[SCSI] libsas: introduce sas_drain_work()

When an lldd invokes ->notify_port_event() it can trigger a chain of libsas
events to:

  1/ form the port and find the direct attached device

  2/ if the attached device is an expander perform domain discovery

A call to flush_workqueue() will only flush the initial port formation work.
Currently libsas users need to call scsi_flush_work() up to the max depth of
chain (which will grow from 2 to 3 when ata discovery is moved to its own
discovery event).  Instead of open coding multiple calls switch to use
drain_workqueue() to flush sas work.

drain_workqueue() does not handle new work submitted during the drain so
libsas needs a bit of infrastructure to hold off unchained work submissions
while a drain is in flight.  A lldd ->notify() event is considered 'unchained'
while a sas_discover_event() is 'chained'.  As Tejun notes:

  "For now, I think it would be best to add private wrapper in libsas to
   support deferring unchained work items while draining."

Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>

drivers/scsi/aic94xx/aic94xx_init.c

@@ -971,7 +971,7 @@ static int asd_scan_finished(struct Scsi_Host *shost, unsigned long time)
 	if (time < HZ)
 		return 0;
 	/* Wait for discovery to finish */
-	scsi_flush_work(shost);
+	sas_drain_work(SHOST_TO_SAS_HA(shost));
 	return 1;
 }
 

drivers/scsi/isci/host.c

@@ -650,15 +650,13 @@ static void isci_host_start_complete(struct isci_host *ihost, enum sci_status co
 
 int isci_host_scan_finished(struct Scsi_Host *shost, unsigned long time)
 {
-	struct isci_host *ihost = SHOST_TO_SAS_HA(shost)->lldd_ha;
+	struct sas_ha_struct *ha = SHOST_TO_SAS_HA(shost);
+	struct isci_host *ihost = ha->lldd_ha;
 
 	if (test_bit(IHOST_START_PENDING, &ihost->flags))
 		return 0;
 
-	/* todo: use sas_flush_discovery once it is upstream */
-	scsi_flush_work(shost);
-
-	scsi_flush_work(shost);
+	sas_drain_work(ha);
 
 	dev_dbg(&ihost->pdev->dev,
 		"%s: ihost->status = %d, time = %ld\n",

drivers/scsi/libsas/sas_discover.c

@@ -367,6 +367,25 @@ static void sas_revalidate_domain(struct work_struct *work)
 
 /* ---------- Events ---------- */
 
+static void sas_chain_work(struct sas_ha_struct *ha, struct work_struct *work)
+{
+	/* chained work is not subject to SA_HA_DRAINING or SAS_HA_REGISTERED */
+	scsi_queue_work(ha->core.shost, work);
+}
+
+static void sas_chain_event(int event, unsigned long *pending,
+			    struct work_struct *work,
+			    struct sas_ha_struct *ha)
+{
+	if (!test_and_set_bit(event, pending)) {
+		unsigned long flags;
+
+		spin_lock_irqsave(&ha->state_lock, flags);
+		sas_chain_work(ha, work);
+		spin_unlock_irqrestore(&ha->state_lock, flags);
+	}
+}
+
 int sas_discover_event(struct asd_sas_port *port, enum discover_event ev)
 {
 	struct sas_discovery *disc;
@@ -377,7 +396,7 @@ int sas_discover_event(struct asd_sas_port *port, enum discover_event ev)
 
 	BUG_ON(ev >= DISC_NUM_EVENTS);
 
-	sas_queue_event(ev, &disc->pending, &disc->disc_work[ev].work, port->ha);
+	sas_chain_event(ev, &disc->pending, &disc->disc_work[ev].work, port->ha);
 
 	return 0;
 }

drivers/scsi/libsas/sas_event.c

@@ -22,10 +22,65 @@
  *
  */
 
+#include <linux/export.h>
 #include <scsi/scsi_host.h>
 #include "sas_internal.h"
 #include "sas_dump.h"
 
+static void sas_queue_work(struct sas_ha_struct *ha, struct work_struct *work)
+{
+	if (!test_bit(SAS_HA_REGISTERED, &ha->state))
+		return;
+
+	if (test_bit(SAS_HA_DRAINING, &ha->state))
+		list_add(&work->entry, &ha->defer_q);
+	else
+		scsi_queue_work(ha->core.shost, work);
+}
+
+static void sas_queue_event(int event, unsigned long *pending,
+			    struct work_struct *work,
+			    struct sas_ha_struct *ha)
+{
+	if (!test_and_set_bit(event, pending)) {
+		unsigned long flags;
+
+		spin_lock_irqsave(&ha->state_lock, flags);
+		sas_queue_work(ha, work);
+		spin_unlock_irqrestore(&ha->state_lock, flags);
+	}
+}
+
+int sas_drain_work(struct sas_ha_struct *ha)
+{
+	struct workqueue_struct *wq = ha->core.shost->work_q;
+	struct work_struct *w, *_w;
+	int err;
+
+	err = mutex_lock_interruptible(&ha->drain_mutex);
+	if (err)
+		return err;
+
+	set_bit(SAS_HA_DRAINING, &ha->state);
+	/* flush submitters */
+	spin_lock_irq(&ha->state_lock);
+	spin_unlock_irq(&ha->state_lock);
+
+	drain_workqueue(wq);
+
+	spin_lock_irq(&ha->state_lock);
+	clear_bit(SAS_HA_DRAINING, &ha->state);
+	list_for_each_entry_safe(w, _w, &ha->defer_q, entry) {
+		list_del_init(&w->entry);
+		sas_queue_work(ha, w);
+	}
+	spin_unlock_irq(&ha->state_lock);
+	mutex_unlock(&ha->drain_mutex);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(sas_drain_work);
+
 static void notify_ha_event(struct sas_ha_struct *sas_ha, enum ha_event event)
 {
 	BUG_ON(event >= HA_NUM_EVENTS);

drivers/scsi/libsas/sas_init.c

@@ -114,6 +114,8 @@ int sas_register_ha(struct sas_ha_struct *sas_ha)
 
 	set_bit(SAS_HA_REGISTERED, &sas_ha->state);
 	spin_lock_init(&sas_ha->state_lock);
+	mutex_init(&sas_ha->drain_mutex);
+	INIT_LIST_HEAD(&sas_ha->defer_q);
 
 	error = sas_register_phys(sas_ha);
 	if (error) {
@@ -157,12 +159,13 @@ int sas_unregister_ha(struct sas_ha_struct *sas_ha)
 {
 	unsigned long flags;
 
-	/* Set the state to unregistered to avoid further
-	 * events to be queued */
+	/* Set the state to unregistered to avoid further unchained
+	 * events to be queued
+	 */
 	spin_lock_irqsave(&sas_ha->state_lock, flags);
 	clear_bit(SAS_HA_REGISTERED, &sas_ha->state);
 	spin_unlock_irqrestore(&sas_ha->state_lock, flags);
-	scsi_flush_work(sas_ha->core.shost);
+	sas_drain_work(sas_ha);
 
 	sas_unregister_ports(sas_ha);
 

drivers/scsi/libsas/sas_internal.h

@@ -92,20 +92,6 @@ static inline int sas_smp_host_handler(struct Scsi_Host *shost,
 }
 #endif
 
-static inline void sas_queue_event(int event, unsigned long *pending,
-				   struct work_struct *work,
-				   struct sas_ha_struct *sas_ha)
-{
-	if (!test_and_set_bit(event, pending)) {
-		unsigned long flags;
-
-		spin_lock_irqsave(&sas_ha->state_lock, flags);
-		if (test_bit(SAS_HA_REGISTERED, &sas_ha->state))
-			scsi_queue_work(sas_ha->core.shost, work);
-		spin_unlock_irqrestore(&sas_ha->state_lock, flags);
-	}
-}
-
 static inline void sas_fill_in_rphy(struct domain_device *dev,
 				    struct sas_rphy *rphy)
 {

drivers/scsi/mvsas/mv_sas.c

@@ -308,7 +308,7 @@ int mvs_scan_finished(struct Scsi_Host *shost, unsigned long time)
 	if (mvs_prv->scan_finished == 0)
 		return 0;
 
-	scsi_flush_work(shost);
+	sas_drain_work(sha);
 	return 1;
 }
 

drivers/scsi/pm8001/pm8001_sas.c

@@ -256,12 +256,14 @@ void pm8001_scan_start(struct Scsi_Host *shost)
 
 int pm8001_scan_finished(struct Scsi_Host *shost, unsigned long time)
 {
+	struct sas_ha_struct *ha = SHOST_TO_SAS_HA(shost);
+
 	/* give the phy enabling interrupt event time to come in (1s
 	* is empirically about all it takes) */
 	if (time < HZ)
 		return 0;
 	/* Wait for discovery to finish */
-	scsi_flush_work(shost);
+	sas_drain_work(ha);
 	return 1;
 }
 

include/scsi/libsas.h

@@ -330,13 +330,16 @@ struct sas_ha_event {
 
 enum sas_ha_state {
 	SAS_HA_REGISTERED,
+	SAS_HA_DRAINING,
 };
 
 struct sas_ha_struct {
 /* private: */
 	struct sas_ha_event ha_events[HA_NUM_EVENTS];
 	unsigned long	 pending;
 
+	struct list_head  defer_q; /* work queued while draining */
+	struct mutex	  drain_mutex;
 	unsigned long	  state;
 	spinlock_t 	  state_lock;
 
@@ -657,6 +660,7 @@ int sas_eh_bus_reset_handler(struct scsi_cmnd *cmd);
 extern void sas_target_destroy(struct scsi_target *);
 extern int sas_slave_alloc(struct scsi_device *);
 extern int sas_ioctl(struct scsi_device *sdev, int cmd, void __user *arg);
+extern int sas_drain_work(struct sas_ha_struct *ha);
 
 extern int sas_smp_handler(struct Scsi_Host *shost, struct sas_rphy *rphy,
 			   struct request *req);