Merge git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6: (102 commits)
  [SCSI] scsi_dh: fix kconfig related build errors
  [SCSI] sym53c8xx: Fix bogus sym_que_entry re-implementation of container_of
  [SCSI] scsi_cmnd.h: remove double inclusion of linux/blkdev.h
  [SCSI] make struct scsi_{host,target}_type static
  [SCSI] fix locking in host use of blk_plug_device()
  [SCSI] zfcp: Cleanup external header file
  [SCSI] zfcp: Cleanup code in zfcp_erp.c
  [SCSI] zfcp: zfcp_fsf cleanup.
  [SCSI] zfcp: consolidate sysfs things into one file.
  [SCSI] zfcp: Cleanup of code in zfcp_aux.c
  [SCSI] zfcp: Cleanup of code in zfcp_scsi.c
  [SCSI] zfcp: Move status accessors from zfcp to SCSI include file.
  [SCSI] zfcp: Small QDIO cleanups
  [SCSI] zfcp: Adapter reopen for large number of unsolicited status
  [SCSI] zfcp: Fix error checking for ELS ADISC requests
  [SCSI] zfcp: wait until adapter is finished with ERP during auto-port
  [SCSI] ibmvfc: IBM Power Virtual Fibre Channel Adapter Client Driver
  [SCSI] sg: Add target reset support
  [SCSI] lib: Add support for the T10 (SCSI) Data Integrity Field CRC
  [SCSI] sd: Move scsi_disk() accessor function to sd.h
  ...

1 files changed, 90 insertions, 73 deletions

diff --git a/drivers/md/dm-mpath.c b/drivers/md/dm-mpath.c
index e7ee59e655d..9f7302d4878 100644
--- a/drivers/md/dm-mpath.c
+++ b/drivers/md/dm-mpath.c
@@ -7,7 +7,6 @@
 
 #include "dm.h"
 #include "dm-path-selector.h"
-#include "dm-hw-handler.h"
 #include "dm-bio-list.h"
 #include "dm-bio-record.h"
 #include "dm-uevent.h"
@@ -20,6 +19,7 @@
 #include <linux/slab.h>
 #include <linux/time.h>
 #include <linux/workqueue.h>
+#include <scsi/scsi_dh.h>
 #include <asm/atomic.h>
 
 #define DM_MSG_PREFIX "multipath"
@@ -61,7 +61,8 @@ struct multipath {
 
 	spinlock_t lock;
 
-	struct hw_handler hw_handler;
+	const char *hw_handler_name;
+	struct work_struct activate_path;
 	unsigned nr_priority_groups;
 	struct list_head priority_groups;
 	unsigned pg_init_required;	/* pg_init needs calling? */
@@ -106,9 +107,10 @@ typedef int (*action_fn) (struct pgpath *pgpath);
 
 static struct kmem_cache *_mpio_cache;
 
-static struct workqueue_struct *kmultipathd;
+static struct workqueue_struct *kmultipathd, *kmpath_handlerd;
 static void process_queued_ios(struct work_struct *work);
 static void trigger_event(struct work_struct *work);
+static void activate_path(struct work_struct *work);
 
 
 /*-----------------------------------------------
@@ -178,6 +180,7 @@ static struct multipath *alloc_multipath(struct dm_target *ti)
 		m->queue_io = 1;
 		INIT_WORK(&m->process_queued_ios, process_queued_ios);
 		INIT_WORK(&m->trigger_event, trigger_event);
+		INIT_WORK(&m->activate_path, activate_path);
 		m->mpio_pool = mempool_create_slab_pool(MIN_IOS, _mpio_cache);
 		if (!m->mpio_pool) {
 			kfree(m);
@@ -193,18 +196,13 @@ static struct multipath *alloc_multipath(struct dm_target *ti)
 static void free_multipath(struct multipath *m)
 {
 	struct priority_group *pg, *tmp;
-	struct hw_handler *hwh = &m->hw_handler;
 
 	list_for_each_entry_safe(pg, tmp, &m->priority_groups, list) {
 		list_del(&pg->list);
 		free_priority_group(pg, m->ti);
 	}
 
-	if (hwh->type) {
-		hwh->type->destroy(hwh);
-		dm_put_hw_handler(hwh->type);
-	}
-
+	kfree(m->hw_handler_name);
 	mempool_destroy(m->mpio_pool);
 	kfree(m);
 }
@@ -216,12 +214,10 @@ static void free_multipath(struct multipath *m)
 
 static void __switch_pg(struct multipath *m, struct pgpath *pgpath)
 {
-	struct hw_handler *hwh = &m->hw_handler;
-
 	m->current_pg = pgpath->pg;
 
 	/* Must we initialise the PG first, and queue I/O till it's ready? */
-	if (hwh->type && hwh->type->pg_init) {
+	if (m->hw_handler_name) {
 		m->pg_init_required = 1;
 		m->queue_io = 1;
 	} else {
@@ -409,7 +405,6 @@ static void process_queued_ios(struct work_struct *work)
 {
 	struct multipath *m =
 		container_of(work, struct multipath, process_queued_ios);
-	struct hw_handler *hwh = &m->hw_handler;
 	struct pgpath *pgpath = NULL;
 	unsigned init_required = 0, must_queue = 1;
 	unsigned long flags;
@@ -439,7 +434,7 @@ out:
 	spin_unlock_irqrestore(&m->lock, flags);
 
 	if (init_required)
-		hwh->type->pg_init(hwh, pgpath->pg->bypassed, &pgpath->path);
+		queue_work(kmpath_handlerd, &m->activate_path);
 
 	if (!must_queue)
 		dispatch_queued_ios(m);
@@ -652,8 +647,6 @@ static struct priority_group *parse_priority_group(struct arg_set *as,
 
 static int parse_hw_handler(struct arg_set *as, struct multipath *m)
 {
-	int r;
-	struct hw_handler_type *hwht;
 	unsigned hw_argc;
 	struct dm_target *ti = m->ti;
 
@@ -661,30 +654,20 @@ static int parse_hw_handler(struct arg_set *as, struct multipath *m)
 		{0, 1024, "invalid number of hardware handler args"},
 	};
 
-	r = read_param(_params, shift(as), &hw_argc, &ti->error);
-	if (r)
+	if (read_param(_params, shift(as), &hw_argc, &ti->error))
 		return -EINVAL;
 
 	if (!hw_argc)
 		return 0;
 
-	hwht = dm_get_hw_handler(shift(as));
-	if (!hwht) {
+	m->hw_handler_name = kstrdup(shift(as), GFP_KERNEL);
+	request_module("scsi_dh_%s", m->hw_handler_name);
+	if (scsi_dh_handler_exist(m->hw_handler_name) == 0) {
 		ti->error = "unknown hardware handler type";
+		kfree(m->hw_handler_name);
+		m->hw_handler_name = NULL;
 		return -EINVAL;
 	}
-
-	m->hw_handler.md = dm_table_get_md(ti->table);
-	dm_put(m->hw_handler.md);
-
-	r = hwht->create(&m->hw_handler, hw_argc - 1, as->argv);
-	if (r) {
-		dm_put_hw_handler(hwht);
-		ti->error = "hardware handler constructor failed";
-		return r;
-	}
-
-	m->hw_handler.type = hwht;
 	consume(as, hw_argc - 1);
 
 	return 0;
@@ -808,6 +791,7 @@ static void multipath_dtr(struct dm_target *ti)
 {
 	struct multipath *m = (struct multipath *) ti->private;
 
+	flush_workqueue(kmpath_handlerd);
 	flush_workqueue(kmultipathd);
 	free_multipath(m);
 }
@@ -1025,52 +1009,85 @@ static int pg_init_limit_reached(struct multipath *m, struct pgpath *pgpath)
 	return limit_reached;
 }
 
-/*
- * pg_init must call this when it has completed its initialisation
- */
-void dm_pg_init_complete(struct dm_path *path, unsigned err_flags)
+static void pg_init_done(struct dm_path *path, int errors)
 {
 	struct pgpath *pgpath = path_to_pgpath(path);
 	struct priority_group *pg = pgpath->pg;
 	struct multipath *m = pg->m;
 	unsigned long flags;
 
-	/*
-	 * If requested, retry pg_init until maximum number of retries exceeded.
-	 * If retry not requested and PG already bypassed, always fail the path.
-	 */
-	if (err_flags & MP_RETRY) {
-		if (pg_init_limit_reached(m, pgpath))
-			err_flags |= MP_FAIL_PATH;
-	} else if (err_flags && pg->bypassed)
-		err_flags |= MP_FAIL_PATH;
-
-	if (err_flags & MP_FAIL_PATH)
+	/* device or driver problems */
+	switch (errors) {
+	case SCSI_DH_OK:
+		break;
+	case SCSI_DH_NOSYS:
+		if (!m->hw_handler_name) {
+			errors = 0;
+			break;
+		}
+		DMERR("Cannot failover device because scsi_dh_%s was not "
+		      "loaded.", m->hw_handler_name);
+		/*
+		 * Fail path for now, so we do not ping pong
+		 */
 		fail_path(pgpath);
-
-	if (err_flags & MP_BYPASS_PG)
+		break;
+	case SCSI_DH_DEV_TEMP_BUSY:
+		/*
+		 * Probably doing something like FW upgrade on the
+		 * controller so try the other pg.
+		 */
 		bypass_pg(m, pg, 1);
+		break;
+	/* TODO: For SCSI_DH_RETRY we should wait a couple seconds */
+	case SCSI_DH_RETRY:
+	case SCSI_DH_IMM_RETRY:
+	case SCSI_DH_RES_TEMP_UNAVAIL:
+		if (pg_init_limit_reached(m, pgpath))
+			fail_path(pgpath);
+		errors = 0;
+		break;
+	default:
+		/*
+		 * We probably do not want to fail the path for a device
+		 * error, but this is what the old dm did. In future
+		 * patches we can do more advanced handling.
+		 */
+		fail_path(pgpath);
+	}
 
 	spin_lock_irqsave(&m->lock, flags);
-	if (err_flags & ~MP_RETRY) {
+	if (errors) {
+		DMERR("Could not failover device. Error %d.", errors);
 		m->current_pgpath = NULL;
 		m->current_pg = NULL;
-	} else if (!m->pg_init_required)
+	} else if (!m->pg_init_required) {
 		m->queue_io = 0;
+		pg->bypassed = 0;
+	}
 
 	m->pg_init_in_progress = 0;
 	queue_work(kmultipathd, &m->process_queued_ios);
 	spin_unlock_irqrestore(&m->lock, flags);
 }
 
+static void activate_path(struct work_struct *work)
+{
+	int ret;
+	struct multipath *m =
+		container_of(work, struct multipath, activate_path);
+	struct dm_path *path = &m->current_pgpath->path;
+
+	ret = scsi_dh_activate(bdev_get_queue(path->dev->bdev));
+	pg_init_done(path, ret);
+}
+
 /*
  * end_io handling
  */
 static int do_end_io(struct multipath *m, struct bio *bio,
 		     int error, struct dm_mpath_io *mpio)
 {
-	struct hw_handler *hwh = &m->hw_handler;
-	unsigned err_flags = MP_FAIL_PATH;	/* Default behavior */
 	unsigned long flags;
 
 	if (!error)
@@ -1097,19 +1114,8 @@ static int do_end_io(struct multipath *m, struct bio *bio,
 	}
 	spin_unlock_irqrestore(&m->lock, flags);
 
-	if (hwh->type && hwh->type->error)
-		err_flags = hwh->type->error(hwh, bio);
-
-	if (mpio->pgpath) {
-		if (err_flags & MP_FAIL_PATH)
-			fail_path(mpio->pgpath);
-
-		if (err_flags & MP_BYPASS_PG)
-			bypass_pg(m, mpio->pgpath->pg, 1);
-	}
-
-	if (err_flags & MP_ERROR_IO)
-		return -EIO;
+	if (mpio->pgpath)
+		fail_path(mpio->pgpath);
 
       requeue:
 	dm_bio_restore(&mpio->details, bio);
@@ -1194,7 +1200,6 @@ static int multipath_status(struct dm_target *ti, status_type_t type,
 	int sz = 0;
 	unsigned long flags;
 	struct multipath *m = (struct multipath *) ti->private;
-	struct hw_handler *hwh = &m->hw_handler;
 	struct priority_group *pg;
 	struct pgpath *p;
 	unsigned pg_num;
@@ -1214,12 +1219,10 @@ static int multipath_status(struct dm_target *ti, status_type_t type,
 			DMEMIT("pg_init_retries %u ", m->pg_init_retries);
 	}
 
-	if (hwh->type && hwh->type->status)
-		sz += hwh->type->status(hwh, type, result + sz, maxlen - sz);
-	else if (!hwh->type || type == STATUSTYPE_INFO)
+	if (!m->hw_handler_name || type == STATUSTYPE_INFO)
 		DMEMIT("0 ");
 	else
-		DMEMIT("1 %s ", hwh->type->name);
+		DMEMIT("1 %s ", m->hw_handler_name);
 
 	DMEMIT("%u ", m->nr_priority_groups);
 
@@ -1422,6 +1425,21 @@ static int __init dm_multipath_init(void)
 		return -ENOMEM;
 	}
 
+	/*
+	 * A separate workqueue is used to handle the device handlers
+	 * to avoid overloading existing workqueue. Overloading the
+	 * old workqueue would also create a bottleneck in the
+	 * path of the storage hardware device activation.
+	 */
+	kmpath_handlerd = create_singlethread_workqueue("kmpath_handlerd");
+	if (!kmpath_handlerd) {
+		DMERR("failed to create workqueue kmpath_handlerd");
+		destroy_workqueue(kmultipathd);
+		dm_unregister_target(&multipath_target);
+		kmem_cache_destroy(_mpio_cache);
+		return -ENOMEM;
+	}
+
 	DMINFO("version %u.%u.%u loaded",
 	       multipath_target.version[0], multipath_target.version[1],
 	       multipath_target.version[2]);
@@ -1433,6 +1451,7 @@ static void __exit dm_multipath_exit(void)
 {
 	int r;
 
+	destroy_workqueue(kmpath_handlerd);
 	destroy_workqueue(kmultipathd);
 
 	r = dm_unregister_target(&multipath_target);
@@ -1441,8 +1460,6 @@ static void __exit dm_multipath_exit(void)
 	kmem_cache_destroy(_mpio_cache);
 }
 
-EXPORT_SYMBOL_GPL(dm_pg_init_complete);
-
 module_init(dm_multipath_init);
 module_exit(dm_multipath_exit);