12 files changed, 240 insertions, 46 deletions

diff --git a/drivers/block/aoe/aoe.h b/drivers/block/aoe/aoe.h
index 175649468c95..025c41d3cb33 100644
--- a/drivers/block/aoe/aoe.h
+++ b/drivers/block/aoe/aoe.h
@@ -1,5 +1,5 @@
-/* Copyright (c) 2012 Coraid, Inc.  See COPYING for GPL terms. */
-#define VERSION "81"
+/* Copyright (c) 2013 Coraid, Inc.  See COPYING for GPL terms. */
+#define VERSION "83"
 #define AOE_MAJOR 152
 #define DEVICE_NAME "aoe"
 
@@ -196,9 +196,11 @@ struct ktstate {
 	struct completion rendez;
 	struct task_struct *task;
 	wait_queue_head_t *waitq;
-	int (*fn) (void);
-	char *name;
+	int (*fn) (int);
+	char name[12];
 	spinlock_t *lock;
+	int id;
+	int active;
 };
 
 int aoeblk_init(void);
@@ -222,6 +224,7 @@ int aoecmd_init(void);
 struct sk_buff *aoecmd_ata_id(struct aoedev *);
 void aoe_freetframe(struct frame *);
 void aoe_flush_iocq(void);
+void aoe_flush_iocq_by_index(int);
 void aoe_end_request(struct aoedev *, struct request *, int);
 int aoe_ktstart(struct ktstate *k);
 void aoe_ktstop(struct ktstate *k);
diff --git a/drivers/block/aoe/aoecmd.c b/drivers/block/aoe/aoecmd.c
index fc803ecbbce4..99cb944a002d 100644
--- a/drivers/block/aoe/aoecmd.c
+++ b/drivers/block/aoe/aoecmd.c
@@ -1,4 +1,4 @@
-/* Copyright (c) 2012 Coraid, Inc.  See COPYING for GPL terms. */
+/* Copyright (c) 2013 Coraid, Inc.  See COPYING for GPL terms. */
 /*
  * aoecmd.c
  * Filesystem request handling methods
@@ -35,14 +35,27 @@ module_param(aoe_maxout, int, 0644);
 MODULE_PARM_DESC(aoe_maxout,
 	"Only aoe_maxout outstanding packets for every MAC on eX.Y.");
 
-static wait_queue_head_t ktiowq;
-static struct ktstate kts;
+/* The number of online cpus during module initialization gives us a
+ * convenient heuristic cap on the parallelism used for ktio threads
+ * doing I/O completion.  It is not important that the cap equal the
+ * actual number of running CPUs at any given time, but because of CPU
+ * hotplug, we take care to use ncpus instead of using
+ * num_online_cpus() after module initialization.
+ */
+static int ncpus;
+
+/* mutex lock used for synchronization while thread spawning */
+static DEFINE_MUTEX(ktio_spawn_lock);
+
+static wait_queue_head_t *ktiowq;
+static struct ktstate *kts;
 
 /* io completion queue */
-static struct {
+struct iocq_ktio {
 	struct list_head head;
 	spinlock_t lock;
-} iocq;
+};
+static struct iocq_ktio *iocq;
 
 static struct page *empty_page;
 
@@ -1278,23 +1291,36 @@ out:
  * Returns true iff responses needing processing remain.
  */
 static int
-ktio(void)
+ktio(int id)
 {
 	struct frame *f;
 	struct list_head *pos;
 	int i;
+	int actual_id;
 
 	for (i = 0; ; ++i) {
 		if (i == MAXIOC)
 			return 1;
-		if (list_empty(&iocq.head))
+		if (list_empty(&iocq[id].head))
 			return 0;
-		pos = iocq.head.next;
+		pos = iocq[id].head.next;
 		list_del(pos);
-		spin_unlock_irq(&iocq.lock);
 		f = list_entry(pos, struct frame, head);
+		spin_unlock_irq(&iocq[id].lock);
 		ktiocomplete(f);
-		spin_lock_irq(&iocq.lock);
+
+		/* Figure out if extra threads are required. */
+		actual_id = f->t->d->aoeminor % ncpus;
+
+		if (!kts[actual_id].active) {
+			BUG_ON(id != 0);
+			mutex_lock(&ktio_spawn_lock);
+			if (!kts[actual_id].active
+				&& aoe_ktstart(&kts[actual_id]) == 0)
+				kts[actual_id].active = 1;
+			mutex_unlock(&ktio_spawn_lock);
+		}
+		spin_lock_irq(&iocq[id].lock);
 	}
 }
 
@@ -1311,7 +1337,7 @@ kthread(void *vp)
 	complete(&k->rendez);	/* tell spawner we're running */
 	do {
 		spin_lock_irq(k->lock);
-		more = k->fn();
+		more = k->fn(k->id);
 		if (!more) {
 			add_wait_queue(k->waitq, &wait);
 			__set_current_state(TASK_INTERRUPTIBLE);
@@ -1340,7 +1366,7 @@ aoe_ktstart(struct ktstate *k)
 	struct task_struct *task;
 
 	init_completion(&k->rendez);
-	task = kthread_run(kthread, k, k->name);
+	task = kthread_run(kthread, k, "%s", k->name);
 	if (task == NULL || IS_ERR(task))
 		return -ENOMEM;
 	k->task = task;
@@ -1353,13 +1379,24 @@ aoe_ktstart(struct ktstate *k)
 static void
 ktcomplete(struct frame *f, struct sk_buff *skb)
 {
+	int id;
 	ulong flags;
 
 	f->r_skb = skb;
-	spin_lock_irqsave(&iocq.lock, flags);
-	list_add_tail(&f->head, &iocq.head);
-	spin_unlock_irqrestore(&iocq.lock, flags);
-	wake_up(&ktiowq);
+	id = f->t->d->aoeminor % ncpus;
+	spin_lock_irqsave(&iocq[id].lock, flags);
+	if (!kts[id].active) {
+		spin_unlock_irqrestore(&iocq[id].lock, flags);
+		/* The thread with id has not been spawned yet,
+		 * so delegate the work to the main thread and
+		 * try spawning a new thread.
+		 */
+		id = 0;
+		spin_lock_irqsave(&iocq[id].lock, flags);
+	}
+	list_add_tail(&f->head, &iocq[id].head);
+	spin_unlock_irqrestore(&iocq[id].lock, flags);
+	wake_up(&ktiowq[id]);
 }
 
 struct sk_buff *
@@ -1706,6 +1743,17 @@ aoe_failbuf(struct aoedev *d, struct buf *buf)
 void
 aoe_flush_iocq(void)
 {
+	int i;
+
+	for (i = 0; i < ncpus; i++) {
+		if (kts[i].active)
+			aoe_flush_iocq_by_index(i);
+	}
+}
+
+void
+aoe_flush_iocq_by_index(int id)
+{
 	struct frame *f;
 	struct aoedev *d;
 	LIST_HEAD(flist);
@@ -1713,9 +1761,9 @@ aoe_flush_iocq(void)
 	struct sk_buff *skb;
 	ulong flags;
 
-	spin_lock_irqsave(&iocq.lock, flags);
-	list_splice_init(&iocq.head, &flist);
-	spin_unlock_irqrestore(&iocq.lock, flags);
+	spin_lock_irqsave(&iocq[id].lock, flags);
+	list_splice_init(&iocq[id].head, &flist);
+	spin_unlock_irqrestore(&iocq[id].lock, flags);
 	while (!list_empty(&flist)) {
 		pos = flist.next;
 		list_del(pos);
@@ -1738,6 +1786,8 @@ int __init
 aoecmd_init(void)
 {
 	void *p;
+	int i;
+	int ret;
 
 	/* get_zeroed_page returns page with ref count 1 */
 	p = (void *) get_zeroed_page(GFP_KERNEL | __GFP_REPEAT);
@@ -1745,22 +1795,72 @@ aoecmd_init(void)
 		return -ENOMEM;
 	empty_page = virt_to_page(p);
 
-	INIT_LIST_HEAD(&iocq.head);
-	spin_lock_init(&iocq.lock);
-	init_waitqueue_head(&ktiowq);
-	kts.name = "aoe_ktio";
-	kts.fn = ktio;
-	kts.waitq = &ktiowq;
-	kts.lock = &iocq.lock;
-	return aoe_ktstart(&kts);
+	ncpus = num_online_cpus();
+
+	iocq = kcalloc(ncpus, sizeof(struct iocq_ktio), GFP_KERNEL);
+	if (!iocq)
+		return -ENOMEM;
+
+	kts = kcalloc(ncpus, sizeof(struct ktstate), GFP_KERNEL);
+	if (!kts) {
+		ret = -ENOMEM;
+		goto kts_fail;
+	}
+
+	ktiowq = kcalloc(ncpus, sizeof(wait_queue_head_t), GFP_KERNEL);
+	if (!ktiowq) {
+		ret = -ENOMEM;
+		goto ktiowq_fail;
+	}
+
+	mutex_init(&ktio_spawn_lock);
+
+	for (i = 0; i < ncpus; i++) {
+		INIT_LIST_HEAD(&iocq[i].head);
+		spin_lock_init(&iocq[i].lock);
+		init_waitqueue_head(&ktiowq[i]);
+		snprintf(kts[i].name, sizeof(kts[i].name), "aoe_ktio%d", i);
+		kts[i].fn = ktio;
+		kts[i].waitq = &ktiowq[i];
+		kts[i].lock = &iocq[i].lock;
+		kts[i].id = i;
+		kts[i].active = 0;
+	}
+	kts[0].active = 1;
+	if (aoe_ktstart(&kts[0])) {
+		ret = -ENOMEM;
+		goto ktstart_fail;
+	}
+	return 0;
+
+ktstart_fail:
+	kfree(ktiowq);
+ktiowq_fail:
+	kfree(kts);
+kts_fail:
+	kfree(iocq);
+
+	return ret;
 }
 
 void
 aoecmd_exit(void)
 {
-	aoe_ktstop(&kts);
+	int i;
+
+	for (i = 0; i < ncpus; i++)
+		if (kts[i].active)
+			aoe_ktstop(&kts[i]);
+
 	aoe_flush_iocq();
 
+	/* Free up the iocq and thread speicific configuration
+	* allocated during startup.
+	*/
+	kfree(iocq);
+	kfree(kts);
+	kfree(ktiowq);
+
 	free_page((unsigned long) page_address(empty_page));
 	empty_page = NULL;
 }
diff --git a/drivers/block/aoe/aoedev.c b/drivers/block/aoe/aoedev.c
index 98f2965778b9..784c92e038d1 100644
--- a/drivers/block/aoe/aoedev.c
+++ b/drivers/block/aoe/aoedev.c
@@ -1,4 +1,4 @@
-/* Copyright (c) 2012 Coraid, Inc.  See COPYING for GPL terms. */
+/* Copyright (c) 2013 Coraid, Inc.  See COPYING for GPL terms. */
 /*
  * aoedev.c
  * AoE device utility functions; maintains device list.
@@ -518,7 +518,6 @@ void
 aoedev_exit(void)
 {
 	flush_scheduled_work();
-	aoe_flush_iocq();
 	flush(NULL, 0, EXITING);
 }
 
diff --git a/drivers/block/aoe/aoenet.c b/drivers/block/aoe/aoenet.c
index 71d3ea8d3006..63773a90581d 100644
--- a/drivers/block/aoe/aoenet.c
+++ b/drivers/block/aoe/aoenet.c
@@ -1,4 +1,4 @@
-/* Copyright (c) 2012 Coraid, Inc.  See COPYING for GPL terms. */
+/* Copyright (c) 2013 Coraid, Inc.  See COPYING for GPL terms. */
 /*
  * aoenet.c
  * Ethernet portion of AoE driver
@@ -52,7 +52,7 @@ static struct sk_buff_head skbtxq;
 
 /* enters with txlock held */
 static int
-tx(void) __must_hold(&txlock)
+tx(int id) __must_hold(&txlock)
 {
 	struct sk_buff *skb;
 	struct net_device *ifp;
@@ -205,7 +205,8 @@ aoenet_init(void)
 	kts.lock = &txlock;
 	kts.fn = tx;
 	kts.waitq = &txwq;
-	kts.name = "aoe_tx";
+	kts.id = 0;
+	snprintf(kts.name, sizeof(kts.name), "aoe_tx%d", kts.id);
 	if (aoe_ktstart(&kts))
 		return -EAGAIN;
 	dev_add_pack(&aoe_pt);
diff --git a/drivers/block/cryptoloop.c b/drivers/block/cryptoloop.c
index 8b6bb764b0a3..99e773cb70d0 100644
--- a/drivers/block/cryptoloop.c
+++ b/drivers/block/cryptoloop.c
@@ -25,9 +25,9 @@
 #include <linux/string.h>
 #include <linux/crypto.h>
 #include <linux/blkdev.h>
-#include <linux/loop.h>
 #include <linux/scatterlist.h>
 #include <asm/uaccess.h>
+#include "loop.h"
 
 MODULE_LICENSE("GPL");
 MODULE_DESCRIPTION("loop blockdevice transferfunction adaptor / CryptoAPI");
diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index d92d50fd84b7..40e715531aa6 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -63,7 +63,6 @@
 #include <linux/init.h>
 #include <linux/swap.h>
 #include <linux/slab.h>
-#include <linux/loop.h>
 #include <linux/compat.h>
 #include <linux/suspend.h>
 #include <linux/freezer.h>
@@ -76,6 +75,7 @@
 #include <linux/sysfs.h>
 #include <linux/miscdevice.h>
 #include <linux/falloc.h>
+#include "loop.h"
 
 #include <asm/uaccess.h>
 
diff --git a/drivers/block/loop.h b/drivers/block/loop.h
new file mode 100644
index 000000000000..90df5d6485b6
--- /dev/null
+++ b/drivers/block/loop.h
@@ -0,0 +1,85 @@
+/*
+ * loop.h
+ *
+ * Written by Theodore Ts'o, 3/29/93.
+ *
+ * Copyright 1993 by Theodore Ts'o.  Redistribution of this file is
+ * permitted under the GNU General Public License.
+ */
+#ifndef _LINUX_LOOP_H
+#define _LINUX_LOOP_H
+
+#include <linux/bio.h>
+#include <linux/blkdev.h>
+#include <linux/spinlock.h>
+#include <linux/mutex.h>
+#include <uapi/linux/loop.h>
+
+/* Possible states of device */
+enum {
+	Lo_unbound,
+	Lo_bound,
+	Lo_rundown,
+};
+
+struct loop_func_table;
+
+struct loop_device {
+	int		lo_number;
+	int		lo_refcnt;
+	loff_t		lo_offset;
+	loff_t		lo_sizelimit;
+	int		lo_flags;
+	int		(*transfer)(struct loop_device *, int cmd,
+				    struct page *raw_page, unsigned raw_off,
+				    struct page *loop_page, unsigned loop_off,
+				    int size, sector_t real_block);
+	char		lo_file_name[LO_NAME_SIZE];
+	char		lo_crypt_name[LO_NAME_SIZE];
+	char		lo_encrypt_key[LO_KEY_SIZE];
+	int		lo_encrypt_key_size;
+	struct loop_func_table *lo_encryption;
+	__u32           lo_init[2];
+	kuid_t		lo_key_owner;	/* Who set the key */
+	int		(*ioctl)(struct loop_device *, int cmd, 
+				 unsigned long arg); 
+
+	struct file *	lo_backing_file;
+	struct block_device *lo_device;
+	unsigned	lo_blocksize;
+	void		*key_data; 
+
+	gfp_t		old_gfp_mask;
+
+	spinlock_t		lo_lock;
+	struct bio_list		lo_bio_list;
+	unsigned int		lo_bio_count;
+	int			lo_state;
+	struct mutex		lo_ctl_mutex;
+	struct task_struct	*lo_thread;
+	wait_queue_head_t	lo_event;
+	/* wait queue for incoming requests */
+	wait_queue_head_t	lo_req_wait;
+
+	struct request_queue	*lo_queue;
+	struct gendisk		*lo_disk;
+};
+
+/* Support for loadable transfer modules */
+struct loop_func_table {
+	int number;	/* filter type */ 
+	int (*transfer)(struct loop_device *lo, int cmd,
+			struct page *raw_page, unsigned raw_off,
+			struct page *loop_page, unsigned loop_off,
+			int size, sector_t real_block);
+	int (*init)(struct loop_device *, const struct loop_info64 *); 
+	/* release is called from loop_unregister_transfer or clr_fd */
+	int (*release)(struct loop_device *); 
+	int (*ioctl)(struct loop_device *, int cmd, unsigned long arg);
+	struct module *owner;
+}; 
+
+int loop_register_transfer(struct loop_func_table *funcs);
+int loop_unregister_transfer(int number); 
+
+#endif
diff --git a/drivers/block/mtip32xx/mtip32xx.c b/drivers/block/mtip32xx/mtip32xx.c
index 20dd52a2f92f..952dbfe22126 100644
--- a/drivers/block/mtip32xx/mtip32xx.c
+++ b/drivers/block/mtip32xx/mtip32xx.c
@@ -4087,7 +4087,8 @@ skip_create_disk:
 start_service_thread:
 	sprintf(thd_name, "mtip_svc_thd_%02d", index);
 	dd->mtip_svc_handler = kthread_create_on_node(mtip_service_thread,
-						dd, dd->numa_node, thd_name);
+						dd, dd->numa_node, "%s",
+						thd_name);
 
 	if (IS_ERR(dd->mtip_svc_handler)) {
 		dev_err(&dd->pdev->dev, "service thread failed to start\n");
diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c
index 037288e7874d..2dc3b5153f0d 100644
--- a/drivers/block/nbd.c
+++ b/drivers/block/nbd.c
@@ -623,8 +623,10 @@ static int __nbd_ioctl(struct block_device *bdev, struct nbd_device *nbd,
 		if (!nbd->sock)
 			return -EINVAL;
 
+		nbd->disconnect = 1;
+
 		nbd_send_req(nbd, &sreq);
-                return 0;
+		return 0;
 	}
  
 	case NBD_CLEAR_SOCK: {
@@ -654,6 +656,7 @@ static int __nbd_ioctl(struct block_device *bdev, struct nbd_device *nbd,
 				nbd->sock = SOCKET_I(inode);
 				if (max_part > 0)
 					bdev->bd_invalidated = 1;
+				nbd->disconnect = 0; /* we're connected now */
 				return 0;
 			} else {
 				fput(file);
@@ -714,7 +717,8 @@ static int __nbd_ioctl(struct block_device *bdev, struct nbd_device *nbd,
 		else
 			blk_queue_flush(nbd->disk->queue, 0);
 
-		thread = kthread_create(nbd_thread, nbd, nbd->disk->disk_name);
+		thread = kthread_create(nbd_thread, nbd, "%s",
+					nbd->disk->disk_name);
 		if (IS_ERR(thread)) {
 			mutex_lock(&nbd->tx_lock);
 			return PTR_ERR(thread);
@@ -742,6 +746,8 @@ static int __nbd_ioctl(struct block_device *bdev, struct nbd_device *nbd,
 		set_capacity(nbd->disk, 0);
 		if (max_part > 0)
 			ioctl_by_bdev(bdev, BLKRRPART, 0);
+		if (nbd->disconnect) /* user requested, ignore socket errors */
+			return 0;
 		return nbd->harderror;
 	}
 
@@ -750,7 +756,6 @@ static int __nbd_ioctl(struct block_device *bdev, struct nbd_device *nbd,
 		 * This is for compatibility only.  The queue is always cleared
 		 * by NBD_DO_IT or NBD_CLEAR_SOCK.
 		 */
-		BUG_ON(!nbd->sock && !list_empty(&nbd->queue_head));
 		return 0;
 
 	case NBD_PRINT_DEBUG:
diff --git a/drivers/block/swim.c b/drivers/block/swim.c
index 2f445b7a174e..8ed6ccb748cf 100644
--- a/drivers/block/swim.c
+++ b/drivers/block/swim.c
@@ -893,7 +893,7 @@ static int swim_probe(struct platform_device *dev)
 
 	swim_base = ioremap(res->start, resource_size(res));
 	if (!swim_base) {
-		return -ENOMEM;
+		ret = -ENOMEM;
 		goto out_release_io;
 	}
 
diff --git a/drivers/block/virtio_blk.c b/drivers/block/virtio_blk.c
index 64723953e1c9..5cdf88b7ad9e 100644
--- a/drivers/block/virtio_blk.c
+++ b/drivers/block/virtio_blk.c
@@ -20,7 +20,7 @@ module_param(use_bio, bool, S_IRUGO);
 static int major;
 static DEFINE_IDA(vd_index_ida);
 
-struct workqueue_struct *virtblk_wq;
+static struct workqueue_struct *virtblk_wq;
 
 struct virtio_blk
 {
diff --git a/drivers/block/xen-blkback/xenbus.c b/drivers/block/xen-blkback/xenbus.c
index 8bfd1bcf95ec..04608a6502d7 100644
--- a/drivers/block/xen-blkback/xenbus.c
+++ b/drivers/block/xen-blkback/xenbus.c
@@ -93,7 +93,7 @@ static void xen_update_blkif_status(struct xen_blkif *blkif)
 	}
 	invalidate_inode_pages2(blkif->vbd.bdev->bd_inode->i_mapping);
 
-	blkif->xenblkd = kthread_run(xen_blkif_schedule, blkif, name);
+	blkif->xenblkd = kthread_run(xen_blkif_schedule, blkif, "%s", name);
 	if (IS_ERR(blkif->xenblkd)) {
 		err = PTR_ERR(blkif->xenblkd);
 		blkif->xenblkd = NULL;