Merge tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi

Pull SCSI updates from James Bottomley:
 "This consists of the usual driver updates (ufs, target, tcmu,
  smartpqi, lpfc, zfcp, qla2xxx, mpt3sas, pm80xx).

  The major core change is using a sbitmap instead of an atomic for
  queue tracking"

* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (412 commits)
  scsi: target: tcm_fc: Fix a kernel-doc header
  scsi: target: Shorten ALUA error messages
  scsi: target: Fix two format specifiers
  scsi: target: Compare explicitly with SAM_STAT_GOOD
  scsi: sd: Introduce a new local variable in sd_check_events()
  scsi: dc395x: Open-code status_byte(u8) calls
  scsi: 53c700: Open-code status_byte(u8) calls
  scsi: smartpqi: Remove unused functions
  scsi: qla4xxx: Remove an unused function
  scsi: myrs: Remove unused functions
  scsi: myrb: Remove unused functions
  scsi: mpt3sas: Fix two kernel-doc headers
  scsi: fcoe: Suppress a compiler warning
  scsi: libfc: Fix a format specifier
  scsi: aacraid: Remove an unused function
  scsi: core: Introduce enum scsi_disposition
  scsi: core: Modify the scsi_send_eh_cmnd() return value for the SDEV_BLOCK case
  scsi: core: Rename scsi_softirq_done() into scsi_complete()
  scsi: core: Remove an incorrect comment
  scsi: core: Make the scsi_alloc_sgtables() documentation more accurate
  ...

1 files changed, 213 insertions, 87 deletions

diff --git a/drivers/target/target_core_transport.c b/drivers/target/target_core_transport.c
index 5ecb9f18a53d..8fbfe75c5744 100644
--- a/drivers/target/target_core_transport.c
+++ b/drivers/target/target_core_transport.c
@@ -41,6 +41,7 @@
 #include <trace/events/target.h>
 
 static struct workqueue_struct *target_completion_wq;
+static struct workqueue_struct *target_submission_wq;
 static struct kmem_cache *se_sess_cache;
 struct kmem_cache *se_ua_cache;
 struct kmem_cache *t10_pr_reg_cache;
@@ -129,8 +130,15 @@ int init_se_kmem_caches(void)
 	if (!target_completion_wq)
 		goto out_free_lba_map_mem_cache;
 
+	target_submission_wq = alloc_workqueue("target_submission",
+					       WQ_MEM_RECLAIM, 0);
+	if (!target_submission_wq)
+		goto out_free_completion_wq;
+
 	return 0;
 
+out_free_completion_wq:
+	destroy_workqueue(target_completion_wq);
 out_free_lba_map_mem_cache:
 	kmem_cache_destroy(t10_alua_lba_map_mem_cache);
 out_free_lba_map_cache:
@@ -153,6 +161,7 @@ out:
 
 void release_se_kmem_caches(void)
 {
+	destroy_workqueue(target_submission_wq);
 	destroy_workqueue(target_completion_wq);
 	kmem_cache_destroy(se_sess_cache);
 	kmem_cache_destroy(se_ua_cache);
@@ -848,7 +857,8 @@ static bool target_cmd_interrupted(struct se_cmd *cmd)
 /* May be called from interrupt context so must not sleep. */
 void target_complete_cmd(struct se_cmd *cmd, u8 scsi_status)
 {
-	int success;
+	struct se_wwn *wwn = cmd->se_sess->se_tpg->se_tpg_wwn;
+	int success, cpu;
 	unsigned long flags;
 
 	if (target_cmd_interrupted(cmd))
@@ -875,7 +885,13 @@ void target_complete_cmd(struct se_cmd *cmd, u8 scsi_status)
 
 	INIT_WORK(&cmd->work, success ? target_complete_ok_work :
 		  target_complete_failure_work);
-	queue_work_on(cmd->cpuid, target_completion_wq, &cmd->work);
+
+	if (wwn->cmd_compl_affinity == SE_COMPL_AFFINITY_CPUID)
+		cpu = cmd->cpuid;
+	else
+		cpu = wwn->cmd_compl_affinity;
+
+	queue_work_on(cpu, target_completion_wq, &cmd->work);
 }
 EXPORT_SYMBOL(target_complete_cmd);
 
@@ -1304,7 +1320,7 @@ target_check_max_data_sg_nents(struct se_cmd *cmd, struct se_device *dev,
  * Compare the data buffer size from the CDB with the data buffer limit from the transport
  * header. Set @cmd->residual_count and SCF_OVERFLOW_BIT or SCF_UNDERFLOW_BIT if necessary.
  *
- * Note: target drivers set @cmd->data_length by calling transport_init_se_cmd().
+ * Note: target drivers set @cmd->data_length by calling __target_init_cmd().
  *
  * Return: TCM_NO_SENSE
  */
@@ -1371,7 +1387,7 @@ target_cmd_size_check(struct se_cmd *cmd, unsigned int size)
  *
  * Preserves the value of @cmd->tag.
  */
-void transport_init_se_cmd(
+void __target_init_cmd(
 	struct se_cmd *cmd,
 	const struct target_core_fabric_ops *tfo,
 	struct se_session *se_sess,
@@ -1382,7 +1398,6 @@ void transport_init_se_cmd(
 {
 	INIT_LIST_HEAD(&cmd->se_delayed_node);
 	INIT_LIST_HEAD(&cmd->se_qf_node);
-	INIT_LIST_HEAD(&cmd->se_cmd_list);
 	INIT_LIST_HEAD(&cmd->state_list);
 	init_completion(&cmd->t_transport_stop_comp);
 	cmd->free_compl = NULL;
@@ -1391,6 +1406,7 @@ void transport_init_se_cmd(
 	INIT_WORK(&cmd->work, NULL);
 	kref_init(&cmd->cmd_kref);
 
+	cmd->t_task_cdb = &cmd->__t_task_cdb[0];
 	cmd->se_tfo = tfo;
 	cmd->se_sess = se_sess;
 	cmd->data_length = data_length;
@@ -1404,7 +1420,7 @@ void transport_init_se_cmd(
 
 	cmd->state_active = false;
 }
-EXPORT_SYMBOL(transport_init_se_cmd);
+EXPORT_SYMBOL(__target_init_cmd);
 
 static sense_reason_t
 transport_check_alloc_task_attr(struct se_cmd *cmd)
@@ -1428,11 +1444,10 @@ transport_check_alloc_task_attr(struct se_cmd *cmd)
 }
 
 sense_reason_t
-target_cmd_init_cdb(struct se_cmd *cmd, unsigned char *cdb)
+target_cmd_init_cdb(struct se_cmd *cmd, unsigned char *cdb, gfp_t gfp)
 {
 	sense_reason_t ret;
 
-	cmd->t_task_cdb = &cmd->__t_task_cdb[0];
 	/*
 	 * Ensure that the received CDB is less than the max (252 + 8) bytes
 	 * for VARIABLE_LENGTH_CMD
@@ -1450,8 +1465,7 @@ target_cmd_init_cdb(struct se_cmd *cmd, unsigned char *cdb)
 	 * setup the pointer from __t_task_cdb to t_task_cdb.
 	 */
 	if (scsi_command_size(cdb) > sizeof(cmd->__t_task_cdb)) {
-		cmd->t_task_cdb = kzalloc(scsi_command_size(cdb),
-						GFP_KERNEL);
+		cmd->t_task_cdb = kzalloc(scsi_command_size(cdb), gfp);
 		if (!cmd->t_task_cdb) {
 			pr_err("Unable to allocate cmd->t_task_cdb"
 				" %u > sizeof(cmd->__t_task_cdb): %lu ops\n",
@@ -1573,46 +1587,31 @@ transport_generic_map_mem_to_cmd(struct se_cmd *cmd, struct scatterlist *sgl,
 }
 
 /**
- * target_submit_cmd_map_sgls - lookup unpacked lun and submit uninitialized
- * 			 se_cmd + use pre-allocated SGL memory.
- *
- * @se_cmd: command descriptor to submit
+ * target_init_cmd - initialize se_cmd
+ * @se_cmd: command descriptor to init
  * @se_sess: associated se_sess for endpoint
- * @cdb: pointer to SCSI CDB
  * @sense: pointer to SCSI sense buffer
  * @unpacked_lun: unpacked LUN to reference for struct se_lun
  * @data_length: fabric expected data transfer length
  * @task_attr: SAM task attribute
  * @data_dir: DMA data direction
  * @flags: flags for command submission from target_sc_flags_tables
- * @sgl: struct scatterlist memory for unidirectional mapping
- * @sgl_count: scatterlist count for unidirectional mapping
- * @sgl_bidi: struct scatterlist memory for bidirectional READ mapping
- * @sgl_bidi_count: scatterlist count for bidirectional READ mapping
- * @sgl_prot: struct scatterlist memory protection information
- * @sgl_prot_count: scatterlist count for protection information
  *
  * Task tags are supported if the caller has set @se_cmd->tag.
  *
- * Returns non zero to signal active I/O shutdown failure.  All other
- * setup exceptions will be returned as a SCSI CHECK_CONDITION response,
- * but still return zero here.
+ * Returns:
+ *	- less than zero to signal active I/O shutdown failure.
+ *	- zero on success.
  *
- * This may only be called from process context, and also currently
- * assumes internal allocation of fabric payload buffer by target-core.
+ * If the fabric driver calls target_stop_session, then it must check the
+ * return code and handle failures. This will never fail for other drivers,
+ * and the return code can be ignored.
  */
-int target_submit_cmd_map_sgls(struct se_cmd *se_cmd, struct se_session *se_sess,
-		unsigned char *cdb, unsigned char *sense, u64 unpacked_lun,
-		u32 data_length, int task_attr, int data_dir, int flags,
-		struct scatterlist *sgl, u32 sgl_count,
-		struct scatterlist *sgl_bidi, u32 sgl_bidi_count,
-		struct scatterlist *sgl_prot, u32 sgl_prot_count)
+int target_init_cmd(struct se_cmd *se_cmd, struct se_session *se_sess,
+		    unsigned char *sense, u64 unpacked_lun,
+		    u32 data_length, int task_attr, int data_dir, int flags)
 {
 	struct se_portal_group *se_tpg;
-	sense_reason_t rc;
-	int ret;
-
-	might_sleep();
 
 	se_tpg = se_sess->se_tpg;
 	BUG_ON(!se_tpg);
@@ -1621,52 +1620,71 @@ int target_submit_cmd_map_sgls(struct se_cmd *se_cmd, struct se_session *se_sess
 	if (flags & TARGET_SCF_USE_CPUID)
 		se_cmd->se_cmd_flags |= SCF_USE_CPUID;
 	/*
+	 * Signal bidirectional data payloads to target-core
+	 */
+	if (flags & TARGET_SCF_BIDI_OP)
+		se_cmd->se_cmd_flags |= SCF_BIDI;
+
+	if (flags & TARGET_SCF_UNKNOWN_SIZE)
+		se_cmd->unknown_data_length = 1;
+	/*
 	 * Initialize se_cmd for target operation.  From this point
 	 * exceptions are handled by sending exception status via
 	 * target_core_fabric_ops->queue_status() callback
 	 */
-	transport_init_se_cmd(se_cmd, se_tpg->se_tpg_tfo, se_sess,
-				data_length, data_dir, task_attr, sense,
-				unpacked_lun);
+	__target_init_cmd(se_cmd, se_tpg->se_tpg_tfo, se_sess, data_length,
+			  data_dir, task_attr, sense, unpacked_lun);
 
-	if (flags & TARGET_SCF_UNKNOWN_SIZE)
-		se_cmd->unknown_data_length = 1;
 	/*
 	 * Obtain struct se_cmd->cmd_kref reference. A second kref_get here is
 	 * necessary for fabrics using TARGET_SCF_ACK_KREF that expect a second
 	 * kref_put() to happen during fabric packet acknowledgement.
 	 */
-	ret = target_get_sess_cmd(se_cmd, flags & TARGET_SCF_ACK_KREF);
-	if (ret)
-		return ret;
-	/*
-	 * Signal bidirectional data payloads to target-core
-	 */
-	if (flags & TARGET_SCF_BIDI_OP)
-		se_cmd->se_cmd_flags |= SCF_BIDI;
+	return target_get_sess_cmd(se_cmd, flags & TARGET_SCF_ACK_KREF);
+}
+EXPORT_SYMBOL_GPL(target_init_cmd);
 
-	rc = target_cmd_init_cdb(se_cmd, cdb);
-	if (rc) {
-		transport_send_check_condition_and_sense(se_cmd, rc, 0);
-		target_put_sess_cmd(se_cmd);
-		return 0;
-	}
+/**
+ * target_submit_prep - prepare cmd for submission
+ * @se_cmd: command descriptor to prep
+ * @cdb: pointer to SCSI CDB
+ * @sgl: struct scatterlist memory for unidirectional mapping
+ * @sgl_count: scatterlist count for unidirectional mapping
+ * @sgl_bidi: struct scatterlist memory for bidirectional READ mapping
+ * @sgl_bidi_count: scatterlist count for bidirectional READ mapping
+ * @sgl_prot: struct scatterlist memory protection information
+ * @sgl_prot_count: scatterlist count for protection information
+ * @gfp: gfp allocation type
+ *
+ * Returns:
+ *	- less than zero to signal failure.
+ *	- zero on success.
+ *
+ * If failure is returned, lio will the callers queue_status to complete
+ * the cmd.
+ */
+int target_submit_prep(struct se_cmd *se_cmd, unsigned char *cdb,
+		       struct scatterlist *sgl, u32 sgl_count,
+		       struct scatterlist *sgl_bidi, u32 sgl_bidi_count,
+		       struct scatterlist *sgl_prot, u32 sgl_prot_count,
+		       gfp_t gfp)
+{
+	sense_reason_t rc;
+
+	rc = target_cmd_init_cdb(se_cmd, cdb, gfp);
+	if (rc)
+		goto send_cc_direct;
 
 	/*
 	 * Locate se_lun pointer and attach it to struct se_cmd
 	 */
 	rc = transport_lookup_cmd_lun(se_cmd);
-	if (rc) {
-		transport_send_check_condition_and_sense(se_cmd, rc, 0);
-		target_put_sess_cmd(se_cmd);
-		return 0;
-	}
+	if (rc)
+		goto send_cc_direct;
 
 	rc = target_cmd_parse_cdb(se_cmd);
-	if (rc != 0) {
-		transport_generic_request_failure(se_cmd, rc);
-		return 0;
-	}
+	if (rc != 0)
+		goto generic_fail;
 
 	/*
 	 * Save pointers for SGLs containing protection information,
@@ -1686,6 +1704,41 @@ int target_submit_cmd_map_sgls(struct se_cmd *se_cmd, struct se_session *se_sess
 	if (sgl_count != 0) {
 		BUG_ON(!sgl);
 
+		rc = transport_generic_map_mem_to_cmd(se_cmd, sgl, sgl_count,
+				sgl_bidi, sgl_bidi_count);
+		if (rc != 0)
+			goto generic_fail;
+	}
+
+	return 0;
+
+send_cc_direct:
+	transport_send_check_condition_and_sense(se_cmd, rc, 0);
+	target_put_sess_cmd(se_cmd);
+	return -EIO;
+
+generic_fail:
+	transport_generic_request_failure(se_cmd, rc);
+	return -EIO;
+}
+EXPORT_SYMBOL_GPL(target_submit_prep);
+
+/**
+ * target_submit - perform final initialization and submit cmd to LIO core
+ * @se_cmd: command descriptor to submit
+ *
+ * target_submit_prep must have been called on the cmd, and this must be
+ * called from process context.
+ */
+void target_submit(struct se_cmd *se_cmd)
+{
+	struct scatterlist *sgl = se_cmd->t_data_sg;
+	unsigned char *buf = NULL;
+
+	might_sleep();
+
+	if (se_cmd->t_data_nents != 0) {
+		BUG_ON(!sgl);
 		/*
 		 * A work-around for tcm_loop as some userspace code via
 		 * scsi-generic do not memset their associated read buffers,
@@ -1696,8 +1749,6 @@ int target_submit_cmd_map_sgls(struct se_cmd *se_cmd, struct se_session *se_sess
 		 */
 		if (!(se_cmd->se_cmd_flags & SCF_SCSI_DATA_CDB) &&
 		     se_cmd->data_direction == DMA_FROM_DEVICE) {
-			unsigned char *buf = NULL;
-
 			if (sgl)
 				buf = kmap(sg_page(sgl)) + sgl->offset;
 
@@ -1707,12 +1758,6 @@ int target_submit_cmd_map_sgls(struct se_cmd *se_cmd, struct se_session *se_sess
 			}
 		}
 
-		rc = transport_generic_map_mem_to_cmd(se_cmd, sgl, sgl_count,
-				sgl_bidi, sgl_bidi_count);
-		if (rc != 0) {
-			transport_generic_request_failure(se_cmd, rc);
-			return 0;
-		}
 	}
 
 	/*
@@ -1722,9 +1767,8 @@ int target_submit_cmd_map_sgls(struct se_cmd *se_cmd, struct se_session *se_sess
 	core_alua_check_nonop_delay(se_cmd);
 
 	transport_handle_cdb_direct(se_cmd);
-	return 0;
 }
-EXPORT_SYMBOL(target_submit_cmd_map_sgls);
+EXPORT_SYMBOL_GPL(target_submit);
 
 /**
  * target_submit_cmd - lookup unpacked lun and submit uninitialized se_cmd
@@ -1741,25 +1785,109 @@ EXPORT_SYMBOL(target_submit_cmd_map_sgls);
  *
  * Task tags are supported if the caller has set @se_cmd->tag.
  *
- * Returns non zero to signal active I/O shutdown failure.  All other
- * setup exceptions will be returned as a SCSI CHECK_CONDITION response,
- * but still return zero here.
- *
  * This may only be called from process context, and also currently
  * assumes internal allocation of fabric payload buffer by target-core.
  *
  * It also assumes interal target core SGL memory allocation.
+ *
+ * This function must only be used by drivers that do their own
+ * sync during shutdown and does not use target_stop_session. If there
+ * is a failure this function will call into the fabric driver's
+ * queue_status with a CHECK_CONDITION.
  */
-int target_submit_cmd(struct se_cmd *se_cmd, struct se_session *se_sess,
+void target_submit_cmd(struct se_cmd *se_cmd, struct se_session *se_sess,
 		unsigned char *cdb, unsigned char *sense, u64 unpacked_lun,
 		u32 data_length, int task_attr, int data_dir, int flags)
 {
-	return target_submit_cmd_map_sgls(se_cmd, se_sess, cdb, sense,
-			unpacked_lun, data_length, task_attr, data_dir,
-			flags, NULL, 0, NULL, 0, NULL, 0);
+	int rc;
+
+	rc = target_init_cmd(se_cmd, se_sess, sense, unpacked_lun, data_length,
+			     task_attr, data_dir, flags);
+	WARN(rc, "Invalid target_submit_cmd use. Driver must not use target_stop_session or call target_init_cmd directly.\n");
+	if (rc)
+		return;
+
+	if (target_submit_prep(se_cmd, cdb, NULL, 0, NULL, 0, NULL, 0,
+			       GFP_KERNEL))
+		return;
+
+	target_submit(se_cmd);
 }
 EXPORT_SYMBOL(target_submit_cmd);
 
+
+static struct se_dev_plug *target_plug_device(struct se_device *se_dev)
+{
+	struct se_dev_plug *se_plug;
+
+	if (!se_dev->transport->plug_device)
+		return NULL;
+
+	se_plug = se_dev->transport->plug_device(se_dev);
+	if (!se_plug)
+		return NULL;
+
+	se_plug->se_dev = se_dev;
+	/*
+	 * We have a ref to the lun at this point, but the cmds could
+	 * complete before we unplug, so grab a ref to the se_device so we
+	 * can call back into the backend.
+	 */
+	config_group_get(&se_dev->dev_group);
+	return se_plug;
+}
+
+static void target_unplug_device(struct se_dev_plug *se_plug)
+{
+	struct se_device *se_dev = se_plug->se_dev;
+
+	se_dev->transport->unplug_device(se_plug);
+	config_group_put(&se_dev->dev_group);
+}
+
+void target_queued_submit_work(struct work_struct *work)
+{
+	struct se_cmd_queue *sq = container_of(work, struct se_cmd_queue, work);
+	struct se_cmd *se_cmd, *next_cmd;
+	struct se_dev_plug *se_plug = NULL;
+	struct se_device *se_dev = NULL;
+	struct llist_node *cmd_list;
+
+	cmd_list = llist_del_all(&sq->cmd_list);
+	if (!cmd_list)
+		/* Previous call took what we were queued to submit */
+		return;
+
+	cmd_list = llist_reverse_order(cmd_list);
+	llist_for_each_entry_safe(se_cmd, next_cmd, cmd_list, se_cmd_list) {
+		if (!se_dev) {
+			se_dev = se_cmd->se_dev;
+			se_plug = target_plug_device(se_dev);
+		}
+
+		target_submit(se_cmd);
+	}
+
+	if (se_plug)
+		target_unplug_device(se_plug);
+}
+
+/**
+ * target_queue_submission - queue the cmd to run on the LIO workqueue
+ * @se_cmd: command descriptor to submit
+ */
+void target_queue_submission(struct se_cmd *se_cmd)
+{
+	struct se_device *se_dev = se_cmd->se_dev;
+	int cpu = se_cmd->cpuid;
+	struct se_cmd_queue *sq;
+
+	sq = &se_dev->queues[cpu].sq;
+	llist_add(&se_cmd->se_cmd_list, &sq->cmd_list);
+	queue_work_on(cpu, target_submission_wq, &sq->work);
+}
+EXPORT_SYMBOL_GPL(target_queue_submission);
+
 static void target_complete_tmr_failure(struct work_struct *work)
 {
 	struct se_cmd *se_cmd = container_of(work, struct se_cmd, work);
@@ -1799,8 +1927,8 @@ int target_submit_tmr(struct se_cmd *se_cmd, struct se_session *se_sess,
 	se_tpg = se_sess->se_tpg;
 	BUG_ON(!se_tpg);
 
-	transport_init_se_cmd(se_cmd, se_tpg->se_tpg_tfo, se_sess,
-			      0, DMA_NONE, TCM_SIMPLE_TAG, sense, unpacked_lun);
+	__target_init_cmd(se_cmd, se_tpg->se_tpg_tfo, se_sess,
+			  0, DMA_NONE, TCM_SIMPLE_TAG, sense, unpacked_lun);
 	/*
 	 * FIXME: Currently expect caller to handle se_cmd->se_tmr_req
 	 * allocation failure.
@@ -2778,9 +2906,7 @@ int target_get_sess_cmd(struct se_cmd *se_cmd, bool ack_kref)
 	 * invocations before se_cmd descriptor release.
 	 */
 	if (ack_kref) {
-		if (!kref_get_unless_zero(&se_cmd->cmd_kref))
-			return -EINVAL;
-
+		kref_get(&se_cmd->cmd_kref);
 		se_cmd->se_cmd_flags |= SCF_ACK_KREF;
 	}