrbp,
 				       struct ufs_hw_queue *hwq)
 {
+	struct scsi_cmnd *cmd = lrbp->cmd;
 	unsigned long flags;
 
 	if (hba->monitor.enabled) {
@@ -2365,11 +2368,11 @@ static inline void ufshcd_send_command(struct ufs_hba *hba,
 		lrbp->compl_time_stamp = ktime_set(0, 0);
 		lrbp->compl_time_stamp_local_clock = 0;
 	}
-	if (lrbp->cmd) {
-		ufshcd_add_command_trace(hba, lrbp->cmd, UFS_CMD_SEND);
+	if (cmd) {
+		ufshcd_add_command_trace(hba, cmd, UFS_CMD_SEND);
 		ufshcd_clk_scaling_start_busy(hba);
-		if (unlikely(ufshcd_should_inform_monitor(hba, lrbp)))
-			ufshcd_start_monitor(hba, lrbp);
+		if (unlikely(ufshcd_should_inform_monitor(hba, cmd)))
+			ufshcd_start_monitor(hba, cmd);
 	}
 
 	if (hba->mcq_enabled) {
@@ -2385,8 +2388,7 @@ static inline void ufshcd_send_command(struct ufs_hba *hba,
 	} else {
 		spin_lock_irqsave(&hba->outstanding_lock, flags);
 		if (hba->vops && hba->vops->setup_xfer_req)
-			hba->vops->setup_xfer_req(hba, lrbp->task_tag,
-						  !!lrbp->cmd);
+			hba->vops->setup_xfer_req(hba, lrbp->task_tag, !!cmd);
 		__set_bit(lrbp->task_tag, &hba->outstanding_reqs);
 		ufshcd_writel(hba, 1 << lrbp->task_tag,
 			      REG_UTP_TRANSFER_REQ_DOOR_BELL);
@@ -5620,19 +5622,17 @@ void ufshcd_release_scsi_cmd(struct ufs_hba *hba,
 void ufshcd_compl_one_cqe(struct ufs_hba *hba, int task_tag,
 			  struct cq_entry *cqe)
 {
-	struct ufshcd_lrb *lrbp;
-	struct scsi_cmnd *cmd;
+	struct ufshcd_lrb *lrbp = &hba->lrb[task_tag];
+	struct scsi_cmnd *cmd = lrbp->cmd;
 	enum utp_ocs ocs;
 
-	lrbp = &hba->lrb[task_tag];
 	if (hba->monitor.enabled) {
 		lrbp->compl_time_stamp = ktime_get();
 		lrbp->compl_time_stamp_local_clock = local_clock();
 	}
-	cmd = lrbp->cmd;
 	if (cmd) {
-		if (unlikely(ufshcd_should_inform_monitor(hba, lrbp)))
-			ufshcd_update_monitor(hba, lrbp);
+		if (unlikely(ufshcd_should_inform_monitor(hba, cmd)))
+			ufshcd_update_monitor(hba, cmd);
 		ufshcd_add_command_trace(hba, cmd, UFS_CMD_COMP);
 		cmd->result = ufshcd_transfer_rsp_status(hba, lrbp, cqe);
 		ufshcd_release_scsi_cmd(hba, lrbp);

RE: [PATCH v5 14/28] ufs: core: Change the monitor function argument types

[Avri Altman]

> Pass a SCSI command pointer instead of a struct ufshcd_lrb pointer. This patch
> prepares for combining the SCSI command and ufshcd_lrb data structures into
> a single data structure.
> 
> Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Avri Altman <avri.altman@sandisk.com>

> ---
>  drivers/ufs/core/ufshcd.c | 44 +++++++++++++++++++--------------------
>  1 file changed, 22 insertions(+), 22 deletions(-)
> 
> diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c index
> be8fd3b15d5f..ec1c8bdf07e6 100644
> --- a/drivers/ufs/core/ufshcd.c
> +++ b/drivers/ufs/core/ufshcd.c
> @@ -2294,19 +2294,20 @@ static inline int ufshcd_monitor_opcode2dir(u8
> opcode)
> 
>  /* Must only be called for SCSI commands. */  static inline bool
> ufshcd_should_inform_monitor(struct ufs_hba *hba,
> -                                               struct ufshcd_lrb *lrbp)
> +                                               struct scsi_cmnd *cmd)
>  {
>         const struct ufs_hba_monitor *m = &hba->monitor;
> +       struct request *rq = scsi_cmd_to_rq(cmd);
> +       struct ufshcd_lrb *lrbp = &hba->lrb[rq->tag];
> 
> -       return (m->enabled &&
> -               (!m->chunk_size || m->chunk_size == lrbp->cmd->sdb.length) &&
> -               ktime_before(hba->monitor.enabled_ts, lrbp->issue_time_stamp));
> +       return m->enabled &&
> +              (!m->chunk_size || m->chunk_size == cmd->sdb.length) &&
> +              ktime_before(hba->monitor.enabled_ts,
> + lrbp->issue_time_stamp);
>  }
> 
> -static void ufshcd_start_monitor(struct ufs_hba *hba,
> -                                const struct ufshcd_lrb *lrbp)
> +static void ufshcd_start_monitor(struct ufs_hba *hba, struct scsi_cmnd
> +*cmd)
>  {
> -       int dir = ufshcd_monitor_opcode2dir(*lrbp->cmd->cmnd);
> +       int dir = ufshcd_monitor_opcode2dir(cmd->cmnd[0]);
>         unsigned long flags;
> 
>         spin_lock_irqsave(hba->host->host_lock, flags); @@ -2315,14 +2316,15
> @@ static void ufshcd_start_monitor(struct ufs_hba *hba,
>         spin_unlock_irqrestore(hba->host->host_lock, flags);  }
> 
> -static void ufshcd_update_monitor(struct ufs_hba *hba, const struct
> ufshcd_lrb *lrbp)
> +static void ufshcd_update_monitor(struct ufs_hba *hba, struct scsi_cmnd
> +*cmd)
>  {
> -       int dir = ufshcd_monitor_opcode2dir(*lrbp->cmd->cmnd);
> +       struct request *req = scsi_cmd_to_rq(cmd);
> +       struct ufshcd_lrb *lrbp = &hba->lrb[req->tag];
> +       int dir = ufshcd_monitor_opcode2dir(cmd->cmnd[0]);
>         unsigned long flags;
> 
>         spin_lock_irqsave(hba->host->host_lock, flags);
>         if (dir >= 0 && hba->monitor.nr_queued[dir] > 0) {
> -               const struct request *req = scsi_cmd_to_rq(lrbp->cmd);
>                 struct ufs_hba_monitor *m = &hba->monitor;
>                 ktime_t now, inc, lat;
> 
> @@ -2357,6 +2359,7 @@ static inline void ufshcd_send_command(struct
> ufs_hba *hba,
>                                        struct ufshcd_lrb *lrbp,
>                                        struct ufs_hw_queue *hwq)  {
> +       struct scsi_cmnd *cmd = lrbp->cmd;
>         unsigned long flags;
> 
>         if (hba->monitor.enabled) {
> @@ -2365,11 +2368,11 @@ static inline void ufshcd_send_command(struct
> ufs_hba *hba,
>                 lrbp->compl_time_stamp = ktime_set(0, 0);
>                 lrbp->compl_time_stamp_local_clock = 0;
>         }
> -       if (lrbp->cmd) {
> -               ufshcd_add_command_trace(hba, lrbp->cmd, UFS_CMD_SEND);
> +       if (cmd) {
> +               ufshcd_add_command_trace(hba, cmd, UFS_CMD_SEND);
>                 ufshcd_clk_scaling_start_busy(hba);
> -               if (unlikely(ufshcd_should_inform_monitor(hba, lrbp)))
> -                       ufshcd_start_monitor(hba, lrbp);
> +               if (unlikely(ufshcd_should_inform_monitor(hba, cmd)))
> +                       ufshcd_start_monitor(hba, cmd);
>         }
> 
>         if (hba->mcq_enabled) {
> @@ -2385,8 +2388,7 @@ static inline void ufshcd_send_command(struct
> ufs_hba *hba,
>         } else {
>                 spin_lock_irqsave(&hba->outstanding_lock, flags);
>                 if (hba->vops && hba->vops->setup_xfer_req)
> -                       hba->vops->setup_xfer_req(hba, lrbp->task_tag,
> -                                                 !!lrbp->cmd);
> +                       hba->vops->setup_xfer_req(hba, lrbp->task_tag,
> + !!cmd);
>                 __set_bit(lrbp->task_tag, &hba->outstanding_reqs);
>                 ufshcd_writel(hba, 1 << lrbp->task_tag,
>                               REG_UTP_TRANSFER_REQ_DOOR_BELL); @@ -5620,19
> +5622,17 @@ void ufshcd_release_scsi_cmd(struct ufs_hba *hba,  void
> ufshcd_compl_one_cqe(struct ufs_hba *hba, int task_tag,
>                           struct cq_entry *cqe)  {
> -       struct ufshcd_lrb *lrbp;
> -       struct scsi_cmnd *cmd;
> +       struct ufshcd_lrb *lrbp = &hba->lrb[task_tag];
> +       struct scsi_cmnd *cmd = lrbp->cmd;
>         enum utp_ocs ocs;
> 
> -       lrbp = &hba->lrb[task_tag];
>         if (hba->monitor.enabled) {
>                 lrbp->compl_time_stamp = ktime_get();
>                 lrbp->compl_time_stamp_local_clock = local_clock();
>         }
> -       cmd = lrbp->cmd;
>         if (cmd) {
> -               if (unlikely(ufshcd_should_inform_monitor(hba, lrbp)))
> -                       ufshcd_update_monitor(hba, lrbp);
> +               if (unlikely(ufshcd_should_inform_monitor(hba, cmd)))
> +                       ufshcd_update_monitor(hba, cmd);
>                 ufshcd_add_command_trace(hba, cmd, UFS_CMD_COMP);
>                 cmd->result = ufshcd_transfer_rsp_status(hba, lrbp, cqe);
>                 ufshcd_release_scsi_cmd(hba, lrbp);

[PATCH v5 15/28] ufs: core: Rework ufshcd_mcq_compl_pending_transfer()

[Bart Van Assche]

Replace a tag loop with blk_mq_tagset_busy_iter(). This patch prepares
for removing the hba->lrb[] array.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufshcd.c | 80 ++++++++++++++++++++++-----------------
 1 file changed, 46 insertions(+), 34 deletions(-)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index ec1c8bdf07e6..eea9e707ab4b 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -5717,6 +5717,48 @@ static int ufshcd_poll(struct Scsi_Host *shost, unsigned int queue_num)
 	return completed_reqs != 0;
 }
 
+static bool ufshcd_mcq_force_compl_one(struct request *rq, void *priv)
+{
+	struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(rq);
+	struct scsi_device *sdev = rq->q->queuedata;
+	struct Scsi_Host *shost = sdev->host;
+	struct ufs_hba *hba = shost_priv(shost);
+	struct ufshcd_lrb *lrbp = &hba->lrb[rq->tag];
+	struct ufs_hw_queue *hwq = ufshcd_mcq_req_to_hwq(hba, rq);
+
+	if (!hwq)
+		return true;
+
+	ufshcd_mcq_compl_all_cqes_lock(hba, hwq);
+
+	/*
+	 * For those cmds of which the cqes are not present in the cq, complete
+	 * them explicitly.
+	 */
+	scoped_guard(spinlock_irqsave, &hwq->cq_lock) {
+		if (!test_bit(SCMD_STATE_COMPLETE, &cmd->state)) {
+			set_host_byte(cmd, DID_REQUEUE);
+			ufshcd_release_scsi_cmd(hba, lrbp);
+			scsi_done(cmd);
+		}
+	}
+
+	return true;
+}
+
+static bool ufshcd_mcq_compl_one(struct request *rq, void *priv)
+{
+	struct scsi_device *sdev = rq->q->queuedata;
+	struct Scsi_Host *shost = sdev->host;
+	struct ufs_hba *hba = shost_priv(shost);
+	struct ufs_hw_queue *hwq = ufshcd_mcq_req_to_hwq(hba, rq);
+
+	if (hwq)
+		ufshcd_mcq_poll_cqe_lock(hba, hwq);
+
+	return true;
+}
+
 /**
  * ufshcd_mcq_compl_pending_transfer - MCQ mode function. It is
  * invoked from the error handler context or ufshcd_host_reset_and_restore()
@@ -5731,40 +5773,10 @@ static int ufshcd_poll(struct Scsi_Host *shost, unsigned int queue_num)
 static void ufshcd_mcq_compl_pending_transfer(struct ufs_hba *hba,
 					      bool force_compl)
 {
-	struct ufs_hw_queue *hwq;
-	struct ufshcd_lrb *lrbp;
-	struct scsi_cmnd *cmd;
-	unsigned long flags;
-	int tag;
-
-	for (tag = 0; tag < hba->nutrs; tag++) {
-		lrbp = &hba->lrb[tag];
-		cmd = lrbp->cmd;
-		if (!ufshcd_cmd_inflight(cmd) ||
-		    test_bit(SCMD_STATE_COMPLETE, &cmd->state))
-			continue;
-
-		hwq = ufshcd_mcq_req_to_hwq(hba, scsi_cmd_to_rq(cmd));
-		if (!hwq)
-			continue;
-
-		if (force_compl) {
-			ufshcd_mcq_compl_all_cqes_lock(hba, hwq);
-			/*
-			 * For those cmds of which the cqes are not present
-			 * in the cq, complete them explicitly.
-			 */
-			spin_lock_irqsave(&hwq->cq_lock, flags);
-			if (cmd && !test_bit(SCMD_STATE_COMPLETE, &cmd->state)) {
-				set_host_byte(cmd, DID_REQUEUE);
-				ufshcd_release_scsi_cmd(hba, lrbp);
-				scsi_done(cmd);
-			}
-			spin_unlock_irqrestore(&hwq->cq_lock, flags);
-		} else {
-			ufshcd_mcq_poll_cqe_lock(hba, hwq);
-		}
-	}
+	blk_mq_tagset_busy_iter(&hba->host->tag_set,
+				force_compl ? ufshcd_mcq_force_compl_one :
+					      ufshcd_mcq_compl_one,
+				NULL);
 }
 
 /**

RE: [PATCH v5 15/28] ufs: core: Rework ufshcd_mcq_compl_pending_transfer()

[Avri Altman]

> 
> Replace a tag loop with blk_mq_tagset_busy_iter(). This patch prepares for
> removing the hba->lrb[] array.
> 
> Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Avri Altman <avri.altman@sandisk.com>

> ---
>  drivers/ufs/core/ufshcd.c | 80 ++++++++++++++++++++++-----------------
>  1 file changed, 46 insertions(+), 34 deletions(-)
> 
> diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c index
> ec1c8bdf07e6..eea9e707ab4b 100644
> --- a/drivers/ufs/core/ufshcd.c
> +++ b/drivers/ufs/core/ufshcd.c
> @@ -5717,6 +5717,48 @@ static int ufshcd_poll(struct Scsi_Host *shost,
> unsigned int queue_num)
>         return completed_reqs != 0;
>  }
> 
> +static bool ufshcd_mcq_force_compl_one(struct request *rq, void *priv)
> +{
> +       struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(rq);
> +       struct scsi_device *sdev = rq->q->queuedata;
> +       struct Scsi_Host *shost = sdev->host;
> +       struct ufs_hba *hba = shost_priv(shost);
> +       struct ufshcd_lrb *lrbp = &hba->lrb[rq->tag];
> +       struct ufs_hw_queue *hwq = ufshcd_mcq_req_to_hwq(hba, rq);
> +
> +       if (!hwq)
> +               return true;
> +
> +       ufshcd_mcq_compl_all_cqes_lock(hba, hwq);
> +
> +       /*
> +        * For those cmds of which the cqes are not present in the cq, complete
> +        * them explicitly.
> +        */
> +       scoped_guard(spinlock_irqsave, &hwq->cq_lock) {
> +               if (!test_bit(SCMD_STATE_COMPLETE, &cmd->state)) {
> +                       set_host_byte(cmd, DID_REQUEUE);
> +                       ufshcd_release_scsi_cmd(hba, lrbp);
> +                       scsi_done(cmd);
> +               }
> +       }
> +
> +       return true;
> +}
> +
> +static bool ufshcd_mcq_compl_one(struct request *rq, void *priv) {
> +       struct scsi_device *sdev = rq->q->queuedata;
> +       struct Scsi_Host *shost = sdev->host;
> +       struct ufs_hba *hba = shost_priv(shost);
> +       struct ufs_hw_queue *hwq = ufshcd_mcq_req_to_hwq(hba, rq);
> +
> +       if (hwq)
> +               ufshcd_mcq_poll_cqe_lock(hba, hwq);
> +
> +       return true;
> +}
> +
>  /**
>   * ufshcd_mcq_compl_pending_transfer - MCQ mode function. It is
>   * invoked from the error handler context or ufshcd_host_reset_and_restore()
> @@ -5731,40 +5773,10 @@ static int ufshcd_poll(struct Scsi_Host *shost,
> unsigned int queue_num)  static void
> ufshcd_mcq_compl_pending_transfer(struct ufs_hba *hba,
>                                               bool force_compl)  {
> -       struct ufs_hw_queue *hwq;
> -       struct ufshcd_lrb *lrbp;
> -       struct scsi_cmnd *cmd;
> -       unsigned long flags;
> -       int tag;
> -
> -       for (tag = 0; tag < hba->nutrs; tag++) {
> -               lrbp = &hba->lrb[tag];
> -               cmd = lrbp->cmd;
> -               if (!ufshcd_cmd_inflight(cmd) ||
> -                   test_bit(SCMD_STATE_COMPLETE, &cmd->state))
> -                       continue;
> -
> -               hwq = ufshcd_mcq_req_to_hwq(hba, scsi_cmd_to_rq(cmd));
> -               if (!hwq)
> -                       continue;
> -
> -               if (force_compl) {
> -                       ufshcd_mcq_compl_all_cqes_lock(hba, hwq);
> -                       /*
> -                        * For those cmds of which the cqes are not present
> -                        * in the cq, complete them explicitly.
> -                        */
> -                       spin_lock_irqsave(&hwq->cq_lock, flags);
> -                       if (cmd && !test_bit(SCMD_STATE_COMPLETE, &cmd->state)) {
> -                               set_host_byte(cmd, DID_REQUEUE);
> -                               ufshcd_release_scsi_cmd(hba, lrbp);
> -                               scsi_done(cmd);
> -                       }
> -                       spin_unlock_irqrestore(&hwq->cq_lock, flags);
> -               } else {
> -                       ufshcd_mcq_poll_cqe_lock(hba, hwq);
> -               }
> -       }
> +       blk_mq_tagset_busy_iter(&hba->host->tag_set,
> +                               force_compl ? ufshcd_mcq_force_compl_one :
> +                                             ufshcd_mcq_compl_one,
> +                               NULL);
>  }
> 
>  /**

[PATCH v5 16/28] ufs: core: Rework ufshcd_eh_device_reset_handler()

[Bart Van Assche]

Merge the MCQ mode and legacy mode loops into a single loop. This patch
prepares for optimizing the hot path by removing the direct hba->lrb[]
accesses from ufshcd_eh_device_reset_handler().

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufshcd.c | 84 ++++++++++++++++++---------------------
 1 file changed, 38 insertions(+), 46 deletions(-)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index eea9e707ab4b..67607f2d10a2 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -7538,6 +7538,36 @@ int ufshcd_advanced_rpmb_req_handler(struct ufs_hba *hba, struct utp_upiu_req *r
 	return err ? : result;
 }
 
+static bool ufshcd_clear_lu_cmds(struct request *req, void *priv)
+{
+	struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(req);
+	struct scsi_device *sdev = cmd->device;
+	struct Scsi_Host *shost = sdev->host;
+	struct ufs_hba *hba = shost_priv(shost);
+	const u64 lun = *(u64 *)priv;
+	const u32 tag = req->tag;
+
+	if (sdev->lun != lun)
+		return true;
+
+	if (ufshcd_clear_cmd(hba, tag) < 0) {
+		dev_err(hba->dev, "%s: failed to clear request %d\n", __func__,
+			tag);
+		return true;
+	}
+
+	if (hba->mcq_enabled) {
+		struct ufs_hw_queue *hwq = ufshcd_mcq_req_to_hwq(hba, req);
+
+		if (hwq)
+			ufshcd_mcq_poll_cqe_lock(hba, hwq);
+		return true;
+	}
+
+	ufshcd_compl_one_cqe(hba, tag, NULL);
+	return true;
+}
+
 /**
  * ufshcd_eh_device_reset_handler() - Reset a single logical unit.
  * @cmd: SCSI command pointer
@@ -7546,12 +7576,8 @@ int ufshcd_advanced_rpmb_req_handler(struct ufs_hba *hba, struct utp_upiu_req *r
  */
 static int ufshcd_eh_device_reset_handler(struct scsi_cmnd *cmd)
 {
-	unsigned long flags, pending_reqs = 0, not_cleared = 0;
 	struct Scsi_Host *host;
 	struct ufs_hba *hba;
-	struct ufs_hw_queue *hwq;
-	struct ufshcd_lrb *lrbp;
-	u32 pos, not_cleared_mask = 0;
 	int err;
 	u8 resp = 0xF, lun;
 
@@ -7560,50 +7586,16 @@ static int ufshcd_eh_device_reset_handler(struct scsi_cmnd *cmd)
 
 	lun = ufshcd_scsi_to_upiu_lun(cmd->device->lun);
 	err = ufshcd_issue_tm_cmd(hba, lun, 0, UFS_LOGICAL_RESET, &resp);
-	if (err || resp != UPIU_TASK_MANAGEMENT_FUNC_COMPL) {
-		if (!err)
-			err = resp;
-		goto out;
-	}
-
-	if (hba->mcq_enabled) {
-		for (pos = 0; pos < hba->nutrs; pos++) {
-			lrbp = &hba->lrb[pos];
-			if (ufshcd_cmd_inflight(lrbp->cmd) &&
-			    lrbp->lun == lun) {
-				ufshcd_clear_cmd(hba, pos);
-				hwq = ufshcd_mcq_req_to_hwq(hba, scsi_cmd_to_rq(lrbp->cmd));
-				ufshcd_mcq_poll_cqe_lock(hba, hwq);
-			}
-		}
-		err = 0;
-		goto out;
-	}
-
-	/* clear the commands that were pending for corresponding LUN */
-	spin_lock_irqsave(&hba->outstanding_lock, flags);
-	for_each_set_bit(pos, &hba->outstanding_reqs, hba->nutrs)
-		if (hba->lrb[pos].lun == lun)
-			__set_bit(pos, &pending_reqs);
-	hba->outstanding_reqs &= ~pending_reqs;
-	spin_unlock_irqrestore(&hba->outstanding_lock, flags);
-
-	for_each_set_bit(pos, &pending_reqs, hba->nutrs) {
-		if (ufshcd_clear_cmd(hba, pos) < 0) {
-			spin_lock_irqsave(&hba->outstanding_lock, flags);
-			not_cleared = 1U << pos &
-				ufshcd_readl(hba, REG_UTP_TRANSFER_REQ_DOOR_BELL);
-			hba->outstanding_reqs |= not_cleared;
-			not_cleared_mask |= not_cleared;
-			spin_unlock_irqrestore(&hba->outstanding_lock, flags);
-
-			dev_err(hba->dev, "%s: failed to clear request %d\n",
-				__func__, pos);
-		}
+	if (err) {
+	} else if (resp != UPIU_TASK_MANAGEMENT_FUNC_COMPL) {
+		err = resp;
+	} else {
+		/* clear the commands that were pending for corresponding LUN */
+		blk_mq_tagset_busy_iter(&hba->host->tag_set,
+					ufshcd_clear_lu_cmds,
+					&cmd->device->lun);
 	}
-	__ufshcd_transfer_req_compl(hba, pending_reqs & ~not_cleared_mask);
 
-out:
 	hba->req_abort_count = 0;
 	ufshcd_update_evt_hist(hba, UFS_EVT_DEV_RESET, (u32)err);
 	if (!err) {

RE: [PATCH v5 16/28] ufs: core: Rework ufshcd_eh_device_reset_handler()

[Avri Altman]

> Merge the MCQ mode and legacy mode loops into a single loop. This patch
> prepares for optimizing the hot path by removing the direct hba->lrb[]
> accesses from ufshcd_eh_device_reset_handler().
> 
> Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Avri Altman <avri.altman@sandisk.com>

> ---
>  drivers/ufs/core/ufshcd.c | 84 ++++++++++++++++++---------------------
>  1 file changed, 38 insertions(+), 46 deletions(-)
> 
> diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c index
> eea9e707ab4b..67607f2d10a2 100644
> --- a/drivers/ufs/core/ufshcd.c
> +++ b/drivers/ufs/core/ufshcd.c
> @@ -7538,6 +7538,36 @@ int ufshcd_advanced_rpmb_req_handler(struct
> ufs_hba *hba, struct utp_upiu_req *r
>         return err ? : result;
>  }
> 
> +static bool ufshcd_clear_lu_cmds(struct request *req, void *priv) {
> +       struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(req);
> +       struct scsi_device *sdev = cmd->device;
> +       struct Scsi_Host *shost = sdev->host;
> +       struct ufs_hba *hba = shost_priv(shost);
> +       const u64 lun = *(u64 *)priv;
> +       const u32 tag = req->tag;
> +
> +       if (sdev->lun != lun)
> +               return true;
> +
> +       if (ufshcd_clear_cmd(hba, tag) < 0) {
> +               dev_err(hba->dev, "%s: failed to clear request %d\n", __func__,
> +                       tag);
> +               return true;
> +       }
> +
> +       if (hba->mcq_enabled) {
> +               struct ufs_hw_queue *hwq = ufshcd_mcq_req_to_hwq(hba,
> + req);
> +
> +               if (hwq)
> +                       ufshcd_mcq_poll_cqe_lock(hba, hwq);
> +               return true;
> +       }
> +
> +       ufshcd_compl_one_cqe(hba, tag, NULL);
> +       return true;
> +}
> +
>  /**
>   * ufshcd_eh_device_reset_handler() - Reset a single logical unit.
>   * @cmd: SCSI command pointer
> @@ -7546,12 +7576,8 @@ int ufshcd_advanced_rpmb_req_handler(struct
> ufs_hba *hba, struct utp_upiu_req *r
>   */
>  static int ufshcd_eh_device_reset_handler(struct scsi_cmnd *cmd)  {
> -       unsigned long flags, pending_reqs = 0, not_cleared = 0;
>         struct Scsi_Host *host;
>         struct ufs_hba *hba;
> -       struct ufs_hw_queue *hwq;
> -       struct ufshcd_lrb *lrbp;
> -       u32 pos, not_cleared_mask = 0;
>         int err;
>         u8 resp = 0xF, lun;
> 
> @@ -7560,50 +7586,16 @@ static int ufshcd_eh_device_reset_handler(struct
> scsi_cmnd *cmd)
> 
>         lun = ufshcd_scsi_to_upiu_lun(cmd->device->lun);
>         err = ufshcd_issue_tm_cmd(hba, lun, 0, UFS_LOGICAL_RESET, &resp);
> -       if (err || resp != UPIU_TASK_MANAGEMENT_FUNC_COMPL) {
> -               if (!err)
> -                       err = resp;
> -               goto out;
> -       }
> -
> -       if (hba->mcq_enabled) {
> -               for (pos = 0; pos < hba->nutrs; pos++) {
> -                       lrbp = &hba->lrb[pos];
> -                       if (ufshcd_cmd_inflight(lrbp->cmd) &&
> -                           lrbp->lun == lun) {
> -                               ufshcd_clear_cmd(hba, pos);
> -                               hwq = ufshcd_mcq_req_to_hwq(hba, scsi_cmd_to_rq(lrbp-
> >cmd));
> -                               ufshcd_mcq_poll_cqe_lock(hba, hwq);
> -                       }
> -               }
> -               err = 0;
> -               goto out;
> -       }
> -
> -       /* clear the commands that were pending for corresponding LUN */
> -       spin_lock_irqsave(&hba->outstanding_lock, flags);
> -       for_each_set_bit(pos, &hba->outstanding_reqs, hba->nutrs)
> -               if (hba->lrb[pos].lun == lun)
> -                       __set_bit(pos, &pending_reqs);
> -       hba->outstanding_reqs &= ~pending_reqs;
> -       spin_unlock_irqrestore(&hba->outstanding_lock, flags);
> -
> -       for_each_set_bit(pos, &pending_reqs, hba->nutrs) {
> -               if (ufshcd_clear_cmd(hba, pos) < 0) {
> -                       spin_lock_irqsave(&hba->outstanding_lock, flags);
> -                       not_cleared = 1U << pos &
> -                               ufshcd_readl(hba, REG_UTP_TRANSFER_REQ_DOOR_BELL);
> -                       hba->outstanding_reqs |= not_cleared;
> -                       not_cleared_mask |= not_cleared;
> -                       spin_unlock_irqrestore(&hba->outstanding_lock, flags);
> -
> -                       dev_err(hba->dev, "%s: failed to clear request %d\n",
> -                               __func__, pos);
> -               }
> +       if (err) {
> +       } else if (resp != UPIU_TASK_MANAGEMENT_FUNC_COMPL) {
> +               err = resp;
> +       } else {
> +               /* clear the commands that were pending for corresponding LUN */
> +               blk_mq_tagset_busy_iter(&hba->host->tag_set,
> +                                       ufshcd_clear_lu_cmds,
> +                                       &cmd->device->lun);
>         }
> -       __ufshcd_transfer_req_compl(hba, pending_reqs & ~not_cleared_mask);
> 
> -out:
>         hba->req_abort_count = 0;
>         ufshcd_update_evt_hist(hba, UFS_EVT_DEV_RESET, (u32)err);
>         if (!err) {

[PATCH v5 17/28] ufs: core: Rework the SCSI host queue depth calculation code

[Bart Van Assche]

Prepare for allocating the SCSI host earlier by making the SCSI host
queue depth independent of the queue depth supported by the UFS device.
This patch may increase the queue depth of the UFS SCSI host.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufs-mcq.c     | 20 +++++---------------
 drivers/ufs/core/ufshcd-priv.h |  2 +-
 drivers/ufs/core/ufshcd.c      | 17 +++++++++++++++--
 3 files changed, 21 insertions(+), 18 deletions(-)

diff --git a/drivers/ufs/core/ufs-mcq.c b/drivers/ufs/core/ufs-mcq.c
index 4eb4f1af7800..b99e482ad8da 100644
--- a/drivers/ufs/core/ufs-mcq.c
+++ b/drivers/ufs/core/ufs-mcq.c
@@ -130,17 +130,15 @@ unsigned int ufshcd_mcq_queue_cfg_addr(struct ufs_hba *hba)
 EXPORT_SYMBOL_GPL(ufshcd_mcq_queue_cfg_addr);
 
 /**
- * ufshcd_mcq_decide_queue_depth - decide the queue depth
+ * ufshcd_get_hba_mac - Maximum number of commands supported by the host
+ *	controller.
  * @hba: per adapter instance
  *
- * Return: queue-depth on success, non-zero on error
+ * Return: queue depth on success; negative upon error.
  *
- * MAC - Max. Active Command of the Host Controller (HC)
- * HC wouldn't send more than this commands to the device.
- * Calculates and adjusts the queue depth based on the depth
- * supported by the HC and ufs device.
+ * MAC = Maximum number of Active Commands supported by the Host Controller.
  */
-int ufshcd_mcq_decide_queue_depth(struct ufs_hba *hba)
+int ufshcd_get_hba_mac(struct ufs_hba *hba)
 {
 	int mac;
 
@@ -160,14 +158,6 @@ int ufshcd_mcq_decide_queue_depth(struct ufs_hba *hba)
 	if (mac < 0)
 		goto err;
 
-	WARN_ON_ONCE(!hba->dev_info.bqueuedepth);
-	/*
-	 * max. value of bqueuedepth = 256, mac is host dependent.
-	 * It is mandatory for UFS device to define bQueueDepth if
-	 * shared queuing architecture is enabled.
-	 */
-	return min_t(int, mac, hba->dev_info.bqueuedepth);
-
 err:
 	dev_err(hba->dev, "Failed to get mac, err=%d\n", mac);
 	return mac;
diff --git a/drivers/ufs/core/ufshcd-priv.h b/drivers/ufs/core/ufshcd-priv.h
index d0a2c963a27d..ec1bb818bc73 100644
--- a/drivers/ufs/core/ufshcd-priv.h
+++ b/drivers/ufs/core/ufshcd-priv.h
@@ -65,7 +65,7 @@ void ufshcd_compl_one_cqe(struct ufs_hba *hba, int task_tag,
 			  struct cq_entry *cqe);
 int ufshcd_mcq_init(struct ufs_hba *hba);
 void ufshcd_mcq_disable(struct ufs_hba *hba);
-int ufshcd_mcq_decide_queue_depth(struct ufs_hba *hba);
+int ufshcd_get_hba_mac(struct ufs_hba *hba);
 int ufshcd_mcq_memory_alloc(struct ufs_hba *hba);
 struct ufs_hw_queue *ufshcd_mcq_req_to_hwq(struct ufs_hba *hba,
 					   struct request *req);
diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index 67607f2d10a2..419f05efd566 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -8438,10 +8438,11 @@ static void ufs_init_rtc(struct ufs_hba *hba, u8 *desc_buf)
 
 static int ufs_get_device_desc(struct ufs_hba *hba)
 {
+	struct ufs_dev_info *dev_info = &hba->dev_info;
+	struct Scsi_Host *shost = hba->host;
 	int err;
 	u8 model_index;
 	u8 *desc_buf;
-	struct ufs_dev_info *dev_info = &hba->dev_info;
 
 	desc_buf = kzalloc(QUERY_DESC_MAX_SIZE, GFP_KERNEL);
 	if (!desc_buf) {
@@ -8469,6 +8470,18 @@ static int ufs_get_device_desc(struct ufs_hba *hba)
 				      desc_buf[DEVICE_DESC_PARAM_SPEC_VER + 1];
 	dev_info->bqueuedepth = desc_buf[DEVICE_DESC_PARAM_Q_DPTH];
 
+	/*
+	 * According to the UFS standard, the UFS device queue depth
+	 * (bQueueDepth) must be in the range 1..255 if the shared queueing
+	 * architecture is supported. bQueueDepth is zero if the shared queueing
+	 * architecture is not supported.
+	 */
+	if (dev_info->bqueuedepth)
+		shost->cmd_per_lun = min(hba->nutrs, dev_info->bqueuedepth) -
+				     UFSHCD_NUM_RESERVED;
+	else
+		shost->cmd_per_lun = shost->can_queue;
+
 	dev_info->rtt_cap = desc_buf[DEVICE_DESC_PARAM_RTT_CAP];
 
 	dev_info->hid_sup = get_unaligned_be32(desc_buf +
@@ -8875,7 +8888,7 @@ static int ufshcd_alloc_mcq(struct ufs_hba *hba)
 	int ret;
 	int old_nutrs = hba->nutrs;
 
-	ret = ufshcd_mcq_decide_queue_depth(hba);
+	ret = ufshcd_get_hba_mac(hba);
 	if (ret < 0)
 		return ret;
 

[PATCH v5 18/28] ufs: core: Allocate the SCSI host earlier

[Bart Van Assche]

Call ufshcd_add_scsi_host() before any UPIU commands are sent to the UFS
device. This patch prepares for letting ufshcd_add_scsi_host() allocate
memory for both SCSI and UPIU commands.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufshcd.c | 17 ++++++++++++-----
 1 file changed, 12 insertions(+), 5 deletions(-)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index 419f05efd566..7ea18204b5fb 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -10581,6 +10581,9 @@ static int ufshcd_add_scsi_host(struct ufs_hba *hba)
 {
 	int err;
 
+	WARN_ON_ONCE(!hba->host->can_queue);
+	WARN_ON_ONCE(!hba->host->cmd_per_lun);
+
 	if (is_mcq_supported(hba)) {
 		ufshcd_mcq_enable(hba);
 		err = ufshcd_alloc_mcq(hba);
@@ -10731,7 +10734,11 @@ int ufshcd_init(struct ufs_hba *hba, void __iomem *mmio_base, unsigned int irq)
 	ufshcd_host_memory_configure(hba);
 
 	host->can_queue = hba->nutrs - UFSHCD_NUM_RESERVED;
-	host->cmd_per_lun = hba->nutrs - UFSHCD_NUM_RESERVED;
+	/*
+	 * Set the queue depth for WLUNs. ufs_get_device_desc() will increase
+	 * host->cmd_per_lun to a larger value.
+	 */
+	host->cmd_per_lun = 1;
 	host->max_id = UFSHCD_MAX_ID;
 	host->max_lun = UFS_MAX_LUNS;
 	host->max_channel = UFSHCD_MAX_CHANNEL;
@@ -10822,6 +10829,10 @@ int ufshcd_init(struct ufs_hba *hba, void __iomem *mmio_base, unsigned int irq)
 			    FIELD_PREP(UFSHCI_AHIBERN8_SCALE_MASK, 3);
 	}
 
+	err = ufshcd_add_scsi_host(hba);
+	if (err)
+		goto out_disable;
+
 	/* Hold auto suspend until async scan completes */
 	pm_runtime_get_sync(dev);
 
@@ -10872,10 +10883,6 @@ int ufshcd_init(struct ufs_hba *hba, void __iomem *mmio_base, unsigned int irq)
 	if (err)
 		goto out_disable;
 
-	err = ufshcd_add_scsi_host(hba);
-	if (err)
-		goto out_disable;
-
 	async_schedule(ufshcd_async_scan, hba);
 	ufs_sysfs_add_nodes(hba->dev);
 

[PATCH v5 19/28] ufs: core: Call ufshcd_init_lrb() later

[Bart Van Assche]

Call ufshcd_init_lrb() from inside ufshcd_setup_dev_cmd() instead of
ufshcd_host_memory_configure(). This patch prepares for calling
ufshcd_host_memory_configure() before the information is available
that is required to call ufshcd_setup_dev_cmd().

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufshcd.c | 53 ++++++++++++++++++++-------------------
 1 file changed, 27 insertions(+), 26 deletions(-)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index 7ea18204b5fb..bb14af688e98 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -2903,8 +2903,32 @@ static void ufshcd_comp_scsi_upiu(struct ufs_hba *hba, struct ufshcd_lrb *lrbp)
 	ufshcd_prepare_utp_scsi_cmd_upiu(lrbp, upiu_flags);
 }
 
-static void __ufshcd_setup_cmd(struct ufshcd_lrb *lrbp, struct scsi_cmnd *cmd, u8 lun, int tag)
+static void ufshcd_init_lrb(struct ufs_hba *hba, struct ufshcd_lrb *lrb, int i)
+{
+	struct utp_transfer_cmd_desc *cmd_descp =
+		(void *)hba->ucdl_base_addr + i * ufshcd_get_ucd_size(hba);
+	struct utp_transfer_req_desc *utrdlp = hba->utrdl_base_addr;
+	dma_addr_t cmd_desc_element_addr =
+		hba->ucdl_dma_addr + i * ufshcd_get_ucd_size(hba);
+	u16 response_offset = le16_to_cpu(utrdlp[i].response_upiu_offset);
+	u16 prdt_offset = le16_to_cpu(utrdlp[i].prd_table_offset);
+
+	lrb->utr_descriptor_ptr = utrdlp + i;
+	lrb->utrd_dma_addr =
+		hba->utrdl_dma_addr + i * sizeof(struct utp_transfer_req_desc);
+	lrb->ucd_req_ptr = (struct utp_upiu_req *)cmd_descp->command_upiu;
+	lrb->ucd_req_dma_addr = cmd_desc_element_addr;
+	lrb->ucd_rsp_ptr = (struct utp_upiu_rsp *)cmd_descp->response_upiu;
+	lrb->ucd_rsp_dma_addr = cmd_desc_element_addr + response_offset;
+	lrb->ucd_prdt_ptr = (struct ufshcd_sg_entry *)cmd_descp->prd_table;
+	lrb->ucd_prdt_dma_addr = cmd_desc_element_addr + prdt_offset;
+}
+
+static void __ufshcd_setup_cmd(struct ufs_hba *hba, struct ufshcd_lrb *lrbp,
+			       struct scsi_cmnd *cmd, u8 lun, int tag)
 {
+	ufshcd_init_lrb(hba, lrbp, tag);
+
 	memset(lrbp->ucd_req_ptr, 0, sizeof(*lrbp->ucd_req_ptr));
 
 	lrbp->cmd = cmd;
@@ -2916,7 +2940,7 @@ static void __ufshcd_setup_cmd(struct ufshcd_lrb *lrbp, struct scsi_cmnd *cmd, u
 static void ufshcd_setup_scsi_cmd(struct ufs_hba *hba, struct ufshcd_lrb *lrbp,
 				  struct scsi_cmnd *cmd, u8 lun, int tag)
 {
-	__ufshcd_setup_cmd(lrbp, cmd, lun, tag);
+	__ufshcd_setup_cmd(hba, lrbp, cmd, lun, tag);
 	lrbp->intr_cmd = !ufshcd_is_intr_aggr_allowed(hba);
 	lrbp->req_abort_skip = false;
 
@@ -2971,27 +2995,6 @@ static void ufshcd_map_queues(struct Scsi_Host *shost)
 	}
 }
 
-static void ufshcd_init_lrb(struct ufs_hba *hba, struct ufshcd_lrb *lrb, int i)
-{
-	struct utp_transfer_cmd_desc *cmd_descp = (void *)hba->ucdl_base_addr +
-		i * ufshcd_get_ucd_size(hba);
-	struct utp_transfer_req_desc *utrdlp = hba->utrdl_base_addr;
-	dma_addr_t cmd_desc_element_addr = hba->ucdl_dma_addr +
-		i * ufshcd_get_ucd_size(hba);
-	u16 response_offset = le16_to_cpu(utrdlp[i].response_upiu_offset);
-	u16 prdt_offset = le16_to_cpu(utrdlp[i].prd_table_offset);
-
-	lrb->utr_descriptor_ptr = utrdlp + i;
-	lrb->utrd_dma_addr = hba->utrdl_dma_addr +
-		i * sizeof(struct utp_transfer_req_desc);
-	lrb->ucd_req_ptr = (struct utp_upiu_req *)cmd_descp->command_upiu;
-	lrb->ucd_req_dma_addr = cmd_desc_element_addr;
-	lrb->ucd_rsp_ptr = (struct utp_upiu_rsp *)cmd_descp->response_upiu;
-	lrb->ucd_rsp_dma_addr = cmd_desc_element_addr + response_offset;
-	lrb->ucd_prdt_ptr = (struct ufshcd_sg_entry *)cmd_descp->prd_table;
-	lrb->ucd_prdt_dma_addr = cmd_desc_element_addr + prdt_offset;
-}
-
 /**
  * ufshcd_queuecommand - main entry point for SCSI requests
  * @host: SCSI host pointer
@@ -3084,7 +3087,7 @@ static int ufshcd_queuecommand(struct Scsi_Host *host, struct scsi_cmnd *cmd)
 static void ufshcd_setup_dev_cmd(struct ufs_hba *hba, struct ufshcd_lrb *lrbp,
 			     enum dev_cmd_type cmd_type, u8 lun, int tag)
 {
-	__ufshcd_setup_cmd(lrbp, NULL, lun, tag);
+	__ufshcd_setup_cmd(hba, lrbp, NULL, lun, tag);
 	lrbp->intr_cmd = true; /* No interrupt aggregation */
 	hba->dev_cmd.type = cmd_type;
 }
@@ -4042,8 +4045,6 @@ static void ufshcd_host_memory_configure(struct ufs_hba *hba)
 			utrdlp[i].response_upiu_length =
 				cpu_to_le16(ALIGNED_UPIU_SIZE >> 2);
 		}
-
-		ufshcd_init_lrb(hba, &hba->lrb[i], i);
 	}
 }
 

RE: [PATCH v5 19/28] ufs: core: Call ufshcd_init_lrb() later

[Avri Altman]

> 
> Call ufshcd_init_lrb() from inside ufshcd_setup_dev_cmd() instead of
> ufshcd_host_memory_configure(). This patch prepares for calling
> ufshcd_host_memory_configure() before the information is available that is
> required to call ufshcd_setup_dev_cmd().
> 
> Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Avri Altman <avri.altman@sandisk.com>

> ---
>  drivers/ufs/core/ufshcd.c | 53 ++++++++++++++++++++-------------------
>  1 file changed, 27 insertions(+), 26 deletions(-)
> 
> diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c index
> 7ea18204b5fb..bb14af688e98 100644
> --- a/drivers/ufs/core/ufshcd.c
> +++ b/drivers/ufs/core/ufshcd.c
> @@ -2903,8 +2903,32 @@ static void ufshcd_comp_scsi_upiu(struct ufs_hba
> *hba, struct ufshcd_lrb *lrbp)
>         ufshcd_prepare_utp_scsi_cmd_upiu(lrbp, upiu_flags);  }
> 
> -static void __ufshcd_setup_cmd(struct ufshcd_lrb *lrbp, struct scsi_cmnd
> *cmd, u8 lun, int tag)
> +static void ufshcd_init_lrb(struct ufs_hba *hba, struct ufshcd_lrb
> +*lrb, int i) {
> +       struct utp_transfer_cmd_desc *cmd_descp =
> +               (void *)hba->ucdl_base_addr + i * ufshcd_get_ucd_size(hba);
> +       struct utp_transfer_req_desc *utrdlp = hba->utrdl_base_addr;
> +       dma_addr_t cmd_desc_element_addr =
> +               hba->ucdl_dma_addr + i * ufshcd_get_ucd_size(hba);
> +       u16 response_offset = le16_to_cpu(utrdlp[i].response_upiu_offset);
> +       u16 prdt_offset = le16_to_cpu(utrdlp[i].prd_table_offset);
> +
> +       lrb->utr_descriptor_ptr = utrdlp + i;
> +       lrb->utrd_dma_addr =
> +               hba->utrdl_dma_addr + i * sizeof(struct utp_transfer_req_desc);
> +       lrb->ucd_req_ptr = (struct utp_upiu_req *)cmd_descp->command_upiu;
> +       lrb->ucd_req_dma_addr = cmd_desc_element_addr;
> +       lrb->ucd_rsp_ptr = (struct utp_upiu_rsp *)cmd_descp->response_upiu;
> +       lrb->ucd_rsp_dma_addr = cmd_desc_element_addr + response_offset;
> +       lrb->ucd_prdt_ptr = (struct ufshcd_sg_entry *)cmd_descp->prd_table;
> +       lrb->ucd_prdt_dma_addr = cmd_desc_element_addr + prdt_offset; }
> +
> +static void __ufshcd_setup_cmd(struct ufs_hba *hba, struct ufshcd_lrb
> *lrbp,
> +                              struct scsi_cmnd *cmd, u8 lun, int tag)
>  {
> +       ufshcd_init_lrb(hba, lrbp, tag);
> +
>         memset(lrbp->ucd_req_ptr, 0, sizeof(*lrbp->ucd_req_ptr));
> 
>         lrbp->cmd = cmd;
> @@ -2916,7 +2940,7 @@ static void __ufshcd_setup_cmd(struct ufshcd_lrb
> *lrbp, struct scsi_cmnd *cmd, u  static void ufshcd_setup_scsi_cmd(struct
> ufs_hba *hba, struct ufshcd_lrb *lrbp,
>                                   struct scsi_cmnd *cmd, u8 lun, int tag)  {
> -       __ufshcd_setup_cmd(lrbp, cmd, lun, tag);
> +       __ufshcd_setup_cmd(hba, lrbp, cmd, lun, tag);
>         lrbp->intr_cmd = !ufshcd_is_intr_aggr_allowed(hba);
>         lrbp->req_abort_skip = false;
> 
> @@ -2971,27 +2995,6 @@ static void ufshcd_map_queues(struct Scsi_Host
> *shost)
>         }
>  }
> 
> -static void ufshcd_init_lrb(struct ufs_hba *hba, struct ufshcd_lrb *lrb, int i) -{
> -       struct utp_transfer_cmd_desc *cmd_descp = (void *)hba-
> >ucdl_base_addr +
> -               i * ufshcd_get_ucd_size(hba);
> -       struct utp_transfer_req_desc *utrdlp = hba->utrdl_base_addr;
> -       dma_addr_t cmd_desc_element_addr = hba->ucdl_dma_addr +
> -               i * ufshcd_get_ucd_size(hba);
> -       u16 response_offset = le16_to_cpu(utrdlp[i].response_upiu_offset);
> -       u16 prdt_offset = le16_to_cpu(utrdlp[i].prd_table_offset);
> -
> -       lrb->utr_descriptor_ptr = utrdlp + i;
> -       lrb->utrd_dma_addr = hba->utrdl_dma_addr +
> -               i * sizeof(struct utp_transfer_req_desc);
> -       lrb->ucd_req_ptr = (struct utp_upiu_req *)cmd_descp->command_upiu;
> -       lrb->ucd_req_dma_addr = cmd_desc_element_addr;
> -       lrb->ucd_rsp_ptr = (struct utp_upiu_rsp *)cmd_descp->response_upiu;
> -       lrb->ucd_rsp_dma_addr = cmd_desc_element_addr + response_offset;
> -       lrb->ucd_prdt_ptr = (struct ufshcd_sg_entry *)cmd_descp->prd_table;
> -       lrb->ucd_prdt_dma_addr = cmd_desc_element_addr + prdt_offset;
> -}
> -
>  /**
>   * ufshcd_queuecommand - main entry point for SCSI requests
>   * @host: SCSI host pointer
> @@ -3084,7 +3087,7 @@ static int ufshcd_queuecommand(struct Scsi_Host
> *host, struct scsi_cmnd *cmd)  static void ufshcd_setup_dev_cmd(struct
> ufs_hba *hba, struct ufshcd_lrb *lrbp,
>                              enum dev_cmd_type cmd_type, u8 lun, int tag)  {
> -       __ufshcd_setup_cmd(lrbp, NULL, lun, tag);
> +       __ufshcd_setup_cmd(hba, lrbp, NULL, lun, tag);
>         lrbp->intr_cmd = true; /* No interrupt aggregation */
>         hba->dev_cmd.type = cmd_type;
>  }
> @@ -4042,8 +4045,6 @@ static void ufshcd_host_memory_configure(struct
> ufs_hba *hba)
>                         utrdlp[i].response_upiu_length =
>                                 cpu_to_le16(ALIGNED_UPIU_SIZE >> 2);
>                 }
> -
> -               ufshcd_init_lrb(hba, &hba->lrb[i], i);
>         }
>  }

[PATCH v5 20/28] ufs: core: Use hba->reserved_slot

[Bart Van Assche]

Use hba->reserved_slot instead of open-coding it. This patch prepares
for changing the value of hba->reserved_slot.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufs-mcq.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/ufs/core/ufs-mcq.c b/drivers/ufs/core/ufs-mcq.c
index b99e482ad8da..9b6674d6f21c 100644
--- a/drivers/ufs/core/ufs-mcq.c
+++ b/drivers/ufs/core/ufs-mcq.c
@@ -536,7 +536,7 @@ int ufshcd_mcq_sq_cleanup(struct ufs_hba *hba, int task_tag)
 	if (hba->quirks & UFSHCD_QUIRK_MCQ_BROKEN_RTC)
 		return -ETIMEDOUT;
 
-	if (task_tag != hba->nutrs - UFSHCD_NUM_RESERVED) {
+	if (task_tag != hba->reserved_slot) {
 		if (!cmd)
 			return -EINVAL;
 		hwq = ufshcd_mcq_req_to_hwq(hba, scsi_cmd_to_rq(cmd));

[PATCH v5 21/28] ufs: core: Make the reserved slot a reserved request

[Bart Van Assche]

Instead of letting the SCSI core allocate hba->nutrs - 1 commands, let
the SCSI core allocate hba->nutrs commands, set the number of reserved
tags to 1 and use the reserved tag for device management commands. This
patch changes the 'reserved slot' from hba->nutrs - 1 into 0 because
the block layer reserves the smallest tags for reserved commands.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufshcd.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index bb14af688e98..1dbb5bdf84f2 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -2475,7 +2475,7 @@ static inline int ufshcd_hba_capabilities(struct ufs_hba *hba)
 	hba->nutrs = (hba->capabilities & MASK_TRANSFER_REQUESTS_SLOTS_SDB) + 1;
 	hba->nutmrs =
 	((hba->capabilities & MASK_TASK_MANAGEMENT_REQUEST_SLOTS) >> 16) + 1;
-	hba->reserved_slot = hba->nutrs - 1;
+	hba->reserved_slot = 0;
 
 	hba->nortt = FIELD_GET(MASK_NUMBER_OUTSTANDING_RTT, hba->capabilities) + 1;
 
@@ -8915,7 +8915,6 @@ static int ufshcd_alloc_mcq(struct ufs_hba *hba)
 		goto err;
 
 	hba->host->can_queue = hba->nutrs - UFSHCD_NUM_RESERVED;
-	hba->reserved_slot = hba->nutrs - UFSHCD_NUM_RESERVED;
 
 	return 0;
 err:
@@ -10740,6 +10739,7 @@ int ufshcd_init(struct ufs_hba *hba, void __iomem *mmio_base, unsigned int irq)
 	 * host->cmd_per_lun to a larger value.
 	 */
 	host->cmd_per_lun = 1;
+	host->nr_reserved_cmds = UFSHCD_NUM_RESERVED;
 	host->max_id = UFSHCD_MAX_ID;
 	host->max_lun = UFS_MAX_LUNS;
 	host->max_channel = UFSHCD_MAX_CHANNEL;

[PATCH v5 22/28] ufs: core: Do not clear driver-private command data

[Bart Van Assche]

Tell the SCSI core to skip the memset() call that clears driver-private
data because __ufshcd_setup_cmd() performs all necessary initialization.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufshcd.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index 1dbb5bdf84f2..a54359289db5 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -2995,6 +2995,15 @@ static void ufshcd_map_queues(struct Scsi_Host *shost)
 	}
 }
 
+/*
+ * The only purpose of this function is to make the SCSI core skip the memset()
+ * call for the private command data.
+ */
+static int ufshcd_init_cmd_priv(struct Scsi_Host *host, struct scsi_cmnd *cmd)
+{
+	return 0;
+}
+
 /**
  * ufshcd_queuecommand - main entry point for SCSI requests
  * @host: SCSI host pointer
@@ -9157,6 +9166,7 @@ static const struct scsi_host_template ufshcd_driver_template = {
 	.name			= UFSHCD,
 	.proc_name		= UFSHCD,
 	.map_queues		= ufshcd_map_queues,
+	.init_cmd_priv		= ufshcd_init_cmd_priv,
 	.queuecommand		= ufshcd_queuecommand,
 	.mq_poll		= ufshcd_poll,
 	.sdev_init		= ufshcd_sdev_init,

[PATCH v5 23/28] ufs: core: Optimize the hot path

[Bart Van Assche]

Set .cmd_size in the SCSI host template such that the SCSI core makes
struct scsi_cmnd and struct ufshcd_lrb adjacent. Convert the cmd->lrbp
and lrbp->cmd memory loads into pointer offset calculations. Remove the
data structure members that became superfluous, namely ufshcd_lrb.cmd
and ufs_hba.lrb.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufs-mcq.c       |   9 +-
 drivers/ufs/core/ufshcd-crypto.h |  18 ++-
 drivers/ufs/core/ufshcd-priv.h   |  41 +++++-
 drivers/ufs/core/ufshcd.c        | 232 ++++++++++++++++---------------
 include/ufs/ufshcd.h             |   5 -
 5 files changed, 177 insertions(+), 128 deletions(-)

diff --git a/drivers/ufs/core/ufs-mcq.c b/drivers/ufs/core/ufs-mcq.c
index 9b6674d6f21c..c6c6cca400de 100644
--- a/drivers/ufs/core/ufs-mcq.c
+++ b/drivers/ufs/core/ufs-mcq.c
@@ -526,8 +526,8 @@ static int ufshcd_mcq_sq_start(struct ufs_hba *hba, struct ufs_hw_queue *hwq)
  */
 int ufshcd_mcq_sq_cleanup(struct ufs_hba *hba, int task_tag)
 {
-	struct ufshcd_lrb *lrbp = &hba->lrb[task_tag];
-	struct scsi_cmnd *cmd = lrbp->cmd;
+	struct scsi_cmnd *cmd = ufshcd_tag_to_cmd(hba, task_tag);
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	struct ufs_hw_queue *hwq;
 	void __iomem *reg, *opr_sqd_base;
 	u32 nexus, id, val;
@@ -612,7 +612,8 @@ static void ufshcd_mcq_nullify_sqe(struct utp_transfer_req_desc *utrd)
 static bool ufshcd_mcq_sqe_search(struct ufs_hba *hba,
 				  struct ufs_hw_queue *hwq, int task_tag)
 {
-	struct ufshcd_lrb *lrbp = &hba->lrb[task_tag];
+	struct scsi_cmnd *cmd = ufshcd_tag_to_cmd(hba, task_tag);
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	struct utp_transfer_req_desc *utrd;
 	__le64  cmd_desc_base_addr;
 	bool ret = false;
@@ -663,7 +664,7 @@ int ufshcd_mcq_abort(struct scsi_cmnd *cmd)
 	struct Scsi_Host *host = cmd->device->host;
 	struct ufs_hba *hba = shost_priv(host);
 	int tag = scsi_cmd_to_rq(cmd)->tag;
-	struct ufshcd_lrb *lrbp = &hba->lrb[tag];
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	struct ufs_hw_queue *hwq;
 	int err;
 
diff --git a/drivers/ufs/core/ufshcd-crypto.h b/drivers/ufs/core/ufshcd-crypto.h
index 89bb97c14c15..c148a5194378 100644
--- a/drivers/ufs/core/ufshcd-crypto.h
+++ b/drivers/ufs/core/ufshcd-crypto.h
@@ -38,10 +38,10 @@ ufshcd_prepare_req_desc_hdr_crypto(struct ufshcd_lrb *lrbp,
 }
 
 static inline int ufshcd_crypto_fill_prdt(struct ufs_hba *hba,
-					  struct ufshcd_lrb *lrbp)
+					  struct scsi_cmnd *cmd)
 {
-	struct scsi_cmnd *cmd = lrbp->cmd;
 	const struct bio_crypt_ctx *crypt_ctx = scsi_cmd_to_rq(cmd)->crypt_ctx;
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 
 	if (crypt_ctx && hba->vops && hba->vops->fill_crypto_prdt)
 		return hba->vops->fill_crypto_prdt(hba, crypt_ctx,
@@ -51,17 +51,19 @@ static inline int ufshcd_crypto_fill_prdt(struct ufs_hba *hba,
 }
 
 static inline void ufshcd_crypto_clear_prdt(struct ufs_hba *hba,
-					    struct ufshcd_lrb *lrbp)
+					    struct scsi_cmnd *cmd)
 {
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
+
 	if (!(hba->quirks & UFSHCD_QUIRK_KEYS_IN_PRDT))
 		return;
 
-	if (!(scsi_cmd_to_rq(lrbp->cmd)->crypt_ctx))
+	if (!(scsi_cmd_to_rq(cmd)->crypt_ctx))
 		return;
 
 	/* Zeroize the PRDT because it can contain cryptographic keys. */
 	memzero_explicit(lrbp->ucd_prdt_ptr,
-			 ufshcd_sg_entry_size(hba) * scsi_sg_count(lrbp->cmd));
+			 ufshcd_sg_entry_size(hba) * scsi_sg_count(cmd));
 }
 
 bool ufshcd_crypto_enable(struct ufs_hba *hba);
@@ -82,13 +84,15 @@ ufshcd_prepare_req_desc_hdr_crypto(struct ufshcd_lrb *lrbp,
 				   struct request_desc_header *h) { }
 
 static inline int ufshcd_crypto_fill_prdt(struct ufs_hba *hba,
-					  struct ufshcd_lrb *lrbp)
+					  struct scsi_cmnd *cmd)
 {
 	return 0;
 }
 
 static inline void ufshcd_crypto_clear_prdt(struct ufs_hba *hba,
-					    struct ufshcd_lrb *lrbp) { }
+					    struct scsi_cmnd *cmd)
+{
+}
 
 static inline bool ufshcd_crypto_enable(struct ufs_hba *hba)
 {
diff --git a/drivers/ufs/core/ufshcd-priv.h b/drivers/ufs/core/ufshcd-priv.h
index ec1bb818bc73..3222c4d3ceb4 100644
--- a/drivers/ufs/core/ufshcd-priv.h
+++ b/drivers/ufs/core/ufshcd-priv.h
@@ -75,8 +75,7 @@ bool ufshcd_cmd_inflight(struct scsi_cmnd *cmd);
 int ufshcd_mcq_sq_cleanup(struct ufs_hba *hba, int task_tag);
 int ufshcd_mcq_abort(struct scsi_cmnd *cmd);
 int ufshcd_try_to_abort_task(struct ufs_hba *hba, int tag);
-void ufshcd_release_scsi_cmd(struct ufs_hba *hba,
-			     struct ufshcd_lrb *lrbp);
+void ufshcd_release_scsi_cmd(struct ufs_hba *hba, struct scsi_cmnd *cmd);
 
 #define SD_ASCII_STD true
 #define SD_RAW false
@@ -361,6 +360,44 @@ static inline bool ufs_is_valid_unit_desc_lun(struct ufs_dev_info *dev_info, u8
 	return lun == UFS_UPIU_RPMB_WLUN || (lun < dev_info->max_lu_supported);
 }
 
+/*
+ * Convert a block layer tag into a SCSI command pointer. This function is
+ * called once per I/O completion path and is also called from error paths.
+ */
+static inline struct scsi_cmnd *ufshcd_tag_to_cmd(struct ufs_hba *hba, u32 tag)
+{
+	struct blk_mq_tags *tags = hba->host->tag_set.shared_tags;
+	struct request *rq;
+
+	/*
+	 * Handle reserved tags differently because the UFS driver does not
+	 * call blk_mq_alloc_request() for allocating reserved requests.
+	 * Allocating reserved tags with blk_mq_alloc_request() would require
+	 * the following:
+	 * - Allocate an additional request queue from &hba->host->tag_set for
+	 *   allocating reserved requests from.
+	 * - For that request queue, allocate a SCSI device.
+	 * - Calling blk_mq_alloc_request(hba->dev_mgmt_queue, REQ_OP_DRV_OUT,
+	 *   BLK_MQ_REQ_RESERVED) for allocating a reserved request and
+	 *   blk_mq_free_request() for freeing reserved requests.
+	 * - Set the .device pointer for these reserved requests.
+	 * - Submit reserved requests with blk_execute_rq().
+	 * - Modify ufshcd_queuecommand() such that it handles reserved requests
+	 *   in another way than SCSI requests.
+	 * - Modify ufshcd_compl_one_cqe() such that it calls scsi_done() for
+	 *   device management commands.
+	 * - Modify all callback functions called by blk_mq_tagset_busy_iter()
+	 *   calls in the UFS driver and skip device management commands.
+	 */
+	rq = tag < UFSHCD_NUM_RESERVED ? tags->static_rqs[tag] :
+					 blk_mq_tag_to_rq(tags, tag);
+
+	if (WARN_ON_ONCE(!rq))
+		return NULL;
+
+	return blk_mq_rq_to_pdu(rq);
+}
+
 static inline void ufshcd_inc_sq_tail(struct ufs_hw_queue *q)
 	__must_hold(&q->sq_lock)
 {
diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index a54359289db5..76375dbcaf5d 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -28,6 +28,7 @@
 #include <scsi/scsi_dbg.h>
 #include <scsi/scsi_driver.h>
 #include <scsi/scsi_eh.h>
+#include <scsi/scsi_tcq.h>
 #include "ufshcd-priv.h"
 #include <ufs/ufs_quirks.h>
 #include <ufs/unipro.h>
@@ -488,7 +489,7 @@ static void ufshcd_add_command_trace(struct ufs_hba *hba, struct scsi_cmnd *cmd,
 	u32 hwq_id = 0;
 	struct request *rq = scsi_cmd_to_rq(cmd);
 	unsigned int tag = rq->tag;
-	struct ufshcd_lrb *lrbp = &hba->lrb[tag];
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	int transfer_len = -1;
 
 	/* trace UPIU also */
@@ -599,14 +600,13 @@ static void ufshcd_print_evt_hist(struct ufs_hba *hba)
 	ufshcd_vops_dbg_register_dump(hba);
 }
 
-static
-void ufshcd_print_tr(struct ufs_hba *hba, int tag, bool pr_prdt)
+static void ufshcd_print_tr(struct ufs_hba *hba, struct scsi_cmnd *cmd,
+			    bool pr_prdt)
 {
-	const struct ufshcd_lrb *lrbp;
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
+	const int tag = lrbp->task_tag;
 	int prdt_length;
 
-	lrbp = &hba->lrb[tag];
-
 	if (hba->monitor.enabled) {
 		dev_err(hba->dev, "UPIU[%d] - issue time %lld us\n", tag,
 			div_u64(lrbp->issue_time_stamp_local_clock, 1000));
@@ -649,7 +649,7 @@ static bool ufshcd_print_tr_iter(struct request *req, void *priv)
 	struct Scsi_Host *shost = sdev->host;
 	struct ufs_hba *hba = shost_priv(shost);
 
-	ufshcd_print_tr(hba, req->tag, *(bool *)priv);
+	ufshcd_print_tr(hba, blk_mq_rq_to_pdu(req), *(bool *)priv);
 
 	return true;
 }
@@ -2297,8 +2297,7 @@ static inline bool ufshcd_should_inform_monitor(struct ufs_hba *hba,
 						struct scsi_cmnd *cmd)
 {
 	const struct ufs_hba_monitor *m = &hba->monitor;
-	struct request *rq = scsi_cmd_to_rq(cmd);
-	struct ufshcd_lrb *lrbp = &hba->lrb[rq->tag];
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 
 	return m->enabled &&
 	       (!m->chunk_size || m->chunk_size == cmd->sdb.length) &&
@@ -2319,7 +2318,7 @@ static void ufshcd_start_monitor(struct ufs_hba *hba, struct scsi_cmnd *cmd)
 static void ufshcd_update_monitor(struct ufs_hba *hba, struct scsi_cmnd *cmd)
 {
 	struct request *req = scsi_cmd_to_rq(cmd);
-	struct ufshcd_lrb *lrbp = &hba->lrb[req->tag];
+	const struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	int dir = ufshcd_monitor_opcode2dir(cmd->cmnd[0]);
 	unsigned long flags;
 
@@ -2349,17 +2348,26 @@ static void ufshcd_update_monitor(struct ufs_hba *hba, struct scsi_cmnd *cmd)
 	spin_unlock_irqrestore(hba->host->host_lock, flags);
 }
 
+/*
+ * Returns %true if @cmd represents a SCSI command that is in flight and %false
+ * if it represents a device management command.
+ */
+static bool ufshcd_is_scsi_cmd(struct scsi_cmnd *cmd)
+{
+	return blk_mq_request_started(scsi_cmd_to_rq(cmd));
+}
+
 /**
  * ufshcd_send_command - Send SCSI or device management commands
  * @hba: per adapter instance
- * @lrbp: Local reference block of SCSI command
+ * @cmd: SCSI command or device management command pointer
  * @hwq: pointer to hardware queue instance
  */
 static inline void ufshcd_send_command(struct ufs_hba *hba,
-				       struct ufshcd_lrb *lrbp,
+				       struct scsi_cmnd *cmd,
 				       struct ufs_hw_queue *hwq)
 {
-	struct scsi_cmnd *cmd = lrbp->cmd;
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	unsigned long flags;
 
 	if (hba->monitor.enabled) {
@@ -2368,7 +2376,7 @@ static inline void ufshcd_send_command(struct ufs_hba *hba,
 		lrbp->compl_time_stamp = ktime_set(0, 0);
 		lrbp->compl_time_stamp_local_clock = 0;
 	}
-	if (cmd) {
+	if (ufshcd_is_scsi_cmd(cmd)) {
 		ufshcd_add_command_trace(hba, cmd, UFS_CMD_SEND);
 		ufshcd_clk_scaling_start_busy(hba);
 		if (unlikely(ufshcd_should_inform_monitor(hba, cmd)))
@@ -2388,7 +2396,8 @@ static inline void ufshcd_send_command(struct ufs_hba *hba,
 	} else {
 		spin_lock_irqsave(&hba->outstanding_lock, flags);
 		if (hba->vops && hba->vops->setup_xfer_req)
-			hba->vops->setup_xfer_req(hba, lrbp->task_tag, !!cmd);
+			hba->vops->setup_xfer_req(hba, lrbp->task_tag,
+						  ufshcd_is_scsi_cmd(cmd));
 		__set_bit(lrbp->task_tag, &hba->outstanding_reqs);
 		ufshcd_writel(hba, 1 << lrbp->task_tag,
 			      REG_UTP_TRANSFER_REQ_DOOR_BELL);
@@ -2398,11 +2407,12 @@ static inline void ufshcd_send_command(struct ufs_hba *hba,
 
 /**
  * ufshcd_copy_sense_data - Copy sense data in case of check condition
- * @lrbp: pointer to local reference block
+ * @cmd: SCSI command
  */
-static inline void ufshcd_copy_sense_data(struct ufshcd_lrb *lrbp)
+static inline void ufshcd_copy_sense_data(struct scsi_cmnd *cmd)
 {
-	u8 *const sense_buffer = lrbp->cmd->sense_buffer;
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
+	u8 *const sense_buffer = cmd->sense_buffer;
 	u16 resp_len;
 	int len;
 
@@ -2707,13 +2717,13 @@ static void ufshcd_sgl_to_prdt(struct ufs_hba *hba, struct ufshcd_lrb *lrbp, int
 /**
  * ufshcd_map_sg - Map scatter-gather list to prdt
  * @hba: per adapter instance
- * @lrbp: pointer to local reference block
+ * @cmd: SCSI command
  *
  * Return: 0 in case of success, non-zero value in case of failure.
  */
-static int ufshcd_map_sg(struct ufs_hba *hba, struct ufshcd_lrb *lrbp)
+static int ufshcd_map_sg(struct ufs_hba *hba, struct scsi_cmnd *cmd)
 {
-	struct scsi_cmnd *cmd = lrbp->cmd;
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	int sg_segments = scsi_dma_map(cmd);
 
 	if (sg_segments < 0)
@@ -2721,7 +2731,7 @@ static int ufshcd_map_sg(struct ufs_hba *hba, struct ufshcd_lrb *lrbp)
 
 	ufshcd_sgl_to_prdt(hba, lrbp, sg_segments, scsi_sglist(cmd));
 
-	return ufshcd_crypto_fill_prdt(hba, lrbp);
+	return ufshcd_crypto_fill_prdt(hba, cmd);
 }
 
 /**
@@ -2780,13 +2790,13 @@ ufshcd_prepare_req_desc_hdr(struct ufs_hba *hba, struct ufshcd_lrb *lrbp,
 /**
  * ufshcd_prepare_utp_scsi_cmd_upiu() - fills the utp_transfer_req_desc,
  * for scsi commands
- * @lrbp: local reference block pointer
+ * @cmd: SCSI command
  * @upiu_flags: flags
  */
-static
-void ufshcd_prepare_utp_scsi_cmd_upiu(struct ufshcd_lrb *lrbp, u8 upiu_flags)
+static void ufshcd_prepare_utp_scsi_cmd_upiu(struct scsi_cmnd *cmd,
+					     u8 upiu_flags)
 {
-	struct scsi_cmnd *cmd = lrbp->cmd;
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	struct utp_upiu_req *ucd_req_ptr = lrbp->ucd_req_ptr;
 	unsigned short cdb_len;
 
@@ -2889,22 +2899,25 @@ static int ufshcd_compose_devman_upiu(struct ufs_hba *hba,
  * ufshcd_comp_scsi_upiu - UFS Protocol Information Unit(UPIU)
  *			   for SCSI Purposes
  * @hba: per adapter instance
- * @lrbp: pointer to local reference block
+ * @cmd: SCSI command
  */
-static void ufshcd_comp_scsi_upiu(struct ufs_hba *hba, struct ufshcd_lrb *lrbp)
+static void ufshcd_comp_scsi_upiu(struct ufs_hba *hba, struct scsi_cmnd *cmd)
 {
-	struct request *rq = scsi_cmd_to_rq(lrbp->cmd);
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
+	struct request *rq = scsi_cmd_to_rq(cmd);
 	unsigned int ioprio_class = IOPRIO_PRIO_CLASS(req_get_ioprio(rq));
 	u8 upiu_flags;
 
-	ufshcd_prepare_req_desc_hdr(hba, lrbp, &upiu_flags, lrbp->cmd->sc_data_direction, 0);
+	ufshcd_prepare_req_desc_hdr(hba, lrbp, &upiu_flags,
+				    cmd->sc_data_direction, 0);
 	if (ioprio_class == IOPRIO_CLASS_RT)
 		upiu_flags |= UPIU_CMD_FLAGS_CP;
-	ufshcd_prepare_utp_scsi_cmd_upiu(lrbp, upiu_flags);
+	ufshcd_prepare_utp_scsi_cmd_upiu(cmd, upiu_flags);
 }
 
-static void ufshcd_init_lrb(struct ufs_hba *hba, struct ufshcd_lrb *lrb, int i)
+static void ufshcd_init_lrb(struct ufs_hba *hba, struct scsi_cmnd *cmd)
 {
+	const int i = scsi_cmd_to_rq(cmd)->tag;
 	struct utp_transfer_cmd_desc *cmd_descp =
 		(void *)hba->ucdl_base_addr + i * ufshcd_get_ucd_size(hba);
 	struct utp_transfer_req_desc *utrdlp = hba->utrdl_base_addr;
@@ -2912,6 +2925,7 @@ static void ufshcd_init_lrb(struct ufs_hba *hba, struct ufshcd_lrb *lrb, int i)
 		hba->ucdl_dma_addr + i * ufshcd_get_ucd_size(hba);
 	u16 response_offset = le16_to_cpu(utrdlp[i].response_upiu_offset);
 	u16 prdt_offset = le16_to_cpu(utrdlp[i].prd_table_offset);
+	struct ufshcd_lrb *lrb = scsi_cmd_priv(cmd);
 
 	lrb->utr_descriptor_ptr = utrdlp + i;
 	lrb->utrd_dma_addr =
@@ -2924,27 +2938,30 @@ static void ufshcd_init_lrb(struct ufs_hba *hba, struct ufshcd_lrb *lrb, int i)
 	lrb->ucd_prdt_dma_addr = cmd_desc_element_addr + prdt_offset;
 }
 
-static void __ufshcd_setup_cmd(struct ufs_hba *hba, struct ufshcd_lrb *lrbp,
-			       struct scsi_cmnd *cmd, u8 lun, int tag)
+static void __ufshcd_setup_cmd(struct ufs_hba *hba, struct scsi_cmnd *cmd,
+			       u8 lun, int tag)
 {
-	ufshcd_init_lrb(hba, lrbp, tag);
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
+
+	ufshcd_init_lrb(hba, cmd);
 
 	memset(lrbp->ucd_req_ptr, 0, sizeof(*lrbp->ucd_req_ptr));
 
-	lrbp->cmd = cmd;
 	lrbp->task_tag = tag;
 	lrbp->lun = lun;
 	ufshcd_prepare_lrbp_crypto(cmd ? scsi_cmd_to_rq(cmd) : NULL, lrbp);
 }
 
-static void ufshcd_setup_scsi_cmd(struct ufs_hba *hba, struct ufshcd_lrb *lrbp,
-				  struct scsi_cmnd *cmd, u8 lun, int tag)
+static void ufshcd_setup_scsi_cmd(struct ufs_hba *hba, struct scsi_cmnd *cmd,
+				  u8 lun, int tag)
 {
-	__ufshcd_setup_cmd(hba, lrbp, cmd, lun, tag);
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
+
+	__ufshcd_setup_cmd(hba, cmd, lun, tag);
 	lrbp->intr_cmd = !ufshcd_is_intr_aggr_allowed(hba);
 	lrbp->req_abort_skip = false;
 
-	ufshcd_comp_scsi_upiu(hba, lrbp);
+	ufshcd_comp_scsi_upiu(hba, cmd);
 }
 
 /**
@@ -3015,7 +3032,6 @@ static int ufshcd_queuecommand(struct Scsi_Host *host, struct scsi_cmnd *cmd)
 {
 	struct ufs_hba *hba = shost_priv(host);
 	int tag = scsi_cmd_to_rq(cmd)->tag;
-	struct ufshcd_lrb *lrbp;
 	int err = 0;
 	struct ufs_hw_queue *hwq = NULL;
 
@@ -3066,11 +3082,10 @@ static int ufshcd_queuecommand(struct Scsi_Host *host, struct scsi_cmnd *cmd)
 
 	ufshcd_hold(hba);
 
-	lrbp = &hba->lrb[tag];
-
-	ufshcd_setup_scsi_cmd(hba, lrbp, cmd, ufshcd_scsi_to_upiu_lun(cmd->device->lun), tag);
+	ufshcd_setup_scsi_cmd(hba, cmd,
+			      ufshcd_scsi_to_upiu_lun(cmd->device->lun), tag);
 
-	err = ufshcd_map_sg(hba, lrbp);
+	err = ufshcd_map_sg(hba, cmd);
 	if (err) {
 		ufshcd_release(hba);
 		goto out;
@@ -3079,7 +3094,7 @@ static int ufshcd_queuecommand(struct Scsi_Host *host, struct scsi_cmnd *cmd)
 	if (hba->mcq_enabled)
 		hwq = ufshcd_mcq_req_to_hwq(hba, scsi_cmd_to_rq(cmd));
 
-	ufshcd_send_command(hba, lrbp, hwq);
+	ufshcd_send_command(hba, cmd, hwq);
 
 out:
 	if (ufs_trigger_eh(hba)) {
@@ -3093,10 +3108,12 @@ static int ufshcd_queuecommand(struct Scsi_Host *host, struct scsi_cmnd *cmd)
 	return err;
 }
 
-static void ufshcd_setup_dev_cmd(struct ufs_hba *hba, struct ufshcd_lrb *lrbp,
-			     enum dev_cmd_type cmd_type, u8 lun, int tag)
+static void ufshcd_setup_dev_cmd(struct ufs_hba *hba, struct scsi_cmnd *cmd,
+				 enum dev_cmd_type cmd_type, u8 lun, int tag)
 {
-	__ufshcd_setup_cmd(hba, lrbp, NULL, lun, tag);
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
+
+	__ufshcd_setup_cmd(hba, cmd, lun, tag);
 	lrbp->intr_cmd = true; /* No interrupt aggregation */
 	hba->dev_cmd.type = cmd_type;
 }
@@ -3104,10 +3121,12 @@ static void ufshcd_setup_dev_cmd(struct ufs_hba *hba, struct ufshcd_lrb *lrbp,
 /*
  * Return: 0 upon success; < 0 upon failure.
  */
-static int ufshcd_compose_dev_cmd(struct ufs_hba *hba,
-		struct ufshcd_lrb *lrbp, enum dev_cmd_type cmd_type, int tag)
+static int ufshcd_compose_dev_cmd(struct ufs_hba *hba, struct scsi_cmnd *cmd,
+				  enum dev_cmd_type cmd_type, int tag)
 {
-	ufshcd_setup_dev_cmd(hba, lrbp, cmd_type, 0, tag);
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
+
+	ufshcd_setup_dev_cmd(hba, cmd, cmd_type, 0, tag);
 
 	return ufshcd_compose_devman_upiu(hba, lrbp);
 }
@@ -3318,13 +3337,14 @@ static void ufshcd_dev_man_unlock(struct ufs_hba *hba)
 /*
  * Return: 0 upon success; < 0 upon failure.
  */
-static int ufshcd_issue_dev_cmd(struct ufs_hba *hba, struct ufshcd_lrb *lrbp,
-			  const u32 tag, int timeout)
+static int ufshcd_issue_dev_cmd(struct ufs_hba *hba, struct scsi_cmnd *cmd,
+				const u32 tag, int timeout)
 {
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	int err;
 
 	ufshcd_add_query_upiu_trace(hba, UFS_QUERY_SEND, lrbp->ucd_req_ptr);
-	ufshcd_send_command(hba, lrbp, hba->dev_cmd_queue);
+	ufshcd_send_command(hba, cmd, hba->dev_cmd_queue);
 	err = ufshcd_wait_for_dev_cmd(hba, lrbp, timeout);
 
 	ufshcd_add_query_upiu_trace(hba, err ? UFS_QUERY_ERR : UFS_QUERY_COMP,
@@ -3348,17 +3368,17 @@ static int ufshcd_exec_dev_cmd(struct ufs_hba *hba,
 		enum dev_cmd_type cmd_type, int timeout)
 {
 	const u32 tag = hba->reserved_slot;
-	struct ufshcd_lrb *lrbp = &hba->lrb[tag];
+	struct scsi_cmnd *cmd = ufshcd_tag_to_cmd(hba, tag);
 	int err;
 
 	/* Protects use of hba->reserved_slot. */
 	lockdep_assert_held(&hba->dev_cmd.lock);
 
-	err = ufshcd_compose_dev_cmd(hba, lrbp, cmd_type, tag);
+	err = ufshcd_compose_dev_cmd(hba, cmd, cmd_type, tag);
 	if (unlikely(err))
 		return err;
 
-	return ufshcd_issue_dev_cmd(hba, lrbp, tag, timeout);
+	return ufshcd_issue_dev_cmd(hba, cmd, tag, timeout);
 }
 
 /**
@@ -3985,14 +4005,6 @@ static int ufshcd_memory_alloc(struct ufs_hba *hba)
 	}
 
 skip_utmrdl:
-	/* Allocate memory for local reference block */
-	hba->lrb = devm_kcalloc(hba->dev,
-				hba->nutrs, sizeof(struct ufshcd_lrb),
-				GFP_KERNEL);
-	if (!hba->lrb) {
-		dev_err(hba->dev, "LRB Memory allocation failed\n");
-		goto out;
-	}
 	return 0;
 out:
 	return -ENOMEM;
@@ -5403,19 +5415,18 @@ static void ufshcd_sdev_destroy(struct scsi_device *sdev)
 
 /**
  * ufshcd_scsi_cmd_status - Update SCSI command result based on SCSI status
- * @lrbp: pointer to local reference block of completed command
+ * @cmd: SCSI command
  * @scsi_status: SCSI command status
  *
  * Return: value base on SCSI command status.
  */
-static inline int
-ufshcd_scsi_cmd_status(struct ufshcd_lrb *lrbp, int scsi_status)
+static inline int ufshcd_scsi_cmd_status(struct scsi_cmnd *cmd, int scsi_status)
 {
 	int result = 0;
 
 	switch (scsi_status) {
 	case SAM_STAT_CHECK_CONDITION:
-		ufshcd_copy_sense_data(lrbp);
+		ufshcd_copy_sense_data(cmd);
 		fallthrough;
 	case SAM_STAT_GOOD:
 		result |= DID_OK << 16 | scsi_status;
@@ -5423,7 +5434,7 @@ ufshcd_scsi_cmd_status(struct ufshcd_lrb *lrbp, int scsi_status)
 	case SAM_STAT_TASK_SET_FULL:
 	case SAM_STAT_BUSY:
 	case SAM_STAT_TASK_ABORTED:
-		ufshcd_copy_sense_data(lrbp);
+		ufshcd_copy_sense_data(cmd);
 		result |= scsi_status;
 		break;
 	default:
@@ -5437,15 +5448,16 @@ ufshcd_scsi_cmd_status(struct ufshcd_lrb *lrbp, int scsi_status)
 /**
  * ufshcd_transfer_rsp_status - Get overall status of the response
  * @hba: per adapter instance
- * @lrbp: pointer to local reference block of completed command
+ * @cmd: SCSI command
  * @cqe: pointer to the completion queue entry
  *
  * Return: result of the command to notify SCSI midlayer.
  */
-static inline int
-ufshcd_transfer_rsp_status(struct ufs_hba *hba, struct ufshcd_lrb *lrbp,
-			   struct cq_entry *cqe)
+static inline int ufshcd_transfer_rsp_status(struct ufs_hba *hba,
+					     struct scsi_cmnd *cmd,
+					     struct cq_entry *cqe)
 {
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	int result = 0;
 	int scsi_status;
 	enum utp_ocs ocs;
@@ -5459,7 +5471,7 @@ ufshcd_transfer_rsp_status(struct ufs_hba *hba, struct ufshcd_lrb *lrbp,
 	 * not set either flag.
 	 */
 	if (resid && !(upiu_flags & UPIU_RSP_FLAG_OVERFLOW))
-		scsi_set_resid(lrbp->cmd, resid);
+		scsi_set_resid(cmd, resid);
 
 	/* overall command status of utrd */
 	ocs = ufshcd_get_tr_ocs(lrbp, cqe);
@@ -5480,7 +5492,7 @@ ufshcd_transfer_rsp_status(struct ufs_hba *hba, struct ufshcd_lrb *lrbp,
 			 * to notify the SCSI midlayer of the command status
 			 */
 			scsi_status = lrbp->ucd_rsp_ptr->header.status;
-			result = ufshcd_scsi_cmd_status(lrbp, scsi_status);
+			result = ufshcd_scsi_cmd_status(cmd, scsi_status);
 
 			/*
 			 * Currently we are only supporting BKOPs exception
@@ -5543,7 +5555,7 @@ ufshcd_transfer_rsp_status(struct ufs_hba *hba, struct ufshcd_lrb *lrbp,
 
 	if ((host_byte(result) != DID_OK) &&
 	    (host_byte(result) != DID_REQUEUE) && !hba->silence_err_logs)
-		ufshcd_print_tr(hba, lrbp->task_tag, true);
+		ufshcd_print_tr(hba, cmd, true);
 	return result;
 }
 
@@ -5612,13 +5624,10 @@ static irqreturn_t ufshcd_uic_cmd_compl(struct ufs_hba *hba, u32 intr_status)
 }
 
 /* Release the resources allocated for processing a SCSI command. */
-void ufshcd_release_scsi_cmd(struct ufs_hba *hba,
-			     struct ufshcd_lrb *lrbp)
+void ufshcd_release_scsi_cmd(struct ufs_hba *hba, struct scsi_cmnd *cmd)
 {
-	struct scsi_cmnd *cmd = lrbp->cmd;
-
 	scsi_dma_unmap(cmd);
-	ufshcd_crypto_clear_prdt(hba, lrbp);
+	ufshcd_crypto_clear_prdt(hba, cmd);
 	ufshcd_release(hba);
 	ufshcd_clk_scaling_update_busy(hba);
 }
@@ -5632,20 +5641,20 @@ void ufshcd_release_scsi_cmd(struct ufs_hba *hba,
 void ufshcd_compl_one_cqe(struct ufs_hba *hba, int task_tag,
 			  struct cq_entry *cqe)
 {
-	struct ufshcd_lrb *lrbp = &hba->lrb[task_tag];
-	struct scsi_cmnd *cmd = lrbp->cmd;
+	struct scsi_cmnd *cmd = ufshcd_tag_to_cmd(hba, task_tag);
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	enum utp_ocs ocs;
 
 	if (hba->monitor.enabled) {
 		lrbp->compl_time_stamp = ktime_get();
 		lrbp->compl_time_stamp_local_clock = local_clock();
 	}
-	if (cmd) {
+	if (ufshcd_is_scsi_cmd(cmd)) {
 		if (unlikely(ufshcd_should_inform_monitor(hba, cmd)))
 			ufshcd_update_monitor(hba, cmd);
 		ufshcd_add_command_trace(hba, cmd, UFS_CMD_COMP);
-		cmd->result = ufshcd_transfer_rsp_status(hba, lrbp, cqe);
-		ufshcd_release_scsi_cmd(hba, lrbp);
+		cmd->result = ufshcd_transfer_rsp_status(hba, cmd, cqe);
+		ufshcd_release_scsi_cmd(hba, cmd);
 		/* Do not touch lrbp after scsi done */
 		scsi_done(cmd);
 	} else {
@@ -5682,7 +5691,7 @@ static void ufshcd_clear_polled(struct ufs_hba *hba,
 	int tag;
 
 	for_each_set_bit(tag, completed_reqs, hba->nutrs) {
-		struct scsi_cmnd *cmd = hba->lrb[tag].cmd;
+		struct scsi_cmnd *cmd = scsi_host_find_tag(hba->host, tag);
 
 		if (!cmd)
 			continue;
@@ -5733,7 +5742,6 @@ static bool ufshcd_mcq_force_compl_one(struct request *rq, void *priv)
 	struct scsi_device *sdev = rq->q->queuedata;
 	struct Scsi_Host *shost = sdev->host;
 	struct ufs_hba *hba = shost_priv(shost);
-	struct ufshcd_lrb *lrbp = &hba->lrb[rq->tag];
 	struct ufs_hw_queue *hwq = ufshcd_mcq_req_to_hwq(hba, rq);
 
 	if (!hwq)
@@ -5748,7 +5756,7 @@ static bool ufshcd_mcq_force_compl_one(struct request *rq, void *priv)
 	scoped_guard(spinlock_irqsave, &hwq->cq_lock) {
 		if (!test_bit(SCMD_STATE_COMPLETE, &cmd->state)) {
 			set_host_byte(cmd, DID_REQUEUE);
-			ufshcd_release_scsi_cmd(hba, lrbp);
+			ufshcd_release_scsi_cmd(hba, cmd);
 			scsi_done(cmd);
 		}
 	}
@@ -6627,7 +6635,7 @@ static bool ufshcd_abort_one(struct request *rq, void *priv)
 
 	*ret = ufshcd_try_to_abort_task(hba, tag);
 	dev_err(hba->dev, "Aborting tag %d / CDB %#02x %s\n", tag,
-		hba->lrb[tag].cmd ? hba->lrb[tag].cmd->cmnd[0] : -1,
+		ufshcd_is_scsi_cmd(cmd) ? cmd->cmnd[0] : -1,
 		*ret ? "failed" : "succeeded");
 
 	return *ret == 0;
@@ -7344,14 +7352,15 @@ static int ufshcd_issue_devman_upiu_cmd(struct ufs_hba *hba,
 					enum query_opcode desc_op)
 {
 	const u32 tag = hba->reserved_slot;
-	struct ufshcd_lrb *lrbp = &hba->lrb[tag];
+	struct scsi_cmnd *cmd = ufshcd_tag_to_cmd(hba, tag);
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	int err = 0;
 	u8 upiu_flags;
 
 	/* Protects use of hba->reserved_slot. */
 	lockdep_assert_held(&hba->dev_cmd.lock);
 
-	ufshcd_setup_dev_cmd(hba, lrbp, cmd_type, 0, tag);
+	ufshcd_setup_dev_cmd(hba, cmd, cmd_type, 0, tag);
 
 	ufshcd_prepare_req_desc_hdr(hba, lrbp, &upiu_flags, DMA_NONE, 0);
 
@@ -7376,7 +7385,7 @@ static int ufshcd_issue_devman_upiu_cmd(struct ufs_hba *hba,
 	 * bound to fail since dev_cmd.query and dev_cmd.type were left empty.
 	 * read the response directly ignoring all errors.
 	 */
-	ufshcd_issue_dev_cmd(hba, lrbp, tag, dev_cmd_timeout);
+	ufshcd_issue_dev_cmd(hba, cmd, tag, dev_cmd_timeout);
 
 	/* just copy the upiu response as it is */
 	memcpy(rsp_upiu, lrbp->ucd_rsp_ptr, sizeof(*rsp_upiu));
@@ -7490,7 +7499,8 @@ int ufshcd_advanced_rpmb_req_handler(struct ufs_hba *hba, struct utp_upiu_req *r
 			 enum dma_data_direction dir)
 {
 	const u32 tag = hba->reserved_slot;
-	struct ufshcd_lrb *lrbp = &hba->lrb[tag];
+	struct scsi_cmnd *cmd = ufshcd_tag_to_cmd(hba, tag);
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	int err = 0;
 	int result;
 	u8 upiu_flags;
@@ -7501,7 +7511,8 @@ int ufshcd_advanced_rpmb_req_handler(struct ufs_hba *hba, struct utp_upiu_req *r
 	/* Protects use of hba->reserved_slot. */
 	ufshcd_dev_man_lock(hba);
 
-	ufshcd_setup_dev_cmd(hba, lrbp, DEV_CMD_TYPE_RPMB, UFS_UPIU_RPMB_WLUN, tag);
+	ufshcd_setup_dev_cmd(hba, cmd, DEV_CMD_TYPE_RPMB, UFS_UPIU_RPMB_WLUN,
+			     tag);
 
 	ufshcd_prepare_req_desc_hdr(hba, lrbp, &upiu_flags, DMA_NONE, ehs);
 
@@ -7518,7 +7529,7 @@ int ufshcd_advanced_rpmb_req_handler(struct ufs_hba *hba, struct utp_upiu_req *r
 
 	memset(lrbp->ucd_rsp_ptr, 0, sizeof(struct utp_upiu_rsp));
 
-	err = ufshcd_issue_dev_cmd(hba, lrbp, tag, ADVANCED_RPMB_REQ_TIMEOUT);
+	err = ufshcd_issue_dev_cmd(hba, cmd, tag, ADVANCED_RPMB_REQ_TIMEOUT);
 
 	if (!err) {
 		/* Just copy the upiu response as it is */
@@ -7619,11 +7630,12 @@ static int ufshcd_eh_device_reset_handler(struct scsi_cmnd *cmd)
 
 static void ufshcd_set_req_abort_skip(struct ufs_hba *hba, unsigned long bitmap)
 {
-	struct ufshcd_lrb *lrbp;
 	int tag;
 
 	for_each_set_bit(tag, &bitmap, hba->nutrs) {
-		lrbp = &hba->lrb[tag];
+		struct scsi_cmnd *cmd = ufshcd_tag_to_cmd(hba, tag);
+		struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
+
 		lrbp->req_abort_skip = true;
 	}
 }
@@ -7631,7 +7643,7 @@ static void ufshcd_set_req_abort_skip(struct ufs_hba *hba, unsigned long bitmap)
 /**
  * ufshcd_try_to_abort_task - abort a specific task
  * @hba: Pointer to adapter instance
- * @tag: Task tag/index to be aborted
+ * @tag: Tag of the task to be aborted
  *
  * Abort the pending command in device by sending UFS_ABORT_TASK task management
  * command, and in host controller by clearing the door-bell register. There can
@@ -7643,7 +7655,8 @@ static void ufshcd_set_req_abort_skip(struct ufs_hba *hba, unsigned long bitmap)
  */
 int ufshcd_try_to_abort_task(struct ufs_hba *hba, int tag)
 {
-	struct ufshcd_lrb *lrbp = &hba->lrb[tag];
+	struct scsi_cmnd *cmd = ufshcd_tag_to_cmd(hba, tag);
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	int err;
 	int poll_cnt;
 	u8 resp = 0xF;
@@ -7665,7 +7678,7 @@ int ufshcd_try_to_abort_task(struct ufs_hba *hba, int tag)
 				hba->dev,
 				"%s: cmd with tag %d not pending in the device.\n",
 				__func__, tag);
-			if (!ufshcd_cmd_inflight(lrbp->cmd)) {
+			if (!ufshcd_cmd_inflight(cmd)) {
 				dev_info(hba->dev,
 					 "%s: cmd with tag=%d completed.\n",
 					 __func__, tag);
@@ -7713,7 +7726,7 @@ static int ufshcd_abort(struct scsi_cmnd *cmd)
 	struct Scsi_Host *host = cmd->device->host;
 	struct ufs_hba *hba = shost_priv(host);
 	int tag = scsi_cmd_to_rq(cmd)->tag;
-	struct ufshcd_lrb *lrbp = &hba->lrb[tag];
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	unsigned long flags;
 	int err = FAILED;
 	bool outstanding;
@@ -7748,9 +7761,9 @@ static int ufshcd_abort(struct scsi_cmnd *cmd)
 		ufshcd_print_evt_hist(hba);
 		ufshcd_print_host_state(hba);
 		ufshcd_print_pwr_info(hba);
-		ufshcd_print_tr(hba, tag, true);
+		ufshcd_print_tr(hba, cmd, true);
 	} else {
-		ufshcd_print_tr(hba, tag, false);
+		ufshcd_print_tr(hba, cmd, false);
 	}
 	hba->req_abort_count++;
 
@@ -7794,7 +7807,7 @@ static int ufshcd_abort(struct scsi_cmnd *cmd)
 		goto release;
 	}
 
-	err = ufshcd_try_to_abort_task(hba, tag);
+	err = ufshcd_try_to_abort_task(hba, lrbp->task_tag);
 	if (err) {
 		dev_err(hba->dev, "%s: failed with err %d\n", __func__, err);
 		ufshcd_set_req_abort_skip(hba, hba->outstanding_reqs);
@@ -7811,7 +7824,7 @@ static int ufshcd_abort(struct scsi_cmnd *cmd)
 	spin_unlock_irqrestore(&hba->outstanding_lock, flags);
 
 	if (outstanding)
-		ufshcd_release_scsi_cmd(hba, lrbp);
+		ufshcd_release_scsi_cmd(hba, cmd);
 
 	err = SUCCESS;
 
@@ -8889,8 +8902,6 @@ static void ufshcd_release_sdb_queue(struct ufs_hba *hba, int nutrs)
 	utrdl_size = sizeof(struct utp_transfer_req_desc) * nutrs;
 	dmam_free_coherent(hba->dev, utrdl_size, hba->utrdl_base_addr,
 			   hba->utrdl_dma_addr);
-
-	devm_kfree(hba->dev, hba->lrb);
 }
 
 static int ufshcd_alloc_mcq(struct ufs_hba *hba)
@@ -9166,6 +9177,7 @@ static const struct scsi_host_template ufshcd_driver_template = {
 	.name			= UFSHCD,
 	.proc_name		= UFSHCD,
 	.map_queues		= ufshcd_map_queues,
+	.cmd_size		= sizeof(struct ufshcd_lrb),
 	.init_cmd_priv		= ufshcd_init_cmd_priv,
 	.queuecommand		= ufshcd_queuecommand,
 	.mq_poll		= ufshcd_poll,
diff --git a/include/ufs/ufshcd.h b/include/ufs/ufshcd.h
index ea0021f067c9..89dcac34c957 100644
--- a/include/ufs/ufshcd.h
+++ b/include/ufs/ufshcd.h
@@ -161,7 +161,6 @@ struct ufs_pm_lvl_states {
  * @ucd_prdt_dma_addr: PRDT dma address for debug
  * @ucd_rsp_dma_addr: UPIU response dma address for debug
  * @ucd_req_dma_addr: UPIU request dma address for debug
- * @cmd: pointer to SCSI command
  * @scsi_status: SCSI status of the command
  * @command_type: SCSI, UFS, Query.
  * @task_tag: Task tag of the command
@@ -186,7 +185,6 @@ struct ufshcd_lrb {
 	dma_addr_t ucd_rsp_dma_addr;
 	dma_addr_t ucd_prdt_dma_addr;
 
-	struct scsi_cmnd *cmd;
 	int scsi_status;
 
 	int command_type;
@@ -833,7 +831,6 @@ enum ufshcd_mcq_opr {
  * @spm_lvl: desired UFS power management level during system PM.
  * @pm_op_in_progress: whether or not a PM operation is in progress.
  * @ahit: value of Auto-Hibernate Idle Timer register.
- * @lrb: local reference block
  * @outstanding_tasks: Bits representing outstanding task requests
  * @outstanding_lock: Protects @outstanding_reqs.
  * @outstanding_reqs: Bits representing outstanding transfer requests
@@ -975,8 +972,6 @@ struct ufs_hba {
 	/* Auto-Hibernate Idle Timer register value */
 	u32 ahit;
 
-	struct ufshcd_lrb *lrb;
-
 	unsigned long outstanding_tasks;
 	spinlock_t outstanding_lock;
 	unsigned long outstanding_reqs;

[PATCH v5 24/28] ufs: core: Pass a SCSI pointer instead of an LRB pointer

[Bart Van Assche]

Pass a pointer to a SCSI command between functions instead of an LRB
pointer. This change prepares for removing the ufshcd_lrb task_tag
member.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufshcd.c | 21 +++++++++++----------
 1 file changed, 11 insertions(+), 10 deletions(-)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index 76375dbcaf5d..02e3e4c4bf3e 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -2821,12 +2821,13 @@ static void ufshcd_prepare_utp_scsi_cmd_upiu(struct scsi_cmnd *cmd,
 /**
  * ufshcd_prepare_utp_query_req_upiu() - fill the utp_transfer_req_desc for query request
  * @hba: UFS hba
- * @lrbp: local reference block pointer
+ * @cmd: SCSI command pointer
  * @upiu_flags: flags
  */
 static void ufshcd_prepare_utp_query_req_upiu(struct ufs_hba *hba,
-				struct ufshcd_lrb *lrbp, u8 upiu_flags)
+				struct scsi_cmnd *cmd, u8 upiu_flags)
 {
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	struct utp_upiu_req *ucd_req_ptr = lrbp->ucd_req_ptr;
 	struct ufs_query *query = &hba->dev_cmd.query;
 	u16 len = be16_to_cpu(query->request.upiu_req.length);
@@ -2855,8 +2856,9 @@ static void ufshcd_prepare_utp_query_req_upiu(struct ufs_hba *hba,
 		memcpy(ucd_req_ptr + 1, query->descriptor, len);
 }
 
-static inline void ufshcd_prepare_utp_nop_upiu(struct ufshcd_lrb *lrbp)
+static inline void ufshcd_prepare_utp_nop_upiu(struct scsi_cmnd *cmd)
 {
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	struct utp_upiu_req *ucd_req_ptr = lrbp->ucd_req_ptr;
 
 	memset(ucd_req_ptr, 0, sizeof(struct utp_upiu_req));
@@ -2871,22 +2873,23 @@ static inline void ufshcd_prepare_utp_nop_upiu(struct ufshcd_lrb *lrbp)
  * ufshcd_compose_devman_upiu - UFS Protocol Information Unit(UPIU)
  *			     for Device Management Purposes
  * @hba: per adapter instance
- * @lrbp: pointer to local reference block
+ * @cmd: SCSI command pointer
  *
  * Return: 0 upon success; < 0 upon failure.
  */
 static int ufshcd_compose_devman_upiu(struct ufs_hba *hba,
-				      struct ufshcd_lrb *lrbp)
+				      struct scsi_cmnd *cmd)
 {
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	u8 upiu_flags;
 	int ret = 0;
 
 	ufshcd_prepare_req_desc_hdr(hba, lrbp, &upiu_flags, DMA_NONE, 0);
 
 	if (hba->dev_cmd.type == DEV_CMD_TYPE_QUERY)
-		ufshcd_prepare_utp_query_req_upiu(hba, lrbp, upiu_flags);
+		ufshcd_prepare_utp_query_req_upiu(hba, cmd, upiu_flags);
 	else if (hba->dev_cmd.type == DEV_CMD_TYPE_NOP)
-		ufshcd_prepare_utp_nop_upiu(lrbp);
+		ufshcd_prepare_utp_nop_upiu(cmd);
 	else
 		ret = -EINVAL;
 
@@ -3124,11 +3127,9 @@ static void ufshcd_setup_dev_cmd(struct ufs_hba *hba, struct scsi_cmnd *cmd,
 static int ufshcd_compose_dev_cmd(struct ufs_hba *hba, struct scsi_cmnd *cmd,
 				  enum dev_cmd_type cmd_type, int tag)
 {
-	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
-
 	ufshcd_setup_dev_cmd(hba, cmd, cmd_type, 0, tag);
 
-	return ufshcd_compose_devman_upiu(hba, lrbp);
+	return ufshcd_compose_devman_upiu(hba, cmd);
 }
 
 /*

[PATCH v5 25/28] ufs: core: Remove the ufshcd_lrb task_tag member

[Bart Van Assche]

Remove the ufshcd_lrb task_tag member and use scsi_cmd_to_rq(cmd)->tag
instead. Use rq->tag instead of lrbp->task_tag. This patch reduces the
size of struct ufshcd_lrb.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufshcd.c | 62 +++++++++++++++++++--------------------
 include/ufs/ufshcd.h      |  1 -
 2 files changed, 30 insertions(+), 33 deletions(-)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index 02e3e4c4bf3e..7193505c030d 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -604,7 +604,7 @@ static void ufshcd_print_tr(struct ufs_hba *hba, struct scsi_cmnd *cmd,
 			    bool pr_prdt)
 {
 	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
-	const int tag = lrbp->task_tag;
+	const int tag = scsi_cmd_to_rq(cmd)->tag;
 	int prdt_length;
 
 	if (hba->monitor.enabled) {
@@ -2368,6 +2368,7 @@ static inline void ufshcd_send_command(struct ufs_hba *hba,
 				       struct ufs_hw_queue *hwq)
 {
 	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
+	const int tag = scsi_cmd_to_rq(cmd)->tag;
 	unsigned long flags;
 
 	if (hba->monitor.enabled) {
@@ -2396,11 +2397,10 @@ static inline void ufshcd_send_command(struct ufs_hba *hba,
 	} else {
 		spin_lock_irqsave(&hba->outstanding_lock, flags);
 		if (hba->vops && hba->vops->setup_xfer_req)
-			hba->vops->setup_xfer_req(hba, lrbp->task_tag,
+			hba->vops->setup_xfer_req(hba, tag,
 						  ufshcd_is_scsi_cmd(cmd));
-		__set_bit(lrbp->task_tag, &hba->outstanding_reqs);
-		ufshcd_writel(hba, 1 << lrbp->task_tag,
-			      REG_UTP_TRANSFER_REQ_DOOR_BELL);
+		__set_bit(tag, &hba->outstanding_reqs);
+		ufshcd_writel(hba, 1 << tag, REG_UTP_TRANSFER_REQ_DOOR_BELL);
 		spin_unlock_irqrestore(&hba->outstanding_lock, flags);
 	}
 }
@@ -2797,6 +2797,7 @@ static void ufshcd_prepare_utp_scsi_cmd_upiu(struct scsi_cmnd *cmd,
 					     u8 upiu_flags)
 {
 	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
+	const int tag = scsi_cmd_to_rq(cmd)->tag;
 	struct utp_upiu_req *ucd_req_ptr = lrbp->ucd_req_ptr;
 	unsigned short cdb_len;
 
@@ -2804,11 +2805,11 @@ static void ufshcd_prepare_utp_scsi_cmd_upiu(struct scsi_cmnd *cmd,
 		.transaction_code = UPIU_TRANSACTION_COMMAND,
 		.flags = upiu_flags,
 		.lun = lrbp->lun,
-		.task_tag = lrbp->task_tag,
+		.task_tag = tag,
 		.command_set_type = UPIU_COMMAND_SET_TYPE_SCSI,
 	};
 
-	WARN_ON_ONCE(ucd_req_ptr->header.task_tag != lrbp->task_tag);
+	WARN_ON_ONCE(ucd_req_ptr->header.task_tag != tag);
 
 	ucd_req_ptr->sc.exp_data_transfer_len = cpu_to_be32(cmd->sdb.length);
 
@@ -2829,6 +2830,7 @@ static void ufshcd_prepare_utp_query_req_upiu(struct ufs_hba *hba,
 {
 	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	struct utp_upiu_req *ucd_req_ptr = lrbp->ucd_req_ptr;
+	const int tag = scsi_cmd_to_rq(cmd)->tag;
 	struct ufs_query *query = &hba->dev_cmd.query;
 	u16 len = be16_to_cpu(query->request.upiu_req.length);
 
@@ -2837,7 +2839,7 @@ static void ufshcd_prepare_utp_query_req_upiu(struct ufs_hba *hba,
 		.transaction_code = UPIU_TRANSACTION_QUERY_REQ,
 		.flags = upiu_flags,
 		.lun = lrbp->lun,
-		.task_tag = lrbp->task_tag,
+		.task_tag = tag,
 		.query_function = query->request.query_func,
 		/* Data segment length only need for WRITE_DESC */
 		.data_segment_length =
@@ -2860,12 +2862,13 @@ static inline void ufshcd_prepare_utp_nop_upiu(struct scsi_cmnd *cmd)
 {
 	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	struct utp_upiu_req *ucd_req_ptr = lrbp->ucd_req_ptr;
+	const int tag = scsi_cmd_to_rq(cmd)->tag;
 
 	memset(ucd_req_ptr, 0, sizeof(struct utp_upiu_req));
 
 	ucd_req_ptr->header = (struct utp_upiu_header){
 		.transaction_code = UPIU_TRANSACTION_NOP_OUT,
-		.task_tag = lrbp->task_tag,
+		.task_tag = tag,
 	};
 }
 
@@ -2950,7 +2953,6 @@ static void __ufshcd_setup_cmd(struct ufs_hba *hba, struct scsi_cmnd *cmd,
 
 	memset(lrbp->ucd_req_ptr, 0, sizeof(*lrbp->ucd_req_ptr));
 
-	lrbp->task_tag = tag;
 	lrbp->lun = lun;
 	ufshcd_prepare_lrbp_crypto(cmd ? scsi_cmd_to_rq(cmd) : NULL, lrbp);
 }
@@ -3246,6 +3248,8 @@ ufshcd_dev_cmd_completion(struct ufs_hba *hba, struct ufshcd_lrb *lrbp)
 static int ufshcd_wait_for_dev_cmd(struct ufs_hba *hba,
 		struct ufshcd_lrb *lrbp, int max_timeout)
 {
+	struct scsi_cmnd *cmd = (struct scsi_cmnd *)lrbp - 1;
+	const int tag = scsi_cmd_to_rq(cmd)->tag;
 	unsigned long time_left = msecs_to_jiffies(max_timeout);
 	unsigned long flags;
 	bool pending;
@@ -3262,18 +3266,18 @@ static int ufshcd_wait_for_dev_cmd(struct ufs_hba *hba,
 	} else {
 		err = -ETIMEDOUT;
 		dev_dbg(hba->dev, "%s: dev_cmd request timedout, tag %d\n",
-			__func__, lrbp->task_tag);
+			__func__, tag);
 
 		/* MCQ mode */
 		if (hba->mcq_enabled) {
 			/* successfully cleared the command, retry if needed */
-			if (ufshcd_clear_cmd(hba, lrbp->task_tag) == 0)
+			if (ufshcd_clear_cmd(hba, tag) == 0)
 				err = -EAGAIN;
 			return err;
 		}
 
 		/* SDB mode */
-		if (ufshcd_clear_cmd(hba, lrbp->task_tag) == 0) {
+		if (ufshcd_clear_cmd(hba, tag) == 0) {
 			/* successfully cleared the command, retry if needed */
 			err = -EAGAIN;
 			/*
@@ -3282,11 +3286,9 @@ static int ufshcd_wait_for_dev_cmd(struct ufs_hba *hba,
 			 * variable.
 			 */
 			spin_lock_irqsave(&hba->outstanding_lock, flags);
-			pending = test_bit(lrbp->task_tag,
-					   &hba->outstanding_reqs);
+			pending = test_bit(tag, &hba->outstanding_reqs);
 			if (pending)
-				__clear_bit(lrbp->task_tag,
-					    &hba->outstanding_reqs);
+				__clear_bit(tag, &hba->outstanding_reqs);
 			spin_unlock_irqrestore(&hba->outstanding_lock, flags);
 
 			if (!pending) {
@@ -3299,11 +3301,10 @@ static int ufshcd_wait_for_dev_cmd(struct ufs_hba *hba,
 			}
 		} else {
 			dev_err(hba->dev, "%s: failed to clear tag %d\n",
-				__func__, lrbp->task_tag);
+				__func__, tag);
 
 			spin_lock_irqsave(&hba->outstanding_lock, flags);
-			pending = test_bit(lrbp->task_tag,
-					   &hba->outstanding_reqs);
+			pending = test_bit(tag, &hba->outstanding_reqs);
 			spin_unlock_irqrestore(&hba->outstanding_lock, flags);
 
 			if (!pending) {
@@ -5459,6 +5460,7 @@ static inline int ufshcd_transfer_rsp_status(struct ufs_hba *hba,
 					     struct cq_entry *cqe)
 {
 	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
+	const int tag = scsi_cmd_to_rq(cmd)->tag;
 	int result = 0;
 	int scsi_status;
 	enum utp_ocs ocs;
@@ -5530,10 +5532,8 @@ static inline int ufshcd_transfer_rsp_status(struct ufs_hba *hba,
 	case OCS_ABORTED:
 	case OCS_INVALID_COMMAND_STATUS:
 		result |= DID_REQUEUE << 16;
-		dev_warn(hba->dev,
-				"OCS %s from controller for tag %d\n",
-				(ocs == OCS_ABORTED ? "aborted" : "invalid"),
-				lrbp->task_tag);
+		dev_warn(hba->dev, "OCS %s from controller for tag %d\n",
+			 ocs == OCS_ABORTED ? "aborted" : "invalid", tag);
 		break;
 	case OCS_INVALID_CMD_TABLE_ATTR:
 	case OCS_INVALID_PRDT_ATTR:
@@ -5546,9 +5546,8 @@ static inline int ufshcd_transfer_rsp_status(struct ufs_hba *hba,
 	case OCS_GENERAL_CRYPTO_ERROR:
 	default:
 		result |= DID_ERROR << 16;
-		dev_err(hba->dev,
-				"OCS error from controller = %x for tag %d\n",
-				ocs, lrbp->task_tag);
+		dev_err(hba->dev, "OCS error from controller = %x for tag %d\n",
+			ocs, tag);
 		ufshcd_print_evt_hist(hba);
 		ufshcd_print_host_state(hba);
 		break;
@@ -7663,8 +7662,8 @@ int ufshcd_try_to_abort_task(struct ufs_hba *hba, int tag)
 	u8 resp = 0xF;
 
 	for (poll_cnt = 100; poll_cnt; poll_cnt--) {
-		err = ufshcd_issue_tm_cmd(hba, lrbp->lun, lrbp->task_tag,
-				UFS_QUERY_TASK, &resp);
+		err = ufshcd_issue_tm_cmd(hba, lrbp->lun, tag, UFS_QUERY_TASK,
+					  &resp);
 		if (!err && resp == UPIU_TASK_MANAGEMENT_FUNC_SUCCEEDED) {
 			/* cmd pending in the device */
 			dev_err(hba->dev, "%s: cmd pending in the device. tag = %d\n",
@@ -7697,8 +7696,7 @@ int ufshcd_try_to_abort_task(struct ufs_hba *hba, int tag)
 	if (!poll_cnt)
 		return -EBUSY;
 
-	err = ufshcd_issue_tm_cmd(hba, lrbp->lun, lrbp->task_tag,
-			UFS_ABORT_TASK, &resp);
+	err = ufshcd_issue_tm_cmd(hba, lrbp->lun, tag, UFS_ABORT_TASK, &resp);
 	if (err || resp != UPIU_TASK_MANAGEMENT_FUNC_COMPL) {
 		if (!err) {
 			err = resp; /* service response error */
@@ -7808,7 +7806,7 @@ static int ufshcd_abort(struct scsi_cmnd *cmd)
 		goto release;
 	}
 
-	err = ufshcd_try_to_abort_task(hba, lrbp->task_tag);
+	err = ufshcd_try_to_abort_task(hba, tag);
 	if (err) {
 		dev_err(hba->dev, "%s: failed with err %d\n", __func__, err);
 		ufshcd_set_req_abort_skip(hba, hba->outstanding_reqs);
diff --git a/include/ufs/ufshcd.h b/include/ufs/ufshcd.h
index 89dcac34c957..9eb69241e40f 100644
--- a/include/ufs/ufshcd.h
+++ b/include/ufs/ufshcd.h
@@ -188,7 +188,6 @@ struct ufshcd_lrb {
 	int scsi_status;
 
 	int command_type;
-	int task_tag;
 	u8 lun; /* UPIU LUN id field is only 8-bit wide */
 	bool intr_cmd;
 	bool req_abort_skip;

[PATCH v5 26/28] ufs: core: Make blk_mq_tagset_busy_iter() skip reserved requests

[Bart Van Assche]

A later patch will convert hba->reserved_slot into a reserved tag. Make
blk_mq_tagset_busy_iter() skip reserved requests such that device
management commands are skipped.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufshcd.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index 7193505c030d..a3c4f8cb2f37 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -649,7 +649,8 @@ static bool ufshcd_print_tr_iter(struct request *req, void *priv)
 	struct Scsi_Host *shost = sdev->host;
 	struct ufs_hba *hba = shost_priv(shost);
 
-	ufshcd_print_tr(hba, blk_mq_rq_to_pdu(req), *(bool *)priv);
+	if (!blk_mq_is_reserved_rq(req))
+		ufshcd_print_tr(hba, blk_mq_rq_to_pdu(req), *(bool *)priv);
 
 	return true;
 }
@@ -5744,7 +5745,7 @@ static bool ufshcd_mcq_force_compl_one(struct request *rq, void *priv)
 	struct ufs_hba *hba = shost_priv(shost);
 	struct ufs_hw_queue *hwq = ufshcd_mcq_req_to_hwq(hba, rq);
 
-	if (!hwq)
+	if (blk_mq_is_reserved_rq(rq) || !hwq)
 		return true;
 
 	ufshcd_mcq_compl_all_cqes_lock(hba, hwq);
@@ -5771,7 +5772,7 @@ static bool ufshcd_mcq_compl_one(struct request *rq, void *priv)
 	struct ufs_hba *hba = shost_priv(shost);
 	struct ufs_hw_queue *hwq = ufshcd_mcq_req_to_hwq(hba, rq);
 
-	if (hwq)
+	if (!blk_mq_is_reserved_rq(rq) && hwq)
 		ufshcd_mcq_poll_cqe_lock(hba, hwq);
 
 	return true;
@@ -6633,6 +6634,9 @@ static bool ufshcd_abort_one(struct request *rq, void *priv)
 	struct Scsi_Host *shost = sdev->host;
 	struct ufs_hba *hba = shost_priv(shost);
 
+	if (blk_mq_is_reserved_rq(rq))
+		return true;
+
 	*ret = ufshcd_try_to_abort_task(hba, tag);
 	dev_err(hba->dev, "Aborting tag %d / CDB %#02x %s\n", tag,
 		ufshcd_is_scsi_cmd(cmd) ? cmd->cmnd[0] : -1,
@@ -7568,7 +7572,7 @@ static bool ufshcd_clear_lu_cmds(struct request *req, void *priv)
 	const u64 lun = *(u64 *)priv;
 	const u32 tag = req->tag;
 
-	if (sdev->lun != lun)
+	if (blk_mq_is_reserved_rq(req) || sdev->lun != lun)
 		return true;
 
 	if (ufshcd_clear_cmd(hba, tag) < 0) {

[PATCH v5 27/28] ufs: core: Move code out of ufshcd_wait_for_dev_cmd()

[Bart Van Assche]

The ufshcd_dev_cmd_completion() call is useful for some but not for all
ufshcd_wait_for_dev_cmd() callers. Hence, remove the
ufshcd_dev_cmd_completion() call from ufshcd_wait_for_dev_cmd() and move
it past the ufshcd_issue_dev_cmd() calls where appropriate. This makes
it easier to detect timeout errors for UPIU frames submitted through the
BSG interface.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufshcd.c | 21 ++++++++++++---------
 1 file changed, 12 insertions(+), 9 deletions(-)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index a3c4f8cb2f37..b6ee8bfbb718 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -3262,8 +3262,6 @@ static int ufshcd_wait_for_dev_cmd(struct ufs_hba *hba,
 
 	if (likely(time_left)) {
 		err = ufshcd_get_tr_ocs(lrbp, NULL);
-		if (!err)
-			err = ufshcd_dev_cmd_completion(hba, lrbp);
 	} else {
 		err = -ETIMEDOUT;
 		dev_dbg(hba->dev, "%s: dev_cmd request timedout, tag %d\n",
@@ -3372,6 +3370,7 @@ static int ufshcd_exec_dev_cmd(struct ufs_hba *hba,
 {
 	const u32 tag = hba->reserved_slot;
 	struct scsi_cmnd *cmd = ufshcd_tag_to_cmd(hba, tag);
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	int err;
 
 	/* Protects use of hba->reserved_slot. */
@@ -3381,7 +3380,11 @@ static int ufshcd_exec_dev_cmd(struct ufs_hba *hba,
 	if (unlikely(err))
 		return err;
 
-	return ufshcd_issue_dev_cmd(hba, cmd, tag, timeout);
+	err = ufshcd_issue_dev_cmd(hba, cmd, tag, timeout);
+	if (err)
+		return err;
+
+	return ufshcd_dev_cmd_completion(hba, lrbp);
 }
 
 /**
@@ -7384,12 +7387,9 @@ static int ufshcd_issue_devman_upiu_cmd(struct ufs_hba *hba,
 
 	memset(lrbp->ucd_rsp_ptr, 0, sizeof(struct utp_upiu_rsp));
 
-	/*
-	 * ignore the returning value here - ufshcd_check_query_response is
-	 * bound to fail since dev_cmd.query and dev_cmd.type were left empty.
-	 * read the response directly ignoring all errors.
-	 */
-	ufshcd_issue_dev_cmd(hba, cmd, tag, dev_cmd_timeout);
+	err = ufshcd_issue_dev_cmd(hba, cmd, tag, dev_cmd_timeout);
+	if (err)
+		return err;
 
 	/* just copy the upiu response as it is */
 	memcpy(rsp_upiu, lrbp->ucd_rsp_ptr, sizeof(*rsp_upiu));
@@ -7534,7 +7534,10 @@ int ufshcd_advanced_rpmb_req_handler(struct ufs_hba *hba, struct utp_upiu_req *r
 	memset(lrbp->ucd_rsp_ptr, 0, sizeof(struct utp_upiu_rsp));
 
 	err = ufshcd_issue_dev_cmd(hba, cmd, tag, ADVANCED_RPMB_REQ_TIMEOUT);
+	if (err)
+		return err;
 
+	err = ufshcd_dev_cmd_completion(hba, lrbp);
 	if (!err) {
 		/* Just copy the upiu response as it is */
 		memcpy(rsp_upiu, lrbp->ucd_rsp_ptr, sizeof(*rsp_upiu));

RE: [PATCH v5 27/28] ufs: core: Move code out of ufshcd_wait_for_dev_cmd()

[Avri Altman]

> The ufshcd_dev_cmd_completion() call is useful for some but not for all
> ufshcd_wait_for_dev_cmd() callers. Hence, remove the
> ufshcd_dev_cmd_completion() call from ufshcd_wait_for_dev_cmd() and
> move it past the ufshcd_issue_dev_cmd() calls where appropriate. This makes
> it easier to detect timeout errors for UPIU frames submitted through the BSG
> interface.
> 
> Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Avri Altman <avri.altman@sandisk.com>

> ---
>  drivers/ufs/core/ufshcd.c | 21 ++++++++++++---------
>  1 file changed, 12 insertions(+), 9 deletions(-)
> 
> diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c index
> a3c4f8cb2f37..b6ee8bfbb718 100644
> --- a/drivers/ufs/core/ufshcd.c
> +++ b/drivers/ufs/core/ufshcd.c
> @@ -3262,8 +3262,6 @@ static int ufshcd_wait_for_dev_cmd(struct ufs_hba
> *hba,
> 
>         if (likely(time_left)) {
>                 err = ufshcd_get_tr_ocs(lrbp, NULL);
> -               if (!err)
> -                       err = ufshcd_dev_cmd_completion(hba, lrbp);
>         } else {
>                 err = -ETIMEDOUT;
>                 dev_dbg(hba->dev, "%s: dev_cmd request timedout, tag %d\n", @@ -
> 3372,6 +3370,7 @@ static int ufshcd_exec_dev_cmd(struct ufs_hba *hba,  {
>         const u32 tag = hba->reserved_slot;
>         struct scsi_cmnd *cmd = ufshcd_tag_to_cmd(hba, tag);
> +       struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
>         int err;
> 
>         /* Protects use of hba->reserved_slot. */ @@ -3381,7 +3380,11 @@ static
> int ufshcd_exec_dev_cmd(struct ufs_hba *hba,
>         if (unlikely(err))
>                 return err;
> 
> -       return ufshcd_issue_dev_cmd(hba, cmd, tag, timeout);
> +       err = ufshcd_issue_dev_cmd(hba, cmd, tag, timeout);
> +       if (err)
> +               return err;
> +
> +       return ufshcd_dev_cmd_completion(hba, lrbp);
>  }
> 
>  /**
> @@ -7384,12 +7387,9 @@ static int ufshcd_issue_devman_upiu_cmd(struct
> ufs_hba *hba,
> 
>         memset(lrbp->ucd_rsp_ptr, 0, sizeof(struct utp_upiu_rsp));
> 
> -       /*
> -        * ignore the returning value here - ufshcd_check_query_response is
> -        * bound to fail since dev_cmd.query and dev_cmd.type were left empty.
> -        * read the response directly ignoring all errors.
> -        */
> -       ufshcd_issue_dev_cmd(hba, cmd, tag, dev_cmd_timeout);
> +       err = ufshcd_issue_dev_cmd(hba, cmd, tag, dev_cmd_timeout);
> +       if (err)
> +               return err;
> 
>         /* just copy the upiu response as it is */
>         memcpy(rsp_upiu, lrbp->ucd_rsp_ptr, sizeof(*rsp_upiu)); @@ -7534,7
> +7534,10 @@ int ufshcd_advanced_rpmb_req_handler(struct ufs_hba *hba,
> struct utp_upiu_req *r
>         memset(lrbp->ucd_rsp_ptr, 0, sizeof(struct utp_upiu_rsp));
> 
>         err = ufshcd_issue_dev_cmd(hba, cmd, tag,
> ADVANCED_RPMB_REQ_TIMEOUT);
> +       if (err)
> +               return err;
> 
> +       err = ufshcd_dev_cmd_completion(hba, lrbp);
>         if (!err) {
>                 /* Just copy the upiu response as it is */
>                 memcpy(rsp_upiu, lrbp->ucd_rsp_ptr, sizeof(*rsp_upiu));

[PATCH v5 28/28] ufs: core: Switch to scsi_get_internal_cmd()

[Bart Van Assche]

Instead of storing the tag of the reserved command in hba->reserved_slot,
use scsi_get_internal_cmd() and scsi_put_internal_cmd() to allocate the
tag for the reserved command dynamically. Add
ufshcd_queue_reserved_command() for submitting reserved commands. Add
support in ufshcd_abort() for device management commands. Use
blk_execute_rq() for submitting reserved commands. Remove the code and
data structures that became superfluous. This includes
ufshcd_wait_for_dev_cmd(), hba->reserved_slot and ufs_dev_cmd.complete.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/ufs/core/ufs-mcq.c     |  19 ++-
 drivers/ufs/core/ufshcd-priv.h |  25 +---
 drivers/ufs/core/ufshcd.c      | 224 ++++++++++++++++-----------------
 include/ufs/ufshcd.h           |   6 -
 4 files changed, 119 insertions(+), 155 deletions(-)

diff --git a/drivers/ufs/core/ufs-mcq.c b/drivers/ufs/core/ufs-mcq.c
index c6c6cca400de..9303687e38a8 100644
--- a/drivers/ufs/core/ufs-mcq.c
+++ b/drivers/ufs/core/ufs-mcq.c
@@ -471,9 +471,6 @@ int ufshcd_mcq_init(struct ufs_hba *hba)
 		mutex_init(&hwq->sq_mutex);
 	}
 
-	/* The very first HW queue serves device commands */
-	hba->dev_cmd_queue = &hba->uhq[0];
-
 	host->host_tagset = 1;
 	return 0;
 }
@@ -528,6 +525,7 @@ int ufshcd_mcq_sq_cleanup(struct ufs_hba *hba, int task_tag)
 {
 	struct scsi_cmnd *cmd = ufshcd_tag_to_cmd(hba, task_tag);
 	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
+	struct request *rq = scsi_cmd_to_rq(cmd);
 	struct ufs_hw_queue *hwq;
 	void __iomem *reg, *opr_sqd_base;
 	u32 nexus, id, val;
@@ -536,15 +534,12 @@ int ufshcd_mcq_sq_cleanup(struct ufs_hba *hba, int task_tag)
 	if (hba->quirks & UFSHCD_QUIRK_MCQ_BROKEN_RTC)
 		return -ETIMEDOUT;
 
-	if (task_tag != hba->reserved_slot) {
-		if (!cmd)
-			return -EINVAL;
-		hwq = ufshcd_mcq_req_to_hwq(hba, scsi_cmd_to_rq(cmd));
-		if (!hwq)
-			return 0;
-	} else {
-		hwq = hba->dev_cmd_queue;
-	}
+	if (!cmd)
+		return -EINVAL;
+
+	hwq = ufshcd_mcq_req_to_hwq(hba, rq);
+	if (!hwq)
+		return 0;
 
 	id = hwq->id;
 
diff --git a/drivers/ufs/core/ufshcd-priv.h b/drivers/ufs/core/ufshcd-priv.h
index 3222c4d3ceb4..35c3277a4373 100644
--- a/drivers/ufs/core/ufshcd-priv.h
+++ b/drivers/ufs/core/ufshcd-priv.h
@@ -367,30 +367,7 @@ static inline bool ufs_is_valid_unit_desc_lun(struct ufs_dev_info *dev_info, u8
 static inline struct scsi_cmnd *ufshcd_tag_to_cmd(struct ufs_hba *hba, u32 tag)
 {
 	struct blk_mq_tags *tags = hba->host->tag_set.shared_tags;
-	struct request *rq;
-
-	/*
-	 * Handle reserved tags differently because the UFS driver does not
-	 * call blk_mq_alloc_request() for allocating reserved requests.
-	 * Allocating reserved tags with blk_mq_alloc_request() would require
-	 * the following:
-	 * - Allocate an additional request queue from &hba->host->tag_set for
-	 *   allocating reserved requests from.
-	 * - For that request queue, allocate a SCSI device.
-	 * - Calling blk_mq_alloc_request(hba->dev_mgmt_queue, REQ_OP_DRV_OUT,
-	 *   BLK_MQ_REQ_RESERVED) for allocating a reserved request and
-	 *   blk_mq_free_request() for freeing reserved requests.
-	 * - Set the .device pointer for these reserved requests.
-	 * - Submit reserved requests with blk_execute_rq().
-	 * - Modify ufshcd_queuecommand() such that it handles reserved requests
-	 *   in another way than SCSI requests.
-	 * - Modify ufshcd_compl_one_cqe() such that it calls scsi_done() for
-	 *   device management commands.
-	 * - Modify all callback functions called by blk_mq_tagset_busy_iter()
-	 *   calls in the UFS driver and skip device management commands.
-	 */
-	rq = tag < UFSHCD_NUM_RESERVED ? tags->static_rqs[tag] :
-					 blk_mq_tag_to_rq(tags, tag);
+	struct request *rq = blk_mq_tag_to_rq(tags, tag);
 
 	if (WARN_ON_ONCE(!rq))
 		return NULL;
diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index b6ee8bfbb718..c7206de67c73 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -2355,7 +2355,7 @@ static void ufshcd_update_monitor(struct ufs_hba *hba, struct scsi_cmnd *cmd)
  */
 static bool ufshcd_is_scsi_cmd(struct scsi_cmnd *cmd)
 {
-	return blk_mq_request_started(scsi_cmd_to_rq(cmd));
+	return !blk_mq_is_reserved_rq(scsi_cmd_to_rq(cmd));
 }
 
 /**
@@ -2486,7 +2486,6 @@ static inline int ufshcd_hba_capabilities(struct ufs_hba *hba)
 	hba->nutrs = (hba->capabilities & MASK_TRANSFER_REQUESTS_SLOTS_SDB) + 1;
 	hba->nutmrs =
 	((hba->capabilities & MASK_TASK_MANAGEMENT_REQUEST_SLOTS) >> 16) + 1;
-	hba->reserved_slot = 0;
 
 	hba->nortt = FIELD_GET(MASK_NUMBER_OUTSTANDING_RTT, hba->capabilities) + 1;
 
@@ -3114,6 +3113,20 @@ static int ufshcd_queuecommand(struct Scsi_Host *host, struct scsi_cmnd *cmd)
 	return err;
 }
 
+static int ufshcd_queue_reserved_command(struct Scsi_Host *host,
+					 struct scsi_cmnd *cmd)
+{
+	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
+	struct request *rq = scsi_cmd_to_rq(cmd);
+	struct ufs_hba *hba = shost_priv(host);
+	struct ufs_hw_queue *hwq =
+		hba->mcq_enabled ? ufshcd_mcq_req_to_hwq(hba, rq) : NULL;
+
+	ufshcd_add_query_upiu_trace(hba, UFS_QUERY_SEND, lrbp->ucd_req_ptr);
+	ufshcd_send_command(hba, cmd, hwq);
+	return 0;
+}
+
 static void ufshcd_setup_dev_cmd(struct ufs_hba *hba, struct scsi_cmnd *cmd,
 				 enum dev_cmd_type cmd_type, u8 lun, int tag)
 {
@@ -3243,84 +3256,6 @@ ufshcd_dev_cmd_completion(struct ufs_hba *hba, struct ufshcd_lrb *lrbp)
 	return err;
 }
 
-/*
- * Return: 0 upon success; < 0 upon failure.
- */
-static int ufshcd_wait_for_dev_cmd(struct ufs_hba *hba,
-		struct ufshcd_lrb *lrbp, int max_timeout)
-{
-	struct scsi_cmnd *cmd = (struct scsi_cmnd *)lrbp - 1;
-	const int tag = scsi_cmd_to_rq(cmd)->tag;
-	unsigned long time_left = msecs_to_jiffies(max_timeout);
-	unsigned long flags;
-	bool pending;
-	int err;
-
-retry:
-	time_left = wait_for_completion_timeout(&hba->dev_cmd.complete,
-						time_left);
-
-	if (likely(time_left)) {
-		err = ufshcd_get_tr_ocs(lrbp, NULL);
-	} else {
-		err = -ETIMEDOUT;
-		dev_dbg(hba->dev, "%s: dev_cmd request timedout, tag %d\n",
-			__func__, tag);
-
-		/* MCQ mode */
-		if (hba->mcq_enabled) {
-			/* successfully cleared the command, retry if needed */
-			if (ufshcd_clear_cmd(hba, tag) == 0)
-				err = -EAGAIN;
-			return err;
-		}
-
-		/* SDB mode */
-		if (ufshcd_clear_cmd(hba, tag) == 0) {
-			/* successfully cleared the command, retry if needed */
-			err = -EAGAIN;
-			/*
-			 * Since clearing the command succeeded we also need to
-			 * clear the task tag bit from the outstanding_reqs
-			 * variable.
-			 */
-			spin_lock_irqsave(&hba->outstanding_lock, flags);
-			pending = test_bit(tag, &hba->outstanding_reqs);
-			if (pending)
-				__clear_bit(tag, &hba->outstanding_reqs);
-			spin_unlock_irqrestore(&hba->outstanding_lock, flags);
-
-			if (!pending) {
-				/*
-				 * The completion handler ran while we tried to
-				 * clear the command.
-				 */
-				time_left = 1;
-				goto retry;
-			}
-		} else {
-			dev_err(hba->dev, "%s: failed to clear tag %d\n",
-				__func__, tag);
-
-			spin_lock_irqsave(&hba->outstanding_lock, flags);
-			pending = test_bit(tag, &hba->outstanding_reqs);
-			spin_unlock_irqrestore(&hba->outstanding_lock, flags);
-
-			if (!pending) {
-				/*
-				 * The completion handler ran while we tried to
-				 * clear the command.
-				 */
-				time_left = 1;
-				goto retry;
-			}
-		}
-	}
-
-	WARN_ONCE(err > 0, "Incorrect return value %d > 0\n", err);
-	return err;
-}
-
 static void ufshcd_dev_man_lock(struct ufs_hba *hba)
 {
 	ufshcd_hold(hba);
@@ -3335,23 +3270,45 @@ static void ufshcd_dev_man_unlock(struct ufs_hba *hba)
 	ufshcd_release(hba);
 }
 
+static struct scsi_cmnd *ufshcd_get_dev_mgmt_cmd(struct ufs_hba *hba)
+{
+	/*
+	 * The caller must hold this lock to guarantee that the NOWAIT
+	 * allocation will succeed.
+	 */
+	lockdep_assert_held(&hba->dev_cmd.lock);
+
+	/*
+	 * Allocate from hardware queue 0 instead of the hardware queue
+	 * corresponding to the current CPU because some UFSHCI controllers
+	 * only support hardware queue 0 for device management commands.
+	 */
+	return scsi_get_internal_cmd_hctx(
+		hba->host->pseudo_sdev, DMA_TO_DEVICE,
+		BLK_MQ_REQ_RESERVED | BLK_MQ_REQ_NOWAIT, 0);
+}
+
+static void ufshcd_put_dev_mgmt_cmd(struct scsi_cmnd *cmd)
+{
+	scsi_put_internal_cmd(cmd);
+}
+
 /*
- * Return: 0 upon success; < 0 upon failure.
+ * Return: 0 upon success; < 0 upon timeout; > 0 in case the UFS device
+ * reported an OCS error.
  */
 static int ufshcd_issue_dev_cmd(struct ufs_hba *hba, struct scsi_cmnd *cmd,
 				const u32 tag, int timeout)
 {
 	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
-	int err;
-
-	ufshcd_add_query_upiu_trace(hba, UFS_QUERY_SEND, lrbp->ucd_req_ptr);
-	ufshcd_send_command(hba, cmd, hba->dev_cmd_queue);
-	err = ufshcd_wait_for_dev_cmd(hba, lrbp, timeout);
-
-	ufshcd_add_query_upiu_trace(hba, err ? UFS_QUERY_ERR : UFS_QUERY_COMP,
-				    (struct utp_upiu_req *)lrbp->ucd_rsp_ptr);
+	struct request *rq = scsi_cmd_to_rq(cmd);
+	blk_status_t sts;
 
-	return err;
+	rq->timeout = timeout;
+	sts = blk_execute_rq(rq, true);
+	if (sts != BLK_STS_OK)
+		return blk_status_to_errno(sts);
+	return lrbp->utr_descriptor_ptr->header.ocs;
 }
 
 /**
@@ -3360,7 +3317,8 @@ static int ufshcd_issue_dev_cmd(struct ufs_hba *hba, struct scsi_cmnd *cmd,
  * @cmd_type: specifies the type (NOP, Query...)
  * @timeout: timeout in milliseconds
  *
- * Return: 0 upon success; < 0 upon failure.
+ * Return: 0 upon success; < 0 upon timeout; > 0 in case the UFS device
+ * reported an OCS error.
  *
  * NOTE: Since there is only one available tag for device management commands,
  * it is expected you hold the hba->dev_cmd.lock mutex.
@@ -3368,23 +3326,31 @@ static int ufshcd_issue_dev_cmd(struct ufs_hba *hba, struct scsi_cmnd *cmd,
 static int ufshcd_exec_dev_cmd(struct ufs_hba *hba,
 		enum dev_cmd_type cmd_type, int timeout)
 {
-	const u32 tag = hba->reserved_slot;
-	struct scsi_cmnd *cmd = ufshcd_tag_to_cmd(hba, tag);
+	struct scsi_cmnd *cmd = ufshcd_get_dev_mgmt_cmd(hba);
 	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
+	u32 tag;
 	int err;
 
-	/* Protects use of hba->reserved_slot. */
+	/* Protects use of hba->dev_cmd. */
 	lockdep_assert_held(&hba->dev_cmd.lock);
 
+	if (WARN_ON_ONCE(!cmd))
+		return -ENOMEM;
+
+	tag = scsi_cmd_to_rq(cmd)->tag;
+
 	err = ufshcd_compose_dev_cmd(hba, cmd, cmd_type, tag);
 	if (unlikely(err))
-		return err;
+		goto out;
 
 	err = ufshcd_issue_dev_cmd(hba, cmd, tag, timeout);
-	if (err)
-		return err;
+	if (err == 0)
+		err = ufshcd_dev_cmd_completion(hba, lrbp);
+
+out:
+	ufshcd_put_dev_mgmt_cmd(cmd);
 
-	return ufshcd_dev_cmd_completion(hba, lrbp);
+	return err;
 }
 
 /**
@@ -5649,6 +5615,10 @@ void ufshcd_compl_one_cqe(struct ufs_hba *hba, int task_tag,
 	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	enum utp_ocs ocs;
 
+	if (WARN_ONCE(!cmd, "cqe->command_desc_base_addr = %#llx\n",
+		      le64_to_cpu(cqe->command_desc_base_addr)))
+		return;
+
 	if (hba->monitor.enabled) {
 		lrbp->compl_time_stamp = ktime_get();
 		lrbp->compl_time_stamp_local_clock = local_clock();
@@ -5659,15 +5629,21 @@ void ufshcd_compl_one_cqe(struct ufs_hba *hba, int task_tag,
 		ufshcd_add_command_trace(hba, cmd, UFS_CMD_COMP);
 		cmd->result = ufshcd_transfer_rsp_status(hba, cmd, cqe);
 		ufshcd_release_scsi_cmd(hba, cmd);
-		/* Do not touch lrbp after scsi done */
-		scsi_done(cmd);
 	} else {
 		if (cqe) {
 			ocs = le32_to_cpu(cqe->status) & MASK_OCS;
 			lrbp->utr_descriptor_ptr->header.ocs = ocs;
+		} else {
+			ocs = lrbp->utr_descriptor_ptr->header.ocs;
 		}
-		complete(&hba->dev_cmd.complete);
+		ufshcd_add_query_upiu_trace(
+			hba,
+			ocs == OCS_SUCCESS ? UFS_QUERY_COMP : UFS_QUERY_ERR,
+			(struct utp_upiu_req *)lrbp->ucd_rsp_ptr);
+		cmd->result = 0;
 	}
+	/* Do not touch lrbp after scsi_done() has been called. */
+	scsi_done(cmd);
 }
 
 /**
@@ -7358,15 +7334,20 @@ static int ufshcd_issue_devman_upiu_cmd(struct ufs_hba *hba,
 					enum dev_cmd_type cmd_type,
 					enum query_opcode desc_op)
 {
-	const u32 tag = hba->reserved_slot;
-	struct scsi_cmnd *cmd = ufshcd_tag_to_cmd(hba, tag);
+	struct scsi_cmnd *cmd = ufshcd_get_dev_mgmt_cmd(hba);
 	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
+	u32 tag;
 	int err = 0;
 	u8 upiu_flags;
 
-	/* Protects use of hba->reserved_slot. */
+	/* Protects use of hba->dev_cmd. */
 	lockdep_assert_held(&hba->dev_cmd.lock);
 
+	if (WARN_ON_ONCE(!cmd))
+		return -ENOMEM;
+
+	tag = scsi_cmd_to_rq(cmd)->tag;
+
 	ufshcd_setup_dev_cmd(hba, cmd, cmd_type, 0, tag);
 
 	ufshcd_prepare_req_desc_hdr(hba, lrbp, &upiu_flags, DMA_NONE, 0);
@@ -7410,6 +7391,8 @@ static int ufshcd_issue_devman_upiu_cmd(struct ufs_hba *hba,
 		}
 	}
 
+	ufshcd_put_dev_mgmt_cmd(cmd);
+
 	return err;
 }
 
@@ -7502,9 +7485,9 @@ int ufshcd_advanced_rpmb_req_handler(struct ufs_hba *hba, struct utp_upiu_req *r
 			 struct ufs_ehs *rsp_ehs, int sg_cnt, struct scatterlist *sg_list,
 			 enum dma_data_direction dir)
 {
-	const u32 tag = hba->reserved_slot;
-	struct scsi_cmnd *cmd = ufshcd_tag_to_cmd(hba, tag);
-	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
+	struct scsi_cmnd *cmd;
+	struct ufshcd_lrb *lrbp;
+	u32 tag;
 	int err = 0;
 	int result;
 	u8 upiu_flags;
@@ -7512,9 +7495,18 @@ int ufshcd_advanced_rpmb_req_handler(struct ufs_hba *hba, struct utp_upiu_req *r
 	u16 ehs_len;
 	int ehs = (hba->capabilities & MASK_EHSLUTRD_SUPPORTED) ? 2 : 0;
 
-	/* Protects use of hba->reserved_slot. */
 	ufshcd_dev_man_lock(hba);
 
+	cmd = ufshcd_get_dev_mgmt_cmd(hba);
+
+	if (WARN_ON_ONCE(!cmd)) {
+		ufshcd_dev_man_unlock(hba);
+		return -ENOMEM;
+	}
+
+	lrbp = scsi_cmd_priv(cmd);
+	tag = scsi_cmd_to_rq(cmd)->tag;
+
 	ufshcd_setup_dev_cmd(hba, cmd, DEV_CMD_TYPE_RPMB, UFS_UPIU_RPMB_WLUN,
 			     tag);
 
@@ -7561,6 +7553,8 @@ int ufshcd_advanced_rpmb_req_handler(struct ufs_hba *hba, struct utp_upiu_req *r
 		}
 	}
 
+	ufshcd_put_dev_mgmt_cmd(cmd);
+
 	ufshcd_dev_man_unlock(hba);
 
 	return err ? : result;
@@ -7731,7 +7725,8 @@ static int ufshcd_abort(struct scsi_cmnd *cmd)
 {
 	struct Scsi_Host *host = cmd->device->host;
 	struct ufs_hba *hba = shost_priv(host);
-	int tag = scsi_cmd_to_rq(cmd)->tag;
+	struct request *rq = scsi_cmd_to_rq(cmd);
+	int tag = rq->tag;
 	struct ufshcd_lrb *lrbp = scsi_cmd_priv(cmd);
 	unsigned long flags;
 	int err = FAILED;
@@ -7761,7 +7756,8 @@ static int ufshcd_abort(struct scsi_cmnd *cmd)
 	 * to reduce repeated printouts. For other aborted requests only print
 	 * basic details.
 	 */
-	scsi_print_command(cmd);
+	if (ufshcd_is_scsi_cmd(cmd))
+		scsi_print_command(cmd);
 	if (!hba->req_abort_count) {
 		ufshcd_update_evt_hist(hba, UFS_EVT_ABORT, tag);
 		ufshcd_print_evt_hist(hba);
@@ -7813,7 +7809,10 @@ static int ufshcd_abort(struct scsi_cmnd *cmd)
 		goto release;
 	}
 
-	err = ufshcd_try_to_abort_task(hba, tag);
+	if (blk_mq_is_reserved_rq(rq))
+		err = ufshcd_clear_cmd(hba, tag);
+	else
+		err = ufshcd_try_to_abort_task(hba, tag);
 	if (err) {
 		dev_err(hba->dev, "%s: failed with err %d\n", __func__, err);
 		ufshcd_set_req_abort_skip(hba, hba->outstanding_reqs);
@@ -9186,6 +9185,7 @@ static const struct scsi_host_template ufshcd_driver_template = {
 	.cmd_size		= sizeof(struct ufshcd_lrb),
 	.init_cmd_priv		= ufshcd_init_cmd_priv,
 	.queuecommand		= ufshcd_queuecommand,
+	.queue_reserved_command	= ufshcd_queue_reserved_command,
 	.mq_poll		= ufshcd_poll,
 	.sdev_init		= ufshcd_sdev_init,
 	.sdev_configure		= ufshcd_sdev_configure,
@@ -10728,8 +10728,6 @@ int ufshcd_init(struct ufs_hba *hba, void __iomem *mmio_base, unsigned int irq)
 						UFS_SLEEP_PWR_MODE,
 						UIC_LINK_HIBERN8_STATE);
 
-	init_completion(&hba->dev_cmd.complete);
-
 	err = ufshcd_hba_init(hba);
 	if (err)
 		goto out_error;
diff --git a/include/ufs/ufshcd.h b/include/ufs/ufshcd.h
index 9eb69241e40f..e07834509182 100644
--- a/include/ufs/ufshcd.h
+++ b/include/ufs/ufshcd.h
@@ -236,13 +236,11 @@ struct ufs_query {
  * struct ufs_dev_cmd - all assosiated fields with device management commands
  * @type: device management command type - Query, NOP OUT
  * @lock: lock to allow one command at a time
- * @complete: internal commands completion
  * @query: Device management query information
  */
 struct ufs_dev_cmd {
 	enum dev_cmd_type type;
 	struct mutex lock;
-	struct completion complete;
 	struct ufs_query query;
 };
 
@@ -838,7 +836,6 @@ enum ufshcd_mcq_opr {
  * @nutrs: Transfer Request Queue depth supported by controller
  * @nortt - Max outstanding RTTs supported by controller
  * @nutmrs: Task Management Queue depth supported by controller
- * @reserved_slot: Used to submit device commands. Protected by @dev_cmd.lock.
  * @ufs_version: UFS Version to which controller complies
  * @vops: pointer to variant specific operations
  * @vps: pointer to variant specific parameters
@@ -929,7 +926,6 @@ enum ufshcd_mcq_opr {
  * @res: array of resource info of MCQ registers
  * @mcq_base: Multi circular queue registers base address
  * @uhq: array of supported hardware queues
- * @dev_cmd_queue: Queue for issuing device management commands
  * @mcq_opr: MCQ operation and runtime registers
  * @ufs_rtc_update_work: A work for UFS RTC periodic update
  * @pm_qos_req: PM QoS request handle
@@ -980,7 +976,6 @@ struct ufs_hba {
 	int nortt;
 	u32 mcq_capabilities;
 	int nutmrs;
-	u32 reserved_slot;
 	u32 ufs_version;
 	const struct ufs_hba_variant_ops *vops;
 	struct ufs_hba_variant_params *vps;
@@ -1098,7 +1093,6 @@ struct ufs_hba {
 	bool mcq_esi_enabled;
 	void __iomem *mcq_base;
 	struct ufs_hw_queue *uhq;
-	struct ufs_hw_queue *dev_cmd_queue;
 	struct ufshcd_mcq_opr_info_t mcq_opr[OPR_MAX];
 
 	struct delayed_work ufs_rtc_update_work;

Re: [PATCH v5 00/28] Optimize the hot path in the UFS driver

[John Garry]

Did you notice this when you tested?

[    0.927347]   dev_size_mb=8, opts=0x0, submit_queues=1, statistics=0
[    0.930065] ------------[ cut here ]------------
[    0.931137] WARNING: CPU: 1 PID: 1 at drivers/scsi/scsi_scan.c:228
scsi_realloc_sdev_budget_map+0x163/0x180
[    0.933292] Modules linked in:
[    0.933948] CPU: 1 UID: 0 PID: 1 Comm: swapper/0 Not tainted
6.17.0-rc1-00143-g67ede8ca8ab5-dirty #1769 PREEMPT(voluntary)
[    0.936317] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009),
BIOS rel-1.16.3-0-ga6ed6b701f0a-prebuilt.qemu.org 04/01/2014
[    0.938683] RIP: 0010:scsi_realloc_sdev_budget_map+0x163/0x180
[    0.939955] Code: 45 39 ee 44 89 ee 48 8b 7c 24 28 41 0f 46 f6 e8
23 df be ff 48 8b 7c 24 08 e8 f9 99 85 ff 48 8b 3c 24 e8 90 ed 8a ff
eb ae 90 <0f> 0b 90 48 83 c4 30 b8 ea ff ff ff 5b 5d 41 5c 41 5d 41 5e
41 5f
[    0.943854] RSP: 0000:ffffb29bc00139c8 EFLAGS: 00010246
[    0.944950] RAX: 0000000000000080 RBX: ffff89bac0a53800 RCX: 
0000000000000005
[    0.946515] RDX: 0000000000000003 RSI: 0000000000000004 RDI: 
ffff89bac0a53800
[    0.948019] RBP: ffff89bac0a4e000 R08: 0000000000000000 R09: 
0000000000000001
[    0.949521] R10: 0000000000000000 R11: 0000000000000008 R12: 
00000000000000c0
[    0.951038] R13: 00000000000000c0 R14: ffff89bac3d09828 R15: 
0000000000000005
[    0.952537] FS:  0000000000000000(0000) GS:ffff89bb47fe3000(0000)
knlGS:0000000000000000
[    0.954258] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    0.955464] CR2: 0000000000000000 CR3: 0000000120e30001 CR4: 
0000000000370ef0
[    0.957001] Call Trace:
[    0.957547]  <TASK>
[    0.958087]  ? attribute_container_add_device+0x4b/0x130
[    0.959295]  scsi_alloc_sdev+0x27f/0x360
[    0.960191]  scsi_probe_and_add_lun+0x1e4/0x2f0
[    0.961203]  __scsi_scan_target+0x104/0x210
[    0.962157]  scsi_scan_channel+0x54/0x90
[    0.962942]  scsi_scan_host_selected+0xcf/0x100
[    0.964420]  scsi_scan_host+0x19f/0x1c0
[    0.965289]  sdebug_driver_probe+0x1c0/0x420
[    0.966317]  really_probe+0xba/0x2b0
[    0.967171]  ? __pfx___device_attach_driver+0x10/0x10
[    0.968308]  __driver_probe_device+0x6e/0x110
[    0.969286]  driver_probe_device+0x1a/0xe0
[    0.970232]  __device_attach_driver+0x7e/0x110
[    0.971276]  bus_for_each_drv+0x8f/0xe0
[    0.972183]  __device_attach+0xa9/0x1a0
[    0.973080]  bus_probe_device+0x88/0xa0
[    0.973973]  device_add+0x61c/0x810
[    0.974805]  sdebug_add_host_helper+0x1ca/0x260
[    0.975843]  scsi_debug_init+0x341/0x7d0
[    0.976770]  ? __pfx_scsi_debug_init+0x10/0x10
[    0.977801]  do_one_initcall+0x55/0x220
[    0.978706]  kernel_init_freeable+0x199/0x2e0
[    0.979744]  ? __pfx_kernel_init+0x10/0x10
[    0.980688]  kernel_init+0x15/0x1c0
[    0.981765]  ret_from_fork+0x80/0xd0
[    0.982710]  ? __pfx_kernel_init+0x10/0x10
[    0.983689]  ret_from_fork_asm+0x1a/0x30
[    0.984742]  </TASK>
[    0.985223] ---[ end trace 0000000000000000 ]---
[    0.986231] scsi_alloc_sdev: Allocation failure during SCSI
scanning, some SCSI devices might not be configured

Re: [PATCH v5 00/28] Optimize the hot path in the UFS driver

[Bart Van Assche]

On 9/26/25 3:08 AM, John Garry wrote:
> Did you notice this when you tested?
Yes, but I only noticed this a few minutes after I posted this patch
series. I queued up the two changes below for when I repost this patch
series. The first change suppresses the kernel warning from your email.
The second change ensures that ufshcd_get_hba_mac() only prints an error
message if it returns a negative value.

Thanks,

Bart.

diff --git a/drivers/scsi/scsi_scan.c b/drivers/scsi/scsi_scan.c
index 511cf84c2f89..5e7ef4655348 100644
--- a/drivers/scsi/scsi_scan.c
+++ b/drivers/scsi/scsi_scan.c
@@ -225,9 +225,6 @@ static int scsi_realloc_sdev_budget_map(struct 
scsi_device *sdev,
  	int ret;
  	struct sbitmap sb_backup;

-	if (WARN_ON_ONCE(!sdev->budget_map.map))
-		return -EINVAL;
-
  	depth = min_t(unsigned int, depth, scsi_device_max_queue_depth(sdev));

  	/*
diff --git a/drivers/ufs/core/ufs-mcq.c b/drivers/ufs/core/ufs-mcq.c
index 9303687e38a8..1fa0f14dd291 100644
--- a/drivers/ufs/core/ufs-mcq.c
+++ b/drivers/ufs/core/ufs-mcq.c
@@ -156,10 +160,7 @@ int ufshcd_get_hba_mac(struct ufs_hba *hba)
  		mac = hba->vops->get_hba_mac(hba);
  	}
  	if (mac < 0)
-		goto err;
-
-err:
-	dev_err(hba->dev, "Failed to get mac, err=%d\n", mac);
+		dev_err(hba->dev, "Failed to get mac, err=%d\n", mac);
  	return mac;
  }