Merge branch 'for-4.5/drivers' of git://git.kernel.dk/linux-block

Pull block driver updates from Jens Axboe:
 "This is the block driver pull request for 4.5, with the exception of
  NVMe, which is in a separate branch and will be posted after this one.

  This pull request contains:

   - A set of bcache stability fixes, which have been acked by Kent.
     These have been used and tested for more than a year by the
     community, so it's about time that they got in.

   - A set of drbd updates from the drbd team (Andreas, Lars, Philipp)
     and Markus Elfring, Oleg Drokin.

   - A set of fixes for xen blkback/front from the usual suspects, (Bob,
     Konrad) as well as community based fixes from Kiri, Julien, and
     Peng.

   - A 2038 time fix for sx8 from Shraddha, with a fix from me.

   - A small mtip32xx cleanup from Zhu Yanjun.

   - A null_blk division fix from Arnd"

* 'for-4.5/drivers' of git://git.kernel.dk/linux-block: (71 commits)
  null_blk: use sector_div instead of do_div
  mtip32xx: restrict variables visible in current code module
  xen/blkfront: Fix crash if backend doesn't follow the right states.
  xen/blkback: Fix two memory leaks.
  xen/blkback: make st_ statistics per ring
  xen/blkfront: Handle non-indirect grant with 64KB pages
  xen-blkfront: Introduce blkif_ring_get_request
  xen-blkback: clear PF_NOFREEZE for xen_blkif_schedule()
  xen/blkback: Free resources if connect_ring failed.
  xen/blocks: Return -EXX instead of -1
  xen/blkback: make pool of persistent grants and free pages per-queue
  xen/blkback: get the number of hardware queues/rings from blkfront
  xen/blkback: pseudo support for multi hardware queues/rings
  xen/blkback: separate ring information out of struct xen_blkif
  xen/blkfront: correct setting for xen_blkif_max_ring_order
  xen/blkfront: make persistent grants pool per-queue
  xen/blkfront: Remove duplicate setting of ->xbdev.
  xen/blkfront: Cleanup of comments, fix unaligned variables, and syntax errors.
  xen/blkfront: negotiate number of queues/rings to be used with backend
  xen/blkfront: split per device io_lock
  ...

MAINTAINERS

@@ -3665,13 +3665,12 @@ F:	drivers/scsi/dpt*
 F:	drivers/scsi/dpt/
 
 DRBD DRIVER
-P:	Philipp Reisner
-P:	Lars Ellenberg
-M:	drbd-dev@lists.linbit.com
-L:	drbd-user@lists.linbit.com
+M:	Philipp Reisner <philipp.reisner@linbit.com>
+M:	Lars Ellenberg <lars.ellenberg@linbit.com>
+L:	drbd-dev@lists.linbit.com
 W:	http://www.drbd.org
-T:	git git://git.drbd.org/linux-2.6-drbd.git drbd
-T:	git git://git.drbd.org/drbd-8.3.git
+T:	git git://git.linbit.com/linux-drbd.git
+T:	git git://git.linbit.com/drbd-8.4.git
 S:	Supported
 F:	drivers/block/drbd/
 F:	lib/lru_cache.c

drivers/block/drbd/drbd_actlog.c

@@ -288,7 +288,162 @@ bool drbd_al_begin_io_prepare(struct drbd_device *device, struct drbd_interval *
 	return need_transaction;
 }
 
-static int al_write_transaction(struct drbd_device *device);
+#if (PAGE_SHIFT + 3) < (AL_EXTENT_SHIFT - BM_BLOCK_SHIFT)
+/* Currently BM_BLOCK_SHIFT, BM_EXT_SHIFT and AL_EXTENT_SHIFT
+ * are still coupled, or assume too much about their relation.
+ * Code below will not work if this is violated.
+ * Will be cleaned up with some followup patch.
+ */
+# error FIXME
+#endif
+
+static unsigned int al_extent_to_bm_page(unsigned int al_enr)
+{
+	return al_enr >>
+		/* bit to page */
+		((PAGE_SHIFT + 3) -
+		/* al extent number to bit */
+		 (AL_EXTENT_SHIFT - BM_BLOCK_SHIFT));
+}
+
+static sector_t al_tr_number_to_on_disk_sector(struct drbd_device *device)
+{
+	const unsigned int stripes = device->ldev->md.al_stripes;
+	const unsigned int stripe_size_4kB = device->ldev->md.al_stripe_size_4k;
+
+	/* transaction number, modulo on-disk ring buffer wrap around */
+	unsigned int t = device->al_tr_number % (device->ldev->md.al_size_4k);
+
+	/* ... to aligned 4k on disk block */
+	t = ((t % stripes) * stripe_size_4kB) + t/stripes;
+
+	/* ... to 512 byte sector in activity log */
+	t *= 8;
+
+	/* ... plus offset to the on disk position */
+	return device->ldev->md.md_offset + device->ldev->md.al_offset + t;
+}
+
+static int __al_write_transaction(struct drbd_device *device, struct al_transaction_on_disk *buffer)
+{
+	struct lc_element *e;
+	sector_t sector;
+	int i, mx;
+	unsigned extent_nr;
+	unsigned crc = 0;
+	int err = 0;
+
+	memset(buffer, 0, sizeof(*buffer));
+	buffer->magic = cpu_to_be32(DRBD_AL_MAGIC);
+	buffer->tr_number = cpu_to_be32(device->al_tr_number);
+
+	i = 0;
+
+	/* Even though no one can start to change this list
+	 * once we set the LC_LOCKED -- from drbd_al_begin_io(),
+	 * lc_try_lock_for_transaction() --, someone may still
+	 * be in the process of changing it. */
+	spin_lock_irq(&device->al_lock);
+	list_for_each_entry(e, &device->act_log->to_be_changed, list) {
+		if (i == AL_UPDATES_PER_TRANSACTION) {
+			i++;
+			break;
+		}
+		buffer->update_slot_nr[i] = cpu_to_be16(e->lc_index);
+		buffer->update_extent_nr[i] = cpu_to_be32(e->lc_new_number);
+		if (e->lc_number != LC_FREE)
+			drbd_bm_mark_for_writeout(device,
+					al_extent_to_bm_page(e->lc_number));
+		i++;
+	}
+	spin_unlock_irq(&device->al_lock);
+	BUG_ON(i > AL_UPDATES_PER_TRANSACTION);
+
+	buffer->n_updates = cpu_to_be16(i);
+	for ( ; i < AL_UPDATES_PER_TRANSACTION; i++) {
+		buffer->update_slot_nr[i] = cpu_to_be16(-1);
+		buffer->update_extent_nr[i] = cpu_to_be32(LC_FREE);
+	}
+
+	buffer->context_size = cpu_to_be16(device->act_log->nr_elements);
+	buffer->context_start_slot_nr = cpu_to_be16(device->al_tr_cycle);
+
+	mx = min_t(int, AL_CONTEXT_PER_TRANSACTION,
+		   device->act_log->nr_elements - device->al_tr_cycle);
+	for (i = 0; i < mx; i++) {
+		unsigned idx = device->al_tr_cycle + i;
+		extent_nr = lc_element_by_index(device->act_log, idx)->lc_number;
+		buffer->context[i] = cpu_to_be32(extent_nr);
+	}
+	for (; i < AL_CONTEXT_PER_TRANSACTION; i++)
+		buffer->context[i] = cpu_to_be32(LC_FREE);
+
+	device->al_tr_cycle += AL_CONTEXT_PER_TRANSACTION;
+	if (device->al_tr_cycle >= device->act_log->nr_elements)
+		device->al_tr_cycle = 0;
+
+	sector = al_tr_number_to_on_disk_sector(device);
+
+	crc = crc32c(0, buffer, 4096);
+	buffer->crc32c = cpu_to_be32(crc);
+
+	if (drbd_bm_write_hinted(device))
+		err = -EIO;
+	else {
+		bool write_al_updates;
+		rcu_read_lock();
+		write_al_updates = rcu_dereference(device->ldev->disk_conf)->al_updates;
+		rcu_read_unlock();
+		if (write_al_updates) {
+			if (drbd_md_sync_page_io(device, device->ldev, sector, WRITE)) {
+				err = -EIO;
+				drbd_chk_io_error(device, 1, DRBD_META_IO_ERROR);
+			} else {
+				device->al_tr_number++;
+				device->al_writ_cnt++;
+			}
+		}
+	}
+
+	return err;
+}
+
+static int al_write_transaction(struct drbd_device *device)
+{
+	struct al_transaction_on_disk *buffer;
+	int err;
+
+	if (!get_ldev(device)) {
+		drbd_err(device, "disk is %s, cannot start al transaction\n",
+			drbd_disk_str(device->state.disk));
+		return -EIO;
+	}
+
+	/* The bitmap write may have failed, causing a state change. */
+	if (device->state.disk < D_INCONSISTENT) {
+		drbd_err(device,
+			"disk is %s, cannot write al transaction\n",
+			drbd_disk_str(device->state.disk));
+		put_ldev(device);
+		return -EIO;
+	}
+
+	/* protects md_io_buffer, al_tr_cycle, ... */
+	buffer = drbd_md_get_buffer(device, __func__);
+	if (!buffer) {
+		drbd_err(device, "disk failed while waiting for md_io buffer\n");
+		put_ldev(device);
+		return -ENODEV;
+	}
+
+	err = __al_write_transaction(device, buffer);
+
+	drbd_md_put_buffer(device);
+	put_ldev(device);
+
+	return err;
+}
+
 
 void drbd_al_begin_io_commit(struct drbd_device *device)
 {
@@ -420,153 +575,6 @@ void drbd_al_complete_io(struct drbd_device *device, struct drbd_interval *i)
 	wake_up(&device->al_wait);
 }
 
-#if (PAGE_SHIFT + 3) < (AL_EXTENT_SHIFT - BM_BLOCK_SHIFT)
-/* Currently BM_BLOCK_SHIFT, BM_EXT_SHIFT and AL_EXTENT_SHIFT
- * are still coupled, or assume too much about their relation.
- * Code below will not work if this is violated.
- * Will be cleaned up with some followup patch.
- */
-# error FIXME
-#endif
-
-static unsigned int al_extent_to_bm_page(unsigned int al_enr)
-{
-	return al_enr >>
-		/* bit to page */
-		((PAGE_SHIFT + 3) -
-		/* al extent number to bit */
-		 (AL_EXTENT_SHIFT - BM_BLOCK_SHIFT));
-}
-
-static sector_t al_tr_number_to_on_disk_sector(struct drbd_device *device)
-{
-	const unsigned int stripes = device->ldev->md.al_stripes;
-	const unsigned int stripe_size_4kB = device->ldev->md.al_stripe_size_4k;
-
-	/* transaction number, modulo on-disk ring buffer wrap around */
-	unsigned int t = device->al_tr_number % (device->ldev->md.al_size_4k);
-
-	/* ... to aligned 4k on disk block */
-	t = ((t % stripes) * stripe_size_4kB) + t/stripes;
-
-	/* ... to 512 byte sector in activity log */
-	t *= 8;
-
-	/* ... plus offset to the on disk position */
-	return device->ldev->md.md_offset + device->ldev->md.al_offset + t;
-}
-
-int al_write_transaction(struct drbd_device *device)
-{
-	struct al_transaction_on_disk *buffer;
-	struct lc_element *e;
-	sector_t sector;
-	int i, mx;
-	unsigned extent_nr;
-	unsigned crc = 0;
-	int err = 0;
-
-	if (!get_ldev(device)) {
-		drbd_err(device, "disk is %s, cannot start al transaction\n",
-			drbd_disk_str(device->state.disk));
-		return -EIO;
-	}
-
-	/* The bitmap write may have failed, causing a state change. */
-	if (device->state.disk < D_INCONSISTENT) {
-		drbd_err(device,
-			"disk is %s, cannot write al transaction\n",
-			drbd_disk_str(device->state.disk));
-		put_ldev(device);
-		return -EIO;
-	}
-
-	/* protects md_io_buffer, al_tr_cycle, ... */
-	buffer = drbd_md_get_buffer(device, __func__);
-	if (!buffer) {
-		drbd_err(device, "disk failed while waiting for md_io buffer\n");
-		put_ldev(device);
-		return -ENODEV;
-	}
-
-	memset(buffer, 0, sizeof(*buffer));
-	buffer->magic = cpu_to_be32(DRBD_AL_MAGIC);
-	buffer->tr_number = cpu_to_be32(device->al_tr_number);
-
-	i = 0;
-
-	/* Even though no one can start to change this list
-	 * once we set the LC_LOCKED -- from drbd_al_begin_io(),
-	 * lc_try_lock_for_transaction() --, someone may still
-	 * be in the process of changing it. */
-	spin_lock_irq(&device->al_lock);
-	list_for_each_entry(e, &device->act_log->to_be_changed, list) {
-		if (i == AL_UPDATES_PER_TRANSACTION) {
-			i++;
-			break;
-		}
-		buffer->update_slot_nr[i] = cpu_to_be16(e->lc_index);
-		buffer->update_extent_nr[i] = cpu_to_be32(e->lc_new_number);
-		if (e->lc_number != LC_FREE)
-			drbd_bm_mark_for_writeout(device,
-					al_extent_to_bm_page(e->lc_number));
-		i++;
-	}
-	spin_unlock_irq(&device->al_lock);
-	BUG_ON(i > AL_UPDATES_PER_TRANSACTION);
-
-	buffer->n_updates = cpu_to_be16(i);
-	for ( ; i < AL_UPDATES_PER_TRANSACTION; i++) {
-		buffer->update_slot_nr[i] = cpu_to_be16(-1);
-		buffer->update_extent_nr[i] = cpu_to_be32(LC_FREE);
-	}
-
-	buffer->context_size = cpu_to_be16(device->act_log->nr_elements);
-	buffer->context_start_slot_nr = cpu_to_be16(device->al_tr_cycle);
-
-	mx = min_t(int, AL_CONTEXT_PER_TRANSACTION,
-		   device->act_log->nr_elements - device->al_tr_cycle);
-	for (i = 0; i < mx; i++) {
-		unsigned idx = device->al_tr_cycle + i;
-		extent_nr = lc_element_by_index(device->act_log, idx)->lc_number;
-		buffer->context[i] = cpu_to_be32(extent_nr);
-	}
-	for (; i < AL_CONTEXT_PER_TRANSACTION; i++)
-		buffer->context[i] = cpu_to_be32(LC_FREE);
-
-	device->al_tr_cycle += AL_CONTEXT_PER_TRANSACTION;
-	if (device->al_tr_cycle >= device->act_log->nr_elements)
-		device->al_tr_cycle = 0;
-
-	sector = al_tr_number_to_on_disk_sector(device);
-
-	crc = crc32c(0, buffer, 4096);
-	buffer->crc32c = cpu_to_be32(crc);
-
-	if (drbd_bm_write_hinted(device))
-		err = -EIO;
-	else {
-		bool write_al_updates;
-		rcu_read_lock();
-		write_al_updates = rcu_dereference(device->ldev->disk_conf)->al_updates;
-		rcu_read_unlock();
-		if (write_al_updates) {
-			if (drbd_md_sync_page_io(device, device->ldev, sector, WRITE)) {
-				err = -EIO;
-				drbd_chk_io_error(device, 1, DRBD_META_IO_ERROR);
-			} else {
-				device->al_tr_number++;
-				device->al_writ_cnt++;
-			}
-		}
-	}
-
-	drbd_md_put_buffer(device);
-	put_ldev(device);
-
-	return err;
-}
-
 static int _try_lc_del(struct drbd_device *device, struct lc_element *al_ext)
 {
 	int rv;
@@ -606,21 +614,24 @@ void drbd_al_shrink(struct drbd_device *device)
 	wake_up(&device->al_wait);
 }
 
-int drbd_initialize_al(struct drbd_device *device, void *buffer)
+int drbd_al_initialize(struct drbd_device *device, void *buffer)
 {
 	struct al_transaction_on_disk *al = buffer;
 	struct drbd_md *md = &device->ldev->md;
-	sector_t al_base = md->md_offset + md->al_offset;
 	int al_size_4k = md->al_stripes * md->al_stripe_size_4k;
 	int i;
 
-	memset(al, 0, 4096);
-	al->magic = cpu_to_be32(DRBD_AL_MAGIC);
-	al->transaction_type = cpu_to_be16(AL_TR_INITIALIZED);
-	al->crc32c = cpu_to_be32(crc32c(0, al, 4096));
+	__al_write_transaction(device, al);
+	/* There may or may not have been a pending transaction. */
+	spin_lock_irq(&device->al_lock);
+	lc_committed(device->act_log);
+	spin_unlock_irq(&device->al_lock);
 
-	for (i = 0; i < al_size_4k; i++) {
-		int err = drbd_md_sync_page_io(device, device->ldev, al_base + i * 8, WRITE);
+	/* The rest of the transactions will have an empty "updates" list, and
+	 * are written out only to provide the context, and to initialize the
+	 * on-disk ring buffer. */
+	for (i = 1; i < al_size_4k; i++) {
+		int err = __al_write_transaction(device, al);
 		if (err)
 			return err;
 	}

drivers/block/drbd/drbd_bitmap.c

@@ -24,7 +24,7 @@
 
 #define pr_fmt(fmt)	KBUILD_MODNAME ": " fmt
 
-#include <linux/bitops.h>
+#include <linux/bitmap.h>
 #include <linux/vmalloc.h>
 #include <linux/string.h>
 #include <linux/drbd.h>
@@ -479,8 +479,14 @@ void drbd_bm_cleanup(struct drbd_device *device)
  * this masks out the remaining bits.
  * Returns the number of bits cleared.
  */
+#ifndef BITS_PER_PAGE
 #define BITS_PER_PAGE		(1UL << (PAGE_SHIFT + 3))
 #define BITS_PER_PAGE_MASK	(BITS_PER_PAGE - 1)
+#else
+# if BITS_PER_PAGE != (1UL << (PAGE_SHIFT + 3))
+#  error "ambiguous BITS_PER_PAGE"
+# endif
+#endif
 #define BITS_PER_LONG_MASK	(BITS_PER_LONG - 1)
 static int bm_clear_surplus(struct drbd_bitmap *b)
 {
@@ -559,21 +565,19 @@ static unsigned long bm_count_bits(struct drbd_bitmap *b)
 	unsigned long *p_addr;
 	unsigned long bits = 0;
 	unsigned long mask = (1UL << (b->bm_bits & BITS_PER_LONG_MASK)) -1;
-	int idx, i, last_word;
+	int idx, last_word;
 
 	/* all but last page */
 	for (idx = 0; idx < b->bm_number_of_pages - 1; idx++) {
 		p_addr = __bm_map_pidx(b, idx);
-		for (i = 0; i < LWPP; i++)
-			bits += hweight_long(p_addr[i]);
+		bits += bitmap_weight(p_addr, BITS_PER_PAGE);
 		__bm_unmap(p_addr);
 		cond_resched();
 	}
 	/* last (or only) page */
 	last_word = ((b->bm_bits - 1) & BITS_PER_PAGE_MASK) >> LN2_BPL;
 	p_addr = __bm_map_pidx(b, idx);
-	for (i = 0; i < last_word; i++)
-		bits += hweight_long(p_addr[i]);
+	bits += bitmap_weight(p_addr, last_word * BITS_PER_LONG);
 	p_addr[last_word] &= cpu_to_lel(mask);
 	bits += hweight_long(p_addr[last_word]);
 	/* 32bit arch, may have an unused padding long */
@@ -1419,6 +1423,9 @@ static inline void bm_set_full_words_within_one_page(struct drbd_bitmap *b,
 	int bits;
 	int changed = 0;
 	unsigned long *paddr = kmap_atomic(b->bm_pages[page_nr]);
+
+	/* I think it is more cache line friendly to hweight_long then set to ~0UL,
+	 * than to first bitmap_weight() all words, then bitmap_fill() all words */
 	for (i = first_word; i < last_word; i++) {
 		bits = hweight_long(paddr[i]);
 		paddr[i] = ~0UL;
@@ -1628,8 +1635,7 @@ int drbd_bm_e_weight(struct drbd_device *device, unsigned long enr)
 		int n = e-s;
 		p_addr = bm_map_pidx(b, bm_word_to_page_idx(b, s));
 		bm = p_addr + MLPP(s);
-		while (n--)
-			count += hweight_long(*bm++);
+		count += bitmap_weight(bm, n * BITS_PER_LONG);
 		bm_unmap(p_addr);
 	} else {
 		drbd_err(device, "start offset (%d) too large in drbd_bm_e_weight\n", s);

drivers/block/drbd/drbd_debugfs.c

@@ -771,6 +771,13 @@ static int device_data_gen_id_show(struct seq_file *m, void *ignored)
 	return 0;
 }
 
+static int device_ed_gen_id_show(struct seq_file *m, void *ignored)
+{
+	struct drbd_device *device = m->private;
+	seq_printf(m, "0x%016llX\n", (unsigned long long)device->ed_uuid);
+	return 0;
+}
+
 #define drbd_debugfs_device_attr(name)						\
 static int device_ ## name ## _open(struct inode *inode, struct file *file)	\
 {										\
@@ -796,6 +803,7 @@ drbd_debugfs_device_attr(oldest_requests)
 drbd_debugfs_device_attr(act_log_extents)
 drbd_debugfs_device_attr(resync_extents)
 drbd_debugfs_device_attr(data_gen_id)
+drbd_debugfs_device_attr(ed_gen_id)
 
 void drbd_debugfs_device_add(struct drbd_device *device)
 {
@@ -839,6 +847,7 @@ void drbd_debugfs_device_add(struct drbd_device *device)
 	DCF(act_log_extents);
 	DCF(resync_extents);
 	DCF(data_gen_id);
+	DCF(ed_gen_id);
 #undef DCF
 	return;
 
@@ -854,6 +863,7 @@ void drbd_debugfs_device_cleanup(struct drbd_device *device)
 	drbd_debugfs_remove(&device->debugfs_vol_act_log_extents);
 	drbd_debugfs_remove(&device->debugfs_vol_resync_extents);
 	drbd_debugfs_remove(&device->debugfs_vol_data_gen_id);
+	drbd_debugfs_remove(&device->debugfs_vol_ed_gen_id);
 	drbd_debugfs_remove(&device->debugfs_vol);
 }
 

drivers/block/drbd/drbd_int.h

@@ -77,13 +77,6 @@ extern int fault_devs;
 extern char usermode_helper[];
 
 
-/* I don't remember why XCPU ...
- * This is used to wake the asender,
- * and to interrupt sending the sending task
- * on disconnect.
- */
-#define DRBD_SIG SIGXCPU
-
 /* This is used to stop/restart our threads.
  * Cannot use SIGTERM nor SIGKILL, since these
  * are sent out by init on runlevel changes
@@ -292,6 +285,9 @@ struct drbd_device_work {
 
 extern int drbd_wait_misc(struct drbd_device *, struct drbd_interval *);
 
+extern void lock_all_resources(void);
+extern void unlock_all_resources(void);
+
 struct drbd_request {
 	struct drbd_work w;
 	struct drbd_device *device;
@@ -504,7 +500,6 @@ enum {
 
 	MD_NO_FUA,		/* Users wants us to not use FUA/FLUSH on meta data dev */
 
-	SUSPEND_IO,		/* suspend application io */
 	BITMAP_IO,		/* suspend application io;
 				   once no more io in flight, start bitmap io */
 	BITMAP_IO_QUEUED,       /* Started bitmap IO */
@@ -632,12 +627,6 @@ struct bm_io_work {
 	void (*done)(struct drbd_device *device, int rv);
 };
 
-enum write_ordering_e {
-	WO_none,
-	WO_drain_io,
-	WO_bdev_flush,
-};
-
 struct fifo_buffer {
 	unsigned int head_index;
 	unsigned int size;
@@ -650,8 +639,7 @@ extern struct fifo_buffer *fifo_alloc(int fifo_size);
 enum {
 	NET_CONGESTED,		/* The data socket is congested */
 	RESOLVE_CONFLICTS,	/* Set on one node, cleared on the peer! */
-	SEND_PING,		/* whether asender should send a ping asap */
-	SIGNAL_ASENDER,		/* whether asender wants to be interrupted */
+	SEND_PING,
 	GOT_PING_ACK,		/* set when we receive a ping_ack packet, ping_wait gets woken */
 	CONN_WD_ST_CHG_REQ,	/* A cluster wide state change on the connection is active */
 	CONN_WD_ST_CHG_OKAY,
@@ -670,6 +658,8 @@ enum {
 	DEVICE_WORK_PENDING,	/* tell worker that some device has pending work */
 };
 
+enum which_state { NOW, OLD = NOW, NEW };
+
 struct drbd_resource {
 	char *name;
 #ifdef CONFIG_DEBUG_FS
@@ -755,7 +745,8 @@ struct drbd_connection {
 	unsigned long last_reconnect_jif;
 	struct drbd_thread receiver;
 	struct drbd_thread worker;
-	struct drbd_thread asender;
+	struct drbd_thread ack_receiver;
+	struct workqueue_struct *ack_sender;
 
 	/* cached pointers,
 	 * so we can look up the oldest pending requests more quickly.
@@ -774,6 +765,8 @@ struct drbd_connection {
 	struct drbd_thread_timing_details r_timing_details[DRBD_THREAD_DETAILS_HIST];
 
 	struct {
+		unsigned long last_sent_barrier_jif;
+
 		/* whether this sender thread
 		 * has processed a single write yet. */
 		bool seen_any_write_yet;
@@ -788,6 +781,17 @@ struct drbd_connection {
 	} send;
 };
 
+static inline bool has_net_conf(struct drbd_connection *connection)
+{
+	bool has_net_conf;
+
+	rcu_read_lock();
+	has_net_conf = rcu_dereference(connection->net_conf);
+	rcu_read_unlock();
+
+	return has_net_conf;
+}
+
 void __update_timing_details(
 		struct drbd_thread_timing_details *tdp,
 		unsigned int *cb_nr,
@@ -811,6 +815,7 @@ struct drbd_peer_device {
 	struct list_head peer_devices;
 	struct drbd_device *device;
 	struct drbd_connection *connection;
+	struct work_struct send_acks_work;
 #ifdef CONFIG_DEBUG_FS
 	struct dentry *debugfs_peer_dev;
 #endif
@@ -829,6 +834,7 @@ struct drbd_device {
 	struct dentry *debugfs_vol_act_log_extents;
 	struct dentry *debugfs_vol_resync_extents;
 	struct dentry *debugfs_vol_data_gen_id;
+	struct dentry *debugfs_vol_ed_gen_id;
 #endif
 
 	unsigned int vnr;	/* volume number within the connection */
@@ -873,6 +879,7 @@ struct drbd_device {
 	atomic_t rs_pending_cnt; /* RS request/data packets on the wire */
 	atomic_t unacked_cnt;	 /* Need to send replies for */
 	atomic_t local_cnt;	 /* Waiting for local completion */
+	atomic_t suspend_cnt;
 
 	/* Interval tree of pending local requests */
 	struct rb_root read_requests;
@@ -1020,6 +1027,12 @@ static inline struct drbd_peer_device *first_peer_device(struct drbd_device *dev
 	return list_first_entry_or_null(&device->peer_devices, struct drbd_peer_device, peer_devices);
 }
 
+static inline struct drbd_peer_device *
+conn_peer_device(struct drbd_connection *connection, int volume_number)
+{
+	return idr_find(&connection->peer_devices, volume_number);
+}
+
 #define for_each_resource(resource, _resources) \
 	list_for_each_entry(resource, _resources, resources)
 
@@ -1113,7 +1126,7 @@ extern int drbd_send_ov_request(struct drbd_peer_device *, sector_t sector, int
 extern int drbd_send_bitmap(struct drbd_device *device);
 extern void drbd_send_sr_reply(struct drbd_peer_device *, enum drbd_state_rv retcode);
 extern void conn_send_sr_reply(struct drbd_connection *connection, enum drbd_state_rv retcode);
-extern void drbd_free_ldev(struct drbd_backing_dev *ldev);
+extern void drbd_backing_dev_free(struct drbd_device *device, struct drbd_backing_dev *ldev);
 extern void drbd_device_cleanup(struct drbd_device *device);
 void drbd_print_uuids(struct drbd_device *device, const char *text);
 
@@ -1424,7 +1437,7 @@ extern struct bio_set *drbd_md_io_bio_set;
 /* to allocate from that set */
 extern struct bio *bio_alloc_drbd(gfp_t gfp_mask);
 
-extern rwlock_t global_state_lock;
+extern struct mutex resources_mutex;
 
 extern int conn_lowest_minor(struct drbd_connection *connection);
 extern enum drbd_ret_code drbd_create_device(struct drbd_config_context *adm_ctx, unsigned int minor);
@@ -1454,6 +1467,9 @@ extern int is_valid_ar_handle(struct drbd_request *, sector_t);
 
 
 /* drbd_nl.c */
+
+extern struct mutex notification_mutex;
+
 extern void drbd_suspend_io(struct drbd_device *device);
 extern void drbd_resume_io(struct drbd_device *device);
 extern char *ppsize(char *buf, unsigned long long size);
@@ -1536,7 +1552,9 @@ extern void drbd_endio_write_sec_final(struct drbd_peer_request *peer_req);
 
 /* drbd_receiver.c */
 extern int drbd_receiver(struct drbd_thread *thi);
-extern int drbd_asender(struct drbd_thread *thi);
+extern int drbd_ack_receiver(struct drbd_thread *thi);
+extern void drbd_send_ping_wf(struct work_struct *ws);
+extern void drbd_send_acks_wf(struct work_struct *ws);
 extern bool drbd_rs_c_min_rate_throttle(struct drbd_device *device);
 extern bool drbd_rs_should_slow_down(struct drbd_device *device, sector_t sector,
 		bool throttle_if_app_is_waiting);
@@ -1649,7 +1667,7 @@ extern int __drbd_change_sync(struct drbd_device *device, sector_t sector, int s
 #define drbd_rs_failed_io(device, sector, size) \
 	__drbd_change_sync(device, sector, size, RECORD_RS_FAILED)
 extern void drbd_al_shrink(struct drbd_device *device);
-extern int drbd_initialize_al(struct drbd_device *, void *);
+extern int drbd_al_initialize(struct drbd_device *, void *);
 
 /* drbd_nl.c */
 /* state info broadcast */
@@ -1668,6 +1686,29 @@ struct sib_info {
 };
 void drbd_bcast_event(struct drbd_device *device, const struct sib_info *sib);
 
+extern void notify_resource_state(struct sk_buff *,
+				  unsigned int,
+				  struct drbd_resource *,
+				  struct resource_info *,
+				  enum drbd_notification_type);
+extern void notify_device_state(struct sk_buff *,
+				unsigned int,
+				struct drbd_device *,
+				struct device_info *,
+				enum drbd_notification_type);
+extern void notify_connection_state(struct sk_buff *,
+				    unsigned int,
+				    struct drbd_connection *,
+				    struct connection_info *,
+				    enum drbd_notification_type);
+extern void notify_peer_device_state(struct sk_buff *,
+				     unsigned int,
+				     struct drbd_peer_device *,
+				     struct peer_device_info *,
+				     enum drbd_notification_type);
+extern void notify_helper(enum drbd_notification_type, struct drbd_device *,
+			  struct drbd_connection *, const char *, int);
+
 /*
  * inline helper functions
  *************************/
@@ -1694,19 +1735,6 @@ static inline int drbd_peer_req_has_active_page(struct drbd_peer_request *peer_r
 	return 0;
 }
 
-static inline enum drbd_state_rv
-_drbd_set_state(struct drbd_device *device, union drbd_state ns,
-		enum chg_state_flags flags, struct completion *done)
-{
-	enum drbd_state_rv rv;
-
-	read_lock(&global_state_lock);
-	rv = __drbd_set_state(device, ns, flags, done);
-	read_unlock(&global_state_lock);
-
-	return rv;
-}
-
 static inline union drbd_state drbd_read_state(struct drbd_device *device)
 {
 	struct drbd_resource *resource = device->resource;
@@ -1937,16 +1965,21 @@ drbd_device_post_work(struct drbd_device *device, int work_bit)
 
 extern void drbd_flush_workqueue(struct drbd_work_queue *work_queue);
 
-static inline void wake_asender(struct drbd_connection *connection)
+/* To get the ack_receiver out of the blocking network stack,
+ * so it can change its sk_rcvtimeo from idle- to ping-timeout,
+ * and send a ping, we need to send a signal.
+ * Which signal we send is irrelevant. */
+static inline void wake_ack_receiver(struct drbd_connection *connection)
 {
-	if (test_bit(SIGNAL_ASENDER, &connection->flags))
-		force_sig(DRBD_SIG, connection->asender.task);
+	struct task_struct *task = connection->ack_receiver.task;
+	if (task && get_t_state(&connection->ack_receiver) == RUNNING)
+		force_sig(SIGXCPU, task);
 }
 
 static inline void request_ping(struct drbd_connection *connection)
 {
 	set_bit(SEND_PING, &connection->flags);
-	wake_asender(connection);
+	wake_ack_receiver(connection);
 }
 
 extern void *conn_prepare_command(struct drbd_connection *, struct drbd_socket *);
@@ -2230,7 +2263,7 @@ static inline bool may_inc_ap_bio(struct drbd_device *device)
 
 	if (drbd_suspended(device))
 		return false;
-	if (test_bit(SUSPEND_IO, &device->flags))
+	if (atomic_read(&device->suspend_cnt))
 		return false;
 
 	/* to avoid potential deadlock or bitmap corruption,

drivers/block/drbd/drbd_main.c

@@ -117,6 +117,7 @@ module_param_string(usermode_helper, usermode_helper, sizeof(usermode_helper), 0
  */
 struct idr drbd_devices;
 struct list_head drbd_resources;
+struct mutex resources_mutex;
 
 struct kmem_cache *drbd_request_cache;
 struct kmem_cache *drbd_ee_cache;	/* peer requests */
@@ -1435,8 +1436,8 @@ static int we_should_drop_the_connection(struct drbd_connection *connection, str
 	/* long elapsed = (long)(jiffies - device->last_received); */
 
 	drop_it =   connection->meta.socket == sock
-		|| !connection->asender.task
-		|| get_t_state(&connection->asender) != RUNNING
+		|| !connection->ack_receiver.task
+		|| get_t_state(&connection->ack_receiver) != RUNNING
 		|| connection->cstate < C_WF_REPORT_PARAMS;
 
 	if (drop_it)
@@ -1793,15 +1794,6 @@ int drbd_send(struct drbd_connection *connection, struct socket *sock,
 		drbd_update_congested(connection);
 	}
 	do {
-		/* STRANGE
-		 * tcp_sendmsg does _not_ use its size parameter at all ?
-		 *
-		 * -EAGAIN on timeout, -EINTR on signal.
-		 */
-/* THINK
- * do we need to block DRBD_SIG if sock == &meta.socket ??
- * otherwise wake_asender() might interrupt some send_*Ack !
- */
 		rv = kernel_sendmsg(sock, &msg, &iov, 1, size);
 		if (rv == -EAGAIN) {
 			if (we_should_drop_the_connection(connection, sock))
@@ -2000,7 +1992,7 @@ void drbd_device_cleanup(struct drbd_device *device)
 		drbd_bm_cleanup(device);
 	}
 
-	drbd_free_ldev(device->ldev);
+	drbd_backing_dev_free(device, device->ldev);
 	device->ldev = NULL;
 
 	clear_bit(AL_SUSPENDED, &device->flags);
@@ -2179,7 +2171,7 @@ void drbd_destroy_device(struct kref *kref)
 	if (device->this_bdev)
 		bdput(device->this_bdev);
 
-	drbd_free_ldev(device->ldev);
+	drbd_backing_dev_free(device, device->ldev);
 	device->ldev = NULL;
 
 	drbd_release_all_peer_reqs(device);
@@ -2563,7 +2555,7 @@ int set_resource_options(struct drbd_resource *resource, struct res_opts *res_op
 		cpumask_copy(resource->cpu_mask, new_cpu_mask);
 		for_each_connection_rcu(connection, resource) {
 			connection->receiver.reset_cpu_mask = 1;
-			connection->asender.reset_cpu_mask = 1;
+			connection->ack_receiver.reset_cpu_mask = 1;
 			connection->worker.reset_cpu_mask = 1;
 		}
 	}
@@ -2590,7 +2582,7 @@ struct drbd_resource *drbd_create_resource(const char *name)
 	kref_init(&resource->kref);
 	idr_init(&resource->devices);
 	INIT_LIST_HEAD(&resource->connections);
-	resource->write_ordering = WO_bdev_flush;
+	resource->write_ordering = WO_BDEV_FLUSH;
 	list_add_tail_rcu(&resource->resources, &drbd_resources);
 	mutex_init(&resource->conf_update);
 	mutex_init(&resource->adm_mutex);
@@ -2652,8 +2644,8 @@ struct drbd_connection *conn_create(const char *name, struct res_opts *res_opts)
 	connection->receiver.connection = connection;
 	drbd_thread_init(resource, &connection->worker, drbd_worker, "worker");
 	connection->worker.connection = connection;
-	drbd_thread_init(resource, &connection->asender, drbd_asender, "asender");
-	connection->asender.connection = connection;
+	drbd_thread_init(resource, &connection->ack_receiver, drbd_ack_receiver, "ack_recv");
+	connection->ack_receiver.connection = connection;
 
 	kref_init(&connection->kref);
 
@@ -2702,8 +2694,8 @@ static int init_submitter(struct drbd_device *device)
 {
 	/* opencoded create_singlethread_workqueue(),
 	 * to be able to say "drbd%d", ..., minor */
-	device->submit.wq = alloc_workqueue("drbd%u_submit",
-			WQ_UNBOUND | WQ_MEM_RECLAIM, 1, device->minor);
+	device->submit.wq =
+		alloc_ordered_workqueue("drbd%u_submit", WQ_MEM_RECLAIM, device->minor);
 	if (!device->submit.wq)
 		return -ENOMEM;
 
@@ -2820,6 +2812,7 @@ enum drbd_ret_code drbd_create_device(struct drbd_config_context *adm_ctx, unsig
 			goto out_idr_remove_from_resource;
 		}
 		kref_get(&connection->kref);
+		INIT_WORK(&peer_device->send_acks_work, drbd_send_acks_wf);
 	}
 
 	if (init_submitter(device)) {
@@ -2923,7 +2916,7 @@ static int __init drbd_init(void)
 	drbd_proc = NULL; /* play safe for drbd_cleanup */
 	idr_init(&drbd_devices);
 
-	rwlock_init(&global_state_lock);
+	mutex_init(&resources_mutex);
 	INIT_LIST_HEAD(&drbd_resources);
 
 	err = drbd_genl_register();
@@ -2971,18 +2964,6 @@ static int __init drbd_init(void)
 	return err;
 }
 
-void drbd_free_ldev(struct drbd_backing_dev *ldev)
-{
-	if (ldev == NULL)
-		return;
-
-	blkdev_put(ldev->backing_bdev, FMODE_READ | FMODE_WRITE | FMODE_EXCL);
-	blkdev_put(ldev->md_bdev, FMODE_READ | FMODE_WRITE | FMODE_EXCL);
-
-	kfree(ldev->disk_conf);
-	kfree(ldev);
-}
-
 static void drbd_free_one_sock(struct drbd_socket *ds)
 {
 	struct socket *s;
@@ -3277,6 +3258,10 @@ int drbd_md_read(struct drbd_device *device, struct drbd_backing_dev *bdev)
 	 * and read it. */
 	bdev->md.meta_dev_idx = bdev->disk_conf->meta_dev_idx;
 	bdev->md.md_offset = drbd_md_ss(bdev);
+	/* Even for (flexible or indexed) external meta data,
+	 * initially restrict us to the 4k superblock for now.
+	 * Affects the paranoia out-of-range access check in drbd_md_sync_page_io(). */
+	bdev->md.md_size_sect = 8;
 
 	if (drbd_md_sync_page_io(device, bdev, bdev->md.md_offset, READ)) {
 		/* NOTE: can't do normal error processing here as this is
@@ -3578,7 +3563,9 @@ void drbd_queue_bitmap_io(struct drbd_device *device,
 
 	spin_lock_irq(&device->resource->req_lock);
 	set_bit(BITMAP_IO, &device->flags);
-	if (atomic_read(&device->ap_bio_cnt) == 0) {
+	/* don't wait for pending application IO if the caller indicates that
+	 * application IO does not conflict anyways. */
+	if (flags == BM_LOCKED_CHANGE_ALLOWED || atomic_read(&device->ap_bio_cnt) == 0) {
 		if (!test_and_set_bit(BITMAP_IO_QUEUED, &device->flags))
 			drbd_queue_work(&first_peer_device(device)->connection->sender_work,
 					&device->bm_io_work.w);
@@ -3746,6 +3733,27 @@ int drbd_wait_misc(struct drbd_device *device, struct drbd_interval *i)
 	return 0;
 }
 
+void lock_all_resources(void)
+{
+	struct drbd_resource *resource;
+	int __maybe_unused i = 0;
+
+	mutex_lock(&resources_mutex);
+	local_irq_disable();
+	for_each_resource(resource, &drbd_resources)
+		spin_lock_nested(&resource->req_lock, i++);
+}
+
+void unlock_all_resources(void)
+{
+	struct drbd_resource *resource;
+
+	for_each_resource(resource, &drbd_resources)
+		spin_unlock(&resource->req_lock);
+	local_irq_enable();
+	mutex_unlock(&resources_mutex);
+}
+
 #ifdef CONFIG_DRBD_FAULT_INJECTION
 /* Fault insertion support including random number generator shamelessly
  * stolen from kernel/rcutorture.c */

drivers/block/drbd/drbd_nl.c

@@ -36,6 +36,7 @@
 #include "drbd_int.h"
 #include "drbd_protocol.h"
 #include "drbd_req.h"
+#include "drbd_state_change.h"
 #include <asm/unaligned.h>
 #include <linux/drbd_limits.h>
 #include <linux/kthread.h>
@@ -75,11 +76,24 @@ int drbd_adm_get_status(struct sk_buff *skb, struct genl_info *info);
 int drbd_adm_get_timeout_type(struct sk_buff *skb, struct genl_info *info);
 /* .dumpit */
 int drbd_adm_get_status_all(struct sk_buff *skb, struct netlink_callback *cb);
+int drbd_adm_dump_resources(struct sk_buff *skb, struct netlink_callback *cb);
+int drbd_adm_dump_devices(struct sk_buff *skb, struct netlink_callback *cb);
+int drbd_adm_dump_devices_done(struct netlink_callback *cb);
+int drbd_adm_dump_connections(struct sk_buff *skb, struct netlink_callback *cb);
+int drbd_adm_dump_connections_done(struct netlink_callback *cb);
+int drbd_adm_dump_peer_devices(struct sk_buff *skb, struct netlink_callback *cb);
+int drbd_adm_dump_peer_devices_done(struct netlink_callback *cb);
+int drbd_adm_get_initial_state(struct sk_buff *skb, struct netlink_callback *cb);
 
 #include <linux/drbd_genl_api.h>
 #include "drbd_nla.h"
 #include <linux/genl_magic_func.h>
 
+static atomic_t drbd_genl_seq = ATOMIC_INIT(2); /* two. */
+static atomic_t notify_genl_seq = ATOMIC_INIT(2); /* two. */
+
+DEFINE_MUTEX(notification_mutex);
+
 /* used blkdev_get_by_path, to claim our meta data device(s) */
 static char *drbd_m_holder = "Hands off! this is DRBD's meta data device.";
 
@@ -349,6 +363,7 @@ int drbd_khelper(struct drbd_device *device, char *cmd)
 	sib.sib_reason = SIB_HELPER_PRE;
 	sib.helper_name = cmd;
 	drbd_bcast_event(device, &sib);
+	notify_helper(NOTIFY_CALL, device, connection, cmd, 0);
 	ret = call_usermodehelper(usermode_helper, argv, envp, UMH_WAIT_PROC);
 	if (ret)
 		drbd_warn(device, "helper command: %s %s %s exit code %u (0x%x)\n",
@@ -361,6 +376,7 @@ int drbd_khelper(struct drbd_device *device, char *cmd)
 	sib.sib_reason = SIB_HELPER_POST;
 	sib.helper_exit_code = ret;
 	drbd_bcast_event(device, &sib);
+	notify_helper(NOTIFY_RESPONSE, device, connection, cmd, ret);
 
 	if (current == connection->worker.task)
 		clear_bit(CALLBACK_PENDING, &connection->flags);
@@ -388,6 +404,7 @@ static int conn_khelper(struct drbd_connection *connection, char *cmd)
 
 	drbd_info(connection, "helper command: %s %s %s\n", usermode_helper, cmd, resource_name);
 	/* TODO: conn_bcast_event() ?? */
+	notify_helper(NOTIFY_CALL, NULL, connection, cmd, 0);
 
 	ret = call_usermodehelper(usermode_helper, argv, envp, UMH_WAIT_PROC);
 	if (ret)
@@ -399,6 +416,7 @@ static int conn_khelper(struct drbd_connection *connection, char *cmd)
 			  usermode_helper, cmd, resource_name,
 			  (ret >> 8) & 0xff, ret);
 	/* TODO: conn_bcast_event() ?? */
+	notify_helper(NOTIFY_RESPONSE, NULL, connection, cmd, ret);
 
 	if (ret < 0) /* Ignore any ERRNOs we got. */
 		ret = 0;
@@ -847,9 +865,11 @@ char *ppsize(char *buf, unsigned long long size)
  * and can be long lived.
  * This changes an device->flag, is triggered by drbd internals,
  * and should be short-lived. */
+/* It needs to be a counter, since multiple threads might
+   independently suspend and resume IO. */
 void drbd_suspend_io(struct drbd_device *device)
 {
-	set_bit(SUSPEND_IO, &device->flags);
+	atomic_inc(&device->suspend_cnt);
 	if (drbd_suspended(device))
 		return;
 	wait_event(device->misc_wait, !atomic_read(&device->ap_bio_cnt));
@@ -857,8 +877,8 @@ void drbd_suspend_io(struct drbd_device *device)
 
 void drbd_resume_io(struct drbd_device *device)
 {
-	clear_bit(SUSPEND_IO, &device->flags);
-	wake_up(&device->misc_wait);
+	if (atomic_dec_and_test(&device->suspend_cnt))
+		wake_up(&device->misc_wait);
 }
 
 /**
@@ -871,27 +891,32 @@ void drbd_resume_io(struct drbd_device *device)
 enum determine_dev_size
 drbd_determine_dev_size(struct drbd_device *device, enum dds_flags flags, struct resize_parms *rs) __must_hold(local)
 {
-	sector_t prev_first_sect, prev_size; /* previous meta location */
-	sector_t la_size_sect, u_size;
+	struct md_offsets_and_sizes {
+		u64 last_agreed_sect;
+		u64 md_offset;
+		s32 al_offset;
+		s32 bm_offset;
+		u32 md_size_sect;
+
+		u32 al_stripes;
+		u32 al_stripe_size_4k;
+	} prev;
+	sector_t u_size, size;
 	struct drbd_md *md = &device->ldev->md;
-	u32 prev_al_stripe_size_4k;
-	u32 prev_al_stripes;
-	sector_t size;
 	char ppb[10];
 	void *buffer;
 
 	int md_moved, la_size_changed;
 	enum determine_dev_size rv = DS_UNCHANGED;
 
-	/* race:
-	 * application request passes inc_ap_bio,
-	 * but then cannot get an AL-reference.
-	 * this function later may wait on ap_bio_cnt == 0. -> deadlock.
+	/* We may change the on-disk offsets of our meta data below.  Lock out
+	 * anything that may cause meta data IO, to avoid acting on incomplete
+	 * layout changes or scribbling over meta data that is in the process
+	 * of being moved.
 	 *
-	 * to avoid that:
-	 * Suspend IO right here.
-	 * still lock the act_log to not trigger ASSERTs there.
-	 */
+	 * Move is not exactly correct, btw, currently we have all our meta
+	 * data in core memory, to "move" it we just write it all out, there
+	 * are no reads. */
 	drbd_suspend_io(device);
 	buffer = drbd_md_get_buffer(device, __func__); /* Lock meta-data IO */
 	if (!buffer) {
@@ -899,19 +924,17 @@ drbd_determine_dev_size(struct drbd_device *device, enum dds_flags flags, struct
 		return DS_ERROR;
 	}
 
-	/* no wait necessary anymore, actually we could assert that */
-	wait_event(device->al_wait, lc_try_lock(device->act_log));
-
-	prev_first_sect = drbd_md_first_sector(device->ldev);
-	prev_size = device->ldev->md.md_size_sect;
-	la_size_sect = device->ldev->md.la_size_sect;
+	/* remember current offset and sizes */
+	prev.last_agreed_sect = md->la_size_sect;
+	prev.md_offset = md->md_offset;
+	prev.al_offset = md->al_offset;
+	prev.bm_offset = md->bm_offset;
+	prev.md_size_sect = md->md_size_sect;
+	prev.al_stripes = md->al_stripes;
+	prev.al_stripe_size_4k = md->al_stripe_size_4k;
 
 	if (rs) {
 		/* rs is non NULL if we should change the AL layout only */
-
-		prev_al_stripes = md->al_stripes;
-		prev_al_stripe_size_4k = md->al_stripe_size_4k;
-
 		md->al_stripes = rs->al_stripes;
 		md->al_stripe_size_4k = rs->al_stripe_size / 4;
 		md->al_size_4k = (u64)rs->al_stripes * rs->al_stripe_size / 4;
@@ -924,7 +947,7 @@ drbd_determine_dev_size(struct drbd_device *device, enum dds_flags flags, struct
 	rcu_read_unlock();
 	size = drbd_new_dev_size(device, device->ldev, u_size, flags & DDSF_FORCED);
 
-	if (size < la_size_sect) {
+	if (size < prev.last_agreed_sect) {
 		if (rs && u_size == 0) {
 			/* Remove "rs &&" later. This check should always be active, but
 			   right now the receiver expects the permissive behavior */
@@ -945,30 +968,29 @@ drbd_determine_dev_size(struct drbd_device *device, enum dds_flags flags, struct
 		err = drbd_bm_resize(device, size, !(flags & DDSF_NO_RESYNC));
 		if (unlikely(err)) {
 			/* currently there is only one error: ENOMEM! */
-			size = drbd_bm_capacity(device)>>1;
+			size = drbd_bm_capacity(device);
 			if (size == 0) {
 				drbd_err(device, "OUT OF MEMORY! "
 				    "Could not allocate bitmap!\n");
 			} else {
 				drbd_err(device, "BM resizing failed. "
-				    "Leaving size unchanged at size = %lu KB\n",
-				    (unsigned long)size);
+				    "Leaving size unchanged\n");
 			}
 			rv = DS_ERROR;
 		}
 		/* racy, see comments above. */
 		drbd_set_my_capacity(device, size);
-		device->ldev->md.la_size_sect = size;
+		md->la_size_sect = size;
 		drbd_info(device, "size = %s (%llu KB)\n", ppsize(ppb, size>>1),
 		     (unsigned long long)size>>1);
 	}
 	if (rv <= DS_ERROR)
 		goto err_out;
 
-	la_size_changed = (la_size_sect != device->ldev->md.la_size_sect);
+	la_size_changed = (prev.last_agreed_sect != md->la_size_sect);
 
-	md_moved = prev_first_sect != drbd_md_first_sector(device->ldev)
-		|| prev_size	   != device->ldev->md.md_size_sect;
+	md_moved = prev.md_offset    != md->md_offset
+		|| prev.md_size_sect != md->md_size_sect;
 
 	if (la_size_changed || md_moved || rs) {
 		u32 prev_flags;
@@ -977,20 +999,29 @@ drbd_determine_dev_size(struct drbd_device *device, enum dds_flags flags, struct
 		 * Clear the timer, to avoid scary "timer expired!" messages,
 		 * "Superblock" is written out at least twice below, anyways. */
 		del_timer(&device->md_sync_timer);
-		drbd_al_shrink(device); /* All extents inactive. */
 
+		/* We won't change the "al-extents" setting, we just may need
+		 * to move the on-disk location of the activity log ringbuffer.
+		 * Lock for transaction is good enough, it may well be "dirty"
+		 * or even "starving". */
+		wait_event(device->al_wait, lc_try_lock_for_transaction(device->act_log));
+
+		/* mark current on-disk bitmap and activity log as unreliable */
 		prev_flags = md->flags;
-		md->flags &= ~MDF_PRIMARY_IND;
+		md->flags |= MDF_FULL_SYNC | MDF_AL_DISABLED;
 		drbd_md_write(device, buffer);
 
+		drbd_al_initialize(device, buffer);
+
 		drbd_info(device, "Writing the whole bitmap, %s\n",
 			 la_size_changed && md_moved ? "size changed and md moved" :
 			 la_size_changed ? "size changed" : "md moved");
 		/* next line implicitly does drbd_suspend_io()+drbd_resume_io() */
 		drbd_bitmap_io(device, md_moved ? &drbd_bm_write_all : &drbd_bm_write,
 			       "size changed", BM_LOCKED_MASK);
-		drbd_initialize_al(device, buffer);
 
+		/* on-disk bitmap and activity log is authoritative again
+		 * (unless there was an IO error meanwhile...) */
 		md->flags = prev_flags;
 		drbd_md_write(device, buffer);
 
@@ -999,20 +1030,22 @@ drbd_determine_dev_size(struct drbd_device *device, enum dds_flags flags, struct
 				  md->al_stripes, md->al_stripe_size_4k * 4);
 	}
 
-	if (size > la_size_sect)
-		rv = la_size_sect ? DS_GREW : DS_GREW_FROM_ZERO;
-	if (size < la_size_sect)
+	if (size > prev.last_agreed_sect)
+		rv = prev.last_agreed_sect ? DS_GREW : DS_GREW_FROM_ZERO;
+	if (size < prev.last_agreed_sect)
 		rv = DS_SHRUNK;
 
 	if (0) {
 	err_out:
-		if (rs) {
-			md->al_stripes = prev_al_stripes;
-			md->al_stripe_size_4k = prev_al_stripe_size_4k;
-			md->al_size_4k = (u64)prev_al_stripes * prev_al_stripe_size_4k;
-
-			drbd_md_set_sector_offsets(device, device->ldev);
-		}
+		/* restore previous offset and sizes */
+		md->la_size_sect = prev.last_agreed_sect;
+		md->md_offset = prev.md_offset;
+		md->al_offset = prev.al_offset;
+		md->bm_offset = prev.bm_offset;
+		md->md_size_sect = prev.md_size_sect;
+		md->al_stripes = prev.al_stripes;
+		md->al_stripe_size_4k = prev.al_stripe_size_4k;
+		md->al_size_4k = (u64)prev.al_stripes * prev.al_stripe_size_4k;
 	}
 	lc_unlock(device->act_log);
 	wake_up(&device->al_wait);
@@ -1115,8 +1148,7 @@ static int drbd_check_al_size(struct drbd_device *device, struct disk_conf *dc)
 		lc_destroy(n);
 		return -EBUSY;
 	} else {
-		if (t)
-			lc_destroy(t);
+		lc_destroy(t);
 	}
 	drbd_md_mark_dirty(device); /* we changed device->act_log->nr_elemens */
 	return 0;
@@ -1151,21 +1183,20 @@ static void drbd_setup_queue_param(struct drbd_device *device, struct drbd_backi
 	if (b) {
 		struct drbd_connection *connection = first_peer_device(device)->connection;
 
+		blk_queue_max_discard_sectors(q, DRBD_MAX_DISCARD_SECTORS);
+
 		if (blk_queue_discard(b) &&
 		    (connection->cstate < C_CONNECTED || connection->agreed_features & FF_TRIM)) {
-			/* For now, don't allow more than one activity log extent worth of data
-			 * to be discarded in one go. We may need to rework drbd_al_begin_io()
-			 * to allow for even larger discard ranges */
-			blk_queue_max_discard_sectors(q, DRBD_MAX_DISCARD_SECTORS);
-
+			/* We don't care, stacking below should fix it for the local device.
+			 * Whether or not it is a suitable granularity on the remote device
+			 * is not our problem, really. If you care, you need to
+			 * use devices with similar topology on all peers. */
+			q->limits.discard_granularity = 512;
 			queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, q);
-			/* REALLY? Is stacking secdiscard "legal"? */
-			if (blk_queue_secdiscard(b))
-				queue_flag_set_unlocked(QUEUE_FLAG_SECDISCARD, q);
 		} else {
 			blk_queue_max_discard_sectors(q, 0);
 			queue_flag_clear_unlocked(QUEUE_FLAG_DISCARD, q);
-			queue_flag_clear_unlocked(QUEUE_FLAG_SECDISCARD, q);
+			q->limits.discard_granularity = 0;
 		}
 
 		blk_queue_stack_limits(q, b);
@@ -1177,6 +1208,12 @@ static void drbd_setup_queue_param(struct drbd_device *device, struct drbd_backi
 			q->backing_dev_info.ra_pages = b->backing_dev_info.ra_pages;
 		}
 	}
+	/* To avoid confusion, if this queue does not support discard, clear
+	 * max_discard_sectors, which is what lsblk -D reports to the user.  */
+	if (!blk_queue_discard(q)) {
+		blk_queue_max_discard_sectors(q, 0);
+		q->limits.discard_granularity = 0;
+	}
 }
 
 void drbd_reconsider_max_bio_size(struct drbd_device *device, struct drbd_backing_dev *bdev)
@@ -1241,8 +1278,8 @@ static void conn_reconfig_done(struct drbd_connection *connection)
 		connection->cstate == C_STANDALONE;
 	spin_unlock_irq(&connection->resource->req_lock);
 	if (stop_threads) {
-		/* asender is implicitly stopped by receiver
-		 * in conn_disconnect() */
+		/* ack_receiver thread and ack_sender workqueue are implicitly
+		 * stopped by receiver in conn_disconnect() */
 		drbd_thread_stop(&connection->receiver);
 		drbd_thread_stop(&connection->worker);
 	}
@@ -1389,13 +1426,13 @@ int drbd_adm_disk_opts(struct sk_buff *skb, struct genl_info *info)
 		goto fail_unlock;
 	}
 
-	write_lock_irq(&global_state_lock);
+	lock_all_resources();
 	retcode = drbd_resync_after_valid(device, new_disk_conf->resync_after);
 	if (retcode == NO_ERROR) {
 		rcu_assign_pointer(device->ldev->disk_conf, new_disk_conf);
 		drbd_resync_after_changed(device);
 	}
-	write_unlock_irq(&global_state_lock);
+	unlock_all_resources();
 
 	if (retcode != NO_ERROR)
 		goto fail_unlock;
@@ -1418,7 +1455,7 @@ int drbd_adm_disk_opts(struct sk_buff *skb, struct genl_info *info)
 		set_bit(MD_NO_FUA, &device->flags);
 
 	if (write_ordering_changed(old_disk_conf, new_disk_conf))
-		drbd_bump_write_ordering(device->resource, NULL, WO_bdev_flush);
+		drbd_bump_write_ordering(device->resource, NULL, WO_BDEV_FLUSH);
 
 	drbd_md_sync(device);
 
@@ -1449,6 +1486,88 @@ int drbd_adm_disk_opts(struct sk_buff *skb, struct genl_info *info)
 	return 0;
 }
 
+static struct block_device *open_backing_dev(struct drbd_device *device,
+		const char *bdev_path, void *claim_ptr, bool do_bd_link)
+{
+	struct block_device *bdev;
+	int err = 0;
+
+	bdev = blkdev_get_by_path(bdev_path,
+				  FMODE_READ | FMODE_WRITE | FMODE_EXCL, claim_ptr);
+	if (IS_ERR(bdev)) {
+		drbd_err(device, "open(\"%s\") failed with %ld\n",
+				bdev_path, PTR_ERR(bdev));
+		return bdev;
+	}
+
+	if (!do_bd_link)
+		return bdev;
+
+	err = bd_link_disk_holder(bdev, device->vdisk);
+	if (err) {
+		blkdev_put(bdev, FMODE_READ | FMODE_WRITE | FMODE_EXCL);
+		drbd_err(device, "bd_link_disk_holder(\"%s\", ...) failed with %d\n",
+				bdev_path, err);
+		bdev = ERR_PTR(err);
+	}
+	return bdev;
+}
+
+static int open_backing_devices(struct drbd_device *device,
+		struct disk_conf *new_disk_conf,
+		struct drbd_backing_dev *nbc)
+{
+	struct block_device *bdev;
+
+	bdev = open_backing_dev(device, new_disk_conf->backing_dev, device, true);
+	if (IS_ERR(bdev))
+		return ERR_OPEN_DISK;
+	nbc->backing_bdev = bdev;
+
+	/*
+	 * meta_dev_idx >= 0: external fixed size, possibly multiple
+	 * drbd sharing one meta device.  TODO in that case, paranoia
+	 * check that [md_bdev, meta_dev_idx] is not yet used by some
+	 * other drbd minor!  (if you use drbd.conf + drbdadm, that
+	 * should check it for you already; but if you don't, or
+	 * someone fooled it, we need to double check here)
+	 */
+	bdev = open_backing_dev(device, new_disk_conf->meta_dev,
+		/* claim ptr: device, if claimed exclusively; shared drbd_m_holder,
+		 * if potentially shared with other drbd minors */
+			(new_disk_conf->meta_dev_idx < 0) ? (void*)device : (void*)drbd_m_holder,
+		/* avoid double bd_claim_by_disk() for the same (source,target) tuple,
+		 * as would happen with internal metadata. */
+			(new_disk_conf->meta_dev_idx != DRBD_MD_INDEX_FLEX_INT &&
+			 new_disk_conf->meta_dev_idx != DRBD_MD_INDEX_INTERNAL));
+	if (IS_ERR(bdev))
+		return ERR_OPEN_MD_DISK;
+	nbc->md_bdev = bdev;
+	return NO_ERROR;
+}
+
+static void close_backing_dev(struct drbd_device *device, struct block_device *bdev,
+	bool do_bd_unlink)
+{
+	if (!bdev)
+		return;
+	if (do_bd_unlink)
+		bd_unlink_disk_holder(bdev, device->vdisk);
+	blkdev_put(bdev, FMODE_READ | FMODE_WRITE | FMODE_EXCL);
+}
+
+void drbd_backing_dev_free(struct drbd_device *device, struct drbd_backing_dev *ldev)
+{
+	if (ldev == NULL)
+		return;
+
+	close_backing_dev(device, ldev->md_bdev, ldev->md_bdev != ldev->backing_bdev);
+	close_backing_dev(device, ldev->backing_bdev, true);
+
+	kfree(ldev->disk_conf);
+	kfree(ldev);
+}
+
 int drbd_adm_attach(struct sk_buff *skb, struct genl_info *info)
 {
 	struct drbd_config_context adm_ctx;
@@ -1462,7 +1581,6 @@ int drbd_adm_attach(struct sk_buff *skb, struct genl_info *info)
 	sector_t min_md_device_sectors;
 	struct drbd_backing_dev *nbc = NULL; /* new_backing_conf */
 	struct disk_conf *new_disk_conf = NULL;
-	struct block_device *bdev;
 	struct lru_cache *resync_lru = NULL;
 	struct fifo_buffer *new_plan = NULL;
 	union drbd_state ns, os;
@@ -1478,7 +1596,7 @@ int drbd_adm_attach(struct sk_buff *skb, struct genl_info *info)
 	device = adm_ctx.device;
 	mutex_lock(&adm_ctx.resource->adm_mutex);
 	peer_device = first_peer_device(device);
-	connection = peer_device ? peer_device->connection : NULL;
+	connection = peer_device->connection;
 	conn_reconfig_start(connection);
 
 	/* if you want to reconfigure, please tear down first */
@@ -1539,12 +1657,6 @@ int drbd_adm_attach(struct sk_buff *skb, struct genl_info *info)
 		goto fail;
 	}
 
-	write_lock_irq(&global_state_lock);
-	retcode = drbd_resync_after_valid(device, new_disk_conf->resync_after);
-	write_unlock_irq(&global_state_lock);
-	if (retcode != NO_ERROR)
-		goto fail;
-
 	rcu_read_lock();
 	nc = rcu_dereference(connection->net_conf);
 	if (nc) {
@@ -1556,35 +1668,9 @@ int drbd_adm_attach(struct sk_buff *skb, struct genl_info *info)
 	}
 	rcu_read_unlock();
 
-	bdev = blkdev_get_by_path(new_disk_conf->backing_dev,
-				  FMODE_READ | FMODE_WRITE | FMODE_EXCL, device);
-	if (IS_ERR(bdev)) {
-		drbd_err(device, "open(\"%s\") failed with %ld\n", new_disk_conf->backing_dev,
-			PTR_ERR(bdev));
-		retcode = ERR_OPEN_DISK;
-		goto fail;
-	}
-	nbc->backing_bdev = bdev;
-
-	/*
-	 * meta_dev_idx >= 0: external fixed size, possibly multiple
-	 * drbd sharing one meta device.  TODO in that case, paranoia
-	 * check that [md_bdev, meta_dev_idx] is not yet used by some
-	 * other drbd minor!  (if you use drbd.conf + drbdadm, that
-	 * should check it for you already; but if you don't, or
-	 * someone fooled it, we need to double check here)
-	 */
-	bdev = blkdev_get_by_path(new_disk_conf->meta_dev,
-				  FMODE_READ | FMODE_WRITE | FMODE_EXCL,
-				  (new_disk_conf->meta_dev_idx < 0) ?
-				  (void *)device : (void *)drbd_m_holder);
-	if (IS_ERR(bdev)) {
-		drbd_err(device, "open(\"%s\") failed with %ld\n", new_disk_conf->meta_dev,
-			PTR_ERR(bdev));
-		retcode = ERR_OPEN_MD_DISK;
+	retcode = open_backing_devices(device, new_disk_conf, nbc);
+	if (retcode != NO_ERROR)
 		goto fail;
-	}
-	nbc->md_bdev = bdev;
 
 	if ((nbc->backing_bdev == nbc->md_bdev) !=
 	    (new_disk_conf->meta_dev_idx == DRBD_MD_INDEX_INTERNAL ||
@@ -1707,6 +1793,13 @@ int drbd_adm_attach(struct sk_buff *skb, struct genl_info *info)
 		goto force_diskless_dec;
 	}
 
+	lock_all_resources();
+	retcode = drbd_resync_after_valid(device, new_disk_conf->resync_after);
+	if (retcode != NO_ERROR) {
+		unlock_all_resources();
+		goto force_diskless_dec;
+	}
+
 	/* Reset the "barriers don't work" bits here, then force meta data to
 	 * be written, to ensure we determine if barriers are supported. */
 	if (new_disk_conf->md_flushes)
@@ -1727,7 +1820,9 @@ int drbd_adm_attach(struct sk_buff *skb, struct genl_info *info)
 	new_disk_conf = NULL;
 	new_plan = NULL;
 
-	drbd_bump_write_ordering(device->resource, device->ldev, WO_bdev_flush);
+	drbd_resync_after_changed(device);
+	drbd_bump_write_ordering(device->resource, device->ldev, WO_BDEV_FLUSH);
+	unlock_all_resources();
 
 	if (drbd_md_test_flag(device->ldev, MDF_CRASHED_PRIMARY))
 		set_bit(CRASHED_PRIMARY, &device->flags);
@@ -1875,12 +1970,8 @@ int drbd_adm_attach(struct sk_buff *skb, struct genl_info *info)
  fail:
 	conn_reconfig_done(connection);
 	if (nbc) {
-		if (nbc->backing_bdev)
-			blkdev_put(nbc->backing_bdev,
-				   FMODE_READ | FMODE_WRITE | FMODE_EXCL);
-		if (nbc->md_bdev)
-			blkdev_put(nbc->md_bdev,
-				   FMODE_READ | FMODE_WRITE | FMODE_EXCL);
+		close_backing_dev(device, nbc->md_bdev, nbc->md_bdev != nbc->backing_bdev);
+		close_backing_dev(device, nbc->backing_bdev, true);
 		kfree(nbc);
 	}
 	kfree(new_disk_conf);
@@ -1895,6 +1986,7 @@ int drbd_adm_attach(struct sk_buff *skb, struct genl_info *info)
 static int adm_detach(struct drbd_device *device, int force)
 {
 	enum drbd_state_rv retcode;
+	void *buffer;
 	int ret;
 
 	if (force) {
@@ -1905,13 +1997,16 @@ static int adm_detach(struct drbd_device *device, int force)
 	}
 
 	drbd_suspend_io(device); /* so no-one is stuck in drbd_al_begin_io */
-	drbd_md_get_buffer(device, __func__); /* make sure there is no in-flight meta-data IO */
-	retcode = drbd_request_state(device, NS(disk, D_FAILED));
-	drbd_md_put_buffer(device);
+	buffer = drbd_md_get_buffer(device, __func__); /* make sure there is no in-flight meta-data IO */
+	if (buffer) {
+		retcode = drbd_request_state(device, NS(disk, D_FAILED));
+		drbd_md_put_buffer(device);
+	} else /* already <= D_FAILED */
+		retcode = SS_NOTHING_TO_DO;
 	/* D_FAILED will transition to DISKLESS. */
+	drbd_resume_io(device);
 	ret = wait_event_interruptible(device->misc_wait,
 			device->state.disk != D_FAILED);
-	drbd_resume_io(device);
 	if ((int)retcode == (int)SS_IS_DISKLESS)
 		retcode = SS_NOTHING_TO_DO;
 	if (ret)
@@ -2245,8 +2340,31 @@ int drbd_adm_net_opts(struct sk_buff *skb, struct genl_info *info)
 	return 0;
 }
 
+static void connection_to_info(struct connection_info *info,
+			       struct drbd_connection *connection)
+{
+	info->conn_connection_state = connection->cstate;
+	info->conn_role = conn_highest_peer(connection);
+}
+
+static void peer_device_to_info(struct peer_device_info *info,
+				struct drbd_peer_device *peer_device)
+{
+	struct drbd_device *device = peer_device->device;
+
+	info->peer_repl_state =
+		max_t(enum drbd_conns, C_WF_REPORT_PARAMS, device->state.conn);
+	info->peer_disk_state = device->state.pdsk;
+	info->peer_resync_susp_user = device->state.user_isp;
+	info->peer_resync_susp_peer = device->state.peer_isp;
+	info->peer_resync_susp_dependency = device->state.aftr_isp;
+}
+
 int drbd_adm_connect(struct sk_buff *skb, struct genl_info *info)
 {
+	struct connection_info connection_info;
+	enum drbd_notification_type flags;
+	unsigned int peer_devices = 0;
 	struct drbd_config_context adm_ctx;
 	struct drbd_peer_device *peer_device;
 	struct net_conf *old_net_conf, *new_net_conf = NULL;
@@ -2347,6 +2465,22 @@ int drbd_adm_connect(struct sk_buff *skb, struct genl_info *info)
 	connection->peer_addr_len = nla_len(adm_ctx.peer_addr);
 	memcpy(&connection->peer_addr, nla_data(adm_ctx.peer_addr), connection->peer_addr_len);
 
+	idr_for_each_entry(&connection->peer_devices, peer_device, i) {
+		peer_devices++;
+	}
+
+	connection_to_info(&connection_info, connection);
+	flags = (peer_devices--) ? NOTIFY_CONTINUES : 0;
+	mutex_lock(&notification_mutex);
+	notify_connection_state(NULL, 0, connection, &connection_info, NOTIFY_CREATE | flags);
+	idr_for_each_entry(&connection->peer_devices, peer_device, i) {
+		struct peer_device_info peer_device_info;
+
+		peer_device_to_info(&peer_device_info, peer_device);
+		flags = (peer_devices--) ? NOTIFY_CONTINUES : 0;
+		notify_peer_device_state(NULL, 0, peer_device, &peer_device_info, NOTIFY_CREATE | flags);
+	}
+	mutex_unlock(&notification_mutex);
 	mutex_unlock(&adm_ctx.resource->conf_update);
 
 	rcu_read_lock();
@@ -2428,6 +2562,8 @@ static enum drbd_state_rv conn_try_disconnect(struct drbd_connection *connection
 			drbd_err(connection,
 				"unexpected rv2=%d in conn_try_disconnect()\n",
 				rv2);
+		/* Unlike in DRBD 9, the state engine has generated
+		 * NOTIFY_DESTROY events before clearing connection->net_conf. */
 	}
 	return rv;
 }
@@ -2585,6 +2721,7 @@ int drbd_adm_resize(struct sk_buff *skb, struct genl_info *info)
 		mutex_unlock(&device->resource->conf_update);
 		synchronize_rcu();
 		kfree(old_disk_conf);
+		new_disk_conf = NULL;
 	}
 
 	ddsf = (rs.resize_force ? DDSF_FORCED : 0) | (rs.no_resync ? DDSF_NO_RESYNC : 0);
@@ -2618,6 +2755,7 @@ int drbd_adm_resize(struct sk_buff *skb, struct genl_info *info)
 
  fail_ldev:
 	put_ldev(device);
+	kfree(new_disk_conf);
 	goto fail;
 }
 
@@ -2855,7 +2993,30 @@ int drbd_adm_resume_io(struct sk_buff *skb, struct genl_info *info)
 	mutex_lock(&adm_ctx.resource->adm_mutex);
 	device = adm_ctx.device;
 	if (test_bit(NEW_CUR_UUID, &device->flags)) {
-		drbd_uuid_new_current(device);
+		if (get_ldev_if_state(device, D_ATTACHING)) {
+			drbd_uuid_new_current(device);
+			put_ldev(device);
+		} else {
+			/* This is effectively a multi-stage "forced down".
+			 * The NEW_CUR_UUID bit is supposedly only set, if we
+			 * lost the replication connection, and are configured
+			 * to freeze IO and wait for some fence-peer handler.
+			 * So we still don't have a replication connection.
+			 * And now we don't have a local disk either.  After
+			 * resume, we will fail all pending and new IO, because
+			 * we don't have any data anymore.  Which means we will
+			 * eventually be able to terminate all users of this
+			 * device, and then take it down.  By bumping the
+			 * "effective" data uuid, we make sure that you really
+			 * need to tear down before you reconfigure, we will
+			 * the refuse to re-connect or re-attach (because no
+			 * matching real data uuid exists).
+			 */
+			u64 val;
+			get_random_bytes(&val, sizeof(u64));
+			drbd_set_ed_uuid(device, val);
+			drbd_warn(device, "Resumed without access to data; please tear down before attempting to re-configure.\n");
+		}
 		clear_bit(NEW_CUR_UUID, &device->flags);
 	}
 	drbd_suspend_io(device);
@@ -2909,6 +3070,486 @@ static int nla_put_drbd_cfg_context(struct sk_buff *skb,
 	return -EMSGSIZE;
 }
 
+/*
+ * The generic netlink dump callbacks are called outside the genl_lock(), so
+ * they cannot use the simple attribute parsing code which uses global
+ * attribute tables.
+ */
+static struct nlattr *find_cfg_context_attr(const struct nlmsghdr *nlh, int attr)
+{
+	const unsigned hdrlen = GENL_HDRLEN + GENL_MAGIC_FAMILY_HDRSZ;
+	const int maxtype = ARRAY_SIZE(drbd_cfg_context_nl_policy) - 1;
+	struct nlattr *nla;
+
+	nla = nla_find(nlmsg_attrdata(nlh, hdrlen), nlmsg_attrlen(nlh, hdrlen),
+		       DRBD_NLA_CFG_CONTEXT);
+	if (!nla)
+		return NULL;
+	return drbd_nla_find_nested(maxtype, nla, __nla_type(attr));
+}
+
+static void resource_to_info(struct resource_info *, struct drbd_resource *);
+
+int drbd_adm_dump_resources(struct sk_buff *skb, struct netlink_callback *cb)
+{
+	struct drbd_genlmsghdr *dh;
+	struct drbd_resource *resource;
+	struct resource_info resource_info;
+	struct resource_statistics resource_statistics;
+	int err;
+
+	rcu_read_lock();
+	if (cb->args[0]) {
+		for_each_resource_rcu(resource, &drbd_resources)
+			if (resource == (struct drbd_resource *)cb->args[0])
+				goto found_resource;
+		err = 0;  /* resource was probably deleted */
+		goto out;
+	}
+	resource = list_entry(&drbd_resources,
+			      struct drbd_resource, resources);
+
+found_resource:
+	list_for_each_entry_continue_rcu(resource, &drbd_resources, resources) {
+		goto put_result;
+	}
+	err = 0;
+	goto out;
+
+put_result:
+	dh = genlmsg_put(skb, NETLINK_CB(cb->skb).portid,
+			cb->nlh->nlmsg_seq, &drbd_genl_family,
+			NLM_F_MULTI, DRBD_ADM_GET_RESOURCES);
+	err = -ENOMEM;
+	if (!dh)
+		goto out;
+	dh->minor = -1U;
+	dh->ret_code = NO_ERROR;
+	err = nla_put_drbd_cfg_context(skb, resource, NULL, NULL);
+	if (err)
+		goto out;
+	err = res_opts_to_skb(skb, &resource->res_opts, !capable(CAP_SYS_ADMIN));
+	if (err)
+		goto out;
+	resource_to_info(&resource_info, resource);
+	err = resource_info_to_skb(skb, &resource_info, !capable(CAP_SYS_ADMIN));
+	if (err)
+		goto out;
+	resource_statistics.res_stat_write_ordering = resource->write_ordering;
+	err = resource_statistics_to_skb(skb, &resource_statistics, !capable(CAP_SYS_ADMIN));
+	if (err)
+		goto out;
+	cb->args[0] = (long)resource;
+	genlmsg_end(skb, dh);
+	err = 0;
+
+out:
+	rcu_read_unlock();
+	if (err)
+		return err;
+	return skb->len;
+}
+
+static void device_to_statistics(struct device_statistics *s,
+				 struct drbd_device *device)
+{
+	memset(s, 0, sizeof(*s));
+	s->dev_upper_blocked = !may_inc_ap_bio(device);
+	if (get_ldev(device)) {
+		struct drbd_md *md = &device->ldev->md;
+		u64 *history_uuids = (u64 *)s->history_uuids;
+		struct request_queue *q;
+		int n;
+
+		spin_lock_irq(&md->uuid_lock);
+		s->dev_current_uuid = md->uuid[UI_CURRENT];
+		BUILD_BUG_ON(sizeof(s->history_uuids) < UI_HISTORY_END - UI_HISTORY_START + 1);
+		for (n = 0; n < UI_HISTORY_END - UI_HISTORY_START + 1; n++)
+			history_uuids[n] = md->uuid[UI_HISTORY_START + n];
+		for (; n < HISTORY_UUIDS; n++)
+			history_uuids[n] = 0;
+		s->history_uuids_len = HISTORY_UUIDS;
+		spin_unlock_irq(&md->uuid_lock);
+
+		s->dev_disk_flags = md->flags;
+		q = bdev_get_queue(device->ldev->backing_bdev);
+		s->dev_lower_blocked =
+			bdi_congested(&q->backing_dev_info,
+				      (1 << WB_async_congested) |
+				      (1 << WB_sync_congested));
+		put_ldev(device);
+	}
+	s->dev_size = drbd_get_capacity(device->this_bdev);
+	s->dev_read = device->read_cnt;
+	s->dev_write = device->writ_cnt;
+	s->dev_al_writes = device->al_writ_cnt;
+	s->dev_bm_writes = device->bm_writ_cnt;
+	s->dev_upper_pending = atomic_read(&device->ap_bio_cnt);
+	s->dev_lower_pending = atomic_read(&device->local_cnt);
+	s->dev_al_suspended = test_bit(AL_SUSPENDED, &device->flags);
+	s->dev_exposed_data_uuid = device->ed_uuid;
+}
+
+static int put_resource_in_arg0(struct netlink_callback *cb, int holder_nr)
+{
+	if (cb->args[0]) {
+		struct drbd_resource *resource =
+			(struct drbd_resource *)cb->args[0];
+		kref_put(&resource->kref, drbd_destroy_resource);
+	}
+
+	return 0;
+}
+
+int drbd_adm_dump_devices_done(struct netlink_callback *cb) {
+	return put_resource_in_arg0(cb, 7);
+}
+
+static void device_to_info(struct device_info *, struct drbd_device *);
+
+int drbd_adm_dump_devices(struct sk_buff *skb, struct netlink_callback *cb)
+{
+	struct nlattr *resource_filter;
+	struct drbd_resource *resource;
+	struct drbd_device *uninitialized_var(device);
+	int minor, err, retcode;
+	struct drbd_genlmsghdr *dh;
+	struct device_info device_info;
+	struct device_statistics device_statistics;
+	struct idr *idr_to_search;
+
+	resource = (struct drbd_resource *)cb->args[0];
+	if (!cb->args[0] && !cb->args[1]) {
+		resource_filter = find_cfg_context_attr(cb->nlh, T_ctx_resource_name);
+		if (resource_filter) {
+			retcode = ERR_RES_NOT_KNOWN;
+			resource = drbd_find_resource(nla_data(resource_filter));
+			if (!resource)
+				goto put_result;
+			cb->args[0] = (long)resource;
+		}
+	}
+
+	rcu_read_lock();
+	minor = cb->args[1];
+	idr_to_search = resource ? &resource->devices : &drbd_devices;
+	device = idr_get_next(idr_to_search, &minor);
+	if (!device) {
+		err = 0;
+		goto out;
+	}
+	idr_for_each_entry_continue(idr_to_search, device, minor) {
+		retcode = NO_ERROR;
+		goto put_result;  /* only one iteration */
+	}
+	err = 0;
+	goto out;  /* no more devices */
+
+put_result:
+	dh = genlmsg_put(skb, NETLINK_CB(cb->skb).portid,
+			cb->nlh->nlmsg_seq, &drbd_genl_family,
+			NLM_F_MULTI, DRBD_ADM_GET_DEVICES);
+	err = -ENOMEM;
+	if (!dh)
+		goto out;
+	dh->ret_code = retcode;
+	dh->minor = -1U;
+	if (retcode == NO_ERROR) {
+		dh->minor = device->minor;
+		err = nla_put_drbd_cfg_context(skb, device->resource, NULL, device);
+		if (err)
+			goto out;
+		if (get_ldev(device)) {
+			struct disk_conf *disk_conf =
+				rcu_dereference(device->ldev->disk_conf);
+
+			err = disk_conf_to_skb(skb, disk_conf, !capable(CAP_SYS_ADMIN));
+			put_ldev(device);
+			if (err)
+				goto out;
+		}
+		device_to_info(&device_info, device);
+		err = device_info_to_skb(skb, &device_info, !capable(CAP_SYS_ADMIN));
+		if (err)
+			goto out;
+
+		device_to_statistics(&device_statistics, device);
+		err = device_statistics_to_skb(skb, &device_statistics, !capable(CAP_SYS_ADMIN));
+		if (err)
+			goto out;
+		cb->args[1] = minor + 1;
+	}
+	genlmsg_end(skb, dh);
+	err = 0;
+
+out:
+	rcu_read_unlock();
+	if (err)
+		return err;
+	return skb->len;
+}
+
+int drbd_adm_dump_connections_done(struct netlink_callback *cb)
+{
+	return put_resource_in_arg0(cb, 6);
+}
+
+enum { SINGLE_RESOURCE, ITERATE_RESOURCES };
+
+int drbd_adm_dump_connections(struct sk_buff *skb, struct netlink_callback *cb)
+{
+	struct nlattr *resource_filter;
+	struct drbd_resource *resource = NULL, *next_resource;
+	struct drbd_connection *uninitialized_var(connection);
+	int err = 0, retcode;
+	struct drbd_genlmsghdr *dh;
+	struct connection_info connection_info;
+	struct connection_statistics connection_statistics;
+
+	rcu_read_lock();
+	resource = (struct drbd_resource *)cb->args[0];
+	if (!cb->args[0]) {
+		resource_filter = find_cfg_context_attr(cb->nlh, T_ctx_resource_name);
+		if (resource_filter) {
+			retcode = ERR_RES_NOT_KNOWN;
+			resource = drbd_find_resource(nla_data(resource_filter));
+			if (!resource)
+				goto put_result;
+			cb->args[0] = (long)resource;
+			cb->args[1] = SINGLE_RESOURCE;
+		}
+	}
+	if (!resource) {
+		if (list_empty(&drbd_resources))
+			goto out;
+		resource = list_first_entry(&drbd_resources, struct drbd_resource, resources);
+		kref_get(&resource->kref);
+		cb->args[0] = (long)resource;
+		cb->args[1] = ITERATE_RESOURCES;
+	}
+
+    next_resource:
+	rcu_read_unlock();
+	mutex_lock(&resource->conf_update);
+	rcu_read_lock();
+	if (cb->args[2]) {
+		for_each_connection_rcu(connection, resource)
+			if (connection == (struct drbd_connection *)cb->args[2])
+				goto found_connection;
+		/* connection was probably deleted */
+		goto no_more_connections;
+	}
+	connection = list_entry(&resource->connections, struct drbd_connection, connections);
+
+found_connection:
+	list_for_each_entry_continue_rcu(connection, &resource->connections, connections) {
+		if (!has_net_conf(connection))
+			continue;
+		retcode = NO_ERROR;
+		goto put_result;  /* only one iteration */
+	}
+
+no_more_connections:
+	if (cb->args[1] == ITERATE_RESOURCES) {
+		for_each_resource_rcu(next_resource, &drbd_resources) {
+			if (next_resource == resource)
+				goto found_resource;
+		}
+		/* resource was probably deleted */
+	}
+	goto out;
+
+found_resource:
+	list_for_each_entry_continue_rcu(next_resource, &drbd_resources, resources) {
+		mutex_unlock(&resource->conf_update);
+		kref_put(&resource->kref, drbd_destroy_resource);
+		resource = next_resource;
+		kref_get(&resource->kref);
+		cb->args[0] = (long)resource;
+		cb->args[2] = 0;
+		goto next_resource;
+	}
+	goto out;  /* no more resources */
+
+put_result:
+	dh = genlmsg_put(skb, NETLINK_CB(cb->skb).portid,
+			cb->nlh->nlmsg_seq, &drbd_genl_family,
+			NLM_F_MULTI, DRBD_ADM_GET_CONNECTIONS);
+	err = -ENOMEM;
+	if (!dh)
+		goto out;
+	dh->ret_code = retcode;
+	dh->minor = -1U;
+	if (retcode == NO_ERROR) {
+		struct net_conf *net_conf;
+
+		err = nla_put_drbd_cfg_context(skb, resource, connection, NULL);
+		if (err)
+			goto out;
+		net_conf = rcu_dereference(connection->net_conf);
+		if (net_conf) {
+			err = net_conf_to_skb(skb, net_conf, !capable(CAP_SYS_ADMIN));
+			if (err)
+				goto out;
+		}
+		connection_to_info(&connection_info, connection);
+		err = connection_info_to_skb(skb, &connection_info, !capable(CAP_SYS_ADMIN));
+		if (err)
+			goto out;
+		connection_statistics.conn_congested = test_bit(NET_CONGESTED, &connection->flags);
+		err = connection_statistics_to_skb(skb, &connection_statistics, !capable(CAP_SYS_ADMIN));
+		if (err)
+			goto out;
+		cb->args[2] = (long)connection;
+	}
+	genlmsg_end(skb, dh);
+	err = 0;
+
+out:
+	rcu_read_unlock();
+	if (resource)
+		mutex_unlock(&resource->conf_update);
+	if (err)
+		return err;
+	return skb->len;
+}
+
+enum mdf_peer_flag {
+	MDF_PEER_CONNECTED =	1 << 0,
+	MDF_PEER_OUTDATED =	1 << 1,
+	MDF_PEER_FENCING =	1 << 2,
+	MDF_PEER_FULL_SYNC =	1 << 3,
+};
+
+static void peer_device_to_statistics(struct peer_device_statistics *s,
+				      struct drbd_peer_device *peer_device)
+{
+	struct drbd_device *device = peer_device->device;
+
+	memset(s, 0, sizeof(*s));
+	s->peer_dev_received = device->recv_cnt;
+	s->peer_dev_sent = device->send_cnt;
+	s->peer_dev_pending = atomic_read(&device->ap_pending_cnt) +
+			      atomic_read(&device->rs_pending_cnt);
+	s->peer_dev_unacked = atomic_read(&device->unacked_cnt);
+	s->peer_dev_out_of_sync = drbd_bm_total_weight(device) << (BM_BLOCK_SHIFT - 9);
+	s->peer_dev_resync_failed = device->rs_failed << (BM_BLOCK_SHIFT - 9);
+	if (get_ldev(device)) {
+		struct drbd_md *md = &device->ldev->md;
+
+		spin_lock_irq(&md->uuid_lock);
+		s->peer_dev_bitmap_uuid = md->uuid[UI_BITMAP];
+		spin_unlock_irq(&md->uuid_lock);
+		s->peer_dev_flags =
+			(drbd_md_test_flag(device->ldev, MDF_CONNECTED_IND) ?
+				MDF_PEER_CONNECTED : 0) +
+			(drbd_md_test_flag(device->ldev, MDF_CONSISTENT) &&
+			 !drbd_md_test_flag(device->ldev, MDF_WAS_UP_TO_DATE) ?
+				MDF_PEER_OUTDATED : 0) +
+			/* FIXME: MDF_PEER_FENCING? */
+			(drbd_md_test_flag(device->ldev, MDF_FULL_SYNC) ?
+				MDF_PEER_FULL_SYNC : 0);
+		put_ldev(device);
+	}
+}
+
+int drbd_adm_dump_peer_devices_done(struct netlink_callback *cb)
+{
+	return put_resource_in_arg0(cb, 9);
+}
+
+int drbd_adm_dump_peer_devices(struct sk_buff *skb, struct netlink_callback *cb)
+{
+	struct nlattr *resource_filter;
+	struct drbd_resource *resource;
+	struct drbd_device *uninitialized_var(device);
+	struct drbd_peer_device *peer_device = NULL;
+	int minor, err, retcode;
+	struct drbd_genlmsghdr *dh;
+	struct idr *idr_to_search;
+
+	resource = (struct drbd_resource *)cb->args[0];
+	if (!cb->args[0] && !cb->args[1]) {
+		resource_filter = find_cfg_context_attr(cb->nlh, T_ctx_resource_name);
+		if (resource_filter) {
+			retcode = ERR_RES_NOT_KNOWN;
+			resource = drbd_find_resource(nla_data(resource_filter));
+			if (!resource)
+				goto put_result;
+		}
+		cb->args[0] = (long)resource;
+	}
+
+	rcu_read_lock();
+	minor = cb->args[1];
+	idr_to_search = resource ? &resource->devices : &drbd_devices;
+	device = idr_find(idr_to_search, minor);
+	if (!device) {
+next_device:
+		minor++;
+		cb->args[2] = 0;
+		device = idr_get_next(idr_to_search, &minor);
+		if (!device) {
+			err = 0;
+			goto out;
+		}
+	}
+	if (cb->args[2]) {
+		for_each_peer_device(peer_device, device)
+			if (peer_device == (struct drbd_peer_device *)cb->args[2])
+				goto found_peer_device;
+		/* peer device was probably deleted */
+		goto next_device;
+	}
+	/* Make peer_device point to the list head (not the first entry). */
+	peer_device = list_entry(&device->peer_devices, struct drbd_peer_device, peer_devices);
+
+found_peer_device:
+	list_for_each_entry_continue_rcu(peer_device, &device->peer_devices, peer_devices) {
+		if (!has_net_conf(peer_device->connection))
+			continue;
+		retcode = NO_ERROR;
+		goto put_result;  /* only one iteration */
+	}
+	goto next_device;
+
+put_result:
+	dh = genlmsg_put(skb, NETLINK_CB(cb->skb).portid,
+			cb->nlh->nlmsg_seq, &drbd_genl_family,
+			NLM_F_MULTI, DRBD_ADM_GET_PEER_DEVICES);
+	err = -ENOMEM;
+	if (!dh)
+		goto out;
+	dh->ret_code = retcode;
+	dh->minor = -1U;
+	if (retcode == NO_ERROR) {
+		struct peer_device_info peer_device_info;
+		struct peer_device_statistics peer_device_statistics;
+
+		dh->minor = minor;
+		err = nla_put_drbd_cfg_context(skb, device->resource, peer_device->connection, device);
+		if (err)
+			goto out;
+		peer_device_to_info(&peer_device_info, peer_device);
+		err = peer_device_info_to_skb(skb, &peer_device_info, !capable(CAP_SYS_ADMIN));
+		if (err)
+			goto out;
+		peer_device_to_statistics(&peer_device_statistics, peer_device);
+		err = peer_device_statistics_to_skb(skb, &peer_device_statistics, !capable(CAP_SYS_ADMIN));
+		if (err)
+			goto out;
+		cb->args[1] = minor;
+		cb->args[2] = (long)peer_device;
+	}
+	genlmsg_end(skb, dh);
+	err = 0;
+
+out:
+	rcu_read_unlock();
+	if (err)
+		return err;
+	return skb->len;
+}
 /*
  * Return the connection of @resource if @resource has exactly one connection.
  */
@@ -3414,8 +4055,18 @@ drbd_check_resource_name(struct drbd_config_context *adm_ctx)
 	return NO_ERROR;
 }
 
+static void resource_to_info(struct resource_info *info,
+			     struct drbd_resource *resource)
+{
+	info->res_role = conn_highest_role(first_connection(resource));
+	info->res_susp = resource->susp;
+	info->res_susp_nod = resource->susp_nod;
+	info->res_susp_fen = resource->susp_fen;
+}
+
 int drbd_adm_new_resource(struct sk_buff *skb, struct genl_info *info)
 {
+	struct drbd_connection *connection;
 	struct drbd_config_context adm_ctx;
 	enum drbd_ret_code retcode;
 	struct res_opts res_opts;
@@ -3449,13 +4100,33 @@ int drbd_adm_new_resource(struct sk_buff *skb, struct genl_info *info)
 	}
 
 	/* not yet safe for genl_family.parallel_ops */
-	if (!conn_create(adm_ctx.resource_name, &res_opts))
+	mutex_lock(&resources_mutex);
+	connection = conn_create(adm_ctx.resource_name, &res_opts);
+	mutex_unlock(&resources_mutex);
+
+	if (connection) {
+		struct resource_info resource_info;
+
+		mutex_lock(&notification_mutex);
+		resource_to_info(&resource_info, connection->resource);
+		notify_resource_state(NULL, 0, connection->resource,
+				      &resource_info, NOTIFY_CREATE);
+		mutex_unlock(&notification_mutex);
+	} else
 		retcode = ERR_NOMEM;
+
 out:
 	drbd_adm_finish(&adm_ctx, info, retcode);
 	return 0;
 }
 
+static void device_to_info(struct device_info *info,
+			   struct drbd_device *device)
+{
+	info->dev_disk_state = device->state.disk;
+}
+
+
 int drbd_adm_new_minor(struct sk_buff *skb, struct genl_info *info)
 {
 	struct drbd_config_context adm_ctx;
@@ -3490,6 +4161,36 @@ int drbd_adm_new_minor(struct sk_buff *skb, struct genl_info *info)
 
 	mutex_lock(&adm_ctx.resource->adm_mutex);
 	retcode = drbd_create_device(&adm_ctx, dh->minor);
+	if (retcode == NO_ERROR) {
+		struct drbd_device *device;
+		struct drbd_peer_device *peer_device;
+		struct device_info info;
+		unsigned int peer_devices = 0;
+		enum drbd_notification_type flags;
+
+		device = minor_to_device(dh->minor);
+		for_each_peer_device(peer_device, device) {
+			if (!has_net_conf(peer_device->connection))
+				continue;
+			peer_devices++;
+		}
+
+		device_to_info(&info, device);
+		mutex_lock(&notification_mutex);
+		flags = (peer_devices--) ? NOTIFY_CONTINUES : 0;
+		notify_device_state(NULL, 0, device, &info, NOTIFY_CREATE | flags);
+		for_each_peer_device(peer_device, device) {
+			struct peer_device_info peer_device_info;
+
+			if (!has_net_conf(peer_device->connection))
+				continue;
+			peer_device_to_info(&peer_device_info, peer_device);
+			flags = (peer_devices--) ? NOTIFY_CONTINUES : 0;
+			notify_peer_device_state(NULL, 0, peer_device, &peer_device_info,
+						 NOTIFY_CREATE | flags);
+		}
+		mutex_unlock(&notification_mutex);
+	}
 	mutex_unlock(&adm_ctx.resource->adm_mutex);
 out:
 	drbd_adm_finish(&adm_ctx, info, retcode);
@@ -3498,13 +4199,35 @@ int drbd_adm_new_minor(struct sk_buff *skb, struct genl_info *info)
 
 static enum drbd_ret_code adm_del_minor(struct drbd_device *device)
 {
+	struct drbd_peer_device *peer_device;
+
 	if (device->state.disk == D_DISKLESS &&
 	    /* no need to be device->state.conn == C_STANDALONE &&
 	     * we may want to delete a minor from a live replication group.
 	     */
 	    device->state.role == R_SECONDARY) {
+		struct drbd_connection *connection =
+			first_connection(device->resource);
+
 		_drbd_request_state(device, NS(conn, C_WF_REPORT_PARAMS),
 				    CS_VERBOSE + CS_WAIT_COMPLETE);
+
+		/* If the state engine hasn't stopped the sender thread yet, we
+		 * need to flush the sender work queue before generating the
+		 * DESTROY events here. */
+		if (get_t_state(&connection->worker) == RUNNING)
+			drbd_flush_workqueue(&connection->sender_work);
+
+		mutex_lock(&notification_mutex);
+		for_each_peer_device(peer_device, device) {
+			if (!has_net_conf(peer_device->connection))
+				continue;
+			notify_peer_device_state(NULL, 0, peer_device, NULL,
+						 NOTIFY_DESTROY | NOTIFY_CONTINUES);
+		}
+		notify_device_state(NULL, 0, device, NULL, NOTIFY_DESTROY);
+		mutex_unlock(&notification_mutex);
+
 		drbd_delete_device(device);
 		return NO_ERROR;
 	} else
@@ -3541,7 +4264,16 @@ static int adm_del_resource(struct drbd_resource *resource)
 	if (!idr_is_empty(&resource->devices))
 		return ERR_RES_IN_USE;
 
+	/* The state engine has stopped the sender thread, so we don't
+	 * need to flush the sender work queue before generating the
+	 * DESTROY event here. */
+	mutex_lock(&notification_mutex);
+	notify_resource_state(NULL, 0, resource, NULL, NOTIFY_DESTROY);
+	mutex_unlock(&notification_mutex);
+
+	mutex_lock(&resources_mutex);
 	list_del_rcu(&resource->resources);
+	mutex_unlock(&resources_mutex);
 	/* Make sure all threads have actually stopped: state handling only
 	 * does drbd_thread_stop_nowait(). */
 	list_for_each_entry(connection, &resource->connections, connections)
@@ -3637,7 +4369,6 @@ int drbd_adm_del_resource(struct sk_buff *skb, struct genl_info *info)
 
 void drbd_bcast_event(struct drbd_device *device, const struct sib_info *sib)
 {
-	static atomic_t drbd_genl_seq = ATOMIC_INIT(2); /* two. */
 	struct sk_buff *msg;
 	struct drbd_genlmsghdr *d_out;
 	unsigned seq;
@@ -3658,7 +4389,7 @@ void drbd_bcast_event(struct drbd_device *device, const struct sib_info *sib)
 	if (nla_put_status_info(msg, device, sib))
 		goto nla_put_failure;
 	genlmsg_end(msg, d_out);
-	err = drbd_genl_multicast_events(msg, 0);
+	err = drbd_genl_multicast_events(msg, GFP_NOWAIT);
 	/* msg has been consumed or freed in netlink_broadcast() */
 	if (err && err != -ESRCH)
 		goto failed;
@@ -3672,3 +4403,405 @@ void drbd_bcast_event(struct drbd_device *device, const struct sib_info *sib)
 			"Event seq:%u sib_reason:%u\n",
 			err, seq, sib->sib_reason);
 }
+
+static int nla_put_notification_header(struct sk_buff *msg,
+				       enum drbd_notification_type type)
+{
+	struct drbd_notification_header nh = {
+		.nh_type = type,
+	};
+
+	return drbd_notification_header_to_skb(msg, &nh, true);
+}
+
+void notify_resource_state(struct sk_buff *skb,
+			   unsigned int seq,
+			   struct drbd_resource *resource,
+			   struct resource_info *resource_info,
+			   enum drbd_notification_type type)
+{
+	struct resource_statistics resource_statistics;
+	struct drbd_genlmsghdr *dh;
+	bool multicast = false;
+	int err;
+
+	if (!skb) {
+		seq = atomic_inc_return(&notify_genl_seq);
+		skb = genlmsg_new(NLMSG_GOODSIZE, GFP_NOIO);
+		err = -ENOMEM;
+		if (!skb)
+			goto failed;
+		multicast = true;
+	}
+
+	err = -EMSGSIZE;
+	dh = genlmsg_put(skb, 0, seq, &drbd_genl_family, 0, DRBD_RESOURCE_STATE);
+	if (!dh)
+		goto nla_put_failure;
+	dh->minor = -1U;
+	dh->ret_code = NO_ERROR;
+	if (nla_put_drbd_cfg_context(skb, resource, NULL, NULL) ||
+	    nla_put_notification_header(skb, type) ||
+	    ((type & ~NOTIFY_FLAGS) != NOTIFY_DESTROY &&
+	     resource_info_to_skb(skb, resource_info, true)))
+		goto nla_put_failure;
+	resource_statistics.res_stat_write_ordering = resource->write_ordering;
+	err = resource_statistics_to_skb(skb, &resource_statistics, !capable(CAP_SYS_ADMIN));
+	if (err)
+		goto nla_put_failure;
+	genlmsg_end(skb, dh);
+	if (multicast) {
+		err = drbd_genl_multicast_events(skb, GFP_NOWAIT);
+		/* skb has been consumed or freed in netlink_broadcast() */
+		if (err && err != -ESRCH)
+			goto failed;
+	}
+	return;
+
+nla_put_failure:
+	nlmsg_free(skb);
+failed:
+	drbd_err(resource, "Error %d while broadcasting event. Event seq:%u\n",
+			err, seq);
+}
+
+void notify_device_state(struct sk_buff *skb,
+			 unsigned int seq,
+			 struct drbd_device *device,
+			 struct device_info *device_info,
+			 enum drbd_notification_type type)
+{
+	struct device_statistics device_statistics;
+	struct drbd_genlmsghdr *dh;
+	bool multicast = false;
+	int err;
+
+	if (!skb) {
+		seq = atomic_inc_return(&notify_genl_seq);
+		skb = genlmsg_new(NLMSG_GOODSIZE, GFP_NOIO);
+		err = -ENOMEM;
+		if (!skb)
+			goto failed;
+		multicast = true;
+	}
+
+	err = -EMSGSIZE;
+	dh = genlmsg_put(skb, 0, seq, &drbd_genl_family, 0, DRBD_DEVICE_STATE);
+	if (!dh)
+		goto nla_put_failure;
+	dh->minor = device->minor;
+	dh->ret_code = NO_ERROR;
+	if (nla_put_drbd_cfg_context(skb, device->resource, NULL, device) ||
+	    nla_put_notification_header(skb, type) ||
+	    ((type & ~NOTIFY_FLAGS) != NOTIFY_DESTROY &&
+	     device_info_to_skb(skb, device_info, true)))
+		goto nla_put_failure;
+	device_to_statistics(&device_statistics, device);
+	device_statistics_to_skb(skb, &device_statistics, !capable(CAP_SYS_ADMIN));
+	genlmsg_end(skb, dh);
+	if (multicast) {
+		err = drbd_genl_multicast_events(skb, GFP_NOWAIT);
+		/* skb has been consumed or freed in netlink_broadcast() */
+		if (err && err != -ESRCH)
+			goto failed;
+	}
+	return;
+
+nla_put_failure:
+	nlmsg_free(skb);
+failed:
+	drbd_err(device, "Error %d while broadcasting event. Event seq:%u\n",
+		 err, seq);
+}
+
+void notify_connection_state(struct sk_buff *skb,
+			     unsigned int seq,
+			     struct drbd_connection *connection,
+			     struct connection_info *connection_info,
+			     enum drbd_notification_type type)
+{
+	struct connection_statistics connection_statistics;
+	struct drbd_genlmsghdr *dh;
+	bool multicast = false;
+	int err;
+
+	if (!skb) {
+		seq = atomic_inc_return(&notify_genl_seq);
+		skb = genlmsg_new(NLMSG_GOODSIZE, GFP_NOIO);
+		err = -ENOMEM;
+		if (!skb)
+			goto failed;
+		multicast = true;
+	}
+
+	err = -EMSGSIZE;
+	dh = genlmsg_put(skb, 0, seq, &drbd_genl_family, 0, DRBD_CONNECTION_STATE);
+	if (!dh)
+		goto nla_put_failure;
+	dh->minor = -1U;
+	dh->ret_code = NO_ERROR;
+	if (nla_put_drbd_cfg_context(skb, connection->resource, connection, NULL) ||
+	    nla_put_notification_header(skb, type) ||
+	    ((type & ~NOTIFY_FLAGS) != NOTIFY_DESTROY &&
+	     connection_info_to_skb(skb, connection_info, true)))
+		goto nla_put_failure;
+	connection_statistics.conn_congested = test_bit(NET_CONGESTED, &connection->flags);
+	connection_statistics_to_skb(skb, &connection_statistics, !capable(CAP_SYS_ADMIN));
+	genlmsg_end(skb, dh);
+	if (multicast) {
+		err = drbd_genl_multicast_events(skb, GFP_NOWAIT);
+		/* skb has been consumed or freed in netlink_broadcast() */
+		if (err && err != -ESRCH)
+			goto failed;
+	}
+	return;
+
+nla_put_failure:
+	nlmsg_free(skb);
+failed:
+	drbd_err(connection, "Error %d while broadcasting event. Event seq:%u\n",
+		 err, seq);
+}
+
+void notify_peer_device_state(struct sk_buff *skb,
+			      unsigned int seq,
+			      struct drbd_peer_device *peer_device,
+			      struct peer_device_info *peer_device_info,
+			      enum drbd_notification_type type)
+{
+	struct peer_device_statistics peer_device_statistics;
+	struct drbd_resource *resource = peer_device->device->resource;
+	struct drbd_genlmsghdr *dh;
+	bool multicast = false;
+	int err;
+
+	if (!skb) {
+		seq = atomic_inc_return(&notify_genl_seq);
+		skb = genlmsg_new(NLMSG_GOODSIZE, GFP_NOIO);
+		err = -ENOMEM;
+		if (!skb)
+			goto failed;
+		multicast = true;
+	}
+
+	err = -EMSGSIZE;
+	dh = genlmsg_put(skb, 0, seq, &drbd_genl_family, 0, DRBD_PEER_DEVICE_STATE);
+	if (!dh)
+		goto nla_put_failure;
+	dh->minor = -1U;
+	dh->ret_code = NO_ERROR;
+	if (nla_put_drbd_cfg_context(skb, resource, peer_device->connection, peer_device->device) ||
+	    nla_put_notification_header(skb, type) ||
+	    ((type & ~NOTIFY_FLAGS) != NOTIFY_DESTROY &&
+	     peer_device_info_to_skb(skb, peer_device_info, true)))
+		goto nla_put_failure;
+	peer_device_to_statistics(&peer_device_statistics, peer_device);
+	peer_device_statistics_to_skb(skb, &peer_device_statistics, !capable(CAP_SYS_ADMIN));
+	genlmsg_end(skb, dh);
+	if (multicast) {
+		err = drbd_genl_multicast_events(skb, GFP_NOWAIT);
+		/* skb has been consumed or freed in netlink_broadcast() */
+		if (err && err != -ESRCH)
+			goto failed;
+	}
+	return;
+
+nla_put_failure:
+	nlmsg_free(skb);
+failed:
+	drbd_err(peer_device, "Error %d while broadcasting event. Event seq:%u\n",
+		 err, seq);
+}
+
+void notify_helper(enum drbd_notification_type type,
+		   struct drbd_device *device, struct drbd_connection *connection,
+		   const char *name, int status)
+{
+	struct drbd_resource *resource = device ? device->resource : connection->resource;
+	struct drbd_helper_info helper_info;
+	unsigned int seq = atomic_inc_return(&notify_genl_seq);
+	struct sk_buff *skb = NULL;
+	struct drbd_genlmsghdr *dh;
+	int err;
+
+	strlcpy(helper_info.helper_name, name, sizeof(helper_info.helper_name));
+	helper_info.helper_name_len = min(strlen(name), sizeof(helper_info.helper_name));
+	helper_info.helper_status = status;
+
+	skb = genlmsg_new(NLMSG_GOODSIZE, GFP_NOIO);
+	err = -ENOMEM;
+	if (!skb)
+		goto fail;
+
+	err = -EMSGSIZE;
+	dh = genlmsg_put(skb, 0, seq, &drbd_genl_family, 0, DRBD_HELPER);
+	if (!dh)
+		goto fail;
+	dh->minor = device ? device->minor : -1;
+	dh->ret_code = NO_ERROR;
+	mutex_lock(&notification_mutex);
+	if (nla_put_drbd_cfg_context(skb, resource, connection, device) ||
+	    nla_put_notification_header(skb, type) ||
+	    drbd_helper_info_to_skb(skb, &helper_info, true))
+		goto unlock_fail;
+	genlmsg_end(skb, dh);
+	err = drbd_genl_multicast_events(skb, GFP_NOWAIT);
+	skb = NULL;
+	/* skb has been consumed or freed in netlink_broadcast() */
+	if (err && err != -ESRCH)
+		goto unlock_fail;
+	mutex_unlock(&notification_mutex);
+	return;
+
+unlock_fail:
+	mutex_unlock(&notification_mutex);
+fail:
+	nlmsg_free(skb);
+	drbd_err(resource, "Error %d while broadcasting event. Event seq:%u\n",
+		 err, seq);
+}
+
+static void notify_initial_state_done(struct sk_buff *skb, unsigned int seq)
+{
+	struct drbd_genlmsghdr *dh;
+	int err;
+
+	err = -EMSGSIZE;
+	dh = genlmsg_put(skb, 0, seq, &drbd_genl_family, 0, DRBD_INITIAL_STATE_DONE);
+	if (!dh)
+		goto nla_put_failure;
+	dh->minor = -1U;
+	dh->ret_code = NO_ERROR;
+	if (nla_put_notification_header(skb, NOTIFY_EXISTS))
+		goto nla_put_failure;
+	genlmsg_end(skb, dh);
+	return;
+
+nla_put_failure:
+	nlmsg_free(skb);
+	pr_err("Error %d sending event. Event seq:%u\n", err, seq);
+}
+
+static void free_state_changes(struct list_head *list)
+{
+	while (!list_empty(list)) {
+		struct drbd_state_change *state_change =
+			list_first_entry(list, struct drbd_state_change, list);
+		list_del(&state_change->list);
+		forget_state_change(state_change);
+	}
+}
+
+static unsigned int notifications_for_state_change(struct drbd_state_change *state_change)
+{
+	return 1 +
+	       state_change->n_connections +
+	       state_change->n_devices +
+	       state_change->n_devices * state_change->n_connections;
+}
+
+static int get_initial_state(struct sk_buff *skb, struct netlink_callback *cb)
+{
+	struct drbd_state_change *state_change = (struct drbd_state_change *)cb->args[0];
+	unsigned int seq = cb->args[2];
+	unsigned int n;
+	enum drbd_notification_type flags = 0;
+
+	/* There is no need for taking notification_mutex here: it doesn't
+	   matter if the initial state events mix with later state chage
+	   events; we can always tell the events apart by the NOTIFY_EXISTS
+	   flag. */
+
+	cb->args[5]--;
+	if (cb->args[5] == 1) {
+		notify_initial_state_done(skb, seq);
+		goto out;
+	}
+	n = cb->args[4]++;
+	if (cb->args[4] < cb->args[3])
+		flags |= NOTIFY_CONTINUES;
+	if (n < 1) {
+		notify_resource_state_change(skb, seq, state_change->resource,
+					     NOTIFY_EXISTS | flags);
+		goto next;
+	}
+	n--;
+	if (n < state_change->n_connections) {
+		notify_connection_state_change(skb, seq, &state_change->connections[n],
+					       NOTIFY_EXISTS | flags);
+		goto next;
+	}
+	n -= state_change->n_connections;
+	if (n < state_change->n_devices) {
+		notify_device_state_change(skb, seq, &state_change->devices[n],
+					   NOTIFY_EXISTS | flags);
+		goto next;
+	}
+	n -= state_change->n_devices;
+	if (n < state_change->n_devices * state_change->n_connections) {
+		notify_peer_device_state_change(skb, seq, &state_change->peer_devices[n],
+						NOTIFY_EXISTS | flags);
+		goto next;
+	}
+
+next:
+	if (cb->args[4] == cb->args[3]) {
+		struct drbd_state_change *next_state_change =
+			list_entry(state_change->list.next,
+				   struct drbd_state_change, list);
+		cb->args[0] = (long)next_state_change;
+		cb->args[3] = notifications_for_state_change(next_state_change);
+		cb->args[4] = 0;
+	}
+out:
+	return skb->len;
+}
+
+int drbd_adm_get_initial_state(struct sk_buff *skb, struct netlink_callback *cb)
+{
+	struct drbd_resource *resource;
+	LIST_HEAD(head);
+
+	if (cb->args[5] >= 1) {
+		if (cb->args[5] > 1)
+			return get_initial_state(skb, cb);
+		if (cb->args[0]) {
+			struct drbd_state_change *state_change =
+				(struct drbd_state_change *)cb->args[0];
+
+			/* connect list to head */
+			list_add(&head, &state_change->list);
+			free_state_changes(&head);
+		}
+		return 0;
+	}
+
+	cb->args[5] = 2;  /* number of iterations */
+	mutex_lock(&resources_mutex);
+	for_each_resource(resource, &drbd_resources) {
+		struct drbd_state_change *state_change;
+
+		state_change = remember_old_state(resource, GFP_KERNEL);
+		if (!state_change) {
+			if (!list_empty(&head))
+				free_state_changes(&head);
+			mutex_unlock(&resources_mutex);
+			return -ENOMEM;
+		}
+		copy_old_to_new_state_change(state_change);
+		list_add_tail(&state_change->list, &head);
+		cb->args[5] += notifications_for_state_change(state_change);
+	}
+	mutex_unlock(&resources_mutex);
+
+	if (!list_empty(&head)) {
+		struct drbd_state_change *state_change =
+			list_entry(head.next, struct drbd_state_change, list);
+		cb->args[0] = (long)state_change;
+		cb->args[3] = notifications_for_state_change(state_change);
+		list_del(&head);  /* detach list from head */
+	}
+
+	cb->args[2] = cb->nlh->nlmsg_seq;
+	return get_initial_state(skb, cb);
+}

drivers/block/drbd/drbd_proc.c

@@ -245,9 +245,9 @@ static int drbd_seq_show(struct seq_file *seq, void *v)
 	char wp;
 
 	static char write_ordering_chars[] = {
-		[WO_none] = 'n',
-		[WO_drain_io] = 'd',
-		[WO_bdev_flush] = 'f',
+		[WO_NONE] = 'n',
+		[WO_DRAIN_IO] = 'd',
+		[WO_BDEV_FLUSH] = 'f',
 	};
 
 	seq_printf(seq, "version: " REL_VERSION " (api:%d/proto:%d-%d)\n%s\n",

drivers/block/drbd/drbd_protocol.h

@@ -23,7 +23,7 @@ enum drbd_packet {
 	P_AUTH_RESPONSE	      = 0x11,
 	P_STATE_CHG_REQ	      = 0x12,
 
-	/* asender (meta socket */
+	/* (meta socket) */
 	P_PING		      = 0x13,
 	P_PING_ACK	      = 0x14,
 	P_RECV_ACK	      = 0x15, /* Used in protocol B */

drivers/block/drbd/drbd_receiver.c

@@ -215,7 +215,7 @@ static void reclaim_finished_net_peer_reqs(struct drbd_device *device,
 	}
 }
 
-static void drbd_kick_lo_and_reclaim_net(struct drbd_device *device)
+static void drbd_reclaim_net_peer_reqs(struct drbd_device *device)
 {
 	LIST_HEAD(reclaimed);
 	struct drbd_peer_request *peer_req, *t;
@@ -223,11 +223,30 @@ static void drbd_kick_lo_and_reclaim_net(struct drbd_device *device)
 	spin_lock_irq(&device->resource->req_lock);
 	reclaim_finished_net_peer_reqs(device, &reclaimed);
 	spin_unlock_irq(&device->resource->req_lock);
-
 	list_for_each_entry_safe(peer_req, t, &reclaimed, w.list)
 		drbd_free_net_peer_req(device, peer_req);
 }
 
+static void conn_reclaim_net_peer_reqs(struct drbd_connection *connection)
+{
+	struct drbd_peer_device *peer_device;
+	int vnr;
+
+	rcu_read_lock();
+	idr_for_each_entry(&connection->peer_devices, peer_device, vnr) {
+		struct drbd_device *device = peer_device->device;
+		if (!atomic_read(&device->pp_in_use_by_net))
+			continue;
+
+		kref_get(&device->kref);
+		rcu_read_unlock();
+		drbd_reclaim_net_peer_reqs(device);
+		kref_put(&device->kref, drbd_destroy_device);
+		rcu_read_lock();
+	}
+	rcu_read_unlock();
+}
+
 /**
  * drbd_alloc_pages() - Returns @number pages, retries forever (or until signalled)
  * @device:	DRBD device.
@@ -265,10 +284,15 @@ struct page *drbd_alloc_pages(struct drbd_peer_device *peer_device, unsigned int
 	if (atomic_read(&device->pp_in_use) < mxb)
 		page = __drbd_alloc_pages(device, number);
 
+	/* Try to keep the fast path fast, but occasionally we need
+	 * to reclaim the pages we lended to the network stack. */
+	if (page && atomic_read(&device->pp_in_use_by_net) > 512)
+		drbd_reclaim_net_peer_reqs(device);
+
 	while (page == NULL) {
 		prepare_to_wait(&drbd_pp_wait, &wait, TASK_INTERRUPTIBLE);
 
-		drbd_kick_lo_and_reclaim_net(device);
+		drbd_reclaim_net_peer_reqs(device);
 
 		if (atomic_read(&device->pp_in_use) < mxb) {
 			page = __drbd_alloc_pages(device, number);
@@ -1099,7 +1123,15 @@ static int conn_connect(struct drbd_connection *connection)
 		return 0;
 	}
 
-	drbd_thread_start(&connection->asender);
+	drbd_thread_start(&connection->ack_receiver);
+	/* opencoded create_singlethread_workqueue(),
+	 * to be able to use format string arguments */
+	connection->ack_sender =
+		alloc_ordered_workqueue("drbd_as_%s", WQ_MEM_RECLAIM, connection->resource->name);
+	if (!connection->ack_sender) {
+		drbd_err(connection, "Failed to create workqueue ack_sender\n");
+		return 0;
+	}
 
 	mutex_lock(&connection->resource->conf_update);
 	/* The discard_my_data flag is a single-shot modifier to the next
@@ -1178,7 +1210,7 @@ static void drbd_flush(struct drbd_connection *connection)
 	struct drbd_peer_device *peer_device;
 	int vnr;
 
-	if (connection->resource->write_ordering >= WO_bdev_flush) {
+	if (connection->resource->write_ordering >= WO_BDEV_FLUSH) {
 		rcu_read_lock();
 		idr_for_each_entry(&connection->peer_devices, peer_device, vnr) {
 			struct drbd_device *device = peer_device->device;
@@ -1203,7 +1235,7 @@ static void drbd_flush(struct drbd_connection *connection)
 				/* would rather check on EOPNOTSUPP, but that is not reliable.
 				 * don't try again for ANY return value != 0
 				 * if (rv == -EOPNOTSUPP) */
-				drbd_bump_write_ordering(connection->resource, NULL, WO_drain_io);
+				drbd_bump_write_ordering(connection->resource, NULL, WO_DRAIN_IO);
 			}
 			put_ldev(device);
 			kref_put(&device->kref, drbd_destroy_device);
@@ -1299,10 +1331,10 @@ max_allowed_wo(struct drbd_backing_dev *bdev, enum write_ordering_e wo)
 
 	dc = rcu_dereference(bdev->disk_conf);
 
-	if (wo == WO_bdev_flush && !dc->disk_flushes)
-		wo = WO_drain_io;
-	if (wo == WO_drain_io && !dc->disk_drain)
-		wo = WO_none;
+	if (wo == WO_BDEV_FLUSH && !dc->disk_flushes)
+		wo = WO_DRAIN_IO;
+	if (wo == WO_DRAIN_IO && !dc->disk_drain)
+		wo = WO_NONE;
 
 	return wo;
 }
@@ -1319,13 +1351,13 @@ void drbd_bump_write_ordering(struct drbd_resource *resource, struct drbd_backin
 	enum write_ordering_e pwo;
 	int vnr;
 	static char *write_ordering_str[] = {
-		[WO_none] = "none",
-		[WO_drain_io] = "drain",
-		[WO_bdev_flush] = "flush",
+		[WO_NONE] = "none",
+		[WO_DRAIN_IO] = "drain",
+		[WO_BDEV_FLUSH] = "flush",
 	};
 
 	pwo = resource->write_ordering;
-	if (wo != WO_bdev_flush)
+	if (wo != WO_BDEV_FLUSH)
 		wo = min(pwo, wo);
 	rcu_read_lock();
 	idr_for_each_entry(&resource->devices, device, vnr) {
@@ -1343,7 +1375,7 @@ void drbd_bump_write_ordering(struct drbd_resource *resource, struct drbd_backin
 	rcu_read_unlock();
 
 	resource->write_ordering = wo;
-	if (pwo != resource->write_ordering || wo == WO_bdev_flush)
+	if (pwo != resource->write_ordering || wo == WO_BDEV_FLUSH)
 		drbd_info(resource, "Method to ensure write ordering: %s\n", write_ordering_str[resource->write_ordering]);
 }
 
@@ -1380,7 +1412,7 @@ int drbd_submit_peer_request(struct drbd_device *device,
 	if (peer_req->flags & EE_IS_TRIM_USE_ZEROOUT) {
 		/* wait for all pending IO completions, before we start
 		 * zeroing things out. */
-		conn_wait_active_ee_empty(first_peer_device(device)->connection);
+		conn_wait_active_ee_empty(peer_req->peer_device->connection);
 		/* add it to the active list now,
 		 * so we can find it to present it in debugfs */
 		peer_req->submit_jif = jiffies;
@@ -1508,12 +1540,6 @@ static void conn_wait_active_ee_empty(struct drbd_connection *connection)
 	rcu_read_unlock();
 }
 
-static struct drbd_peer_device *
-conn_peer_device(struct drbd_connection *connection, int volume_number)
-{
-	return idr_find(&connection->peer_devices, volume_number);
-}
-
 static int receive_Barrier(struct drbd_connection *connection, struct packet_info *pi)
 {
 	int rv;
@@ -1533,7 +1559,7 @@ static int receive_Barrier(struct drbd_connection *connection, struct packet_inf
 	 * Therefore we must send the barrier_ack after the barrier request was
 	 * completed. */
 	switch (connection->resource->write_ordering) {
-	case WO_none:
+	case WO_NONE:
 		if (rv == FE_RECYCLED)
 			return 0;
 
@@ -1546,8 +1572,8 @@ static int receive_Barrier(struct drbd_connection *connection, struct packet_inf
 			drbd_warn(connection, "Allocation of an epoch failed, slowing down\n");
 			/* Fall through */
 
-	case WO_bdev_flush:
-	case WO_drain_io:
+	case WO_BDEV_FLUSH:
+	case WO_DRAIN_IO:
 		conn_wait_active_ee_empty(connection);
 		drbd_flush(connection);
 
@@ -1752,7 +1778,7 @@ static int recv_dless_read(struct drbd_peer_device *peer_device, struct drbd_req
 }
 
 /*
- * e_end_resync_block() is called in asender context via
+ * e_end_resync_block() is called in ack_sender context via
  * drbd_finish_peer_reqs().
  */
 static int e_end_resync_block(struct drbd_work *w, int unused)
@@ -1926,7 +1952,7 @@ static void restart_conflicting_writes(struct drbd_device *device,
 }
 
 /*
- * e_end_block() is called in asender context via drbd_finish_peer_reqs().
+ * e_end_block() is called in ack_sender context via drbd_finish_peer_reqs().
  */
 static int e_end_block(struct drbd_work *w, int cancel)
 {
@@ -1966,7 +1992,7 @@ static int e_end_block(struct drbd_work *w, int cancel)
 	} else
 		D_ASSERT(device, drbd_interval_empty(&peer_req->i));
 
-	drbd_may_finish_epoch(first_peer_device(device)->connection, peer_req->epoch, EV_PUT + (cancel ? EV_CLEANUP : 0));
+	drbd_may_finish_epoch(peer_device->connection, peer_req->epoch, EV_PUT + (cancel ? EV_CLEANUP : 0));
 
 	return err;
 }
@@ -2098,7 +2124,7 @@ static int wait_for_and_update_peer_seq(struct drbd_peer_device *peer_device, co
 		}
 
 		rcu_read_lock();
-		tp = rcu_dereference(first_peer_device(device)->connection->net_conf)->two_primaries;
+		tp = rcu_dereference(peer_device->connection->net_conf)->two_primaries;
 		rcu_read_unlock();
 
 		if (!tp)
@@ -2217,7 +2243,7 @@ static int handle_write_conflicts(struct drbd_device *device,
 			peer_req->w.cb = superseded ? e_send_superseded :
 						   e_send_retry_write;
 			list_add_tail(&peer_req->w.list, &device->done_ee);
-			wake_asender(connection);
+			queue_work(connection->ack_sender, &peer_req->peer_device->send_acks_work);
 
 			err = -ENOENT;
 			goto out;
@@ -2364,7 +2390,7 @@ static int receive_Data(struct drbd_connection *connection, struct packet_info *
 	if (dp_flags & DP_SEND_RECEIVE_ACK) {
 		/* I really don't like it that the receiver thread
 		 * sends on the msock, but anyways */
-		drbd_send_ack(first_peer_device(device), P_RECV_ACK, peer_req);
+		drbd_send_ack(peer_device, P_RECV_ACK, peer_req);
 	}
 
 	if (tp) {
@@ -4056,7 +4082,7 @@ static int receive_state(struct drbd_connection *connection, struct packet_info
 	os = ns = drbd_read_state(device);
 	spin_unlock_irq(&device->resource->req_lock);
 
-	/* If some other part of the code (asender thread, timeout)
+	/* If some other part of the code (ack_receiver thread, timeout)
 	 * already decided to close the connection again,
 	 * we must not "re-establish" it here. */
 	if (os.conn <= C_TEAR_DOWN)
@@ -4661,8 +4687,12 @@ static void conn_disconnect(struct drbd_connection *connection)
 	 */
 	conn_request_state(connection, NS(conn, C_NETWORK_FAILURE), CS_HARD);
 
-	/* asender does not clean up anything. it must not interfere, either */
-	drbd_thread_stop(&connection->asender);
+	/* ack_receiver does not clean up anything. it must not interfere, either */
+	drbd_thread_stop(&connection->ack_receiver);
+	if (connection->ack_sender) {
+		destroy_workqueue(connection->ack_sender);
+		connection->ack_sender = NULL;
+	}
 	drbd_free_sock(connection);
 
 	rcu_read_lock();
@@ -5431,49 +5461,39 @@ static int got_skip(struct drbd_connection *connection, struct packet_info *pi)
 	return 0;
 }
 
-static int connection_finish_peer_reqs(struct drbd_connection *connection)
+struct meta_sock_cmd {
+	size_t pkt_size;
+	int (*fn)(struct drbd_connection *connection, struct packet_info *);
+};
+
+static void set_rcvtimeo(struct drbd_connection *connection, bool ping_timeout)
 {
-	struct drbd_peer_device *peer_device;
-	int vnr, not_empty = 0;
+	long t;
+	struct net_conf *nc;
 
-	do {
-		clear_bit(SIGNAL_ASENDER, &connection->flags);
-		flush_signals(current);
+	rcu_read_lock();
+	nc = rcu_dereference(connection->net_conf);
+	t = ping_timeout ? nc->ping_timeo : nc->ping_int;
+	rcu_read_unlock();
 
-		rcu_read_lock();
-		idr_for_each_entry(&connection->peer_devices, peer_device, vnr) {
-			struct drbd_device *device = peer_device->device;
-			kref_get(&device->kref);
-			rcu_read_unlock();
-			if (drbd_finish_peer_reqs(device)) {
-				kref_put(&device->kref, drbd_destroy_device);
-				return 1;
-			}
-			kref_put(&device->kref, drbd_destroy_device);
-			rcu_read_lock();
-		}
-		set_bit(SIGNAL_ASENDER, &connection->flags);
+	t *= HZ;
+	if (ping_timeout)
+		t /= 10;
 
-		spin_lock_irq(&connection->resource->req_lock);
-		idr_for_each_entry(&connection->peer_devices, peer_device, vnr) {
-			struct drbd_device *device = peer_device->device;
-			not_empty = !list_empty(&device->done_ee);
-			if (not_empty)
-				break;
-		}
-		spin_unlock_irq(&connection->resource->req_lock);
-		rcu_read_unlock();
-	} while (not_empty);
+	connection->meta.socket->sk->sk_rcvtimeo = t;
+}
 
-	return 0;
+static void set_ping_timeout(struct drbd_connection *connection)
+{
+	set_rcvtimeo(connection, 1);
 }
 
-struct asender_cmd {
-	size_t pkt_size;
-	int (*fn)(struct drbd_connection *connection, struct packet_info *);
-};
+static void set_idle_timeout(struct drbd_connection *connection)
+{
+	set_rcvtimeo(connection, 0);
+}
 
-static struct asender_cmd asender_tbl[] = {
+static struct meta_sock_cmd ack_receiver_tbl[] = {
 	[P_PING]	    = { 0, got_Ping },
 	[P_PING_ACK]	    = { 0, got_PingAck },
 	[P_RECV_ACK]	    = { sizeof(struct p_block_ack), got_BlockAck },
@@ -5493,64 +5513,40 @@ static struct asender_cmd asender_tbl[] = {
 	[P_RETRY_WRITE]	    = { sizeof(struct p_block_ack), got_BlockAck },
 };
 
-int drbd_asender(struct drbd_thread *thi)
+int drbd_ack_receiver(struct drbd_thread *thi)
 {
 	struct drbd_connection *connection = thi->connection;
-	struct asender_cmd *cmd = NULL;
+	struct meta_sock_cmd *cmd = NULL;
 	struct packet_info pi;
+	unsigned long pre_recv_jif;
 	int rv;
 	void *buf    = connection->meta.rbuf;
 	int received = 0;
 	unsigned int header_size = drbd_header_size(connection);
 	int expect   = header_size;
 	bool ping_timeout_active = false;
-	struct net_conf *nc;
-	int ping_timeo, tcp_cork, ping_int;
 	struct sched_param param = { .sched_priority = 2 };
 
 	rv = sched_setscheduler(current, SCHED_RR, &param);
 	if (rv < 0)
-		drbd_err(connection, "drbd_asender: ERROR set priority, ret=%d\n", rv);
+		drbd_err(connection, "drbd_ack_receiver: ERROR set priority, ret=%d\n", rv);
 
 	while (get_t_state(thi) == RUNNING) {
 		drbd_thread_current_set_cpu(thi);
 
-		rcu_read_lock();
-		nc = rcu_dereference(connection->net_conf);
-		ping_timeo = nc->ping_timeo;
-		tcp_cork = nc->tcp_cork;
-		ping_int = nc->ping_int;
-		rcu_read_unlock();
+		conn_reclaim_net_peer_reqs(connection);
 
 		if (test_and_clear_bit(SEND_PING, &connection->flags)) {
 			if (drbd_send_ping(connection)) {
 				drbd_err(connection, "drbd_send_ping has failed\n");
 				goto reconnect;
 			}
-			connection->meta.socket->sk->sk_rcvtimeo = ping_timeo * HZ / 10;
+			set_ping_timeout(connection);
 			ping_timeout_active = true;
 		}
 
-		/* TODO: conditionally cork; it may hurt latency if we cork without
-		   much to send */
-		if (tcp_cork)
-			drbd_tcp_cork(connection->meta.socket);
-		if (connection_finish_peer_reqs(connection)) {
-			drbd_err(connection, "connection_finish_peer_reqs() failed\n");
-			goto reconnect;
-		}
-		/* but unconditionally uncork unless disabled */
-		if (tcp_cork)
-			drbd_tcp_uncork(connection->meta.socket);
-
-		/* short circuit, recv_msg would return EINTR anyways. */
-		if (signal_pending(current))
-			continue;
-
+		pre_recv_jif = jiffies;
 		rv = drbd_recv_short(connection->meta.socket, buf, expect-received, 0);
-		clear_bit(SIGNAL_ASENDER, &connection->flags);
-
-		flush_signals(current);
 
 		/* Note:
 		 * -EINTR	 (on meta) we got a signal
@@ -5562,7 +5558,6 @@ int drbd_asender(struct drbd_thread *thi)
 		 * rv <  expected: "woken" by signal during receive
 		 * rv == 0	 : "connection shut down by peer"
 		 */
-received_more:
 		if (likely(rv > 0)) {
 			received += rv;
 			buf	 += rv;
@@ -5584,8 +5579,7 @@ int drbd_asender(struct drbd_thread *thi)
 		} else if (rv == -EAGAIN) {
 			/* If the data socket received something meanwhile,
 			 * that is good enough: peer is still alive. */
-			if (time_after(connection->last_received,
-				jiffies - connection->meta.socket->sk->sk_rcvtimeo))
+			if (time_after(connection->last_received, pre_recv_jif))
 				continue;
 			if (ping_timeout_active) {
 				drbd_err(connection, "PingAck did not arrive in time.\n");
@@ -5594,6 +5588,10 @@ int drbd_asender(struct drbd_thread *thi)
 			set_bit(SEND_PING, &connection->flags);
 			continue;
 		} else if (rv == -EINTR) {
+			/* maybe drbd_thread_stop(): the while condition will notice.
+			 * maybe woken for send_ping: we'll send a ping above,
+			 * and change the rcvtimeo */
+			flush_signals(current);
 			continue;
 		} else {
 			drbd_err(connection, "sock_recvmsg returned %d\n", rv);
@@ -5603,8 +5601,8 @@ int drbd_asender(struct drbd_thread *thi)
 		if (received == expect && cmd == NULL) {
 			if (decode_header(connection, connection->meta.rbuf, &pi))
 				goto reconnect;
-			cmd = &asender_tbl[pi.cmd];
-			if (pi.cmd >= ARRAY_SIZE(asender_tbl) || !cmd->fn) {
+			cmd = &ack_receiver_tbl[pi.cmd];
+			if (pi.cmd >= ARRAY_SIZE(ack_receiver_tbl) || !cmd->fn) {
 				drbd_err(connection, "Unexpected meta packet %s (0x%04x)\n",
 					 cmdname(pi.cmd), pi.cmd);
 				goto disconnect;
@@ -5627,9 +5625,8 @@ int drbd_asender(struct drbd_thread *thi)
 
 			connection->last_received = jiffies;
 
-			if (cmd == &asender_tbl[P_PING_ACK]) {
-				/* restore idle timeout */
-				connection->meta.socket->sk->sk_rcvtimeo = ping_int * HZ;
+			if (cmd == &ack_receiver_tbl[P_PING_ACK]) {
+				set_idle_timeout(connection);
 				ping_timeout_active = false;
 			}
 
@@ -5638,11 +5635,6 @@ int drbd_asender(struct drbd_thread *thi)
 			expect	 = header_size;
 			cmd	 = NULL;
 		}
-		if (test_bit(SEND_PING, &connection->flags))
-			continue;
-		rv = drbd_recv_short(connection->meta.socket, buf, expect-received, MSG_DONTWAIT);
-		if (rv > 0)
-			goto received_more;
 	}
 
 	if (0) {
@@ -5654,9 +5646,41 @@ int drbd_asender(struct drbd_thread *thi)
 disconnect:
 		conn_request_state(connection, NS(conn, C_DISCONNECTING), CS_HARD);
 	}
-	clear_bit(SIGNAL_ASENDER, &connection->flags);
 
-	drbd_info(connection, "asender terminated\n");
+	drbd_info(connection, "ack_receiver terminated\n");
 
 	return 0;
 }
+
+void drbd_send_acks_wf(struct work_struct *ws)
+{
+	struct drbd_peer_device *peer_device =
+		container_of(ws, struct drbd_peer_device, send_acks_work);
+	struct drbd_connection *connection = peer_device->connection;
+	struct drbd_device *device = peer_device->device;
+	struct net_conf *nc;
+	int tcp_cork, err;
+
+	rcu_read_lock();
+	nc = rcu_dereference(connection->net_conf);
+	tcp_cork = nc->tcp_cork;
+	rcu_read_unlock();
+
+	if (tcp_cork)
+		drbd_tcp_cork(connection->meta.socket);
+
+	err = drbd_finish_peer_reqs(device);
+	kref_put(&device->kref, drbd_destroy_device);
+	/* get is in drbd_endio_write_sec_final(). That is necessary to keep the
+	   struct work_struct send_acks_work alive, which is in the peer_device object */
+
+	if (err) {
+		conn_request_state(connection, NS(conn, C_NETWORK_FAILURE), CS_HARD);
+		return;
+	}
+
+	if (tcp_cork)
+		drbd_tcp_uncork(connection->meta.socket);
+
+	return;
+}

drivers/block/drbd/drbd_req.c

@@ -453,12 +453,12 @@ static void mod_rq_state(struct drbd_request *req, struct bio_and_error *m,
 		kref_get(&req->kref); /* wait for the DONE */
 
 	if (!(s & RQ_NET_SENT) && (set & RQ_NET_SENT)) {
-		/* potentially already completed in the asender thread */
+		/* potentially already completed in the ack_receiver thread */
 		if (!(s & RQ_NET_DONE)) {
 			atomic_add(req->i.size >> 9, &device->ap_in_flight);
 			set_if_null_req_not_net_done(peer_device, req);
 		}
-		if (s & RQ_NET_PENDING)
+		if (req->rq_state & RQ_NET_PENDING)
 			set_if_null_req_ack_pending(peer_device, req);
 	}
 
@@ -1095,6 +1095,24 @@ static bool do_remote_read(struct drbd_request *req)
 	return false;
 }
 
+bool drbd_should_do_remote(union drbd_dev_state s)
+{
+	return s.pdsk == D_UP_TO_DATE ||
+		(s.pdsk >= D_INCONSISTENT &&
+		 s.conn >= C_WF_BITMAP_T &&
+		 s.conn < C_AHEAD);
+	/* Before proto 96 that was >= CONNECTED instead of >= C_WF_BITMAP_T.
+	   That is equivalent since before 96 IO was frozen in the C_WF_BITMAP*
+	   states. */
+}
+
+static bool drbd_should_send_out_of_sync(union drbd_dev_state s)
+{
+	return s.conn == C_AHEAD || s.conn == C_WF_BITMAP_S;
+	/* pdsk = D_INCONSISTENT as a consequence. Protocol 96 check not necessary
+	   since we enter state C_AHEAD only if proto >= 96 */
+}
+
 /* returns number of connections (== 1, for drbd 8.4)
  * expected to actually write this data,
  * which does NOT include those that we are L_AHEAD for. */
@@ -1149,7 +1167,6 @@ drbd_submit_req_private_bio(struct drbd_request *req)
 	 * stable storage, and this is a WRITE, we may not even submit
 	 * this bio. */
 	if (get_ldev(device)) {
-		req->pre_submit_jif = jiffies;
 		if (drbd_insert_fault(device,
 				      rw == WRITE ? DRBD_FAULT_DT_WR
 				    : rw == READ  ? DRBD_FAULT_DT_RD
@@ -1293,6 +1310,7 @@ static void drbd_send_and_submit(struct drbd_device *device, struct drbd_request
 			&device->pending_master_completion[rw == WRITE]);
 	if (req->private_bio) {
 		/* needs to be marked within the same spinlock */
+		req->pre_submit_jif = jiffies;
 		list_add_tail(&req->req_pending_local,
 			&device->pending_completion[rw == WRITE]);
 		_req_mod(req, TO_BE_SUBMITTED);
@@ -1513,6 +1531,78 @@ blk_qc_t drbd_make_request(struct request_queue *q, struct bio *bio)
 	return BLK_QC_T_NONE;
 }
 
+static bool net_timeout_reached(struct drbd_request *net_req,
+		struct drbd_connection *connection,
+		unsigned long now, unsigned long ent,
+		unsigned int ko_count, unsigned int timeout)
+{
+	struct drbd_device *device = net_req->device;
+
+	if (!time_after(now, net_req->pre_send_jif + ent))
+		return false;
+
+	if (time_in_range(now, connection->last_reconnect_jif, connection->last_reconnect_jif + ent))
+		return false;
+
+	if (net_req->rq_state & RQ_NET_PENDING) {
+		drbd_warn(device, "Remote failed to finish a request within %ums > ko-count (%u) * timeout (%u * 0.1s)\n",
+			jiffies_to_msecs(now - net_req->pre_send_jif), ko_count, timeout);
+		return true;
+	}
+
+	/* We received an ACK already (or are using protocol A),
+	 * but are waiting for the epoch closing barrier ack.
+	 * Check if we sent the barrier already.  We should not blame the peer
+	 * for being unresponsive, if we did not even ask it yet. */
+	if (net_req->epoch == connection->send.current_epoch_nr) {
+		drbd_warn(device,
+			"We did not send a P_BARRIER for %ums > ko-count (%u) * timeout (%u * 0.1s); drbd kernel thread blocked?\n",
+			jiffies_to_msecs(now - net_req->pre_send_jif), ko_count, timeout);
+		return false;
+	}
+
+	/* Worst case: we may have been blocked for whatever reason, then
+	 * suddenly are able to send a lot of requests (and epoch separating
+	 * barriers) in quick succession.
+	 * The timestamp of the net_req may be much too old and not correspond
+	 * to the sending time of the relevant unack'ed barrier packet, so
+	 * would trigger a spurious timeout.  The latest barrier packet may
+	 * have a too recent timestamp to trigger the timeout, potentially miss
+	 * a timeout.  Right now we don't have a place to conveniently store
+	 * these timestamps.
+	 * But in this particular situation, the application requests are still
+	 * completed to upper layers, DRBD should still "feel" responsive.
+	 * No need yet to kill this connection, it may still recover.
+	 * If not, eventually we will have queued enough into the network for
+	 * us to block. From that point of view, the timestamp of the last sent
+	 * barrier packet is relevant enough.
+	 */
+	if (time_after(now, connection->send.last_sent_barrier_jif + ent)) {
+		drbd_warn(device, "Remote failed to answer a P_BARRIER (sent at %lu jif; now=%lu jif) within %ums > ko-count (%u) * timeout (%u * 0.1s)\n",
+			connection->send.last_sent_barrier_jif, now,
+			jiffies_to_msecs(now - connection->send.last_sent_barrier_jif), ko_count, timeout);
+		return true;
+	}
+	return false;
+}
+
+/* A request is considered timed out, if
+ * - we have some effective timeout from the configuration,
+ *   with some state restrictions applied,
+ * - the oldest request is waiting for a response from the network
+ *   resp. the local disk,
+ * - the oldest request is in fact older than the effective timeout,
+ * - the connection was established (resp. disk was attached)
+ *   for longer than the timeout already.
+ * Note that for 32bit jiffies and very stable connections/disks,
+ * we may have a wrap around, which is catched by
+ *   !time_in_range(now, last_..._jif, last_..._jif + timeout).
+ *
+ * Side effect: once per 32bit wrap-around interval, which means every
+ * ~198 days with 250 HZ, we have a window where the timeout would need
+ * to expire twice (worst case) to become effective. Good enough.
+ */
+
 void request_timer_fn(unsigned long data)
 {
 	struct drbd_device *device = (struct drbd_device *) data;
@@ -1522,11 +1612,14 @@ void request_timer_fn(unsigned long data)
 	unsigned long oldest_submit_jif;
 	unsigned long ent = 0, dt = 0, et, nt; /* effective timeout = ko_count * timeout */
 	unsigned long now;
+	unsigned int ko_count = 0, timeout = 0;
 
 	rcu_read_lock();
 	nc = rcu_dereference(connection->net_conf);
-	if (nc && device->state.conn >= C_WF_REPORT_PARAMS)
-		ent = nc->timeout * HZ/10 * nc->ko_count;
+	if (nc && device->state.conn >= C_WF_REPORT_PARAMS) {
+		ko_count = nc->ko_count;
+		timeout = nc->timeout;
+	}
 
 	if (get_ldev(device)) { /* implicit state.disk >= D_INCONSISTENT */
 		dt = rcu_dereference(device->ldev->disk_conf)->disk_timeout * HZ / 10;
@@ -1534,6 +1627,8 @@ void request_timer_fn(unsigned long data)
 	}
 	rcu_read_unlock();
 
+
+	ent = timeout * HZ/10 * ko_count;
 	et = min_not_zero(dt, ent);
 
 	if (!et)
@@ -1545,11 +1640,22 @@ void request_timer_fn(unsigned long data)
 	spin_lock_irq(&device->resource->req_lock);
 	req_read = list_first_entry_or_null(&device->pending_completion[0], struct drbd_request, req_pending_local);
 	req_write = list_first_entry_or_null(&device->pending_completion[1], struct drbd_request, req_pending_local);
-	req_peer = connection->req_not_net_done;
+
 	/* maybe the oldest request waiting for the peer is in fact still
-	 * blocking in tcp sendmsg */
-	if (!req_peer && connection->req_next && connection->req_next->pre_send_jif)
-		req_peer = connection->req_next;
+	 * blocking in tcp sendmsg.  That's ok, though, that's handled via the
+	 * socket send timeout, requesting a ping, and bumping ko-count in
+	 * we_should_drop_the_connection().
+	 */
+
+	/* check the oldest request we did successfully sent,
+	 * but which is still waiting for an ACK. */
+	req_peer = connection->req_ack_pending;
+
+	/* if we don't have such request (e.g. protocoll A)
+	 * check the oldest requests which is still waiting on its epoch
+	 * closing barrier ack. */
+	if (!req_peer)
+		req_peer = connection->req_not_net_done;
 
 	/* evaluate the oldest peer request only in one timer! */
 	if (req_peer && req_peer->device != device)
@@ -1566,28 +1672,9 @@ void request_timer_fn(unsigned long data)
 		: req_write ? req_write->pre_submit_jif
 		: req_read ? req_read->pre_submit_jif : now;
 
-	/* The request is considered timed out, if
-	 * - we have some effective timeout from the configuration,
-	 *   with above state restrictions applied,
-	 * - the oldest request is waiting for a response from the network
-	 *   resp. the local disk,
-	 * - the oldest request is in fact older than the effective timeout,
-	 * - the connection was established (resp. disk was attached)
-	 *   for longer than the timeout already.
-	 * Note that for 32bit jiffies and very stable connections/disks,
-	 * we may have a wrap around, which is catched by
-	 *   !time_in_range(now, last_..._jif, last_..._jif + timeout).
-	 *
-	 * Side effect: once per 32bit wrap-around interval, which means every
-	 * ~198 days with 250 HZ, we have a window where the timeout would need
-	 * to expire twice (worst case) to become effective. Good enough.
-	 */
-	if (ent && req_peer &&
-		 time_after(now, req_peer->pre_send_jif + ent) &&
-		!time_in_range(now, connection->last_reconnect_jif, connection->last_reconnect_jif + ent)) {
-		drbd_warn(device, "Remote failed to finish a request within ko-count * timeout\n");
+	if (ent && req_peer && net_timeout_reached(req_peer, connection, now, ent, ko_count, timeout))
 		_conn_request_state(connection, NS(conn, C_TIMEOUT), CS_VERBOSE | CS_HARD);
-	}
+
 	if (dt && oldest_submit_jif != now &&
 		 time_after(now, oldest_submit_jif + dt) &&
 		!time_in_range(now, device->last_reattach_jif, device->last_reattach_jif + dt)) {

drivers/block/drbd/drbd_req.h

@@ -331,21 +331,6 @@ static inline int req_mod(struct drbd_request *req,
 	return rv;
 }
 
-static inline bool drbd_should_do_remote(union drbd_dev_state s)
-{
-	return s.pdsk == D_UP_TO_DATE ||
-		(s.pdsk >= D_INCONSISTENT &&
-		 s.conn >= C_WF_BITMAP_T &&
-		 s.conn < C_AHEAD);
-	/* Before proto 96 that was >= CONNECTED instead of >= C_WF_BITMAP_T.
-	   That is equivalent since before 96 IO was frozen in the C_WF_BITMAP*
-	   states. */
-}
-static inline bool drbd_should_send_out_of_sync(union drbd_dev_state s)
-{
-	return s.conn == C_AHEAD || s.conn == C_WF_BITMAP_S;
-	/* pdsk = D_INCONSISTENT as a consequence. Protocol 96 check not necessary
-	   since we enter state C_AHEAD only if proto >= 96 */
-}
+extern bool drbd_should_do_remote(union drbd_dev_state);
 
 #endif

drivers/block/drbd/drbd_state.c

@@ -29,6 +29,7 @@
 #include "drbd_int.h"
 #include "drbd_protocol.h"
 #include "drbd_req.h"
+#include "drbd_state_change.h"
 
 struct after_state_chg_work {
 	struct drbd_work w;
@@ -37,6 +38,7 @@ struct after_state_chg_work {
 	union drbd_state ns;
 	enum chg_state_flags flags;
 	struct completion *done;
+	struct drbd_state_change *state_change;
 };
 
 enum sanitize_state_warnings {
@@ -48,9 +50,248 @@ enum sanitize_state_warnings {
 	IMPLICITLY_UPGRADED_PDSK,
 };
 
+static void count_objects(struct drbd_resource *resource,
+			  unsigned int *n_devices,
+			  unsigned int *n_connections)
+{
+	struct drbd_device *device;
+	struct drbd_connection *connection;
+	int vnr;
+
+	*n_devices = 0;
+	*n_connections = 0;
+
+	idr_for_each_entry(&resource->devices, device, vnr)
+		(*n_devices)++;
+	for_each_connection(connection, resource)
+		(*n_connections)++;
+}
+
+static struct drbd_state_change *alloc_state_change(unsigned int n_devices, unsigned int n_connections, gfp_t gfp)
+{
+	struct drbd_state_change *state_change;
+	unsigned int size, n;
+
+	size = sizeof(struct drbd_state_change) +
+	       n_devices * sizeof(struct drbd_device_state_change) +
+	       n_connections * sizeof(struct drbd_connection_state_change) +
+	       n_devices * n_connections * sizeof(struct drbd_peer_device_state_change);
+	state_change = kmalloc(size, gfp);
+	if (!state_change)
+		return NULL;
+	state_change->n_devices = n_devices;
+	state_change->n_connections = n_connections;
+	state_change->devices = (void *)(state_change + 1);
+	state_change->connections = (void *)&state_change->devices[n_devices];
+	state_change->peer_devices = (void *)&state_change->connections[n_connections];
+	state_change->resource->resource = NULL;
+	for (n = 0; n < n_devices; n++)
+		state_change->devices[n].device = NULL;
+	for (n = 0; n < n_connections; n++)
+		state_change->connections[n].connection = NULL;
+	return state_change;
+}
+
+struct drbd_state_change *remember_old_state(struct drbd_resource *resource, gfp_t gfp)
+{
+	struct drbd_state_change *state_change;
+	struct drbd_device *device;
+	unsigned int n_devices;
+	struct drbd_connection *connection;
+	unsigned int n_connections;
+	int vnr;
+
+	struct drbd_device_state_change *device_state_change;
+	struct drbd_peer_device_state_change *peer_device_state_change;
+	struct drbd_connection_state_change *connection_state_change;
+
+	/* Caller holds req_lock spinlock.
+	 * No state, no device IDR, no connections lists can change. */
+	count_objects(resource, &n_devices, &n_connections);
+	state_change = alloc_state_change(n_devices, n_connections, gfp);
+	if (!state_change)
+		return NULL;
+
+	kref_get(&resource->kref);
+	state_change->resource->resource = resource;
+	state_change->resource->role[OLD] =
+		conn_highest_role(first_connection(resource));
+	state_change->resource->susp[OLD] = resource->susp;
+	state_change->resource->susp_nod[OLD] = resource->susp_nod;
+	state_change->resource->susp_fen[OLD] = resource->susp_fen;
+
+	connection_state_change = state_change->connections;
+	for_each_connection(connection, resource) {
+		kref_get(&connection->kref);
+		connection_state_change->connection = connection;
+		connection_state_change->cstate[OLD] =
+			connection->cstate;
+		connection_state_change->peer_role[OLD] =
+			conn_highest_peer(connection);
+		connection_state_change++;
+	}
+
+	device_state_change = state_change->devices;
+	peer_device_state_change = state_change->peer_devices;
+	idr_for_each_entry(&resource->devices, device, vnr) {
+		kref_get(&device->kref);
+		device_state_change->device = device;
+		device_state_change->disk_state[OLD] = device->state.disk;
+
+		/* The peer_devices for each device have to be enumerated in
+		   the order of the connections. We may not use for_each_peer_device() here. */
+		for_each_connection(connection, resource) {
+			struct drbd_peer_device *peer_device;
+
+			peer_device = conn_peer_device(connection, device->vnr);
+			peer_device_state_change->peer_device = peer_device;
+			peer_device_state_change->disk_state[OLD] =
+				device->state.pdsk;
+			peer_device_state_change->repl_state[OLD] =
+				max_t(enum drbd_conns,
+				      C_WF_REPORT_PARAMS, device->state.conn);
+			peer_device_state_change->resync_susp_user[OLD] =
+				device->state.user_isp;
+			peer_device_state_change->resync_susp_peer[OLD] =
+				device->state.peer_isp;
+			peer_device_state_change->resync_susp_dependency[OLD] =
+				device->state.aftr_isp;
+			peer_device_state_change++;
+		}
+		device_state_change++;
+	}
+
+	return state_change;
+}
+
+static void remember_new_state(struct drbd_state_change *state_change)
+{
+	struct drbd_resource_state_change *resource_state_change;
+	struct drbd_resource *resource;
+	unsigned int n;
+
+	if (!state_change)
+		return;
+
+	resource_state_change = &state_change->resource[0];
+	resource = resource_state_change->resource;
+
+	resource_state_change->role[NEW] =
+		conn_highest_role(first_connection(resource));
+	resource_state_change->susp[NEW] = resource->susp;
+	resource_state_change->susp_nod[NEW] = resource->susp_nod;
+	resource_state_change->susp_fen[NEW] = resource->susp_fen;
+
+	for (n = 0; n < state_change->n_devices; n++) {
+		struct drbd_device_state_change *device_state_change =
+			&state_change->devices[n];
+		struct drbd_device *device = device_state_change->device;
+
+		device_state_change->disk_state[NEW] = device->state.disk;
+	}
+
+	for (n = 0; n < state_change->n_connections; n++) {
+		struct drbd_connection_state_change *connection_state_change =
+			&state_change->connections[n];
+		struct drbd_connection *connection =
+			connection_state_change->connection;
+
+		connection_state_change->cstate[NEW] = connection->cstate;
+		connection_state_change->peer_role[NEW] =
+			conn_highest_peer(connection);
+	}
+
+	for (n = 0; n < state_change->n_devices * state_change->n_connections; n++) {
+		struct drbd_peer_device_state_change *peer_device_state_change =
+			&state_change->peer_devices[n];
+		struct drbd_device *device =
+			peer_device_state_change->peer_device->device;
+		union drbd_dev_state state = device->state;
+
+		peer_device_state_change->disk_state[NEW] = state.pdsk;
+		peer_device_state_change->repl_state[NEW] =
+			max_t(enum drbd_conns, C_WF_REPORT_PARAMS, state.conn);
+		peer_device_state_change->resync_susp_user[NEW] =
+			state.user_isp;
+		peer_device_state_change->resync_susp_peer[NEW] =
+			state.peer_isp;
+		peer_device_state_change->resync_susp_dependency[NEW] =
+			state.aftr_isp;
+	}
+}
+
+void copy_old_to_new_state_change(struct drbd_state_change *state_change)
+{
+	struct drbd_resource_state_change *resource_state_change = &state_change->resource[0];
+	unsigned int n_device, n_connection, n_peer_device, n_peer_devices;
+
+#define OLD_TO_NEW(x) \
+	(x[NEW] = x[OLD])
+
+	OLD_TO_NEW(resource_state_change->role);
+	OLD_TO_NEW(resource_state_change->susp);
+	OLD_TO_NEW(resource_state_change->susp_nod);
+	OLD_TO_NEW(resource_state_change->susp_fen);
+
+	for (n_connection = 0; n_connection < state_change->n_connections; n_connection++) {
+		struct drbd_connection_state_change *connection_state_change =
+				&state_change->connections[n_connection];
+
+		OLD_TO_NEW(connection_state_change->peer_role);
+		OLD_TO_NEW(connection_state_change->cstate);
+	}
+
+	for (n_device = 0; n_device < state_change->n_devices; n_device++) {
+		struct drbd_device_state_change *device_state_change =
+			&state_change->devices[n_device];
+
+		OLD_TO_NEW(device_state_change->disk_state);
+	}
+
+	n_peer_devices = state_change->n_devices * state_change->n_connections;
+	for (n_peer_device = 0; n_peer_device < n_peer_devices; n_peer_device++) {
+		struct drbd_peer_device_state_change *p =
+			&state_change->peer_devices[n_peer_device];
+
+		OLD_TO_NEW(p->disk_state);
+		OLD_TO_NEW(p->repl_state);
+		OLD_TO_NEW(p->resync_susp_user);
+		OLD_TO_NEW(p->resync_susp_peer);
+		OLD_TO_NEW(p->resync_susp_dependency);
+	}
+
+#undef OLD_TO_NEW
+}
+
+void forget_state_change(struct drbd_state_change *state_change)
+{
+	unsigned int n;
+
+	if (!state_change)
+		return;
+
+	if (state_change->resource->resource)
+		kref_put(&state_change->resource->resource->kref, drbd_destroy_resource);
+	for (n = 0; n < state_change->n_devices; n++) {
+		struct drbd_device *device = state_change->devices[n].device;
+
+		if (device)
+			kref_put(&device->kref, drbd_destroy_device);
+	}
+	for (n = 0; n < state_change->n_connections; n++) {
+		struct drbd_connection *connection =
+			state_change->connections[n].connection;
+
+		if (connection)
+			kref_put(&connection->kref, drbd_destroy_connection);
+	}
+	kfree(state_change);
+}
+
 static int w_after_state_ch(struct drbd_work *w, int unused);
 static void after_state_ch(struct drbd_device *device, union drbd_state os,
-			   union drbd_state ns, enum chg_state_flags flags);
+			   union drbd_state ns, enum chg_state_flags flags,
+			   struct drbd_state_change *);
 static enum drbd_state_rv is_valid_state(struct drbd_device *, union drbd_state);
 static enum drbd_state_rv is_valid_soft_transition(union drbd_state, union drbd_state, struct drbd_connection *);
 static enum drbd_state_rv is_valid_transition(union drbd_state os, union drbd_state ns);
@@ -93,6 +334,7 @@ static enum drbd_role max_role(enum drbd_role role1, enum drbd_role role2)
 		return R_SECONDARY;
 	return R_UNKNOWN;
 }
+
 static enum drbd_role min_role(enum drbd_role role1, enum drbd_role role2)
 {
 	if (role1 == R_UNKNOWN || role2 == R_UNKNOWN)
@@ -937,7 +1179,7 @@ void drbd_resume_al(struct drbd_device *device)
 		drbd_info(device, "Resumed AL updates\n");
 }
 
-/* helper for __drbd_set_state */
+/* helper for _drbd_set_state */
 static void set_ov_position(struct drbd_device *device, enum drbd_conns cs)
 {
 	if (first_peer_device(device)->connection->agreed_pro_version < 90)
@@ -965,17 +1207,17 @@ static void set_ov_position(struct drbd_device *device, enum drbd_conns cs)
 }
 
 /**
- * __drbd_set_state() - Set a new DRBD state
+ * _drbd_set_state() - Set a new DRBD state
  * @device:	DRBD device.
  * @ns:		new state.
  * @flags:	Flags
  * @done:	Optional completion, that will get completed after the after_state_ch() finished
  *
- * Caller needs to hold req_lock, and global_state_lock. Do not call directly.
+ * Caller needs to hold req_lock. Do not call directly.
  */
 enum drbd_state_rv
-__drbd_set_state(struct drbd_device *device, union drbd_state ns,
-	         enum chg_state_flags flags, struct completion *done)
+_drbd_set_state(struct drbd_device *device, union drbd_state ns,
+	        enum chg_state_flags flags, struct completion *done)
 {
 	struct drbd_peer_device *peer_device = first_peer_device(device);
 	struct drbd_connection *connection = peer_device ? peer_device->connection : NULL;
@@ -983,6 +1225,7 @@ __drbd_set_state(struct drbd_device *device, union drbd_state ns,
 	enum drbd_state_rv rv = SS_SUCCESS;
 	enum sanitize_state_warnings ssw;
 	struct after_state_chg_work *ascw;
+	struct drbd_state_change *state_change;
 
 	os = drbd_read_state(device);
 
@@ -1037,6 +1280,9 @@ __drbd_set_state(struct drbd_device *device, union drbd_state ns,
 	if (!is_sync_state(os.conn) && is_sync_state(ns.conn))
 		clear_bit(RS_DONE, &device->flags);
 
+	/* FIXME: Have any flags been set earlier in this function already? */
+	state_change = remember_old_state(device->resource, GFP_ATOMIC);
+
 	/* changes to local_cnt and device flags should be visible before
 	 * changes to state, which again should be visible before anything else
 	 * depending on that change happens. */
@@ -1047,6 +1293,8 @@ __drbd_set_state(struct drbd_device *device, union drbd_state ns,
 	device->resource->susp_fen = ns.susp_fen;
 	smp_wmb();
 
+	remember_new_state(state_change);
+
 	/* put replicated vs not-replicated requests in seperate epochs */
 	if (drbd_should_do_remote((union drbd_dev_state)os.i) !=
 	    drbd_should_do_remote((union drbd_dev_state)ns.i))
@@ -1184,6 +1432,7 @@ __drbd_set_state(struct drbd_device *device, union drbd_state ns,
 		ascw->w.cb = w_after_state_ch;
 		ascw->device = device;
 		ascw->done = done;
+		ascw->state_change = state_change;
 		drbd_queue_work(&connection->sender_work,
 				&ascw->w);
 	} else {
@@ -1199,7 +1448,8 @@ static int w_after_state_ch(struct drbd_work *w, int unused)
 		container_of(w, struct after_state_chg_work, w);
 	struct drbd_device *device = ascw->device;
 
-	after_state_ch(device, ascw->os, ascw->ns, ascw->flags);
+	after_state_ch(device, ascw->os, ascw->ns, ascw->flags, ascw->state_change);
+	forget_state_change(ascw->state_change);
 	if (ascw->flags & CS_WAIT_COMPLETE)
 		complete(ascw->done);
 	kfree(ascw);
@@ -1234,7 +1484,7 @@ int drbd_bitmap_io_from_worker(struct drbd_device *device,
 	D_ASSERT(device, current == first_peer_device(device)->connection->worker.task);
 
 	/* open coded non-blocking drbd_suspend_io(device); */
-	set_bit(SUSPEND_IO, &device->flags);
+	atomic_inc(&device->suspend_cnt);
 
 	drbd_bm_lock(device, why, flags);
 	rv = io_fn(device);
@@ -1245,6 +1495,139 @@ int drbd_bitmap_io_from_worker(struct drbd_device *device,
 	return rv;
 }
 
+void notify_resource_state_change(struct sk_buff *skb,
+				  unsigned int seq,
+				  struct drbd_resource_state_change *resource_state_change,
+				  enum drbd_notification_type type)
+{
+	struct drbd_resource *resource = resource_state_change->resource;
+	struct resource_info resource_info = {
+		.res_role = resource_state_change->role[NEW],
+		.res_susp = resource_state_change->susp[NEW],
+		.res_susp_nod = resource_state_change->susp_nod[NEW],
+		.res_susp_fen = resource_state_change->susp_fen[NEW],
+	};
+
+	notify_resource_state(skb, seq, resource, &resource_info, type);
+}
+
+void notify_connection_state_change(struct sk_buff *skb,
+				    unsigned int seq,
+				    struct drbd_connection_state_change *connection_state_change,
+				    enum drbd_notification_type type)
+{
+	struct drbd_connection *connection = connection_state_change->connection;
+	struct connection_info connection_info = {
+		.conn_connection_state = connection_state_change->cstate[NEW],
+		.conn_role = connection_state_change->peer_role[NEW],
+	};
+
+	notify_connection_state(skb, seq, connection, &connection_info, type);
+}
+
+void notify_device_state_change(struct sk_buff *skb,
+				unsigned int seq,
+				struct drbd_device_state_change *device_state_change,
+				enum drbd_notification_type type)
+{
+	struct drbd_device *device = device_state_change->device;
+	struct device_info device_info = {
+		.dev_disk_state = device_state_change->disk_state[NEW],
+	};
+
+	notify_device_state(skb, seq, device, &device_info, type);
+}
+
+void notify_peer_device_state_change(struct sk_buff *skb,
+				     unsigned int seq,
+				     struct drbd_peer_device_state_change *p,
+				     enum drbd_notification_type type)
+{
+	struct drbd_peer_device *peer_device = p->peer_device;
+	struct peer_device_info peer_device_info = {
+		.peer_repl_state = p->repl_state[NEW],
+		.peer_disk_state = p->disk_state[NEW],
+		.peer_resync_susp_user = p->resync_susp_user[NEW],
+		.peer_resync_susp_peer = p->resync_susp_peer[NEW],
+		.peer_resync_susp_dependency = p->resync_susp_dependency[NEW],
+	};
+
+	notify_peer_device_state(skb, seq, peer_device, &peer_device_info, type);
+}
+
+static void broadcast_state_change(struct drbd_state_change *state_change)
+{
+	struct drbd_resource_state_change *resource_state_change = &state_change->resource[0];
+	bool resource_state_has_changed;
+	unsigned int n_device, n_connection, n_peer_device, n_peer_devices;
+	void (*last_func)(struct sk_buff *, unsigned int, void *,
+			  enum drbd_notification_type) = NULL;
+	void *uninitialized_var(last_arg);
+
+#define HAS_CHANGED(state) ((state)[OLD] != (state)[NEW])
+#define FINAL_STATE_CHANGE(type) \
+	({ if (last_func) \
+		last_func(NULL, 0, last_arg, type); \
+	})
+#define REMEMBER_STATE_CHANGE(func, arg, type) \
+	({ FINAL_STATE_CHANGE(type | NOTIFY_CONTINUES); \
+	   last_func = (typeof(last_func))func; \
+	   last_arg = arg; \
+	 })
+
+	mutex_lock(&notification_mutex);
+
+	resource_state_has_changed =
+	    HAS_CHANGED(resource_state_change->role) ||
+	    HAS_CHANGED(resource_state_change->susp) ||
+	    HAS_CHANGED(resource_state_change->susp_nod) ||
+	    HAS_CHANGED(resource_state_change->susp_fen);
+
+	if (resource_state_has_changed)
+		REMEMBER_STATE_CHANGE(notify_resource_state_change,
+				      resource_state_change, NOTIFY_CHANGE);
+
+	for (n_connection = 0; n_connection < state_change->n_connections; n_connection++) {
+		struct drbd_connection_state_change *connection_state_change =
+				&state_change->connections[n_connection];
+
+		if (HAS_CHANGED(connection_state_change->peer_role) ||
+		    HAS_CHANGED(connection_state_change->cstate))
+			REMEMBER_STATE_CHANGE(notify_connection_state_change,
+					      connection_state_change, NOTIFY_CHANGE);
+	}
+
+	for (n_device = 0; n_device < state_change->n_devices; n_device++) {
+		struct drbd_device_state_change *device_state_change =
+			&state_change->devices[n_device];
+
+		if (HAS_CHANGED(device_state_change->disk_state))
+			REMEMBER_STATE_CHANGE(notify_device_state_change,
+					      device_state_change, NOTIFY_CHANGE);
+	}
+
+	n_peer_devices = state_change->n_devices * state_change->n_connections;
+	for (n_peer_device = 0; n_peer_device < n_peer_devices; n_peer_device++) {
+		struct drbd_peer_device_state_change *p =
+			&state_change->peer_devices[n_peer_device];
+
+		if (HAS_CHANGED(p->disk_state) ||
+		    HAS_CHANGED(p->repl_state) ||
+		    HAS_CHANGED(p->resync_susp_user) ||
+		    HAS_CHANGED(p->resync_susp_peer) ||
+		    HAS_CHANGED(p->resync_susp_dependency))
+			REMEMBER_STATE_CHANGE(notify_peer_device_state_change,
+					      p, NOTIFY_CHANGE);
+	}
+
+	FINAL_STATE_CHANGE(NOTIFY_CHANGE);
+	mutex_unlock(&notification_mutex);
+
+#undef HAS_CHANGED
+#undef FINAL_STATE_CHANGE
+#undef REMEMBER_STATE_CHANGE
+}
+
 /**
  * after_state_ch() - Perform after state change actions that may sleep
  * @device:	DRBD device.
@@ -1253,13 +1636,16 @@ int drbd_bitmap_io_from_worker(struct drbd_device *device,
  * @flags:	Flags
  */
 static void after_state_ch(struct drbd_device *device, union drbd_state os,
-			   union drbd_state ns, enum chg_state_flags flags)
+			   union drbd_state ns, enum chg_state_flags flags,
+			   struct drbd_state_change *state_change)
 {
 	struct drbd_resource *resource = device->resource;
 	struct drbd_peer_device *peer_device = first_peer_device(device);
 	struct drbd_connection *connection = peer_device ? peer_device->connection : NULL;
 	struct sib_info sib;
 
+	broadcast_state_change(state_change);
+
 	sib.sib_reason = SIB_STATE_CHANGE;
 	sib.os = os;
 	sib.ns = ns;
@@ -1377,7 +1763,7 @@ static void after_state_ch(struct drbd_device *device, union drbd_state os,
 	}
 
 	if (ns.pdsk < D_INCONSISTENT && get_ldev(device)) {
-		if (os.peer == R_SECONDARY && ns.peer == R_PRIMARY &&
+		if (os.peer != R_PRIMARY && ns.peer == R_PRIMARY &&
 		    device->ldev->md.uuid[UI_BITMAP] == 0 && ns.disk >= D_UP_TO_DATE) {
 			drbd_uuid_new_current(device);
 			drbd_send_uuids(peer_device);
@@ -1444,7 +1830,7 @@ static void after_state_ch(struct drbd_device *device, union drbd_state os,
 	if (os.disk != D_FAILED && ns.disk == D_FAILED) {
 		enum drbd_io_error_p eh = EP_PASS_ON;
 		int was_io_error = 0;
-		/* corresponding get_ldev was in __drbd_set_state, to serialize
+		/* corresponding get_ldev was in _drbd_set_state, to serialize
 		 * our cleanup here with the transition to D_DISKLESS.
 		 * But is is still not save to dreference ldev here, since
 		 * we might come from an failed Attach before ldev was set. */
@@ -1455,6 +1841,10 @@ static void after_state_ch(struct drbd_device *device, union drbd_state os,
 
 			was_io_error = test_and_clear_bit(WAS_IO_ERROR, &device->flags);
 
+			/* Intentionally call this handler first, before drbd_send_state().
+			 * See: 2932204 drbd: call local-io-error handler early
+			 * People may chose to hard-reset the box from this handler.
+			 * It is useful if this looks like a "regular node crash". */
 			if (was_io_error && eh == EP_CALL_HELPER)
 				drbd_khelper(device, "local-io-error");
 
@@ -1572,6 +1962,7 @@ struct after_conn_state_chg_work {
 	union drbd_state ns_max; /* new, max state, over all devices */
 	enum chg_state_flags flags;
 	struct drbd_connection *connection;
+	struct drbd_state_change *state_change;
 };
 
 static int w_after_conn_state_ch(struct drbd_work *w, int unused)
@@ -1584,6 +1975,8 @@ static int w_after_conn_state_ch(struct drbd_work *w, int unused)
 	struct drbd_peer_device *peer_device;
 	int vnr;
 
+	broadcast_state_change(acscw->state_change);
+	forget_state_change(acscw->state_change);
 	kfree(acscw);
 
 	/* Upon network configuration, we need to start the receiver */
@@ -1593,6 +1986,13 @@ static int w_after_conn_state_ch(struct drbd_work *w, int unused)
 	if (oc == C_DISCONNECTING && ns_max.conn == C_STANDALONE) {
 		struct net_conf *old_conf;
 
+		mutex_lock(&notification_mutex);
+		idr_for_each_entry(&connection->peer_devices, peer_device, vnr)
+			notify_peer_device_state(NULL, 0, peer_device, NULL,
+						 NOTIFY_DESTROY | NOTIFY_CONTINUES);
+		notify_connection_state(NULL, 0, connection, NULL, NOTIFY_DESTROY);
+		mutex_unlock(&notification_mutex);
+
 		mutex_lock(&connection->resource->conf_update);
 		old_conf = connection->net_conf;
 		connection->my_addr_len = 0;
@@ -1759,7 +2159,7 @@ conn_set_state(struct drbd_connection *connection, union drbd_state mask, union
 		if (flags & CS_IGN_OUTD_FAIL && ns.disk == D_OUTDATED && os.disk < D_OUTDATED)
 			ns.disk = os.disk;
 
-		rv = __drbd_set_state(device, ns, flags, NULL);
+		rv = _drbd_set_state(device, ns, flags, NULL);
 		if (rv < SS_SUCCESS)
 			BUG();
 
@@ -1823,6 +2223,7 @@ _conn_request_state(struct drbd_connection *connection, union drbd_state mask, u
 	enum drbd_conns oc = connection->cstate;
 	union drbd_state ns_max, ns_min, os;
 	bool have_mutex = false;
+	struct drbd_state_change *state_change;
 
 	if (mask.conn) {
 		rv = is_valid_conn_transition(oc, val.conn);
@@ -1868,10 +2269,12 @@ _conn_request_state(struct drbd_connection *connection, union drbd_state mask, u
 			goto abort;
 	}
 
+	state_change = remember_old_state(connection->resource, GFP_ATOMIC);
 	conn_old_common_state(connection, &os, &flags);
 	flags |= CS_DC_SUSP;
 	conn_set_state(connection, mask, val, &ns_min, &ns_max, flags);
 	conn_pr_state_change(connection, os, ns_max, flags);
+	remember_new_state(state_change);
 
 	acscw = kmalloc(sizeof(*acscw), GFP_ATOMIC);
 	if (acscw) {
@@ -1882,6 +2285,7 @@ _conn_request_state(struct drbd_connection *connection, union drbd_state mask, u
 		acscw->w.cb = w_after_conn_state_ch;
 		kref_get(&connection->kref);
 		acscw->connection = connection;
+		acscw->state_change = state_change;
 		drbd_queue_work(&connection->sender_work, &acscw->w);
 	} else {
 		drbd_err(connection, "Could not kmalloc an acscw\n");

drivers/block/drbd/drbd_state.h

@@ -122,9 +122,9 @@ extern enum drbd_state_rv
 _drbd_request_state_holding_state_mutex(struct drbd_device *, union drbd_state,
 					union drbd_state, enum chg_state_flags);
 
-extern enum drbd_state_rv __drbd_set_state(struct drbd_device *, union drbd_state,
-					   enum chg_state_flags,
-					   struct completion *done);
+extern enum drbd_state_rv _drbd_set_state(struct drbd_device *, union drbd_state,
+					  enum chg_state_flags,
+					  struct completion *done);
 extern void print_st_err(struct drbd_device *, union drbd_state,
 			union drbd_state, int);
 

drivers/block/drbd/drbd_state_change.h

+#ifndef DRBD_STATE_CHANGE_H
+#define DRBD_STATE_CHANGE_H
+
+struct drbd_resource_state_change {
+	struct drbd_resource *resource;
+	enum drbd_role role[2];
+	bool susp[2];
+	bool susp_nod[2];
+	bool susp_fen[2];
+};
+
+struct drbd_device_state_change {
+	struct drbd_device *device;
+	enum drbd_disk_state disk_state[2];
+};
+
+struct drbd_connection_state_change {
+	struct drbd_connection *connection;
+	enum drbd_conns cstate[2];  /* drbd9: enum drbd_conn_state */
+	enum drbd_role peer_role[2];
+};
+
+struct drbd_peer_device_state_change {
+	struct drbd_peer_device *peer_device;
+	enum drbd_disk_state disk_state[2];
+	enum drbd_conns repl_state[2];  /* drbd9: enum drbd_repl_state */
+	bool resync_susp_user[2];
+	bool resync_susp_peer[2];
+	bool resync_susp_dependency[2];
+};
+
+struct drbd_state_change {
+	struct list_head list;
+	unsigned int n_devices;
+	unsigned int n_connections;
+	struct drbd_resource_state_change resource[1];
+	struct drbd_device_state_change *devices;
+	struct drbd_connection_state_change *connections;
+	struct drbd_peer_device_state_change *peer_devices;
+};
+
+extern struct drbd_state_change *remember_old_state(struct drbd_resource *, gfp_t);
+extern void copy_old_to_new_state_change(struct drbd_state_change *);
+extern void forget_state_change(struct drbd_state_change *);
+
+extern void notify_resource_state_change(struct sk_buff *,
+					 unsigned int,
+					 struct drbd_resource_state_change *,
+					 enum drbd_notification_type type);
+extern void notify_connection_state_change(struct sk_buff *,
+					   unsigned int,
+					   struct drbd_connection_state_change *,
+					   enum drbd_notification_type type);
+extern void notify_device_state_change(struct sk_buff *,
+				       unsigned int,
+				       struct drbd_device_state_change *,
+				       enum drbd_notification_type type);
+extern void notify_peer_device_state_change(struct sk_buff *,
+					    unsigned int,
+					    struct drbd_peer_device_state_change *,
+					    enum drbd_notification_type type);
+
+#endif  /* DRBD_STATE_CHANGE_H */

drivers/block/drbd/drbd_worker.c

@@ -55,13 +55,6 @@ static int make_resync_request(struct drbd_device *, int);
  *
  */
 
-
-/* About the global_state_lock
-   Each state transition on an device holds a read lock. In case we have
-   to evaluate the resync after dependencies, we grab a write lock, because
-   we need stable states on all devices for that.  */
-rwlock_t global_state_lock;
-
 /* used for synchronous meta data and bitmap IO
  * submitted by drbd_md_sync_page_io()
  */
@@ -120,6 +113,7 @@ void drbd_endio_write_sec_final(struct drbd_peer_request *peer_req) __releases(l
 	unsigned long flags = 0;
 	struct drbd_peer_device *peer_device = peer_req->peer_device;
 	struct drbd_device *device = peer_device->device;
+	struct drbd_connection *connection = peer_device->connection;
 	struct drbd_interval i;
 	int do_wake;
 	u64 block_id;
@@ -152,6 +146,12 @@ void drbd_endio_write_sec_final(struct drbd_peer_request *peer_req) __releases(l
 	 * ((peer_req->flags & (EE_WAS_ERROR|EE_IS_TRIM)) == EE_WAS_ERROR) */
 	if (peer_req->flags & EE_WAS_ERROR)
 		__drbd_chk_io_error(device, DRBD_WRITE_ERROR);
+
+	if (connection->cstate >= C_WF_REPORT_PARAMS) {
+		kref_get(&device->kref); /* put is in drbd_send_acks_wf() */
+		if (!queue_work(connection->ack_sender, &peer_device->send_acks_work))
+			kref_put(&device->kref, drbd_destroy_device);
+	}
 	spin_unlock_irqrestore(&device->resource->req_lock, flags);
 
 	if (block_id == ID_SYNCER)
@@ -163,7 +163,6 @@ void drbd_endio_write_sec_final(struct drbd_peer_request *peer_req) __releases(l
 	if (do_al_complete_io)
 		drbd_al_complete_io(device, &i);
 
-	wake_asender(peer_device->connection);
 	put_ldev(device);
 }
 
@@ -195,6 +194,12 @@ void drbd_peer_request_endio(struct bio *bio)
 	}
 }
 
+void drbd_panic_after_delayed_completion_of_aborted_request(struct drbd_device *device)
+{
+	panic("drbd%u %s/%u potential random memory corruption caused by delayed completion of aborted local request\n",
+		device->minor, device->resource->name, device->vnr);
+}
+
 /* read, readA or write requests on R_PRIMARY coming from drbd_make_request
  */
 void drbd_request_endio(struct bio *bio)
@@ -238,7 +243,7 @@ void drbd_request_endio(struct bio *bio)
 			drbd_emerg(device, "delayed completion of aborted local request; disk-timeout may be too aggressive\n");
 
 		if (!bio->bi_error)
-			panic("possible random memory corruption caused by delayed completion of aborted local request\n");
+			drbd_panic_after_delayed_completion_of_aborted_request(device);
 	}
 
 	/* to avoid recursion in __req_mod */
@@ -1291,6 +1296,7 @@ static int drbd_send_barrier(struct drbd_connection *connection)
 	p->barrier = connection->send.current_epoch_nr;
 	p->pad = 0;
 	connection->send.current_epoch_writes = 0;
+	connection->send.last_sent_barrier_jif = jiffies;
 
 	return conn_send_command(connection, sock, P_BARRIER, sizeof(*p), NULL, 0);
 }
@@ -1315,6 +1321,7 @@ static void re_init_if_first_write(struct drbd_connection *connection, unsigned
 		connection->send.seen_any_write_yet = true;
 		connection->send.current_epoch_nr = epoch;
 		connection->send.current_epoch_writes = 0;
+		connection->send.last_sent_barrier_jif = jiffies;
 	}
 }
 
@@ -1456,70 +1463,73 @@ static int _drbd_may_sync_now(struct drbd_device *device)
 }
 
 /**
- * _drbd_pause_after() - Pause resync on all devices that may not resync now
+ * drbd_pause_after() - Pause resync on all devices that may not resync now
  * @device:	DRBD device.
  *
  * Called from process context only (admin command and after_state_ch).
  */
-static int _drbd_pause_after(struct drbd_device *device)
+static bool drbd_pause_after(struct drbd_device *device)
 {
+	bool changed = false;
 	struct drbd_device *odev;
-	int i, rv = 0;
+	int i;
 
 	rcu_read_lock();
 	idr_for_each_entry(&drbd_devices, odev, i) {
 		if (odev->state.conn == C_STANDALONE && odev->state.disk == D_DISKLESS)
 			continue;
-		if (!_drbd_may_sync_now(odev))
-			rv |= (__drbd_set_state(_NS(odev, aftr_isp, 1), CS_HARD, NULL)
-			       != SS_NOTHING_TO_DO);
+		if (!_drbd_may_sync_now(odev) &&
+		    _drbd_set_state(_NS(odev, aftr_isp, 1),
+				    CS_HARD, NULL) != SS_NOTHING_TO_DO)
+			changed = true;
 	}
 	rcu_read_unlock();
 
-	return rv;
+	return changed;
 }
 
 /**
- * _drbd_resume_next() - Resume resync on all devices that may resync now
+ * drbd_resume_next() - Resume resync on all devices that may resync now
  * @device:	DRBD device.
  *
  * Called from process context only (admin command and worker).
  */
-static int _drbd_resume_next(struct drbd_device *device)
+static bool drbd_resume_next(struct drbd_device *device)
 {
+	bool changed = false;
 	struct drbd_device *odev;
-	int i, rv = 0;
+	int i;
 
 	rcu_read_lock();
 	idr_for_each_entry(&drbd_devices, odev, i) {
 		if (odev->state.conn == C_STANDALONE && odev->state.disk == D_DISKLESS)
 			continue;
 		if (odev->state.aftr_isp) {
-			if (_drbd_may_sync_now(odev))
-				rv |= (__drbd_set_state(_NS(odev, aftr_isp, 0),
-							CS_HARD, NULL)
-				       != SS_NOTHING_TO_DO) ;
+			if (_drbd_may_sync_now(odev) &&
+			    _drbd_set_state(_NS(odev, aftr_isp, 0),
+					    CS_HARD, NULL) != SS_NOTHING_TO_DO)
+				changed = true;
 		}
 	}
 	rcu_read_unlock();
-	return rv;
+	return changed;
 }
 
 void resume_next_sg(struct drbd_device *device)
 {
-	write_lock_irq(&global_state_lock);
-	_drbd_resume_next(device);
-	write_unlock_irq(&global_state_lock);
+	lock_all_resources();
+	drbd_resume_next(device);
+	unlock_all_resources();
 }
 
 void suspend_other_sg(struct drbd_device *device)
 {
-	write_lock_irq(&global_state_lock);
-	_drbd_pause_after(device);
-	write_unlock_irq(&global_state_lock);
+	lock_all_resources();
+	drbd_pause_after(device);
+	unlock_all_resources();
 }
 
-/* caller must hold global_state_lock */
+/* caller must lock_all_resources() */
 enum drbd_ret_code drbd_resync_after_valid(struct drbd_device *device, int o_minor)
 {
 	struct drbd_device *odev;
@@ -1557,15 +1567,15 @@ enum drbd_ret_code drbd_resync_after_valid(struct drbd_device *device, int o_min
 	}
 }
 
-/* caller must hold global_state_lock */
+/* caller must lock_all_resources() */
 void drbd_resync_after_changed(struct drbd_device *device)
 {
-	int changes;
+	int changed;
 
 	do {
-		changes  = _drbd_pause_after(device);
-		changes |= _drbd_resume_next(device);
-	} while (changes);
+		changed  = drbd_pause_after(device);
+		changed |= drbd_resume_next(device);
+	} while (changed);
 }
 
 void drbd_rs_controller_reset(struct drbd_device *device)
@@ -1685,19 +1695,14 @@ void drbd_start_resync(struct drbd_device *device, enum drbd_conns side)
 	} else {
 		mutex_lock(device->state_mutex);
 	}
-	clear_bit(B_RS_H_DONE, &device->flags);
 
-	/* req_lock: serialize with drbd_send_and_submit() and others
-	 * global_state_lock: for stable sync-after dependencies */
-	spin_lock_irq(&device->resource->req_lock);
-	write_lock(&global_state_lock);
+	lock_all_resources();
+	clear_bit(B_RS_H_DONE, &device->flags);
 	/* Did some connection breakage or IO error race with us? */
 	if (device->state.conn < C_CONNECTED
 	|| !get_ldev_if_state(device, D_NEGOTIATING)) {
-		write_unlock(&global_state_lock);
-		spin_unlock_irq(&device->resource->req_lock);
-		mutex_unlock(device->state_mutex);
-		return;
+		unlock_all_resources();
+		goto out;
 	}
 
 	ns = drbd_read_state(device);
@@ -1711,7 +1716,7 @@ void drbd_start_resync(struct drbd_device *device, enum drbd_conns side)
 	else /* side == C_SYNC_SOURCE */
 		ns.pdsk = D_INCONSISTENT;
 
-	r = __drbd_set_state(device, ns, CS_VERBOSE, NULL);
+	r = _drbd_set_state(device, ns, CS_VERBOSE, NULL);
 	ns = drbd_read_state(device);
 
 	if (ns.conn < C_CONNECTED)
@@ -1732,7 +1737,7 @@ void drbd_start_resync(struct drbd_device *device, enum drbd_conns side)
 			device->rs_mark_left[i] = tw;
 			device->rs_mark_time[i] = now;
 		}
-		_drbd_pause_after(device);
+		drbd_pause_after(device);
 		/* Forget potentially stale cached per resync extent bit-counts.
 		 * Open coded drbd_rs_cancel_all(device), we already have IRQs
 		 * disabled, and know the disk state is ok. */
@@ -1742,8 +1747,7 @@ void drbd_start_resync(struct drbd_device *device, enum drbd_conns side)
 		device->resync_wenr = LC_FREE;
 		spin_unlock(&device->al_lock);
 	}
-	write_unlock(&global_state_lock);
-	spin_unlock_irq(&device->resource->req_lock);
+	unlock_all_resources();
 
 	if (r == SS_SUCCESS) {
 		wake_up(&device->al_wait); /* for lc_reset() above */
@@ -1807,6 +1811,7 @@ void drbd_start_resync(struct drbd_device *device, enum drbd_conns side)
 		drbd_md_sync(device);
 	}
 	put_ldev(device);
+out:
 	mutex_unlock(device->state_mutex);
 }
 
@@ -1836,7 +1841,7 @@ static void drbd_ldev_destroy(struct drbd_device *device)
 	device->act_log = NULL;
 
 	__acquire(local);
-	drbd_free_ldev(device->ldev);
+	drbd_backing_dev_free(device, device->ldev);
 	device->ldev = NULL;
 	__release(local);
 

drivers/block/mtip32xx/mtip32xx.c

@@ -104,9 +104,9 @@
 /* Device instance number, incremented each time a device is probed. */
 static int instance;
 
-struct list_head online_list;
-struct list_head removing_list;
-spinlock_t dev_lock;
+static struct list_head online_list;
+static struct list_head removing_list;
+static spinlock_t dev_lock;
 
 /*
  * Global variable used to hold the major block device number

drivers/block/null_blk.c

@@ -495,17 +495,17 @@ static int null_lnvm_id(struct nvm_dev *dev, struct nvm_id *id)
 	id->ppaf.ch_offset = 56;
 	id->ppaf.ch_len = 8;
 
-	do_div(size, bs); /* convert size to pages */
-	do_div(size, 256); /* concert size to pgs pr blk */
+	sector_div(size, bs); /* convert size to pages */
+	size >>= 8; /* concert size to pgs pr blk */
 	grp = &id->groups[0];
 	grp->mtype = 0;
 	grp->fmtype = 0;
 	grp->num_ch = 1;
 	grp->num_pg = 256;
 	blksize = size;
-	do_div(size, (1 << 16));
+	size >>= 16;
 	grp->num_lun = size + 1;
-	do_div(blksize, grp->num_lun);
+	sector_div(blksize, grp->num_lun);
 	grp->num_blk = blksize;
 	grp->num_pln = 1;
 

drivers/block/sx8.c

@@ -23,7 +23,7 @@
 #include <linux/workqueue.h>
 #include <linux/bitops.h>
 #include <linux/delay.h>
-#include <linux/time.h>
+#include <linux/ktime.h>
 #include <linux/hdreg.h>
 #include <linux/dma-mapping.h>
 #include <linux/completion.h>
@@ -671,16 +671,15 @@ static int carm_send_special (struct carm_host *host, carm_sspc_t func)
 static unsigned int carm_fill_sync_time(struct carm_host *host,
 					unsigned int idx, void *mem)
 {
-	struct timeval tv;
 	struct carm_msg_sync_time *st = mem;
 
-	do_gettimeofday(&tv);
+	time64_t tv = ktime_get_real_seconds();
 
 	memset(st, 0, sizeof(*st));
 	st->type	= CARM_MSG_MISC;
 	st->subtype	= MISC_SET_TIME;
 	st->handle	= cpu_to_le32(TAG_ENCODE(idx));
-	st->timestamp	= cpu_to_le32(tv.tv_sec);
+	st->timestamp	= cpu_to_le32(tv);
 
 	return sizeof(struct carm_msg_sync_time);
 }

drivers/block/xen-blkback/blkback.c

@@ -83,6 +83,16 @@ module_param_named(max_persistent_grants, xen_blkif_max_pgrants, int, 0644);
 MODULE_PARM_DESC(max_persistent_grants,
                  "Maximum number of grants to map persistently");
 
+/*
+ * Maximum number of rings/queues blkback supports, allow as many queues as there
+ * are CPUs if user has not specified a value.
+ */
+unsigned int xenblk_max_queues;
+module_param_named(max_queues, xenblk_max_queues, uint, 0644);
+MODULE_PARM_DESC(max_queues,
+		 "Maximum number of hardware queues per virtual disk." \
+		 "By default it is the number of online CPUs.");
+
 /*
  * Maximum order of pages to be used for the shared ring between front and
  * backend, 4KB page granularity is used.
@@ -113,71 +123,71 @@ module_param(log_stats, int, 0644);
 /* Number of free pages to remove on each call to gnttab_free_pages */
 #define NUM_BATCH_FREE_PAGES 10
 
-static inline int get_free_page(struct xen_blkif *blkif, struct page **page)
+static inline int get_free_page(struct xen_blkif_ring *ring, struct page **page)
 {
 	unsigned long flags;
 
-	spin_lock_irqsave(&blkif->free_pages_lock, flags);
-	if (list_empty(&blkif->free_pages)) {
-		BUG_ON(blkif->free_pages_num != 0);
-		spin_unlock_irqrestore(&blkif->free_pages_lock, flags);
+	spin_lock_irqsave(&ring->free_pages_lock, flags);
+	if (list_empty(&ring->free_pages)) {
+		BUG_ON(ring->free_pages_num != 0);
+		spin_unlock_irqrestore(&ring->free_pages_lock, flags);
 		return gnttab_alloc_pages(1, page);
 	}
-	BUG_ON(blkif->free_pages_num == 0);
-	page[0] = list_first_entry(&blkif->free_pages, struct page, lru);
+	BUG_ON(ring->free_pages_num == 0);
+	page[0] = list_first_entry(&ring->free_pages, struct page, lru);
 	list_del(&page[0]->lru);
-	blkif->free_pages_num--;
-	spin_unlock_irqrestore(&blkif->free_pages_lock, flags);
+	ring->free_pages_num--;
+	spin_unlock_irqrestore(&ring->free_pages_lock, flags);
 
 	return 0;
 }
 
-static inline void put_free_pages(struct xen_blkif *blkif, struct page **page,
+static inline void put_free_pages(struct xen_blkif_ring *ring, struct page **page,
                                   int num)
 {
 	unsigned long flags;
 	int i;
 
-	spin_lock_irqsave(&blkif->free_pages_lock, flags);
+	spin_lock_irqsave(&ring->free_pages_lock, flags);
 	for (i = 0; i < num; i++)
-		list_add(&page[i]->lru, &blkif->free_pages);
-	blkif->free_pages_num += num;
-	spin_unlock_irqrestore(&blkif->free_pages_lock, flags);
+		list_add(&page[i]->lru, &ring->free_pages);
+	ring->free_pages_num += num;
+	spin_unlock_irqrestore(&ring->free_pages_lock, flags);
 }
 
-static inline void shrink_free_pagepool(struct xen_blkif *blkif, int num)
+static inline void shrink_free_pagepool(struct xen_blkif_ring *ring, int num)
 {
 	/* Remove requested pages in batches of NUM_BATCH_FREE_PAGES */
 	struct page *page[NUM_BATCH_FREE_PAGES];
 	unsigned int num_pages = 0;
 	unsigned long flags;
 
-	spin_lock_irqsave(&blkif->free_pages_lock, flags);
-	while (blkif->free_pages_num > num) {
-		BUG_ON(list_empty(&blkif->free_pages));
-		page[num_pages] = list_first_entry(&blkif->free_pages,
+	spin_lock_irqsave(&ring->free_pages_lock, flags);
+	while (ring->free_pages_num > num) {
+		BUG_ON(list_empty(&ring->free_pages));
+		page[num_pages] = list_first_entry(&ring->free_pages,
 		                                   struct page, lru);
 		list_del(&page[num_pages]->lru);
-		blkif->free_pages_num--;
+		ring->free_pages_num--;
 		if (++num_pages == NUM_BATCH_FREE_PAGES) {
-			spin_unlock_irqrestore(&blkif->free_pages_lock, flags);
+			spin_unlock_irqrestore(&ring->free_pages_lock, flags);
 			gnttab_free_pages(num_pages, page);
-			spin_lock_irqsave(&blkif->free_pages_lock, flags);
+			spin_lock_irqsave(&ring->free_pages_lock, flags);
 			num_pages = 0;
 		}
 	}
-	spin_unlock_irqrestore(&blkif->free_pages_lock, flags);
+	spin_unlock_irqrestore(&ring->free_pages_lock, flags);
 	if (num_pages != 0)
 		gnttab_free_pages(num_pages, page);
 }
 
 #define vaddr(page) ((unsigned long)pfn_to_kaddr(page_to_pfn(page)))
 
-static int do_block_io_op(struct xen_blkif *blkif);
-static int dispatch_rw_block_io(struct xen_blkif *blkif,
+static int do_block_io_op(struct xen_blkif_ring *ring);
+static int dispatch_rw_block_io(struct xen_blkif_ring *ring,
 				struct blkif_request *req,
 				struct pending_req *pending_req);
-static void make_response(struct xen_blkif *blkif, u64 id,
+static void make_response(struct xen_blkif_ring *ring, u64 id,
 			  unsigned short op, int st);
 
 #define foreach_grant_safe(pos, n, rbtree, node) \
@@ -190,7 +200,7 @@ static void make_response(struct xen_blkif *blkif, u64 id,
 
 /*
  * We don't need locking around the persistent grant helpers
- * because blkback uses a single-thread for each backed, so we
+ * because blkback uses a single-thread for each backend, so we
  * can be sure that this functions will never be called recursively.
  *
  * The only exception to that is put_persistent_grant, that can be called
@@ -198,19 +208,20 @@ static void make_response(struct xen_blkif *blkif, u64 id,
  * bit operations to modify the flags of a persistent grant and to count
  * the number of used grants.
  */
-static int add_persistent_gnt(struct xen_blkif *blkif,
+static int add_persistent_gnt(struct xen_blkif_ring *ring,
 			       struct persistent_gnt *persistent_gnt)
 {
 	struct rb_node **new = NULL, *parent = NULL;
 	struct persistent_gnt *this;
+	struct xen_blkif *blkif = ring->blkif;
 
-	if (blkif->persistent_gnt_c >= xen_blkif_max_pgrants) {
+	if (ring->persistent_gnt_c >= xen_blkif_max_pgrants) {
 		if (!blkif->vbd.overflow_max_grants)
 			blkif->vbd.overflow_max_grants = 1;
 		return -EBUSY;
 	}
 	/* Figure out where to put new node */
-	new = &blkif->persistent_gnts.rb_node;
+	new = &ring->persistent_gnts.rb_node;
 	while (*new) {
 		this = container_of(*new, struct persistent_gnt, node);
 
@@ -229,19 +240,19 @@ static int add_persistent_gnt(struct xen_blkif *blkif,
 	set_bit(PERSISTENT_GNT_ACTIVE, persistent_gnt->flags);
 	/* Add new node and rebalance tree. */
 	rb_link_node(&(persistent_gnt->node), parent, new);
-	rb_insert_color(&(persistent_gnt->node), &blkif->persistent_gnts);
-	blkif->persistent_gnt_c++;
-	atomic_inc(&blkif->persistent_gnt_in_use);
+	rb_insert_color(&(persistent_gnt->node), &ring->persistent_gnts);
+	ring->persistent_gnt_c++;
+	atomic_inc(&ring->persistent_gnt_in_use);
 	return 0;
 }
 
-static struct persistent_gnt *get_persistent_gnt(struct xen_blkif *blkif,
+static struct persistent_gnt *get_persistent_gnt(struct xen_blkif_ring *ring,
 						 grant_ref_t gref)
 {
 	struct persistent_gnt *data;
 	struct rb_node *node = NULL;
 
-	node = blkif->persistent_gnts.rb_node;
+	node = ring->persistent_gnts.rb_node;
 	while (node) {
 		data = container_of(node, struct persistent_gnt, node);
 
@@ -255,24 +266,24 @@ static struct persistent_gnt *get_persistent_gnt(struct xen_blkif *blkif,
 				return NULL;
 			}
 			set_bit(PERSISTENT_GNT_ACTIVE, data->flags);
-			atomic_inc(&blkif->persistent_gnt_in_use);
+			atomic_inc(&ring->persistent_gnt_in_use);
 			return data;
 		}
 	}
 	return NULL;
 }
 
-static void put_persistent_gnt(struct xen_blkif *blkif,
+static void put_persistent_gnt(struct xen_blkif_ring *ring,
                                struct persistent_gnt *persistent_gnt)
 {
 	if(!test_bit(PERSISTENT_GNT_ACTIVE, persistent_gnt->flags))
 		pr_alert_ratelimited("freeing a grant already unused\n");
 	set_bit(PERSISTENT_GNT_WAS_ACTIVE, persistent_gnt->flags);
 	clear_bit(PERSISTENT_GNT_ACTIVE, persistent_gnt->flags);
-	atomic_dec(&blkif->persistent_gnt_in_use);
+	atomic_dec(&ring->persistent_gnt_in_use);
 }
 
-static void free_persistent_gnts(struct xen_blkif *blkif, struct rb_root *root,
+static void free_persistent_gnts(struct xen_blkif_ring *ring, struct rb_root *root,
                                  unsigned int num)
 {
 	struct gnttab_unmap_grant_ref unmap[BLKIF_MAX_SEGMENTS_PER_REQUEST];
@@ -303,7 +314,7 @@ static void free_persistent_gnts(struct xen_blkif *blkif, struct rb_root *root,
 			unmap_data.count = segs_to_unmap;
 			BUG_ON(gnttab_unmap_refs_sync(&unmap_data));
 
-			put_free_pages(blkif, pages, segs_to_unmap);
+			put_free_pages(ring, pages, segs_to_unmap);
 			segs_to_unmap = 0;
 		}
 
@@ -320,15 +331,15 @@ void xen_blkbk_unmap_purged_grants(struct work_struct *work)
 	struct page *pages[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 	struct persistent_gnt *persistent_gnt;
 	int segs_to_unmap = 0;
-	struct xen_blkif *blkif = container_of(work, typeof(*blkif), persistent_purge_work);
+	struct xen_blkif_ring *ring = container_of(work, typeof(*ring), persistent_purge_work);
 	struct gntab_unmap_queue_data unmap_data;
 
 	unmap_data.pages = pages;
 	unmap_data.unmap_ops = unmap;
 	unmap_data.kunmap_ops = NULL;
 
-	while(!list_empty(&blkif->persistent_purge_list)) {
-		persistent_gnt = list_first_entry(&blkif->persistent_purge_list,
+	while(!list_empty(&ring->persistent_purge_list)) {
+		persistent_gnt = list_first_entry(&ring->persistent_purge_list,
 		                                  struct persistent_gnt,
 		                                  remove_node);
 		list_del(&persistent_gnt->remove_node);
@@ -343,7 +354,7 @@ void xen_blkbk_unmap_purged_grants(struct work_struct *work)
 		if (++segs_to_unmap == BLKIF_MAX_SEGMENTS_PER_REQUEST) {
 			unmap_data.count = segs_to_unmap;
 			BUG_ON(gnttab_unmap_refs_sync(&unmap_data));
-			put_free_pages(blkif, pages, segs_to_unmap);
+			put_free_pages(ring, pages, segs_to_unmap);
 			segs_to_unmap = 0;
 		}
 		kfree(persistent_gnt);
@@ -351,11 +362,11 @@ void xen_blkbk_unmap_purged_grants(struct work_struct *work)
 	if (segs_to_unmap > 0) {
 		unmap_data.count = segs_to_unmap;
 		BUG_ON(gnttab_unmap_refs_sync(&unmap_data));
-		put_free_pages(blkif, pages, segs_to_unmap);
+		put_free_pages(ring, pages, segs_to_unmap);
 	}
 }
 
-static void purge_persistent_gnt(struct xen_blkif *blkif)
+static void purge_persistent_gnt(struct xen_blkif_ring *ring)
 {
 	struct persistent_gnt *persistent_gnt;
 	struct rb_node *n;
@@ -363,23 +374,23 @@ static void purge_persistent_gnt(struct xen_blkif *blkif)
 	bool scan_used = false, clean_used = false;
 	struct rb_root *root;
 
-	if (blkif->persistent_gnt_c < xen_blkif_max_pgrants ||
-	    (blkif->persistent_gnt_c == xen_blkif_max_pgrants &&
-	    !blkif->vbd.overflow_max_grants)) {
-		return;
+	if (ring->persistent_gnt_c < xen_blkif_max_pgrants ||
+	    (ring->persistent_gnt_c == xen_blkif_max_pgrants &&
+	    !ring->blkif->vbd.overflow_max_grants)) {
+		goto out;
 	}
 
-	if (work_busy(&blkif->persistent_purge_work)) {
+	if (work_busy(&ring->persistent_purge_work)) {
 		pr_alert_ratelimited("Scheduled work from previous purge is still busy, cannot purge list\n");
-		return;
+		goto out;
 	}
 
 	num_clean = (xen_blkif_max_pgrants / 100) * LRU_PERCENT_CLEAN;
-	num_clean = blkif->persistent_gnt_c - xen_blkif_max_pgrants + num_clean;
-	num_clean = min(blkif->persistent_gnt_c, num_clean);
+	num_clean = ring->persistent_gnt_c - xen_blkif_max_pgrants + num_clean;
+	num_clean = min(ring->persistent_gnt_c, num_clean);
 	if ((num_clean == 0) ||
-	    (num_clean > (blkif->persistent_gnt_c - atomic_read(&blkif->persistent_gnt_in_use))))
-		return;
+	    (num_clean > (ring->persistent_gnt_c - atomic_read(&ring->persistent_gnt_in_use))))
+		goto out;
 
 	/*
 	 * At this point, we can assure that there will be no calls
@@ -394,8 +405,8 @@ static void purge_persistent_gnt(struct xen_blkif *blkif)
 
 	pr_debug("Going to purge %u persistent grants\n", num_clean);
 
-	BUG_ON(!list_empty(&blkif->persistent_purge_list));
-	root = &blkif->persistent_gnts;
+	BUG_ON(!list_empty(&ring->persistent_purge_list));
+	root = &ring->persistent_gnts;
 purge_list:
 	foreach_grant_safe(persistent_gnt, n, root, node) {
 		BUG_ON(persistent_gnt->handle ==
@@ -414,7 +425,7 @@ static void purge_persistent_gnt(struct xen_blkif *blkif)
 
 		rb_erase(&persistent_gnt->node, root);
 		list_add(&persistent_gnt->remove_node,
-		         &blkif->persistent_purge_list);
+			 &ring->persistent_purge_list);
 		if (--num_clean == 0)
 			goto finished;
 	}
@@ -435,30 +446,32 @@ static void purge_persistent_gnt(struct xen_blkif *blkif)
 		goto purge_list;
 	}
 
-	blkif->persistent_gnt_c -= (total - num_clean);
-	blkif->vbd.overflow_max_grants = 0;
+	ring->persistent_gnt_c -= (total - num_clean);
+	ring->blkif->vbd.overflow_max_grants = 0;
 
 	/* We can defer this work */
-	schedule_work(&blkif->persistent_purge_work);
+	schedule_work(&ring->persistent_purge_work);
 	pr_debug("Purged %u/%u\n", (total - num_clean), total);
+
+out:
 	return;
 }
 
 /*
  * Retrieve from the 'pending_reqs' a free pending_req structure to be used.
  */
-static struct pending_req *alloc_req(struct xen_blkif *blkif)
+static struct pending_req *alloc_req(struct xen_blkif_ring *ring)
 {
 	struct pending_req *req = NULL;
 	unsigned long flags;
 
-	spin_lock_irqsave(&blkif->pending_free_lock, flags);
-	if (!list_empty(&blkif->pending_free)) {
-		req = list_entry(blkif->pending_free.next, struct pending_req,
+	spin_lock_irqsave(&ring->pending_free_lock, flags);
+	if (!list_empty(&ring->pending_free)) {
+		req = list_entry(ring->pending_free.next, struct pending_req,
 				 free_list);
 		list_del(&req->free_list);
 	}
-	spin_unlock_irqrestore(&blkif->pending_free_lock, flags);
+	spin_unlock_irqrestore(&ring->pending_free_lock, flags);
 	return req;
 }
 
@@ -466,17 +479,17 @@ static struct pending_req *alloc_req(struct xen_blkif *blkif)
  * Return the 'pending_req' structure back to the freepool. We also
  * wake up the thread if it was waiting for a free page.
  */
-static void free_req(struct xen_blkif *blkif, struct pending_req *req)
+static void free_req(struct xen_blkif_ring *ring, struct pending_req *req)
 {
 	unsigned long flags;
 	int was_empty;
 
-	spin_lock_irqsave(&blkif->pending_free_lock, flags);
-	was_empty = list_empty(&blkif->pending_free);
-	list_add(&req->free_list, &blkif->pending_free);
-	spin_unlock_irqrestore(&blkif->pending_free_lock, flags);
+	spin_lock_irqsave(&ring->pending_free_lock, flags);
+	was_empty = list_empty(&ring->pending_free);
+	list_add(&req->free_list, &ring->pending_free);
+	spin_unlock_irqrestore(&ring->pending_free_lock, flags);
 	if (was_empty)
-		wake_up(&blkif->pending_free_wq);
+		wake_up(&ring->pending_free_wq);
 }
 
 /*
@@ -556,10 +569,10 @@ static void xen_vbd_resize(struct xen_blkif *blkif)
 /*
  * Notification from the guest OS.
  */
-static void blkif_notify_work(struct xen_blkif *blkif)
+static void blkif_notify_work(struct xen_blkif_ring *ring)
 {
-	blkif->waiting_reqs = 1;
-	wake_up(&blkif->wq);
+	ring->waiting_reqs = 1;
+	wake_up(&ring->wq);
 }
 
 irqreturn_t xen_blkif_be_int(int irq, void *dev_id)
@@ -572,31 +585,33 @@ irqreturn_t xen_blkif_be_int(int irq, void *dev_id)
  * SCHEDULER FUNCTIONS
  */
 
-static void print_stats(struct xen_blkif *blkif)
+static void print_stats(struct xen_blkif_ring *ring)
 {
 	pr_info("(%s): oo %3llu  |  rd %4llu  |  wr %4llu  |  f %4llu"
 		 "  |  ds %4llu | pg: %4u/%4d\n",
-		 current->comm, blkif->st_oo_req,
-		 blkif->st_rd_req, blkif->st_wr_req,
-		 blkif->st_f_req, blkif->st_ds_req,
-		 blkif->persistent_gnt_c,
+		 current->comm, ring->st_oo_req,
+		 ring->st_rd_req, ring->st_wr_req,
+		 ring->st_f_req, ring->st_ds_req,
+		 ring->persistent_gnt_c,
 		 xen_blkif_max_pgrants);
-	blkif->st_print = jiffies + msecs_to_jiffies(10 * 1000);
-	blkif->st_rd_req = 0;
-	blkif->st_wr_req = 0;
-	blkif->st_oo_req = 0;
-	blkif->st_ds_req = 0;
+	ring->st_print = jiffies + msecs_to_jiffies(10 * 1000);
+	ring->st_rd_req = 0;
+	ring->st_wr_req = 0;
+	ring->st_oo_req = 0;
+	ring->st_ds_req = 0;
 }
 
 int xen_blkif_schedule(void *arg)
 {
-	struct xen_blkif *blkif = arg;
+	struct xen_blkif_ring *ring = arg;
+	struct xen_blkif *blkif = ring->blkif;
 	struct xen_vbd *vbd = &blkif->vbd;
 	unsigned long timeout;
 	int ret;
 
 	xen_blkif_get(blkif);
 
+	set_freezable();
 	while (!kthread_should_stop()) {
 		if (try_to_freeze())
 			continue;
@@ -606,50 +621,50 @@ int xen_blkif_schedule(void *arg)
 		timeout = msecs_to_jiffies(LRU_INTERVAL);
 
 		timeout = wait_event_interruptible_timeout(
-			blkif->wq,
-			blkif->waiting_reqs || kthread_should_stop(),
+			ring->wq,
+			ring->waiting_reqs || kthread_should_stop(),
 			timeout);
 		if (timeout == 0)
 			goto purge_gnt_list;
 		timeout = wait_event_interruptible_timeout(
-			blkif->pending_free_wq,
-			!list_empty(&blkif->pending_free) ||
+			ring->pending_free_wq,
+			!list_empty(&ring->pending_free) ||
 			kthread_should_stop(),
 			timeout);
 		if (timeout == 0)
 			goto purge_gnt_list;
 
-		blkif->waiting_reqs = 0;
+		ring->waiting_reqs = 0;
 		smp_mb(); /* clear flag *before* checking for work */
 
-		ret = do_block_io_op(blkif);
+		ret = do_block_io_op(ring);
 		if (ret > 0)
-			blkif->waiting_reqs = 1;
+			ring->waiting_reqs = 1;
 		if (ret == -EACCES)
-			wait_event_interruptible(blkif->shutdown_wq,
+			wait_event_interruptible(ring->shutdown_wq,
 						 kthread_should_stop());
 
 purge_gnt_list:
 		if (blkif->vbd.feature_gnt_persistent &&
-		    time_after(jiffies, blkif->next_lru)) {
-			purge_persistent_gnt(blkif);
-			blkif->next_lru = jiffies + msecs_to_jiffies(LRU_INTERVAL);
+		    time_after(jiffies, ring->next_lru)) {
+			purge_persistent_gnt(ring);
+			ring->next_lru = jiffies + msecs_to_jiffies(LRU_INTERVAL);
 		}
 
 		/* Shrink if we have more than xen_blkif_max_buffer_pages */
-		shrink_free_pagepool(blkif, xen_blkif_max_buffer_pages);
+		shrink_free_pagepool(ring, xen_blkif_max_buffer_pages);
 
-		if (log_stats && time_after(jiffies, blkif->st_print))
-			print_stats(blkif);
+		if (log_stats && time_after(jiffies, ring->st_print))
+			print_stats(ring);
 	}
 
 	/* Drain pending purge work */
-	flush_work(&blkif->persistent_purge_work);
+	flush_work(&ring->persistent_purge_work);
 
 	if (log_stats)
-		print_stats(blkif);
+		print_stats(ring);
 
-	blkif->xenblkd = NULL;
+	ring->xenblkd = NULL;
 	xen_blkif_put(blkif);
 
 	return 0;
@@ -658,22 +673,22 @@ int xen_blkif_schedule(void *arg)
 /*
  * Remove persistent grants and empty the pool of free pages
  */
-void xen_blkbk_free_caches(struct xen_blkif *blkif)
+void xen_blkbk_free_caches(struct xen_blkif_ring *ring)
 {
 	/* Free all persistent grant pages */
-	if (!RB_EMPTY_ROOT(&blkif->persistent_gnts))
-		free_persistent_gnts(blkif, &blkif->persistent_gnts,
-			blkif->persistent_gnt_c);
+	if (!RB_EMPTY_ROOT(&ring->persistent_gnts))
+		free_persistent_gnts(ring, &ring->persistent_gnts,
+			ring->persistent_gnt_c);
 
-	BUG_ON(!RB_EMPTY_ROOT(&blkif->persistent_gnts));
-	blkif->persistent_gnt_c = 0;
+	BUG_ON(!RB_EMPTY_ROOT(&ring->persistent_gnts));
+	ring->persistent_gnt_c = 0;
 
 	/* Since we are shutting down remove all pages from the buffer */
-	shrink_free_pagepool(blkif, 0 /* All */);
+	shrink_free_pagepool(ring, 0 /* All */);
 }
 
 static unsigned int xen_blkbk_unmap_prepare(
-	struct xen_blkif *blkif,
+	struct xen_blkif_ring *ring,
 	struct grant_page **pages,
 	unsigned int num,
 	struct gnttab_unmap_grant_ref *unmap_ops,
@@ -683,7 +698,7 @@ static unsigned int xen_blkbk_unmap_prepare(
 
 	for (i = 0; i < num; i++) {
 		if (pages[i]->persistent_gnt != NULL) {
-			put_persistent_gnt(blkif, pages[i]->persistent_gnt);
+			put_persistent_gnt(ring, pages[i]->persistent_gnt);
 			continue;
 		}
 		if (pages[i]->handle == BLKBACK_INVALID_HANDLE)
@@ -700,17 +715,18 @@ static unsigned int xen_blkbk_unmap_prepare(
 
 static void xen_blkbk_unmap_and_respond_callback(int result, struct gntab_unmap_queue_data *data)
 {
-	struct pending_req* pending_req = (struct pending_req*) (data->data);
-	struct xen_blkif *blkif = pending_req->blkif;
+	struct pending_req *pending_req = (struct pending_req *)(data->data);
+	struct xen_blkif_ring *ring = pending_req->ring;
+	struct xen_blkif *blkif = ring->blkif;
 
 	/* BUG_ON used to reproduce existing behaviour,
 	   but is this the best way to deal with this? */
 	BUG_ON(result);
 
-	put_free_pages(blkif, data->pages, data->count);
-	make_response(blkif, pending_req->id,
+	put_free_pages(ring, data->pages, data->count);
+	make_response(ring, pending_req->id,
 		      pending_req->operation, pending_req->status);
-	free_req(blkif, pending_req);
+	free_req(ring, pending_req);
 	/*
 	 * Make sure the request is freed before releasing blkif,
 	 * or there could be a race between free_req and the
@@ -723,7 +739,7 @@ static void xen_blkbk_unmap_and_respond_callback(int result, struct gntab_unmap_
 	 * pending_free_wq if there's a drain going on, but it has
 	 * to be taken into account if the current model is changed.
 	 */
-	if (atomic_dec_and_test(&blkif->inflight) && atomic_read(&blkif->drain)) {
+	if (atomic_dec_and_test(&ring->inflight) && atomic_read(&blkif->drain)) {
 		complete(&blkif->drain_complete);
 	}
 	xen_blkif_put(blkif);
@@ -732,11 +748,11 @@ static void xen_blkbk_unmap_and_respond_callback(int result, struct gntab_unmap_
 static void xen_blkbk_unmap_and_respond(struct pending_req *req)
 {
 	struct gntab_unmap_queue_data* work = &req->gnttab_unmap_data;
-	struct xen_blkif *blkif = req->blkif;
+	struct xen_blkif_ring *ring = req->ring;
 	struct grant_page **pages = req->segments;
 	unsigned int invcount;
 
-	invcount = xen_blkbk_unmap_prepare(blkif, pages, req->nr_segs,
+	invcount = xen_blkbk_unmap_prepare(ring, pages, req->nr_segs,
 					   req->unmap, req->unmap_pages);
 
 	work->data = req;
@@ -757,7 +773,7 @@ static void xen_blkbk_unmap_and_respond(struct pending_req *req)
  * of hypercalls, but since this is only used in error paths there's
  * no real need.
  */
-static void xen_blkbk_unmap(struct xen_blkif *blkif,
+static void xen_blkbk_unmap(struct xen_blkif_ring *ring,
                             struct grant_page *pages[],
                             int num)
 {
@@ -768,20 +784,20 @@ static void xen_blkbk_unmap(struct xen_blkif *blkif,
 
 	while (num) {
 		unsigned int batch = min(num, BLKIF_MAX_SEGMENTS_PER_REQUEST);
-		
-		invcount = xen_blkbk_unmap_prepare(blkif, pages, batch,
+
+		invcount = xen_blkbk_unmap_prepare(ring, pages, batch,
 						   unmap, unmap_pages);
 		if (invcount) {
 			ret = gnttab_unmap_refs(unmap, NULL, unmap_pages, invcount);
 			BUG_ON(ret);
-			put_free_pages(blkif, unmap_pages, invcount);
+			put_free_pages(ring, unmap_pages, invcount);
 		}
 		pages += batch;
 		num -= batch;
 	}
 }
 
-static int xen_blkbk_map(struct xen_blkif *blkif,
+static int xen_blkbk_map(struct xen_blkif_ring *ring,
 			 struct grant_page *pages[],
 			 int num, bool ro)
 {
@@ -794,6 +810,7 @@ static int xen_blkbk_map(struct xen_blkif *blkif,
 	int ret = 0;
 	int last_map = 0, map_until = 0;
 	int use_persistent_gnts;
+	struct xen_blkif *blkif = ring->blkif;
 
 	use_persistent_gnts = (blkif->vbd.feature_gnt_persistent);
 
@@ -806,10 +823,11 @@ static int xen_blkbk_map(struct xen_blkif *blkif,
 	for (i = map_until; i < num; i++) {
 		uint32_t flags;
 
-		if (use_persistent_gnts)
+		if (use_persistent_gnts) {
 			persistent_gnt = get_persistent_gnt(
-				blkif,
+				ring,
 				pages[i]->gref);
+		}
 
 		if (persistent_gnt) {
 			/*
@@ -819,7 +837,7 @@ static int xen_blkbk_map(struct xen_blkif *blkif,
 			pages[i]->page = persistent_gnt->page;
 			pages[i]->persistent_gnt = persistent_gnt;
 		} else {
-			if (get_free_page(blkif, &pages[i]->page))
+			if (get_free_page(ring, &pages[i]->page))
 				goto out_of_memory;
 			addr = vaddr(pages[i]->page);
 			pages_to_gnt[segs_to_map] = pages[i]->page;
@@ -852,7 +870,7 @@ static int xen_blkbk_map(struct xen_blkif *blkif,
 			BUG_ON(new_map_idx >= segs_to_map);
 			if (unlikely(map[new_map_idx].status != 0)) {
 				pr_debug("invalid buffer -- could not remap it\n");
-				put_free_pages(blkif, &pages[seg_idx]->page, 1);
+				put_free_pages(ring, &pages[seg_idx]->page, 1);
 				pages[seg_idx]->handle = BLKBACK_INVALID_HANDLE;
 				ret |= 1;
 				goto next;
@@ -862,7 +880,7 @@ static int xen_blkbk_map(struct xen_blkif *blkif,
 			continue;
 		}
 		if (use_persistent_gnts &&
-		    blkif->persistent_gnt_c < xen_blkif_max_pgrants) {
+		    ring->persistent_gnt_c < xen_blkif_max_pgrants) {
 			/*
 			 * We are using persistent grants, the grant is
 			 * not mapped but we might have room for it.
@@ -880,7 +898,7 @@ static int xen_blkbk_map(struct xen_blkif *blkif,
 			persistent_gnt->gnt = map[new_map_idx].ref;
 			persistent_gnt->handle = map[new_map_idx].handle;
 			persistent_gnt->page = pages[seg_idx]->page;
-			if (add_persistent_gnt(blkif,
+			if (add_persistent_gnt(ring,
 			                       persistent_gnt)) {
 				kfree(persistent_gnt);
 				persistent_gnt = NULL;
@@ -888,7 +906,7 @@ static int xen_blkbk_map(struct xen_blkif *blkif,
 			}
 			pages[seg_idx]->persistent_gnt = persistent_gnt;
 			pr_debug("grant %u added to the tree of persistent grants, using %u/%u\n",
-				 persistent_gnt->gnt, blkif->persistent_gnt_c,
+				 persistent_gnt->gnt, ring->persistent_gnt_c,
 				 xen_blkif_max_pgrants);
 			goto next;
 		}
@@ -913,7 +931,7 @@ static int xen_blkbk_map(struct xen_blkif *blkif,
 
 out_of_memory:
 	pr_alert("%s: out of memory\n", __func__);
-	put_free_pages(blkif, pages_to_gnt, segs_to_map);
+	put_free_pages(ring, pages_to_gnt, segs_to_map);
 	return -ENOMEM;
 }
 
@@ -921,7 +939,7 @@ static int xen_blkbk_map_seg(struct pending_req *pending_req)
 {
 	int rc;
 
-	rc = xen_blkbk_map(pending_req->blkif, pending_req->segments,
+	rc = xen_blkbk_map(pending_req->ring, pending_req->segments,
 			   pending_req->nr_segs,
 	                   (pending_req->operation != BLKIF_OP_READ));
 
@@ -934,7 +952,7 @@ static int xen_blkbk_parse_indirect(struct blkif_request *req,
 				    struct phys_req *preq)
 {
 	struct grant_page **pages = pending_req->indirect_pages;
-	struct xen_blkif *blkif = pending_req->blkif;
+	struct xen_blkif_ring *ring = pending_req->ring;
 	int indirect_grefs, rc, n, nseg, i;
 	struct blkif_request_segment *segments = NULL;
 
@@ -945,7 +963,7 @@ static int xen_blkbk_parse_indirect(struct blkif_request *req,
 	for (i = 0; i < indirect_grefs; i++)
 		pages[i]->gref = req->u.indirect.indirect_grefs[i];
 
-	rc = xen_blkbk_map(blkif, pages, indirect_grefs, true);
+	rc = xen_blkbk_map(ring, pages, indirect_grefs, true);
 	if (rc)
 		goto unmap;
 
@@ -977,15 +995,16 @@ static int xen_blkbk_parse_indirect(struct blkif_request *req,
 unmap:
 	if (segments)
 		kunmap_atomic(segments);
-	xen_blkbk_unmap(blkif, pages, indirect_grefs);
+	xen_blkbk_unmap(ring, pages, indirect_grefs);
 	return rc;
 }
 
-static int dispatch_discard_io(struct xen_blkif *blkif,
+static int dispatch_discard_io(struct xen_blkif_ring *ring,
 				struct blkif_request *req)
 {
 	int err = 0;
 	int status = BLKIF_RSP_OKAY;
+	struct xen_blkif *blkif = ring->blkif;
 	struct block_device *bdev = blkif->vbd.bdev;
 	unsigned long secure;
 	struct phys_req preq;
@@ -1002,7 +1021,7 @@ static int dispatch_discard_io(struct xen_blkif *blkif,
 			preq.sector_number + preq.nr_sects, blkif->vbd.pdevice);
 		goto fail_response;
 	}
-	blkif->st_ds_req++;
+	ring->st_ds_req++;
 
 	secure = (blkif->vbd.discard_secure &&
 		 (req->u.discard.flag & BLKIF_DISCARD_SECURE)) ?
@@ -1018,26 +1037,28 @@ static int dispatch_discard_io(struct xen_blkif *blkif,
 	} else if (err)
 		status = BLKIF_RSP_ERROR;
 
-	make_response(blkif, req->u.discard.id, req->operation, status);
+	make_response(ring, req->u.discard.id, req->operation, status);
 	xen_blkif_put(blkif);
 	return err;
 }
 
-static int dispatch_other_io(struct xen_blkif *blkif,
+static int dispatch_other_io(struct xen_blkif_ring *ring,
 			     struct blkif_request *req,
 			     struct pending_req *pending_req)
 {
-	free_req(blkif, pending_req);
-	make_response(blkif, req->u.other.id, req->operation,
+	free_req(ring, pending_req);
+	make_response(ring, req->u.other.id, req->operation,
 		      BLKIF_RSP_EOPNOTSUPP);
 	return -EIO;
 }
 
-static void xen_blk_drain_io(struct xen_blkif *blkif)
+static void xen_blk_drain_io(struct xen_blkif_ring *ring)
 {
+	struct xen_blkif *blkif = ring->blkif;
+
 	atomic_set(&blkif->drain, 1);
 	do {
-		if (atomic_read(&blkif->inflight) == 0)
+		if (atomic_read(&ring->inflight) == 0)
 			break;
 		wait_for_completion_interruptible_timeout(
 				&blkif->drain_complete, HZ);
@@ -1058,12 +1079,12 @@ static void __end_block_io_op(struct pending_req *pending_req, int error)
 	if ((pending_req->operation == BLKIF_OP_FLUSH_DISKCACHE) &&
 	    (error == -EOPNOTSUPP)) {
 		pr_debug("flush diskcache op failed, not supported\n");
-		xen_blkbk_flush_diskcache(XBT_NIL, pending_req->blkif->be, 0);
+		xen_blkbk_flush_diskcache(XBT_NIL, pending_req->ring->blkif->be, 0);
 		pending_req->status = BLKIF_RSP_EOPNOTSUPP;
 	} else if ((pending_req->operation == BLKIF_OP_WRITE_BARRIER) &&
 		    (error == -EOPNOTSUPP)) {
 		pr_debug("write barrier op failed, not supported\n");
-		xen_blkbk_barrier(XBT_NIL, pending_req->blkif->be, 0);
+		xen_blkbk_barrier(XBT_NIL, pending_req->ring->blkif->be, 0);
 		pending_req->status = BLKIF_RSP_EOPNOTSUPP;
 	} else if (error) {
 		pr_debug("Buffer not up-to-date at end of operation,"
@@ -1097,9 +1118,9 @@ static void end_block_io_op(struct bio *bio)
  * and transmute  it to the block API to hand it over to the proper block disk.
  */
 static int
-__do_block_io_op(struct xen_blkif *blkif)
+__do_block_io_op(struct xen_blkif_ring *ring)
 {
-	union blkif_back_rings *blk_rings = &blkif->blk_rings;
+	union blkif_back_rings *blk_rings = &ring->blk_rings;
 	struct blkif_request req;
 	struct pending_req *pending_req;
 	RING_IDX rc, rp;
@@ -1112,7 +1133,7 @@ __do_block_io_op(struct xen_blkif *blkif)
 	if (RING_REQUEST_PROD_OVERFLOW(&blk_rings->common, rp)) {
 		rc = blk_rings->common.rsp_prod_pvt;
 		pr_warn("Frontend provided bogus ring requests (%d - %d = %d). Halting ring processing on dev=%04x\n",
-			rp, rc, rp - rc, blkif->vbd.pdevice);
+			rp, rc, rp - rc, ring->blkif->vbd.pdevice);
 		return -EACCES;
 	}
 	while (rc != rp) {
@@ -1125,14 +1146,14 @@ __do_block_io_op(struct xen_blkif *blkif)
 			break;
 		}
 
-		pending_req = alloc_req(blkif);
+		pending_req = alloc_req(ring);
 		if (NULL == pending_req) {
-			blkif->st_oo_req++;
+			ring->st_oo_req++;
 			more_to_do = 1;
 			break;
 		}
 
-		switch (blkif->blk_protocol) {
+		switch (ring->blkif->blk_protocol) {
 		case BLKIF_PROTOCOL_NATIVE:
 			memcpy(&req, RING_GET_REQUEST(&blk_rings->native, rc), sizeof(req));
 			break;
@@ -1156,16 +1177,16 @@ __do_block_io_op(struct xen_blkif *blkif)
 		case BLKIF_OP_WRITE_BARRIER:
 		case BLKIF_OP_FLUSH_DISKCACHE:
 		case BLKIF_OP_INDIRECT:
-			if (dispatch_rw_block_io(blkif, &req, pending_req))
+			if (dispatch_rw_block_io(ring, &req, pending_req))
 				goto done;
 			break;
 		case BLKIF_OP_DISCARD:
-			free_req(blkif, pending_req);
-			if (dispatch_discard_io(blkif, &req))
+			free_req(ring, pending_req);
+			if (dispatch_discard_io(ring, &req))
 				goto done;
 			break;
 		default:
-			if (dispatch_other_io(blkif, &req, pending_req))
+			if (dispatch_other_io(ring, &req, pending_req))
 				goto done;
 			break;
 		}
@@ -1178,13 +1199,13 @@ __do_block_io_op(struct xen_blkif *blkif)
 }
 
 static int
-do_block_io_op(struct xen_blkif *blkif)
+do_block_io_op(struct xen_blkif_ring *ring)
 {
-	union blkif_back_rings *blk_rings = &blkif->blk_rings;
+	union blkif_back_rings *blk_rings = &ring->blk_rings;
 	int more_to_do;
 
 	do {
-		more_to_do = __do_block_io_op(blkif);
+		more_to_do = __do_block_io_op(ring);
 		if (more_to_do)
 			break;
 
@@ -1197,7 +1218,7 @@ do_block_io_op(struct xen_blkif *blkif)
  * Transmutation of the 'struct blkif_request' to a proper 'struct bio'
  * and call the 'submit_bio' to pass it to the underlying storage.
  */
-static int dispatch_rw_block_io(struct xen_blkif *blkif,
+static int dispatch_rw_block_io(struct xen_blkif_ring *ring,
 				struct blkif_request *req,
 				struct pending_req *pending_req)
 {
@@ -1225,17 +1246,17 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif,
 
 	switch (req_operation) {
 	case BLKIF_OP_READ:
-		blkif->st_rd_req++;
+		ring->st_rd_req++;
 		operation = READ;
 		break;
 	case BLKIF_OP_WRITE:
-		blkif->st_wr_req++;
+		ring->st_wr_req++;
 		operation = WRITE_ODIRECT;
 		break;
 	case BLKIF_OP_WRITE_BARRIER:
 		drain = true;
 	case BLKIF_OP_FLUSH_DISKCACHE:
-		blkif->st_f_req++;
+		ring->st_f_req++;
 		operation = WRITE_FLUSH;
 		break;
 	default:
@@ -1260,7 +1281,7 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif,
 
 	preq.nr_sects      = 0;
 
-	pending_req->blkif     = blkif;
+	pending_req->ring      = ring;
 	pending_req->id        = req->u.rw.id;
 	pending_req->operation = req_operation;
 	pending_req->status    = BLKIF_RSP_OKAY;
@@ -1287,12 +1308,12 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif,
 			goto fail_response;
 	}
 
-	if (xen_vbd_translate(&preq, blkif, operation) != 0) {
+	if (xen_vbd_translate(&preq, ring->blkif, operation) != 0) {
 		pr_debug("access denied: %s of [%llu,%llu] on dev=%04x\n",
 			 operation == READ ? "read" : "write",
 			 preq.sector_number,
 			 preq.sector_number + preq.nr_sects,
-			 blkif->vbd.pdevice);
+			 ring->blkif->vbd.pdevice);
 		goto fail_response;
 	}
 
@@ -1304,7 +1325,7 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif,
 		if (((int)preq.sector_number|(int)seg[i].nsec) &
 		    ((bdev_logical_block_size(preq.bdev) >> 9) - 1)) {
 			pr_debug("Misaligned I/O request from domain %d\n",
-				 blkif->domid);
+				 ring->blkif->domid);
 			goto fail_response;
 		}
 	}
@@ -1313,7 +1334,7 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif,
 	 * issue the WRITE_FLUSH.
 	 */
 	if (drain)
-		xen_blk_drain_io(pending_req->blkif);
+		xen_blk_drain_io(pending_req->ring);
 
 	/*
 	 * If we have failed at this point, we need to undo the M2P override,
@@ -1328,8 +1349,8 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif,
 	 * This corresponding xen_blkif_put is done in __end_block_io_op, or
 	 * below (in "!bio") if we are handling a BLKIF_OP_DISCARD.
 	 */
-	xen_blkif_get(blkif);
-	atomic_inc(&blkif->inflight);
+	xen_blkif_get(ring->blkif);
+	atomic_inc(&ring->inflight);
 
 	for (i = 0; i < nseg; i++) {
 		while ((bio == NULL) ||
@@ -1377,19 +1398,19 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif,
 	blk_finish_plug(&plug);
 
 	if (operation == READ)
-		blkif->st_rd_sect += preq.nr_sects;
+		ring->st_rd_sect += preq.nr_sects;
 	else if (operation & WRITE)
-		blkif->st_wr_sect += preq.nr_sects;
+		ring->st_wr_sect += preq.nr_sects;
 
 	return 0;
 
  fail_flush:
-	xen_blkbk_unmap(blkif, pending_req->segments,
+	xen_blkbk_unmap(ring, pending_req->segments,
 	                pending_req->nr_segs);
  fail_response:
 	/* Haven't submitted any bio's yet. */
-	make_response(blkif, req->u.rw.id, req_operation, BLKIF_RSP_ERROR);
-	free_req(blkif, pending_req);
+	make_response(ring, req->u.rw.id, req_operation, BLKIF_RSP_ERROR);
+	free_req(ring, pending_req);
 	msleep(1); /* back off a bit */
 	return -EIO;
 
@@ -1407,21 +1428,22 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif,
 /*
  * Put a response on the ring on how the operation fared.
  */
-static void make_response(struct xen_blkif *blkif, u64 id,
+static void make_response(struct xen_blkif_ring *ring, u64 id,
 			  unsigned short op, int st)
 {
 	struct blkif_response  resp;
 	unsigned long     flags;
-	union blkif_back_rings *blk_rings = &blkif->blk_rings;
+	union blkif_back_rings *blk_rings;
 	int notify;
 
 	resp.id        = id;
 	resp.operation = op;
 	resp.status    = st;
 
-	spin_lock_irqsave(&blkif->blk_ring_lock, flags);
+	spin_lock_irqsave(&ring->blk_ring_lock, flags);
+	blk_rings = &ring->blk_rings;
 	/* Place on the response ring for the relevant domain. */
-	switch (blkif->blk_protocol) {
+	switch (ring->blkif->blk_protocol) {
 	case BLKIF_PROTOCOL_NATIVE:
 		memcpy(RING_GET_RESPONSE(&blk_rings->native, blk_rings->native.rsp_prod_pvt),
 		       &resp, sizeof(resp));
@@ -1439,9 +1461,9 @@ static void make_response(struct xen_blkif *blkif, u64 id,
 	}
 	blk_rings->common.rsp_prod_pvt++;
 	RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&blk_rings->common, notify);
-	spin_unlock_irqrestore(&blkif->blk_ring_lock, flags);
+	spin_unlock_irqrestore(&ring->blk_ring_lock, flags);
 	if (notify)
-		notify_remote_via_irq(blkif->irq);
+		notify_remote_via_irq(ring->irq);
 }
 
 static int __init xen_blkif_init(void)
@@ -1457,6 +1479,9 @@ static int __init xen_blkif_init(void)
 		xen_blkif_max_ring_order = XENBUS_MAX_RING_GRANT_ORDER;
 	}
 
+	if (xenblk_max_queues == 0)
+		xenblk_max_queues = num_online_cpus();
+
 	rc = xen_blkif_interface_init();
 	if (rc)
 		goto failed_init;

drivers/block/xen-blkback/common.h

@@ -46,6 +46,7 @@
 #include <xen/interface/io/protocols.h>
 
 extern unsigned int xen_blkif_max_ring_order;
+extern unsigned int xenblk_max_queues;
 /*
  * This is the maximum number of segments that would be allowed in indirect
  * requests. This value will also be passed to the frontend.
@@ -269,68 +270,79 @@ struct persistent_gnt {
 	struct list_head remove_node;
 };
 
-struct xen_blkif {
-	/* Unique identifier for this interface. */
-	domid_t			domid;
-	unsigned int		handle;
+/* Per-ring information. */
+struct xen_blkif_ring {
 	/* Physical parameters of the comms window. */
 	unsigned int		irq;
-	/* Comms information. */
-	enum blkif_protocol	blk_protocol;
 	union blkif_back_rings	blk_rings;
 	void			*blk_ring;
-	/* The VBD attached to this interface. */
-	struct xen_vbd		vbd;
-	/* Back pointer to the backend_info. */
-	struct backend_info	*be;
 	/* Private fields. */
 	spinlock_t		blk_ring_lock;
-	atomic_t		refcnt;
 
 	wait_queue_head_t	wq;
-	/* for barrier (drain) requests */
-	struct completion	drain_complete;
-	atomic_t		drain;
 	atomic_t		inflight;
-	/* One thread per one blkif. */
+	/* One thread per blkif ring. */
 	struct task_struct	*xenblkd;
 	unsigned int		waiting_reqs;
 
-	/* tree to store persistent grants */
+	/* List of all 'pending_req' available */
+	struct list_head	pending_free;
+	/* And its spinlock. */
+	spinlock_t		pending_free_lock;
+	wait_queue_head_t	pending_free_wq;
+
+	/* Tree to store persistent grants. */
+	spinlock_t		pers_gnts_lock;
 	struct rb_root		persistent_gnts;
 	unsigned int		persistent_gnt_c;
 	atomic_t		persistent_gnt_in_use;
 	unsigned long           next_lru;
 
-	/* used by the kworker that offload work from the persistent purge */
+	/* Statistics. */
+	unsigned long		st_print;
+	unsigned long long	st_rd_req;
+	unsigned long long	st_wr_req;
+	unsigned long long	st_oo_req;
+	unsigned long long	st_f_req;
+	unsigned long long	st_ds_req;
+	unsigned long long	st_rd_sect;
+	unsigned long long	st_wr_sect;
+
+	/* Used by the kworker that offload work from the persistent purge. */
 	struct list_head	persistent_purge_list;
 	struct work_struct	persistent_purge_work;
 
-	/* buffer of free pages to map grant refs */
+	/* Buffer of free pages to map grant refs. */
 	spinlock_t		free_pages_lock;
 	int			free_pages_num;
 	struct list_head	free_pages;
 
-	/* List of all 'pending_req' available */
-	struct list_head	pending_free;
-	/* And its spinlock. */
-	spinlock_t		pending_free_lock;
-	wait_queue_head_t	pending_free_wq;
-
-	/* statistics */
-	unsigned long		st_print;
-	unsigned long long			st_rd_req;
-	unsigned long long			st_wr_req;
-	unsigned long long			st_oo_req;
-	unsigned long long			st_f_req;
-	unsigned long long			st_ds_req;
-	unsigned long long			st_rd_sect;
-	unsigned long long			st_wr_sect;
-
 	struct work_struct	free_work;
 	/* Thread shutdown wait queue. */
 	wait_queue_head_t	shutdown_wq;
-	unsigned int nr_ring_pages;
+	struct xen_blkif 	*blkif;
+};
+
+struct xen_blkif {
+	/* Unique identifier for this interface. */
+	domid_t			domid;
+	unsigned int		handle;
+	/* Comms information. */
+	enum blkif_protocol	blk_protocol;
+	/* The VBD attached to this interface. */
+	struct xen_vbd		vbd;
+	/* Back pointer to the backend_info. */
+	struct backend_info	*be;
+	atomic_t		refcnt;
+	/* for barrier (drain) requests */
+	struct completion	drain_complete;
+	atomic_t		drain;
+
+	struct work_struct	free_work;
+	unsigned int 		nr_ring_pages;
+	/* All rings for this device. */
+	struct xen_blkif_ring	*rings;
+	unsigned int		nr_rings;
 };
 
 struct seg_buf {
@@ -352,7 +364,7 @@ struct grant_page {
  * response queued for it, with the saved 'id' passed back.
  */
 struct pending_req {
-	struct xen_blkif	*blkif;
+	struct xen_blkif_ring   *ring;
 	u64			id;
 	int			nr_segs;
 	atomic_t		pendcnt;
@@ -394,7 +406,7 @@ int xen_blkif_xenbus_init(void);
 irqreturn_t xen_blkif_be_int(int irq, void *dev_id);
 int xen_blkif_schedule(void *arg);
 int xen_blkif_purge_persistent(void *arg);
-void xen_blkbk_free_caches(struct xen_blkif *blkif);
+void xen_blkbk_free_caches(struct xen_blkif_ring *ring);
 
 int xen_blkbk_flush_diskcache(struct xenbus_transaction xbt,
 			      struct backend_info *be, int state);

drivers/block/xen-blkback/xenbus.c

@@ -86,9 +86,11 @@ static void xen_update_blkif_status(struct xen_blkif *blkif)
 {
 	int err;
 	char name[BLKBACK_NAME_LEN];
+	struct xen_blkif_ring *ring;
+	int i;
 
 	/* Not ready to connect? */
-	if (!blkif->irq || !blkif->vbd.bdev)
+	if (!blkif->rings || !blkif->rings[0].irq || !blkif->vbd.bdev)
 		return;
 
 	/* Already connected? */
@@ -113,13 +115,55 @@ 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, "%s", name);
-	if (IS_ERR(blkif->xenblkd)) {
-		err = PTR_ERR(blkif->xenblkd);
-		blkif->xenblkd = NULL;
-		xenbus_dev_error(blkif->be->dev, err, "start xenblkd");
-		return;
+	for (i = 0; i < blkif->nr_rings; i++) {
+		ring = &blkif->rings[i];
+		ring->xenblkd = kthread_run(xen_blkif_schedule, ring, "%s-%d", name, i);
+		if (IS_ERR(ring->xenblkd)) {
+			err = PTR_ERR(ring->xenblkd);
+			ring->xenblkd = NULL;
+			xenbus_dev_fatal(blkif->be->dev, err,
+					"start %s-%d xenblkd", name, i);
+			goto out;
+		}
+	}
+	return;
+
+out:
+	while (--i >= 0) {
+		ring = &blkif->rings[i];
+		kthread_stop(ring->xenblkd);
+	}
+	return;
+}
+
+static int xen_blkif_alloc_rings(struct xen_blkif *blkif)
+{
+	unsigned int r;
+
+	blkif->rings = kzalloc(blkif->nr_rings * sizeof(struct xen_blkif_ring), GFP_KERNEL);
+	if (!blkif->rings)
+		return -ENOMEM;
+
+	for (r = 0; r < blkif->nr_rings; r++) {
+		struct xen_blkif_ring *ring = &blkif->rings[r];
+
+		spin_lock_init(&ring->blk_ring_lock);
+		init_waitqueue_head(&ring->wq);
+		INIT_LIST_HEAD(&ring->pending_free);
+		INIT_LIST_HEAD(&ring->persistent_purge_list);
+		INIT_WORK(&ring->persistent_purge_work, xen_blkbk_unmap_purged_grants);
+		spin_lock_init(&ring->free_pages_lock);
+		INIT_LIST_HEAD(&ring->free_pages);
+
+		spin_lock_init(&ring->pending_free_lock);
+		init_waitqueue_head(&ring->pending_free_wq);
+		init_waitqueue_head(&ring->shutdown_wq);
+		ring->blkif = blkif;
+		ring->st_print = jiffies;
+		xen_blkif_get(blkif);
 	}
+
+	return 0;
 }
 
 static struct xen_blkif *xen_blkif_alloc(domid_t domid)
@@ -133,41 +177,25 @@ static struct xen_blkif *xen_blkif_alloc(domid_t domid)
 		return ERR_PTR(-ENOMEM);
 
 	blkif->domid = domid;
-	spin_lock_init(&blkif->blk_ring_lock);
 	atomic_set(&blkif->refcnt, 1);
-	init_waitqueue_head(&blkif->wq);
 	init_completion(&blkif->drain_complete);
-	atomic_set(&blkif->drain, 0);
-	blkif->st_print = jiffies;
-	blkif->persistent_gnts.rb_node = NULL;
-	spin_lock_init(&blkif->free_pages_lock);
-	INIT_LIST_HEAD(&blkif->free_pages);
-	INIT_LIST_HEAD(&blkif->persistent_purge_list);
-	blkif->free_pages_num = 0;
-	atomic_set(&blkif->persistent_gnt_in_use, 0);
-	atomic_set(&blkif->inflight, 0);
-	INIT_WORK(&blkif->persistent_purge_work, xen_blkbk_unmap_purged_grants);
-
-	INIT_LIST_HEAD(&blkif->pending_free);
 	INIT_WORK(&blkif->free_work, xen_blkif_deferred_free);
-	spin_lock_init(&blkif->pending_free_lock);
-	init_waitqueue_head(&blkif->pending_free_wq);
-	init_waitqueue_head(&blkif->shutdown_wq);
 
 	return blkif;
 }
 
-static int xen_blkif_map(struct xen_blkif *blkif, grant_ref_t *gref,
+static int xen_blkif_map(struct xen_blkif_ring *ring, grant_ref_t *gref,
 			 unsigned int nr_grefs, unsigned int evtchn)
 {
 	int err;
+	struct xen_blkif *blkif = ring->blkif;
 
 	/* Already connected through? */
-	if (blkif->irq)
+	if (ring->irq)
 		return 0;
 
 	err = xenbus_map_ring_valloc(blkif->be->dev, gref, nr_grefs,
-				     &blkif->blk_ring);
+				     &ring->blk_ring);
 	if (err < 0)
 		return err;
 
@@ -175,24 +203,24 @@ static int xen_blkif_map(struct xen_blkif *blkif, grant_ref_t *gref,
 	case BLKIF_PROTOCOL_NATIVE:
 	{
 		struct blkif_sring *sring;
-		sring = (struct blkif_sring *)blkif->blk_ring;
-		BACK_RING_INIT(&blkif->blk_rings.native, sring,
+		sring = (struct blkif_sring *)ring->blk_ring;
+		BACK_RING_INIT(&ring->blk_rings.native, sring,
 			       XEN_PAGE_SIZE * nr_grefs);
 		break;
 	}
 	case BLKIF_PROTOCOL_X86_32:
 	{
 		struct blkif_x86_32_sring *sring_x86_32;
-		sring_x86_32 = (struct blkif_x86_32_sring *)blkif->blk_ring;
-		BACK_RING_INIT(&blkif->blk_rings.x86_32, sring_x86_32,
+		sring_x86_32 = (struct blkif_x86_32_sring *)ring->blk_ring;
+		BACK_RING_INIT(&ring->blk_rings.x86_32, sring_x86_32,
 			       XEN_PAGE_SIZE * nr_grefs);
 		break;
 	}
 	case BLKIF_PROTOCOL_X86_64:
 	{
 		struct blkif_x86_64_sring *sring_x86_64;
-		sring_x86_64 = (struct blkif_x86_64_sring *)blkif->blk_ring;
-		BACK_RING_INIT(&blkif->blk_rings.x86_64, sring_x86_64,
+		sring_x86_64 = (struct blkif_x86_64_sring *)ring->blk_ring;
+		BACK_RING_INIT(&ring->blk_rings.x86_64, sring_x86_64,
 			       XEN_PAGE_SIZE * nr_grefs);
 		break;
 	}
@@ -202,13 +230,13 @@ static int xen_blkif_map(struct xen_blkif *blkif, grant_ref_t *gref,
 
 	err = bind_interdomain_evtchn_to_irqhandler(blkif->domid, evtchn,
 						    xen_blkif_be_int, 0,
-						    "blkif-backend", blkif);
+						    "blkif-backend", ring);
 	if (err < 0) {
-		xenbus_unmap_ring_vfree(blkif->be->dev, blkif->blk_ring);
-		blkif->blk_rings.common.sring = NULL;
+		xenbus_unmap_ring_vfree(blkif->be->dev, ring->blk_ring);
+		ring->blk_rings.common.sring = NULL;
 		return err;
 	}
-	blkif->irq = err;
+	ring->irq = err;
 
 	return 0;
 }
@@ -216,50 +244,69 @@ static int xen_blkif_map(struct xen_blkif *blkif, grant_ref_t *gref,
 static int xen_blkif_disconnect(struct xen_blkif *blkif)
 {
 	struct pending_req *req, *n;
-	int i = 0, j;
+	unsigned int j, r;
 
-	if (blkif->xenblkd) {
-		kthread_stop(blkif->xenblkd);
-		wake_up(&blkif->shutdown_wq);
-		blkif->xenblkd = NULL;
-	}
+	for (r = 0; r < blkif->nr_rings; r++) {
+		struct xen_blkif_ring *ring = &blkif->rings[r];
+		unsigned int i = 0;
 
-	/* The above kthread_stop() guarantees that at this point we
-	 * don't have any discard_io or other_io requests. So, checking
-	 * for inflight IO is enough.
-	 */
-	if (atomic_read(&blkif->inflight) > 0)
-		return -EBUSY;
+		if (ring->xenblkd) {
+			kthread_stop(ring->xenblkd);
+			wake_up(&ring->shutdown_wq);
+			ring->xenblkd = NULL;
+		}
 
-	if (blkif->irq) {
-		unbind_from_irqhandler(blkif->irq, blkif);
-		blkif->irq = 0;
-	}
+		/* The above kthread_stop() guarantees that at this point we
+		 * don't have any discard_io or other_io requests. So, checking
+		 * for inflight IO is enough.
+		 */
+		if (atomic_read(&ring->inflight) > 0)
+			return -EBUSY;
 
-	if (blkif->blk_rings.common.sring) {
-		xenbus_unmap_ring_vfree(blkif->be->dev, blkif->blk_ring);
-		blkif->blk_rings.common.sring = NULL;
-	}
+		if (ring->irq) {
+			unbind_from_irqhandler(ring->irq, ring);
+			ring->irq = 0;
+		}
 
-	/* Remove all persistent grants and the cache of ballooned pages. */
-	xen_blkbk_free_caches(blkif);
+		if (ring->blk_rings.common.sring) {
+			xenbus_unmap_ring_vfree(blkif->be->dev, ring->blk_ring);
+			ring->blk_rings.common.sring = NULL;
+		}
 
-	/* Check that there is no request in use */
-	list_for_each_entry_safe(req, n, &blkif->pending_free, free_list) {
-		list_del(&req->free_list);
+		/* Remove all persistent grants and the cache of ballooned pages. */
+		xen_blkbk_free_caches(ring);
 
-		for (j = 0; j < MAX_INDIRECT_SEGMENTS; j++)
-			kfree(req->segments[j]);
+		/* Check that there is no request in use */
+		list_for_each_entry_safe(req, n, &ring->pending_free, free_list) {
+			list_del(&req->free_list);
 
-		for (j = 0; j < MAX_INDIRECT_PAGES; j++)
-			kfree(req->indirect_pages[j]);
+			for (j = 0; j < MAX_INDIRECT_SEGMENTS; j++)
+				kfree(req->segments[j]);
 
-		kfree(req);
-		i++;
-	}
+			for (j = 0; j < MAX_INDIRECT_PAGES; j++)
+				kfree(req->indirect_pages[j]);
+
+			kfree(req);
+			i++;
+		}
 
-	WARN_ON(i != (XEN_BLKIF_REQS_PER_PAGE * blkif->nr_ring_pages));
+		BUG_ON(atomic_read(&ring->persistent_gnt_in_use) != 0);
+		BUG_ON(!list_empty(&ring->persistent_purge_list));
+		BUG_ON(!RB_EMPTY_ROOT(&ring->persistent_gnts));
+		BUG_ON(!list_empty(&ring->free_pages));
+		BUG_ON(ring->free_pages_num != 0);
+		BUG_ON(ring->persistent_gnt_c != 0);
+		WARN_ON(i != (XEN_BLKIF_REQS_PER_PAGE * blkif->nr_ring_pages));
+		xen_blkif_put(blkif);
+	}
 	blkif->nr_ring_pages = 0;
+	/*
+	 * blkif->rings was allocated in connect_ring, so we should free it in
+	 * here.
+	 */
+	kfree(blkif->rings);
+	blkif->rings = NULL;
+	blkif->nr_rings = 0;
 
 	return 0;
 }
@@ -271,13 +318,6 @@ static void xen_blkif_free(struct xen_blkif *blkif)
 	xen_vbd_free(&blkif->vbd);
 
 	/* Make sure everything is drained before shutting down */
-	BUG_ON(blkif->persistent_gnt_c != 0);
-	BUG_ON(atomic_read(&blkif->persistent_gnt_in_use) != 0);
-	BUG_ON(blkif->free_pages_num != 0);
-	BUG_ON(!list_empty(&blkif->persistent_purge_list));
-	BUG_ON(!list_empty(&blkif->free_pages));
-	BUG_ON(!RB_EMPTY_ROOT(&blkif->persistent_gnts));
-
 	kmem_cache_free(xen_blkif_cachep, blkif);
 }
 
@@ -296,25 +336,38 @@ int __init xen_blkif_interface_init(void)
  *  sysfs interface for VBD I/O requests
  */
 
-#define VBD_SHOW(name, format, args...)					\
+#define VBD_SHOW_ALLRING(name, format)					\
 	static ssize_t show_##name(struct device *_dev,			\
 				   struct device_attribute *attr,	\
 				   char *buf)				\
 	{								\
 		struct xenbus_device *dev = to_xenbus_device(_dev);	\
 		struct backend_info *be = dev_get_drvdata(&dev->dev);	\
+		struct xen_blkif *blkif = be->blkif;			\
+		unsigned int i;						\
+		unsigned long long result = 0;				\
 									\
-		return sprintf(buf, format, ##args);			\
+		if (!blkif->rings)				\
+			goto out;					\
+									\
+		for (i = 0; i < blkif->nr_rings; i++) {		\
+			struct xen_blkif_ring *ring = &blkif->rings[i];	\
+									\
+			result += ring->st_##name;			\
+		}							\
+									\
+out:									\
+		return sprintf(buf, format, result);			\
 	}								\
 	static DEVICE_ATTR(name, S_IRUGO, show_##name, NULL)
 
-VBD_SHOW(oo_req,  "%llu\n", be->blkif->st_oo_req);
-VBD_SHOW(rd_req,  "%llu\n", be->blkif->st_rd_req);
-VBD_SHOW(wr_req,  "%llu\n", be->blkif->st_wr_req);
-VBD_SHOW(f_req,  "%llu\n", be->blkif->st_f_req);
-VBD_SHOW(ds_req,  "%llu\n", be->blkif->st_ds_req);
-VBD_SHOW(rd_sect, "%llu\n", be->blkif->st_rd_sect);
-VBD_SHOW(wr_sect, "%llu\n", be->blkif->st_wr_sect);
+VBD_SHOW_ALLRING(oo_req,  "%llu\n");
+VBD_SHOW_ALLRING(rd_req,  "%llu\n");
+VBD_SHOW_ALLRING(wr_req,  "%llu\n");
+VBD_SHOW_ALLRING(f_req,  "%llu\n");
+VBD_SHOW_ALLRING(ds_req,  "%llu\n");
+VBD_SHOW_ALLRING(rd_sect, "%llu\n");
+VBD_SHOW_ALLRING(wr_sect, "%llu\n");
 
 static struct attribute *xen_vbdstat_attrs[] = {
 	&dev_attr_oo_req.attr,
@@ -332,6 +385,18 @@ static struct attribute_group xen_vbdstat_group = {
 	.attrs = xen_vbdstat_attrs,
 };
 
+#define VBD_SHOW(name, format, args...)					\
+	static ssize_t show_##name(struct device *_dev,			\
+				   struct device_attribute *attr,	\
+				   char *buf)				\
+	{								\
+		struct xenbus_device *dev = to_xenbus_device(_dev);	\
+		struct backend_info *be = dev_get_drvdata(&dev->dev);	\
+									\
+		return sprintf(buf, format, ##args);			\
+	}								\
+	static DEVICE_ATTR(name, S_IRUGO, show_##name, NULL)
+
 VBD_SHOW(physical_device, "%x:%x\n", be->major, be->minor);
 VBD_SHOW(mode, "%s\n", be->mode);
 
@@ -440,11 +505,11 @@ static int xen_blkbk_remove(struct xenbus_device *dev)
 
 	dev_set_drvdata(&dev->dev, NULL);
 
-	if (be->blkif) {
+	if (be->blkif)
 		xen_blkif_disconnect(be->blkif);
-		xen_blkif_put(be->blkif);
-	}
 
+	/* Put the reference we set in xen_blkif_alloc(). */
+	xen_blkif_put(be->blkif);
 	kfree(be->mode);
 	kfree(be);
 	return 0;
@@ -553,6 +618,12 @@ static int xen_blkbk_probe(struct xenbus_device *dev,
 		goto fail;
 	}
 
+	/* Multi-queue: advertise how many queues are supported by us.*/
+	err = xenbus_printf(XBT_NIL, dev->nodename,
+			    "multi-queue-max-queues", "%u", xenblk_max_queues);
+	if (err)
+		pr_warn("Error writing multi-queue-max-queues\n");
+
 	/* setup back pointer */
 	be->blkif->be = be;
 
@@ -708,8 +779,14 @@ static void frontend_changed(struct xenbus_device *dev,
 		}
 
 		err = connect_ring(be);
-		if (err)
+		if (err) {
+			/*
+			 * Clean up so that memory resources can be used by
+			 * other devices. connect_ring reported already error.
+			 */
+			xen_blkif_disconnect(be->blkif);
 			break;
+		}
 		xen_update_blkif_status(be->blkif);
 		break;
 
@@ -825,50 +902,43 @@ static void connect(struct backend_info *be)
 	xenbus_transaction_end(xbt, 1);
 }
 
-
-static int connect_ring(struct backend_info *be)
+/*
+ * Each ring may have multi pages, depends on "ring-page-order".
+ */
+static int read_per_ring_refs(struct xen_blkif_ring *ring, const char *dir)
 {
-	struct xenbus_device *dev = be->dev;
 	unsigned int ring_ref[XENBUS_MAX_RING_GRANTS];
-	unsigned int evtchn, nr_grefs, ring_page_order;
-	unsigned int pers_grants;
-	char protocol[64] = "";
 	struct pending_req *req, *n;
 	int err, i, j;
+	struct xen_blkif *blkif = ring->blkif;
+	struct xenbus_device *dev = blkif->be->dev;
+	unsigned int ring_page_order, nr_grefs, evtchn;
 
-	pr_debug("%s %s\n", __func__, dev->otherend);
-
-	err = xenbus_scanf(XBT_NIL, dev->otherend, "event-channel", "%u",
+	err = xenbus_scanf(XBT_NIL, dir, "event-channel", "%u",
 			  &evtchn);
 	if (err != 1) {
 		err = -EINVAL;
-		xenbus_dev_fatal(dev, err, "reading %s/event-channel",
-				 dev->otherend);
+		xenbus_dev_fatal(dev, err, "reading %s/event-channel", dir);
 		return err;
 	}
-	pr_info("event-channel %u\n", evtchn);
 
 	err = xenbus_scanf(XBT_NIL, dev->otherend, "ring-page-order", "%u",
 			  &ring_page_order);
 	if (err != 1) {
-		err = xenbus_scanf(XBT_NIL, dev->otherend, "ring-ref",
-				  "%u", &ring_ref[0]);
+		err = xenbus_scanf(XBT_NIL, dir, "ring-ref", "%u", &ring_ref[0]);
 		if (err != 1) {
 			err = -EINVAL;
-			xenbus_dev_fatal(dev, err, "reading %s/ring-ref",
-					 dev->otherend);
+			xenbus_dev_fatal(dev, err, "reading %s/ring-ref", dir);
 			return err;
 		}
 		nr_grefs = 1;
-		pr_info("%s:using single page: ring-ref %d\n", dev->otherend,
-			ring_ref[0]);
 	} else {
 		unsigned int i;
 
 		if (ring_page_order > xen_blkif_max_ring_order) {
 			err = -EINVAL;
 			xenbus_dev_fatal(dev, err, "%s/request %d ring page order exceed max:%d",
-					 dev->otherend, ring_page_order,
+					 dir, ring_page_order,
 					 xen_blkif_max_ring_order);
 			return err;
 		}
@@ -878,52 +948,23 @@ static int connect_ring(struct backend_info *be)
 			char ring_ref_name[RINGREF_NAME_LEN];
 
 			snprintf(ring_ref_name, RINGREF_NAME_LEN, "ring-ref%u", i);
-			err = xenbus_scanf(XBT_NIL, dev->otherend, ring_ref_name,
+			err = xenbus_scanf(XBT_NIL, dir, ring_ref_name,
 					   "%u", &ring_ref[i]);
 			if (err != 1) {
 				err = -EINVAL;
 				xenbus_dev_fatal(dev, err, "reading %s/%s",
-						 dev->otherend, ring_ref_name);
+						 dir, ring_ref_name);
 				return err;
 			}
-			pr_info("ring-ref%u: %u\n", i, ring_ref[i]);
 		}
 	}
-
-	be->blkif->blk_protocol = BLKIF_PROTOCOL_DEFAULT;
-	err = xenbus_gather(XBT_NIL, dev->otherend, "protocol",
-			    "%63s", protocol, NULL);
-	if (err)
-		strcpy(protocol, "unspecified, assuming default");
-	else if (0 == strcmp(protocol, XEN_IO_PROTO_ABI_NATIVE))
-		be->blkif->blk_protocol = BLKIF_PROTOCOL_NATIVE;
-	else if (0 == strcmp(protocol, XEN_IO_PROTO_ABI_X86_32))
-		be->blkif->blk_protocol = BLKIF_PROTOCOL_X86_32;
-	else if (0 == strcmp(protocol, XEN_IO_PROTO_ABI_X86_64))
-		be->blkif->blk_protocol = BLKIF_PROTOCOL_X86_64;
-	else {
-		xenbus_dev_fatal(dev, err, "unknown fe protocol %s", protocol);
-		return -1;
-	}
-	err = xenbus_gather(XBT_NIL, dev->otherend,
-			    "feature-persistent", "%u",
-			    &pers_grants, NULL);
-	if (err)
-		pers_grants = 0;
-
-	be->blkif->vbd.feature_gnt_persistent = pers_grants;
-	be->blkif->vbd.overflow_max_grants = 0;
-	be->blkif->nr_ring_pages = nr_grefs;
-
-	pr_info("ring-pages:%d, event-channel %d, protocol %d (%s) %s\n",
-		nr_grefs, evtchn, be->blkif->blk_protocol, protocol,
-		pers_grants ? "persistent grants" : "");
+	blkif->nr_ring_pages = nr_grefs;
 
 	for (i = 0; i < nr_grefs * XEN_BLKIF_REQS_PER_PAGE; i++) {
 		req = kzalloc(sizeof(*req), GFP_KERNEL);
 		if (!req)
 			goto fail;
-		list_add_tail(&req->free_list, &be->blkif->pending_free);
+		list_add_tail(&req->free_list, &ring->pending_free);
 		for (j = 0; j < MAX_INDIRECT_SEGMENTS; j++) {
 			req->segments[j] = kzalloc(sizeof(*req->segments[0]), GFP_KERNEL);
 			if (!req->segments[j])
@@ -938,7 +979,7 @@ static int connect_ring(struct backend_info *be)
 	}
 
 	/* Map the shared frame, irq etc. */
-	err = xen_blkif_map(be->blkif, ring_ref, nr_grefs, evtchn);
+	err = xen_blkif_map(ring, ring_ref, nr_grefs, evtchn);
 	if (err) {
 		xenbus_dev_fatal(dev, err, "mapping ring-ref port %u", evtchn);
 		return err;
@@ -947,7 +988,7 @@ static int connect_ring(struct backend_info *be)
 	return 0;
 
 fail:
-	list_for_each_entry_safe(req, n, &be->blkif->pending_free, free_list) {
+	list_for_each_entry_safe(req, n, &ring->pending_free, free_list) {
 		list_del(&req->free_list);
 		for (j = 0; j < MAX_INDIRECT_SEGMENTS; j++) {
 			if (!req->segments[j])
@@ -962,6 +1003,93 @@ static int connect_ring(struct backend_info *be)
 		kfree(req);
 	}
 	return -ENOMEM;
+
+}
+
+static int connect_ring(struct backend_info *be)
+{
+	struct xenbus_device *dev = be->dev;
+	unsigned int pers_grants;
+	char protocol[64] = "";
+	int err, i;
+	char *xspath;
+	size_t xspathsize;
+	const size_t xenstore_path_ext_size = 11; /* sufficient for "/queue-NNN" */
+	unsigned int requested_num_queues = 0;
+
+	pr_debug("%s %s\n", __func__, dev->otherend);
+
+	be->blkif->blk_protocol = BLKIF_PROTOCOL_DEFAULT;
+	err = xenbus_gather(XBT_NIL, dev->otherend, "protocol",
+			    "%63s", protocol, NULL);
+	if (err)
+		strcpy(protocol, "unspecified, assuming default");
+	else if (0 == strcmp(protocol, XEN_IO_PROTO_ABI_NATIVE))
+		be->blkif->blk_protocol = BLKIF_PROTOCOL_NATIVE;
+	else if (0 == strcmp(protocol, XEN_IO_PROTO_ABI_X86_32))
+		be->blkif->blk_protocol = BLKIF_PROTOCOL_X86_32;
+	else if (0 == strcmp(protocol, XEN_IO_PROTO_ABI_X86_64))
+		be->blkif->blk_protocol = BLKIF_PROTOCOL_X86_64;
+	else {
+		xenbus_dev_fatal(dev, err, "unknown fe protocol %s", protocol);
+		return -ENOSYS;
+	}
+	err = xenbus_gather(XBT_NIL, dev->otherend,
+			    "feature-persistent", "%u",
+			    &pers_grants, NULL);
+	if (err)
+		pers_grants = 0;
+
+	be->blkif->vbd.feature_gnt_persistent = pers_grants;
+	be->blkif->vbd.overflow_max_grants = 0;
+
+	/*
+	 * Read the number of hardware queues from frontend.
+	 */
+	err = xenbus_scanf(XBT_NIL, dev->otherend, "multi-queue-num-queues",
+			   "%u", &requested_num_queues);
+	if (err < 0) {
+		requested_num_queues = 1;
+	} else {
+		if (requested_num_queues > xenblk_max_queues
+		    || requested_num_queues == 0) {
+			/* Buggy or malicious guest. */
+			xenbus_dev_fatal(dev, err,
+					"guest requested %u queues, exceeding the maximum of %u.",
+					requested_num_queues, xenblk_max_queues);
+			return -ENOSYS;
+		}
+	}
+	be->blkif->nr_rings = requested_num_queues;
+	if (xen_blkif_alloc_rings(be->blkif))
+		return -ENOMEM;
+
+	pr_info("%s: using %d queues, protocol %d (%s) %s\n", dev->nodename,
+		 be->blkif->nr_rings, be->blkif->blk_protocol, protocol,
+		 pers_grants ? "persistent grants" : "");
+
+	if (be->blkif->nr_rings == 1)
+		return read_per_ring_refs(&be->blkif->rings[0], dev->otherend);
+	else {
+		xspathsize = strlen(dev->otherend) + xenstore_path_ext_size;
+		xspath = kmalloc(xspathsize, GFP_KERNEL);
+		if (!xspath) {
+			xenbus_dev_fatal(dev, -ENOMEM, "reading ring references");
+			return -ENOMEM;
+		}
+
+		for (i = 0; i < be->blkif->nr_rings; i++) {
+			memset(xspath, 0, xspathsize);
+			snprintf(xspath, xspathsize, "%s/queue-%u", dev->otherend, i);
+			err = read_per_ring_refs(&be->blkif->rings[i], xspath);
+			if (err) {
+				kfree(xspath);
+				return err;
+			}
+		}
+		kfree(xspath);
+	}
+	return 0;
 }
 
 static const struct xenbus_device_id xen_blkbk_ids[] = {

drivers/block/xen-blkfront.c

@@ -60,6 +60,20 @@
 
 #include <asm/xen/hypervisor.h>
 
+/*
+ * The minimal size of segment supported by the block framework is PAGE_SIZE.
+ * When Linux is using a different page size than Xen, it may not be possible
+ * to put all the data in a single segment.
+ * This can happen when the backend doesn't support indirect descriptor and
+ * therefore the maximum amount of data that a request can carry is
+ * BLKIF_MAX_SEGMENTS_PER_REQUEST * XEN_PAGE_SIZE = 44KB
+ *
+ * Note that we only support one extra request. So the Linux page size
+ * should be <= ( 2 * BLKIF_MAX_SEGMENTS_PER_REQUEST * XEN_PAGE_SIZE) =
+ * 88KB.
+ */
+#define HAS_EXTRA_REQ (BLKIF_MAX_SEGMENTS_PER_REQUEST < XEN_PFN_PER_PAGE)
+
 enum blkif_state {
 	BLKIF_STATE_DISCONNECTED,
 	BLKIF_STATE_CONNECTED,
@@ -72,6 +86,13 @@ struct grant {
 	struct list_head node;
 };
 
+enum blk_req_status {
+	REQ_WAITING,
+	REQ_DONE,
+	REQ_ERROR,
+	REQ_EOPNOTSUPP,
+};
+
 struct blk_shadow {
 	struct blkif_request req;
 	struct request *request;
@@ -79,6 +100,14 @@ struct blk_shadow {
 	struct grant **indirect_grants;
 	struct scatterlist *sg;
 	unsigned int num_sg;
+	enum blk_req_status status;
+
+	#define NO_ASSOCIATED_ID ~0UL
+	/*
+	 * Id of the sibling if we ever need 2 requests when handling a
+	 * block I/O request
+	 */
+	unsigned long associated_id;
 };
 
 struct split_bio {
@@ -99,6 +128,10 @@ static unsigned int xen_blkif_max_segments = 32;
 module_param_named(max, xen_blkif_max_segments, int, S_IRUGO);
 MODULE_PARM_DESC(max, "Maximum amount of segments in indirect requests (default is 32)");
 
+static unsigned int xen_blkif_max_queues = 4;
+module_param_named(max_queues, xen_blkif_max_queues, uint, S_IRUGO);
+MODULE_PARM_DESC(max_queues, "Maximum number of hardware queues/rings used per virtual disk");
+
 /*
  * Maximum order of pages to be used for the shared ring between front and
  * backend, 4KB page granularity is used.
@@ -114,10 +147,35 @@ MODULE_PARM_DESC(max_ring_page_order, "Maximum order of pages to be used for the
 	__CONST_RING_SIZE(blkif, XEN_PAGE_SIZE * XENBUS_MAX_RING_GRANTS)
 
 /*
- * ring-ref%i i=(-1UL) would take 11 characters + 'ring-ref' is 8, so 19
- * characters are enough. Define to 20 to keep consist with backend.
+ * ring-ref%u i=(-1UL) would take 11 characters + 'ring-ref' is 8, so 19
+ * characters are enough. Define to 20 to keep consistent with backend.
  */
 #define RINGREF_NAME_LEN (20)
+/*
+ * queue-%u would take 7 + 10(UINT_MAX) = 17 characters.
+ */
+#define QUEUE_NAME_LEN (17)
+
+/*
+ *  Per-ring info.
+ *  Every blkfront device can associate with one or more blkfront_ring_info,
+ *  depending on how many hardware queues/rings to be used.
+ */
+struct blkfront_ring_info {
+	/* Lock to protect data in every ring buffer. */
+	spinlock_t ring_lock;
+	struct blkif_front_ring ring;
+	unsigned int ring_ref[XENBUS_MAX_RING_GRANTS];
+	unsigned int evtchn, irq;
+	struct work_struct work;
+	struct gnttab_free_callback callback;
+	struct blk_shadow shadow[BLK_MAX_RING_SIZE];
+	struct list_head indirect_pages;
+	struct list_head grants;
+	unsigned int persistent_gnts_c;
+	unsigned long shadow_free;
+	struct blkfront_info *dev_info;
+};
 
 /*
  * We have one of these per vbd, whether ide, scsi or 'other'.  They
@@ -126,25 +184,15 @@ MODULE_PARM_DESC(max_ring_page_order, "Maximum order of pages to be used for the
  */
 struct blkfront_info
 {
-	spinlock_t io_lock;
 	struct mutex mutex;
 	struct xenbus_device *xbdev;
 	struct gendisk *gd;
 	int vdevice;
 	blkif_vdev_t handle;
 	enum blkif_state connected;
-	int ring_ref[XENBUS_MAX_RING_GRANTS];
+	/* Number of pages per ring buffer. */
 	unsigned int nr_ring_pages;
-	struct blkif_front_ring ring;
-	unsigned int evtchn, irq;
 	struct request_queue *rq;
-	struct work_struct work;
-	struct gnttab_free_callback callback;
-	struct blk_shadow shadow[BLK_MAX_RING_SIZE];
-	struct list_head grants;
-	struct list_head indirect_pages;
-	unsigned int persistent_gnts_c;
-	unsigned long shadow_free;
 	unsigned int feature_flush;
 	unsigned int feature_discard:1;
 	unsigned int feature_secdiscard:1;
@@ -155,6 +203,8 @@ struct blkfront_info
 	unsigned int max_indirect_segments;
 	int is_ready;
 	struct blk_mq_tag_set tag_set;
+	struct blkfront_ring_info *rinfo;
+	unsigned int nr_rings;
 };
 
 static unsigned int nr_minors;
@@ -198,38 +248,40 @@ static DEFINE_SPINLOCK(minor_lock);
 
 #define GREFS(_psegs)	((_psegs) * GRANTS_PER_PSEG)
 
-static int blkfront_setup_indirect(struct blkfront_info *info);
-static int blkfront_gather_backend_features(struct blkfront_info *info);
+static int blkfront_setup_indirect(struct blkfront_ring_info *rinfo);
+static void blkfront_gather_backend_features(struct blkfront_info *info);
 
-static int get_id_from_freelist(struct blkfront_info *info)
+static int get_id_from_freelist(struct blkfront_ring_info *rinfo)
 {
-	unsigned long free = info->shadow_free;
-	BUG_ON(free >= BLK_RING_SIZE(info));
-	info->shadow_free = info->shadow[free].req.u.rw.id;
-	info->shadow[free].req.u.rw.id = 0x0fffffee; /* debug */
+	unsigned long free = rinfo->shadow_free;
+
+	BUG_ON(free >= BLK_RING_SIZE(rinfo->dev_info));
+	rinfo->shadow_free = rinfo->shadow[free].req.u.rw.id;
+	rinfo->shadow[free].req.u.rw.id = 0x0fffffee; /* debug */
 	return free;
 }
 
-static int add_id_to_freelist(struct blkfront_info *info,
-			       unsigned long id)
+static int add_id_to_freelist(struct blkfront_ring_info *rinfo,
+			      unsigned long id)
 {
-	if (info->shadow[id].req.u.rw.id != id)
+	if (rinfo->shadow[id].req.u.rw.id != id)
 		return -EINVAL;
-	if (info->shadow[id].request == NULL)
+	if (rinfo->shadow[id].request == NULL)
 		return -EINVAL;
-	info->shadow[id].req.u.rw.id  = info->shadow_free;
-	info->shadow[id].request = NULL;
-	info->shadow_free = id;
+	rinfo->shadow[id].req.u.rw.id  = rinfo->shadow_free;
+	rinfo->shadow[id].request = NULL;
+	rinfo->shadow_free = id;
 	return 0;
 }
 
-static int fill_grant_buffer(struct blkfront_info *info, int num)
+static int fill_grant_buffer(struct blkfront_ring_info *rinfo, int num)
 {
+	struct blkfront_info *info = rinfo->dev_info;
 	struct page *granted_page;
 	struct grant *gnt_list_entry, *n;
 	int i = 0;
 
-	while(i < num) {
+	while (i < num) {
 		gnt_list_entry = kzalloc(sizeof(struct grant), GFP_NOIO);
 		if (!gnt_list_entry)
 			goto out_of_memory;
@@ -244,7 +296,7 @@ static int fill_grant_buffer(struct blkfront_info *info, int num)
 		}
 
 		gnt_list_entry->gref = GRANT_INVALID_REF;
-		list_add(&gnt_list_entry->node, &info->grants);
+		list_add(&gnt_list_entry->node, &rinfo->grants);
 		i++;
 	}
 
@@ -252,7 +304,7 @@ static int fill_grant_buffer(struct blkfront_info *info, int num)
 
 out_of_memory:
 	list_for_each_entry_safe(gnt_list_entry, n,
-	                         &info->grants, node) {
+	                         &rinfo->grants, node) {
 		list_del(&gnt_list_entry->node);
 		if (info->feature_persistent)
 			__free_page(gnt_list_entry->page);
@@ -263,17 +315,17 @@ static int fill_grant_buffer(struct blkfront_info *info, int num)
 	return -ENOMEM;
 }
 
-static struct grant *get_free_grant(struct blkfront_info *info)
+static struct grant *get_free_grant(struct blkfront_ring_info *rinfo)
 {
 	struct grant *gnt_list_entry;
 
-	BUG_ON(list_empty(&info->grants));
-	gnt_list_entry = list_first_entry(&info->grants, struct grant,
+	BUG_ON(list_empty(&rinfo->grants));
+	gnt_list_entry = list_first_entry(&rinfo->grants, struct grant,
 					  node);
 	list_del(&gnt_list_entry->node);
 
 	if (gnt_list_entry->gref != GRANT_INVALID_REF)
-		info->persistent_gnts_c--;
+		rinfo->persistent_gnts_c--;
 
 	return gnt_list_entry;
 }
@@ -289,9 +341,10 @@ static inline void grant_foreign_access(const struct grant *gnt_list_entry,
 
 static struct grant *get_grant(grant_ref_t *gref_head,
 			       unsigned long gfn,
-			       struct blkfront_info *info)
+			       struct blkfront_ring_info *rinfo)
 {
-	struct grant *gnt_list_entry = get_free_grant(info);
+	struct grant *gnt_list_entry = get_free_grant(rinfo);
+	struct blkfront_info *info = rinfo->dev_info;
 
 	if (gnt_list_entry->gref != GRANT_INVALID_REF)
 		return gnt_list_entry;
@@ -312,9 +365,10 @@ static struct grant *get_grant(grant_ref_t *gref_head,
 }
 
 static struct grant *get_indirect_grant(grant_ref_t *gref_head,
-					struct blkfront_info *info)
+					struct blkfront_ring_info *rinfo)
 {
-	struct grant *gnt_list_entry = get_free_grant(info);
+	struct grant *gnt_list_entry = get_free_grant(rinfo);
+	struct blkfront_info *info = rinfo->dev_info;
 
 	if (gnt_list_entry->gref != GRANT_INVALID_REF)
 		return gnt_list_entry;
@@ -326,8 +380,8 @@ static struct grant *get_indirect_grant(grant_ref_t *gref_head,
 		struct page *indirect_page;
 
 		/* Fetch a pre-allocated page to use for indirect grefs */
-		BUG_ON(list_empty(&info->indirect_pages));
-		indirect_page = list_first_entry(&info->indirect_pages,
+		BUG_ON(list_empty(&rinfo->indirect_pages));
+		indirect_page = list_first_entry(&rinfo->indirect_pages,
 						 struct page, lru);
 		list_del(&indirect_page->lru);
 		gnt_list_entry->page = indirect_page;
@@ -403,8 +457,8 @@ static void xlbd_release_minors(unsigned int minor, unsigned int nr)
 
 static void blkif_restart_queue_callback(void *arg)
 {
-	struct blkfront_info *info = (struct blkfront_info *)arg;
-	schedule_work(&info->work);
+	struct blkfront_ring_info *rinfo = (struct blkfront_ring_info *)arg;
+	schedule_work(&rinfo->work);
 }
 
 static int blkif_getgeo(struct block_device *bd, struct hd_geometry *hg)
@@ -456,16 +510,33 @@ static int blkif_ioctl(struct block_device *bdev, fmode_t mode,
 	return 0;
 }
 
-static int blkif_queue_discard_req(struct request *req)
+static unsigned long blkif_ring_get_request(struct blkfront_ring_info *rinfo,
+					    struct request *req,
+					    struct blkif_request **ring_req)
 {
-	struct blkfront_info *info = req->rq_disk->private_data;
+	unsigned long id;
+
+	*ring_req = RING_GET_REQUEST(&rinfo->ring, rinfo->ring.req_prod_pvt);
+	rinfo->ring.req_prod_pvt++;
+
+	id = get_id_from_freelist(rinfo);
+	rinfo->shadow[id].request = req;
+	rinfo->shadow[id].status = REQ_WAITING;
+	rinfo->shadow[id].associated_id = NO_ASSOCIATED_ID;
+
+	(*ring_req)->u.rw.id = id;
+
+	return id;
+}
+
+static int blkif_queue_discard_req(struct request *req, struct blkfront_ring_info *rinfo)
+{
+	struct blkfront_info *info = rinfo->dev_info;
 	struct blkif_request *ring_req;
 	unsigned long id;
 
 	/* Fill out a communications ring structure. */
-	ring_req = RING_GET_REQUEST(&info->ring, info->ring.req_prod_pvt);
-	id = get_id_from_freelist(info);
-	info->shadow[id].request = req;
+	id = blkif_ring_get_request(rinfo, req, &ring_req);
 
 	ring_req->operation = BLKIF_OP_DISCARD;
 	ring_req->u.discard.nr_sectors = blk_rq_sectors(req);
@@ -476,10 +547,8 @@ static int blkif_queue_discard_req(struct request *req)
 	else
 		ring_req->u.discard.flag = 0;
 
-	info->ring.req_prod_pvt++;
-
 	/* Keep a private copy so we can reissue requests when recovering. */
-	info->shadow[id].req = *ring_req;
+	rinfo->shadow[id].req = *ring_req;
 
 	return 0;
 }
@@ -487,7 +556,7 @@ static int blkif_queue_discard_req(struct request *req)
 struct setup_rw_req {
 	unsigned int grant_idx;
 	struct blkif_request_segment *segments;
-	struct blkfront_info *info;
+	struct blkfront_ring_info *rinfo;
 	struct blkif_request *ring_req;
 	grant_ref_t gref_head;
 	unsigned int id;
@@ -495,6 +564,9 @@ struct setup_rw_req {
 	bool need_copy;
 	unsigned int bvec_off;
 	char *bvec_data;
+
+	bool require_extra_req;
+	struct blkif_request *extra_ring_req;
 };
 
 static void blkif_setup_rw_req_grant(unsigned long gfn, unsigned int offset,
@@ -507,8 +579,24 @@ static void blkif_setup_rw_req_grant(unsigned long gfn, unsigned int offset,
 	/* Convenient aliases */
 	unsigned int grant_idx = setup->grant_idx;
 	struct blkif_request *ring_req = setup->ring_req;
-	struct blkfront_info *info = setup->info;
-	struct blk_shadow *shadow = &info->shadow[setup->id];
+	struct blkfront_ring_info *rinfo = setup->rinfo;
+	/*
+	 * We always use the shadow of the first request to store the list
+	 * of grant associated to the block I/O request. This made the
+	 * completion more easy to handle even if the block I/O request is
+	 * split.
+	 */
+	struct blk_shadow *shadow = &rinfo->shadow[setup->id];
+
+	if (unlikely(setup->require_extra_req &&
+		     grant_idx >= BLKIF_MAX_SEGMENTS_PER_REQUEST)) {
+		/*
+		 * We are using the second request, setup grant_idx
+		 * to be the index of the segment array.
+		 */
+		grant_idx -= BLKIF_MAX_SEGMENTS_PER_REQUEST;
+		ring_req = setup->extra_ring_req;
+	}
 
 	if ((ring_req->operation == BLKIF_OP_INDIRECT) &&
 	    (grant_idx % GRANTS_PER_INDIRECT_FRAME == 0)) {
@@ -516,15 +604,19 @@ static void blkif_setup_rw_req_grant(unsigned long gfn, unsigned int offset,
 			kunmap_atomic(setup->segments);
 
 		n = grant_idx / GRANTS_PER_INDIRECT_FRAME;
-		gnt_list_entry = get_indirect_grant(&setup->gref_head, info);
+		gnt_list_entry = get_indirect_grant(&setup->gref_head, rinfo);
 		shadow->indirect_grants[n] = gnt_list_entry;
 		setup->segments = kmap_atomic(gnt_list_entry->page);
 		ring_req->u.indirect.indirect_grefs[n] = gnt_list_entry->gref;
 	}
 
-	gnt_list_entry = get_grant(&setup->gref_head, gfn, info);
+	gnt_list_entry = get_grant(&setup->gref_head, gfn, rinfo);
 	ref = gnt_list_entry->gref;
-	shadow->grants_used[grant_idx] = gnt_list_entry;
+	/*
+	 * All the grants are stored in the shadow of the first
+	 * request. Therefore we have to use the global index.
+	 */
+	shadow->grants_used[setup->grant_idx] = gnt_list_entry;
 
 	if (setup->need_copy) {
 		void *shared_data;
@@ -566,16 +658,36 @@ static void blkif_setup_rw_req_grant(unsigned long gfn, unsigned int offset,
 	(setup->grant_idx)++;
 }
 
-static int blkif_queue_rw_req(struct request *req)
+static void blkif_setup_extra_req(struct blkif_request *first,
+				  struct blkif_request *second)
 {
-	struct blkfront_info *info = req->rq_disk->private_data;
-	struct blkif_request *ring_req;
-	unsigned long id;
+	uint16_t nr_segments = first->u.rw.nr_segments;
+
+	/*
+	 * The second request is only present when the first request uses
+	 * all its segments. It's always the continuity of the first one.
+	 */
+	first->u.rw.nr_segments = BLKIF_MAX_SEGMENTS_PER_REQUEST;
+
+	second->u.rw.nr_segments = nr_segments - BLKIF_MAX_SEGMENTS_PER_REQUEST;
+	second->u.rw.sector_number = first->u.rw.sector_number +
+		(BLKIF_MAX_SEGMENTS_PER_REQUEST * XEN_PAGE_SIZE) / 512;
+
+	second->u.rw.handle = first->u.rw.handle;
+	second->operation = first->operation;
+}
+
+static int blkif_queue_rw_req(struct request *req, struct blkfront_ring_info *rinfo)
+{
+	struct blkfront_info *info = rinfo->dev_info;
+	struct blkif_request *ring_req, *extra_ring_req = NULL;
+	unsigned long id, extra_id = NO_ASSOCIATED_ID;
+	bool require_extra_req = false;
 	int i;
 	struct setup_rw_req setup = {
 		.grant_idx = 0,
 		.segments = NULL,
-		.info = info,
+		.rinfo = rinfo,
 		.need_copy = rq_data_dir(req) && info->feature_persistent,
 	};
 
@@ -584,7 +696,6 @@ static int blkif_queue_rw_req(struct request *req)
 	 * existing persistent grants, or if we have to get new grants,
 	 * as there are not sufficiently many free.
 	 */
-	bool new_persistent_gnts;
 	struct scatterlist *sg;
 	int num_sg, max_grefs, num_grant;
 
@@ -596,41 +707,36 @@ static int blkif_queue_rw_req(struct request *req)
 		 */
 		max_grefs += INDIRECT_GREFS(max_grefs);
 
-	/* Check if we have enough grants to allocate a requests */
-	if (info->persistent_gnts_c < max_grefs) {
-		new_persistent_gnts = 1;
-		if (gnttab_alloc_grant_references(
-		    max_grefs - info->persistent_gnts_c,
-		    &setup.gref_head) < 0) {
+	/*
+	 * We have to reserve 'max_grefs' grants because persistent
+	 * grants are shared by all rings.
+	 */
+	if (max_grefs > 0)
+		if (gnttab_alloc_grant_references(max_grefs, &setup.gref_head) < 0) {
 			gnttab_request_free_callback(
-				&info->callback,
+				&rinfo->callback,
 				blkif_restart_queue_callback,
-				info,
+				rinfo,
 				max_grefs);
 			return 1;
 		}
-	} else
-		new_persistent_gnts = 0;
 
 	/* Fill out a communications ring structure. */
-	ring_req = RING_GET_REQUEST(&info->ring, info->ring.req_prod_pvt);
-	id = get_id_from_freelist(info);
-	info->shadow[id].request = req;
-
-	BUG_ON(info->max_indirect_segments == 0 &&
-	       GREFS(req->nr_phys_segments) > BLKIF_MAX_SEGMENTS_PER_REQUEST);
-	BUG_ON(info->max_indirect_segments &&
-	       GREFS(req->nr_phys_segments) > info->max_indirect_segments);
+	id = blkif_ring_get_request(rinfo, req, &ring_req);
 
-	num_sg = blk_rq_map_sg(req->q, req, info->shadow[id].sg);
+	num_sg = blk_rq_map_sg(req->q, req, rinfo->shadow[id].sg);
 	num_grant = 0;
 	/* Calculate the number of grant used */
-	for_each_sg(info->shadow[id].sg, sg, num_sg, i)
+	for_each_sg(rinfo->shadow[id].sg, sg, num_sg, i)
 	       num_grant += gnttab_count_grant(sg->offset, sg->length);
 
-	ring_req->u.rw.id = id;
-	info->shadow[id].num_sg = num_sg;
-	if (num_grant > BLKIF_MAX_SEGMENTS_PER_REQUEST) {
+	require_extra_req = info->max_indirect_segments == 0 &&
+		num_grant > BLKIF_MAX_SEGMENTS_PER_REQUEST;
+	BUG_ON(!HAS_EXTRA_REQ && require_extra_req);
+
+	rinfo->shadow[id].num_sg = num_sg;
+	if (num_grant > BLKIF_MAX_SEGMENTS_PER_REQUEST &&
+	    likely(!require_extra_req)) {
 		/*
 		 * The indirect operation can only be a BLKIF_OP_READ or
 		 * BLKIF_OP_WRITE
@@ -670,11 +776,31 @@ static int blkif_queue_rw_req(struct request *req)
 			}
 		}
 		ring_req->u.rw.nr_segments = num_grant;
+		if (unlikely(require_extra_req)) {
+			extra_id = blkif_ring_get_request(rinfo, req,
+							  &extra_ring_req);
+			/*
+			 * Only the first request contains the scatter-gather
+			 * list.
+			 */
+			rinfo->shadow[extra_id].num_sg = 0;
+
+			blkif_setup_extra_req(ring_req, extra_ring_req);
+
+			/* Link the 2 requests together */
+			rinfo->shadow[extra_id].associated_id = id;
+			rinfo->shadow[id].associated_id = extra_id;
+		}
 	}
 
 	setup.ring_req = ring_req;
 	setup.id = id;
-	for_each_sg(info->shadow[id].sg, sg, num_sg, i) {
+
+	setup.require_extra_req = require_extra_req;
+	if (unlikely(require_extra_req))
+		setup.extra_ring_req = extra_ring_req;
+
+	for_each_sg(rinfo->shadow[id].sg, sg, num_sg, i) {
 		BUG_ON(sg->offset + sg->length > PAGE_SIZE);
 
 		if (setup.need_copy) {
@@ -694,12 +820,12 @@ static int blkif_queue_rw_req(struct request *req)
 	if (setup.segments)
 		kunmap_atomic(setup.segments);
 
-	info->ring.req_prod_pvt++;
-
 	/* Keep a private copy so we can reissue requests when recovering. */
-	info->shadow[id].req = *ring_req;
+	rinfo->shadow[id].req = *ring_req;
+	if (unlikely(require_extra_req))
+		rinfo->shadow[extra_id].req = *extra_ring_req;
 
-	if (new_persistent_gnts)
+	if (max_grefs > 0)
 		gnttab_free_grant_references(setup.gref_head);
 
 	return 0;
@@ -711,27 +837,25 @@ static int blkif_queue_rw_req(struct request *req)
  *
  * @req: a request struct
  */
-static int blkif_queue_request(struct request *req)
+static int blkif_queue_request(struct request *req, struct blkfront_ring_info *rinfo)
 {
-	struct blkfront_info *info = req->rq_disk->private_data;
-
-	if (unlikely(info->connected != BLKIF_STATE_CONNECTED))
+	if (unlikely(rinfo->dev_info->connected != BLKIF_STATE_CONNECTED))
 		return 1;
 
 	if (unlikely(req->cmd_flags & (REQ_DISCARD | REQ_SECURE)))
-		return blkif_queue_discard_req(req);
+		return blkif_queue_discard_req(req, rinfo);
 	else
-		return blkif_queue_rw_req(req);
+		return blkif_queue_rw_req(req, rinfo);
 }
 
-static inline void flush_requests(struct blkfront_info *info)
+static inline void flush_requests(struct blkfront_ring_info *rinfo)
 {
 	int notify;
 
-	RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&info->ring, notify);
+	RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&rinfo->ring, notify);
 
 	if (notify)
-		notify_remote_via_irq(info->irq);
+		notify_remote_via_irq(rinfo->irq);
 }
 
 static inline bool blkif_request_flush_invalid(struct request *req,
@@ -745,38 +869,50 @@ static inline bool blkif_request_flush_invalid(struct request *req,
 }
 
 static int blkif_queue_rq(struct blk_mq_hw_ctx *hctx,
-			   const struct blk_mq_queue_data *qd)
+			  const struct blk_mq_queue_data *qd)
 {
-	struct blkfront_info *info = qd->rq->rq_disk->private_data;
+	unsigned long flags;
+	struct blkfront_ring_info *rinfo = (struct blkfront_ring_info *)hctx->driver_data;
 
 	blk_mq_start_request(qd->rq);
-	spin_lock_irq(&info->io_lock);
-	if (RING_FULL(&info->ring))
+	spin_lock_irqsave(&rinfo->ring_lock, flags);
+	if (RING_FULL(&rinfo->ring))
 		goto out_busy;
 
-	if (blkif_request_flush_invalid(qd->rq, info))
+	if (blkif_request_flush_invalid(qd->rq, rinfo->dev_info))
 		goto out_err;
 
-	if (blkif_queue_request(qd->rq))
+	if (blkif_queue_request(qd->rq, rinfo))
 		goto out_busy;
 
-	flush_requests(info);
-	spin_unlock_irq(&info->io_lock);
+	flush_requests(rinfo);
+	spin_unlock_irqrestore(&rinfo->ring_lock, flags);
 	return BLK_MQ_RQ_QUEUE_OK;
 
 out_err:
-	spin_unlock_irq(&info->io_lock);
+	spin_unlock_irqrestore(&rinfo->ring_lock, flags);
 	return BLK_MQ_RQ_QUEUE_ERROR;
 
 out_busy:
-	spin_unlock_irq(&info->io_lock);
+	spin_unlock_irqrestore(&rinfo->ring_lock, flags);
 	blk_mq_stop_hw_queue(hctx);
 	return BLK_MQ_RQ_QUEUE_BUSY;
 }
 
+static int blk_mq_init_hctx(struct blk_mq_hw_ctx *hctx, void *data,
+			    unsigned int index)
+{
+	struct blkfront_info *info = (struct blkfront_info *)data;
+
+	BUG_ON(info->nr_rings <= index);
+	hctx->driver_data = &info->rinfo[index];
+	return 0;
+}
+
 static struct blk_mq_ops blkfront_mq_ops = {
 	.queue_rq = blkif_queue_rq,
 	.map_queue = blk_mq_map_queue,
+	.init_hctx = blk_mq_init_hctx,
 };
 
 static int xlvbd_init_blk_queue(struct gendisk *gd, u16 sector_size,
@@ -788,19 +924,28 @@ static int xlvbd_init_blk_queue(struct gendisk *gd, u16 sector_size,
 
 	memset(&info->tag_set, 0, sizeof(info->tag_set));
 	info->tag_set.ops = &blkfront_mq_ops;
-	info->tag_set.nr_hw_queues = 1;
-	info->tag_set.queue_depth =  BLK_RING_SIZE(info);
+	info->tag_set.nr_hw_queues = info->nr_rings;
+	if (HAS_EXTRA_REQ && info->max_indirect_segments == 0) {
+		/*
+		 * When indirect descriptior is not supported, the I/O request
+		 * will be split between multiple request in the ring.
+		 * To avoid problems when sending the request, divide by
+		 * 2 the depth of the queue.
+		 */
+		info->tag_set.queue_depth =  BLK_RING_SIZE(info) / 2;
+	} else
+		info->tag_set.queue_depth = BLK_RING_SIZE(info);
 	info->tag_set.numa_node = NUMA_NO_NODE;
 	info->tag_set.flags = BLK_MQ_F_SHOULD_MERGE | BLK_MQ_F_SG_MERGE;
 	info->tag_set.cmd_size = 0;
 	info->tag_set.driver_data = info;
 
 	if (blk_mq_alloc_tag_set(&info->tag_set))
-		return -1;
+		return -EINVAL;
 	rq = blk_mq_init_queue(&info->tag_set);
 	if (IS_ERR(rq)) {
 		blk_mq_free_tag_set(&info->tag_set);
-		return -1;
+		return PTR_ERR(rq);
 	}
 
 	queue_flag_set_unlocked(QUEUE_FLAG_VIRT, rq);
@@ -1028,7 +1173,7 @@ static int xlvbd_alloc_gendisk(blkif_sector_t capacity,
 
 static void xlvbd_release_gendisk(struct blkfront_info *info)
 {
-	unsigned int minor, nr_minors;
+	unsigned int minor, nr_minors, i;
 
 	if (info->rq == NULL)
 		return;
@@ -1036,11 +1181,15 @@ static void xlvbd_release_gendisk(struct blkfront_info *info)
 	/* No more blkif_request(). */
 	blk_mq_stop_hw_queues(info->rq);
 
-	/* No more gnttab callback work. */
-	gnttab_cancel_free_callback(&info->callback);
+	for (i = 0; i < info->nr_rings; i++) {
+		struct blkfront_ring_info *rinfo = &info->rinfo[i];
 
-	/* Flush gnttab callback work. Must be done with no locks held. */
-	flush_work(&info->work);
+		/* No more gnttab callback work. */
+		gnttab_cancel_free_callback(&rinfo->callback);
+
+		/* Flush gnttab callback work. Must be done with no locks held. */
+		flush_work(&rinfo->work);
+	}
 
 	del_gendisk(info->gd);
 
@@ -1056,88 +1205,87 @@ static void xlvbd_release_gendisk(struct blkfront_info *info)
 	info->gd = NULL;
 }
 
-/* Must be called with io_lock holded */
-static void kick_pending_request_queues(struct blkfront_info *info)
+/* Already hold rinfo->ring_lock. */
+static inline void kick_pending_request_queues_locked(struct blkfront_ring_info *rinfo)
 {
-	if (!RING_FULL(&info->ring))
-		blk_mq_start_stopped_hw_queues(info->rq, true);
+	if (!RING_FULL(&rinfo->ring))
+		blk_mq_start_stopped_hw_queues(rinfo->dev_info->rq, true);
 }
 
-static void blkif_restart_queue(struct work_struct *work)
+static void kick_pending_request_queues(struct blkfront_ring_info *rinfo)
 {
-	struct blkfront_info *info = container_of(work, struct blkfront_info, work);
+	unsigned long flags;
 
-	spin_lock_irq(&info->io_lock);
-	if (info->connected == BLKIF_STATE_CONNECTED)
-		kick_pending_request_queues(info);
-	spin_unlock_irq(&info->io_lock);
+	spin_lock_irqsave(&rinfo->ring_lock, flags);
+	kick_pending_request_queues_locked(rinfo);
+	spin_unlock_irqrestore(&rinfo->ring_lock, flags);
 }
 
-static void blkif_free(struct blkfront_info *info, int suspend)
+static void blkif_restart_queue(struct work_struct *work)
 {
-	struct grant *persistent_gnt;
-	struct grant *n;
-	int i, j, segs;
+	struct blkfront_ring_info *rinfo = container_of(work, struct blkfront_ring_info, work);
 
-	/* Prevent new requests being issued until we fix things up. */
-	spin_lock_irq(&info->io_lock);
-	info->connected = suspend ?
-		BLKIF_STATE_SUSPENDED : BLKIF_STATE_DISCONNECTED;
-	/* No more blkif_request(). */
-	if (info->rq)
-		blk_mq_stop_hw_queues(info->rq);
+	if (rinfo->dev_info->connected == BLKIF_STATE_CONNECTED)
+		kick_pending_request_queues(rinfo);
+}
 
-	/* Remove all persistent grants */
-	if (!list_empty(&info->grants)) {
-		list_for_each_entry_safe(persistent_gnt, n,
-		                         &info->grants, node) {
-			list_del(&persistent_gnt->node);
-			if (persistent_gnt->gref != GRANT_INVALID_REF) {
-				gnttab_end_foreign_access(persistent_gnt->gref,
-				                          0, 0UL);
-				info->persistent_gnts_c--;
-			}
-			if (info->feature_persistent)
-				__free_page(persistent_gnt->page);
-			kfree(persistent_gnt);
-		}
-	}
-	BUG_ON(info->persistent_gnts_c != 0);
+static void blkif_free_ring(struct blkfront_ring_info *rinfo)
+{
+	struct grant *persistent_gnt, *n;
+	struct blkfront_info *info = rinfo->dev_info;
+	int i, j, segs;
 
 	/*
 	 * Remove indirect pages, this only happens when using indirect
 	 * descriptors but not persistent grants
 	 */
-	if (!list_empty(&info->indirect_pages)) {
+	if (!list_empty(&rinfo->indirect_pages)) {
 		struct page *indirect_page, *n;
 
 		BUG_ON(info->feature_persistent);
-		list_for_each_entry_safe(indirect_page, n, &info->indirect_pages, lru) {
+		list_for_each_entry_safe(indirect_page, n, &rinfo->indirect_pages, lru) {
 			list_del(&indirect_page->lru);
 			__free_page(indirect_page);
 		}
 	}
 
+	/* Remove all persistent grants. */
+	if (!list_empty(&rinfo->grants)) {
+		list_for_each_entry_safe(persistent_gnt, n,
+					 &rinfo->grants, node) {
+			list_del(&persistent_gnt->node);
+			if (persistent_gnt->gref != GRANT_INVALID_REF) {
+				gnttab_end_foreign_access(persistent_gnt->gref,
+							  0, 0UL);
+				rinfo->persistent_gnts_c--;
+			}
+			if (info->feature_persistent)
+				__free_page(persistent_gnt->page);
+			kfree(persistent_gnt);
+		}
+	}
+	BUG_ON(rinfo->persistent_gnts_c != 0);
+
 	for (i = 0; i < BLK_RING_SIZE(info); i++) {
 		/*
 		 * Clear persistent grants present in requests already
 		 * on the shared ring
 		 */
-		if (!info->shadow[i].request)
+		if (!rinfo->shadow[i].request)
 			goto free_shadow;
 
-		segs = info->shadow[i].req.operation == BLKIF_OP_INDIRECT ?
-		       info->shadow[i].req.u.indirect.nr_segments :
-		       info->shadow[i].req.u.rw.nr_segments;
+		segs = rinfo->shadow[i].req.operation == BLKIF_OP_INDIRECT ?
+		       rinfo->shadow[i].req.u.indirect.nr_segments :
+		       rinfo->shadow[i].req.u.rw.nr_segments;
 		for (j = 0; j < segs; j++) {
-			persistent_gnt = info->shadow[i].grants_used[j];
+			persistent_gnt = rinfo->shadow[i].grants_used[j];
 			gnttab_end_foreign_access(persistent_gnt->gref, 0, 0UL);
 			if (info->feature_persistent)
 				__free_page(persistent_gnt->page);
 			kfree(persistent_gnt);
 		}
 
-		if (info->shadow[i].req.operation != BLKIF_OP_INDIRECT)
+		if (rinfo->shadow[i].req.operation != BLKIF_OP_INDIRECT)
 			/*
 			 * If this is not an indirect operation don't try to
 			 * free indirect segments
@@ -1145,42 +1293,59 @@ static void blkif_free(struct blkfront_info *info, int suspend)
 			goto free_shadow;
 
 		for (j = 0; j < INDIRECT_GREFS(segs); j++) {
-			persistent_gnt = info->shadow[i].indirect_grants[j];
+			persistent_gnt = rinfo->shadow[i].indirect_grants[j];
 			gnttab_end_foreign_access(persistent_gnt->gref, 0, 0UL);
 			__free_page(persistent_gnt->page);
 			kfree(persistent_gnt);
 		}
 
 free_shadow:
-		kfree(info->shadow[i].grants_used);
-		info->shadow[i].grants_used = NULL;
-		kfree(info->shadow[i].indirect_grants);
-		info->shadow[i].indirect_grants = NULL;
-		kfree(info->shadow[i].sg);
-		info->shadow[i].sg = NULL;
+		kfree(rinfo->shadow[i].grants_used);
+		rinfo->shadow[i].grants_used = NULL;
+		kfree(rinfo->shadow[i].indirect_grants);
+		rinfo->shadow[i].indirect_grants = NULL;
+		kfree(rinfo->shadow[i].sg);
+		rinfo->shadow[i].sg = NULL;
 	}
 
 	/* No more gnttab callback work. */
-	gnttab_cancel_free_callback(&info->callback);
-	spin_unlock_irq(&info->io_lock);
+	gnttab_cancel_free_callback(&rinfo->callback);
 
 	/* Flush gnttab callback work. Must be done with no locks held. */
-	flush_work(&info->work);
+	flush_work(&rinfo->work);
 
 	/* Free resources associated with old device channel. */
 	for (i = 0; i < info->nr_ring_pages; i++) {
-		if (info->ring_ref[i] != GRANT_INVALID_REF) {
-			gnttab_end_foreign_access(info->ring_ref[i], 0, 0);
-			info->ring_ref[i] = GRANT_INVALID_REF;
+		if (rinfo->ring_ref[i] != GRANT_INVALID_REF) {
+			gnttab_end_foreign_access(rinfo->ring_ref[i], 0, 0);
+			rinfo->ring_ref[i] = GRANT_INVALID_REF;
 		}
 	}
-	free_pages((unsigned long)info->ring.sring, get_order(info->nr_ring_pages * PAGE_SIZE));
-	info->ring.sring = NULL;
+	free_pages((unsigned long)rinfo->ring.sring, get_order(info->nr_ring_pages * PAGE_SIZE));
+	rinfo->ring.sring = NULL;
 
-	if (info->irq)
-		unbind_from_irqhandler(info->irq, info);
-	info->evtchn = info->irq = 0;
+	if (rinfo->irq)
+		unbind_from_irqhandler(rinfo->irq, rinfo);
+	rinfo->evtchn = rinfo->irq = 0;
+}
 
+static void blkif_free(struct blkfront_info *info, int suspend)
+{
+	unsigned int i;
+
+	/* Prevent new requests being issued until we fix things up. */
+	info->connected = suspend ?
+		BLKIF_STATE_SUSPENDED : BLKIF_STATE_DISCONNECTED;
+	/* No more blkif_request(). */
+	if (info->rq)
+		blk_mq_stop_hw_queues(info->rq);
+
+	for (i = 0; i < info->nr_rings; i++)
+		blkif_free_ring(&info->rinfo[i]);
+
+	kfree(info->rinfo);
+	info->rinfo = NULL;
+	info->nr_rings = 0;
 }
 
 struct copy_from_grant {
@@ -1209,19 +1374,93 @@ static void blkif_copy_from_grant(unsigned long gfn, unsigned int offset,
 	kunmap_atomic(shared_data);
 }
 
-static void blkif_completion(struct blk_shadow *s, struct blkfront_info *info,
+static enum blk_req_status blkif_rsp_to_req_status(int rsp)
+{
+	switch (rsp)
+	{
+	case BLKIF_RSP_OKAY:
+		return REQ_DONE;
+	case BLKIF_RSP_EOPNOTSUPP:
+		return REQ_EOPNOTSUPP;
+	case BLKIF_RSP_ERROR:
+		/* Fallthrough. */
+	default:
+		return REQ_ERROR;
+	}
+}
+
+/*
+ * Get the final status of the block request based on two ring response
+ */
+static int blkif_get_final_status(enum blk_req_status s1,
+				  enum blk_req_status s2)
+{
+	BUG_ON(s1 == REQ_WAITING);
+	BUG_ON(s2 == REQ_WAITING);
+
+	if (s1 == REQ_ERROR || s2 == REQ_ERROR)
+		return BLKIF_RSP_ERROR;
+	else if (s1 == REQ_EOPNOTSUPP || s2 == REQ_EOPNOTSUPP)
+		return BLKIF_RSP_EOPNOTSUPP;
+	return BLKIF_RSP_OKAY;
+}
+
+static bool blkif_completion(unsigned long *id,
+			     struct blkfront_ring_info *rinfo,
 			     struct blkif_response *bret)
 {
 	int i = 0;
 	struct scatterlist *sg;
 	int num_sg, num_grant;
+	struct blkfront_info *info = rinfo->dev_info;
+	struct blk_shadow *s = &rinfo->shadow[*id];
 	struct copy_from_grant data = {
-		.s = s,
 		.grant_idx = 0,
 	};
 
 	num_grant = s->req.operation == BLKIF_OP_INDIRECT ?
 		s->req.u.indirect.nr_segments : s->req.u.rw.nr_segments;
+
+	/* The I/O request may be split in two. */
+	if (unlikely(s->associated_id != NO_ASSOCIATED_ID)) {
+		struct blk_shadow *s2 = &rinfo->shadow[s->associated_id];
+
+		/* Keep the status of the current response in shadow. */
+		s->status = blkif_rsp_to_req_status(bret->status);
+
+		/* Wait the second response if not yet here. */
+		if (s2->status == REQ_WAITING)
+			return 0;
+
+		bret->status = blkif_get_final_status(s->status,
+						      s2->status);
+
+		/*
+		 * All the grants is stored in the first shadow in order
+		 * to make the completion code simpler.
+		 */
+		num_grant += s2->req.u.rw.nr_segments;
+
+		/*
+		 * The two responses may not come in order. Only the
+		 * first request will store the scatter-gather list.
+		 */
+		if (s2->num_sg != 0) {
+			/* Update "id" with the ID of the first response. */
+			*id = s->associated_id;
+			s = s2;
+		}
+
+		/*
+		 * We don't need anymore the second request, so recycling
+		 * it now.
+		 */
+		if (add_id_to_freelist(rinfo, s->associated_id))
+			WARN(1, "%s: can't recycle the second part (id = %ld) of the request\n",
+			     info->gd->disk_name, s->associated_id);
+	}
+
+	data.s = s;
 	num_sg = s->num_sg;
 
 	if (bret->operation == BLKIF_OP_READ && info->feature_persistent) {
@@ -1252,8 +1491,8 @@ static void blkif_completion(struct blk_shadow *s, struct blkfront_info *info,
 			if (!info->feature_persistent)
 				pr_alert_ratelimited("backed has not unmapped grant: %u\n",
 						     s->grants_used[i]->gref);
-			list_add(&s->grants_used[i]->node, &info->grants);
-			info->persistent_gnts_c++;
+			list_add(&s->grants_used[i]->node, &rinfo->grants);
+			rinfo->persistent_gnts_c++;
 		} else {
 			/*
 			 * If the grant is not mapped by the backend we end the
@@ -1263,7 +1502,7 @@ static void blkif_completion(struct blk_shadow *s, struct blkfront_info *info,
 			 */
 			gnttab_end_foreign_access(s->grants_used[i]->gref, 0, 0UL);
 			s->grants_used[i]->gref = GRANT_INVALID_REF;
-			list_add_tail(&s->grants_used[i]->node, &info->grants);
+			list_add_tail(&s->grants_used[i]->node, &rinfo->grants);
 		}
 	}
 	if (s->req.operation == BLKIF_OP_INDIRECT) {
@@ -1272,8 +1511,8 @@ static void blkif_completion(struct blk_shadow *s, struct blkfront_info *info,
 				if (!info->feature_persistent)
 					pr_alert_ratelimited("backed has not unmapped grant: %u\n",
 							     s->indirect_grants[i]->gref);
-				list_add(&s->indirect_grants[i]->node, &info->grants);
-				info->persistent_gnts_c++;
+				list_add(&s->indirect_grants[i]->node, &rinfo->grants);
+				rinfo->persistent_gnts_c++;
 			} else {
 				struct page *indirect_page;
 
@@ -1284,13 +1523,15 @@ static void blkif_completion(struct blk_shadow *s, struct blkfront_info *info,
 				 */
 				if (!info->feature_persistent) {
 					indirect_page = s->indirect_grants[i]->page;
-					list_add(&indirect_page->lru, &info->indirect_pages);
+					list_add(&indirect_page->lru, &rinfo->indirect_pages);
 				}
 				s->indirect_grants[i]->gref = GRANT_INVALID_REF;
-				list_add_tail(&s->indirect_grants[i]->node, &info->grants);
+				list_add_tail(&s->indirect_grants[i]->node, &rinfo->grants);
 			}
 		}
 	}
+
+	return 1;
 }
 
 static irqreturn_t blkif_interrupt(int irq, void *dev_id)
@@ -1299,24 +1540,22 @@ static irqreturn_t blkif_interrupt(int irq, void *dev_id)
 	struct blkif_response *bret;
 	RING_IDX i, rp;
 	unsigned long flags;
-	struct blkfront_info *info = (struct blkfront_info *)dev_id;
+	struct blkfront_ring_info *rinfo = (struct blkfront_ring_info *)dev_id;
+	struct blkfront_info *info = rinfo->dev_info;
 	int error;
 
-	spin_lock_irqsave(&info->io_lock, flags);
-
-	if (unlikely(info->connected != BLKIF_STATE_CONNECTED)) {
-		spin_unlock_irqrestore(&info->io_lock, flags);
+	if (unlikely(info->connected != BLKIF_STATE_CONNECTED))
 		return IRQ_HANDLED;
-	}
 
+	spin_lock_irqsave(&rinfo->ring_lock, flags);
  again:
-	rp = info->ring.sring->rsp_prod;
+	rp = rinfo->ring.sring->rsp_prod;
 	rmb(); /* Ensure we see queued responses up to 'rp'. */
 
-	for (i = info->ring.rsp_cons; i != rp; i++) {
+	for (i = rinfo->ring.rsp_cons; i != rp; i++) {
 		unsigned long id;
 
-		bret = RING_GET_RESPONSE(&info->ring, i);
+		bret = RING_GET_RESPONSE(&rinfo->ring, i);
 		id   = bret->id;
 		/*
 		 * The backend has messed up and given us an id that we would
@@ -1330,12 +1569,18 @@ static irqreturn_t blkif_interrupt(int irq, void *dev_id)
 			 * the id is busted. */
 			continue;
 		}
-		req  = info->shadow[id].request;
+		req  = rinfo->shadow[id].request;
 
-		if (bret->operation != BLKIF_OP_DISCARD)
-			blkif_completion(&info->shadow[id], info, bret);
+		if (bret->operation != BLKIF_OP_DISCARD) {
+			/*
+			 * We may need to wait for an extra response if the
+			 * I/O request is split in 2
+			 */
+			if (!blkif_completion(&id, rinfo, bret))
+				continue;
+		}
 
-		if (add_id_to_freelist(info, id)) {
+		if (add_id_to_freelist(rinfo, id)) {
 			WARN(1, "%s: response to %s (id %ld) couldn't be recycled!\n",
 			     info->gd->disk_name, op_name(bret->operation), id);
 			continue;
@@ -1364,7 +1609,7 @@ static irqreturn_t blkif_interrupt(int irq, void *dev_id)
 				error = -EOPNOTSUPP;
 			}
 			if (unlikely(bret->status == BLKIF_RSP_ERROR &&
-				     info->shadow[id].req.u.rw.nr_segments == 0)) {
+				     rinfo->shadow[id].req.u.rw.nr_segments == 0)) {
 				printk(KERN_WARNING "blkfront: %s: empty %s op failed\n",
 				       info->gd->disk_name, op_name(bret->operation));
 				error = -EOPNOTSUPP;
@@ -1389,34 +1634,35 @@ static irqreturn_t blkif_interrupt(int irq, void *dev_id)
 		}
 	}
 
-	info->ring.rsp_cons = i;
+	rinfo->ring.rsp_cons = i;
 
-	if (i != info->ring.req_prod_pvt) {
+	if (i != rinfo->ring.req_prod_pvt) {
 		int more_to_do;
-		RING_FINAL_CHECK_FOR_RESPONSES(&info->ring, more_to_do);
+		RING_FINAL_CHECK_FOR_RESPONSES(&rinfo->ring, more_to_do);
 		if (more_to_do)
 			goto again;
 	} else
-		info->ring.sring->rsp_event = i + 1;
+		rinfo->ring.sring->rsp_event = i + 1;
 
-	kick_pending_request_queues(info);
+	kick_pending_request_queues_locked(rinfo);
 
-	spin_unlock_irqrestore(&info->io_lock, flags);
+	spin_unlock_irqrestore(&rinfo->ring_lock, flags);
 
 	return IRQ_HANDLED;
 }
 
 
 static int setup_blkring(struct xenbus_device *dev,
-			 struct blkfront_info *info)
+			 struct blkfront_ring_info *rinfo)
 {
 	struct blkif_sring *sring;
 	int err, i;
+	struct blkfront_info *info = rinfo->dev_info;
 	unsigned long ring_size = info->nr_ring_pages * XEN_PAGE_SIZE;
 	grant_ref_t gref[XENBUS_MAX_RING_GRANTS];
 
 	for (i = 0; i < info->nr_ring_pages; i++)
-		info->ring_ref[i] = GRANT_INVALID_REF;
+		rinfo->ring_ref[i] = GRANT_INVALID_REF;
 
 	sring = (struct blkif_sring *)__get_free_pages(GFP_NOIO | __GFP_HIGH,
 						       get_order(ring_size));
@@ -1425,29 +1671,29 @@ static int setup_blkring(struct xenbus_device *dev,
 		return -ENOMEM;
 	}
 	SHARED_RING_INIT(sring);
-	FRONT_RING_INIT(&info->ring, sring, ring_size);
+	FRONT_RING_INIT(&rinfo->ring, sring, ring_size);
 
-	err = xenbus_grant_ring(dev, info->ring.sring, info->nr_ring_pages, gref);
+	err = xenbus_grant_ring(dev, rinfo->ring.sring, info->nr_ring_pages, gref);
 	if (err < 0) {
 		free_pages((unsigned long)sring, get_order(ring_size));
-		info->ring.sring = NULL;
+		rinfo->ring.sring = NULL;
 		goto fail;
 	}
 	for (i = 0; i < info->nr_ring_pages; i++)
-		info->ring_ref[i] = gref[i];
+		rinfo->ring_ref[i] = gref[i];
 
-	err = xenbus_alloc_evtchn(dev, &info->evtchn);
+	err = xenbus_alloc_evtchn(dev, &rinfo->evtchn);
 	if (err)
 		goto fail;
 
-	err = bind_evtchn_to_irqhandler(info->evtchn, blkif_interrupt, 0,
-					"blkif", info);
+	err = bind_evtchn_to_irqhandler(rinfo->evtchn, blkif_interrupt, 0,
+					"blkif", rinfo);
 	if (err <= 0) {
 		xenbus_dev_fatal(dev, err,
 				 "bind_evtchn_to_irqhandler failed");
 		goto fail;
 	}
-	info->irq = err;
+	rinfo->irq = err;
 
 	return 0;
 fail:
@@ -1455,6 +1701,53 @@ static int setup_blkring(struct xenbus_device *dev,
 	return err;
 }
 
+/*
+ * Write out per-ring/queue nodes including ring-ref and event-channel, and each
+ * ring buffer may have multi pages depending on ->nr_ring_pages.
+ */
+static int write_per_ring_nodes(struct xenbus_transaction xbt,
+				struct blkfront_ring_info *rinfo, const char *dir)
+{
+	int err;
+	unsigned int i;
+	const char *message = NULL;
+	struct blkfront_info *info = rinfo->dev_info;
+
+	if (info->nr_ring_pages == 1) {
+		err = xenbus_printf(xbt, dir, "ring-ref", "%u", rinfo->ring_ref[0]);
+		if (err) {
+			message = "writing ring-ref";
+			goto abort_transaction;
+		}
+	} else {
+		for (i = 0; i < info->nr_ring_pages; i++) {
+			char ring_ref_name[RINGREF_NAME_LEN];
+
+			snprintf(ring_ref_name, RINGREF_NAME_LEN, "ring-ref%u", i);
+			err = xenbus_printf(xbt, dir, ring_ref_name,
+					    "%u", rinfo->ring_ref[i]);
+			if (err) {
+				message = "writing ring-ref";
+				goto abort_transaction;
+			}
+		}
+	}
+
+	err = xenbus_printf(xbt, dir, "event-channel", "%u", rinfo->evtchn);
+	if (err) {
+		message = "writing event-channel";
+		goto abort_transaction;
+	}
+
+	return 0;
+
+abort_transaction:
+	xenbus_transaction_end(xbt, 1);
+	if (message)
+		xenbus_dev_fatal(info->xbdev, err, "%s", message);
+
+	return err;
+}
 
 /* Common code used when first setting up, and when resuming. */
 static int talk_to_blkback(struct xenbus_device *dev,
@@ -1462,8 +1755,8 @@ static int talk_to_blkback(struct xenbus_device *dev,
 {
 	const char *message = NULL;
 	struct xenbus_transaction xbt;
-	int err, i;
-	unsigned int max_page_order = 0;
+	int err;
+	unsigned int i, max_page_order = 0;
 	unsigned int ring_page_order = 0;
 
 	err = xenbus_scanf(XBT_NIL, info->xbdev->otherend,
@@ -1475,10 +1768,14 @@ static int talk_to_blkback(struct xenbus_device *dev,
 		info->nr_ring_pages = 1 << ring_page_order;
 	}
 
-	/* Create shared ring, alloc event channel. */
-	err = setup_blkring(dev, info);
-	if (err)
-		goto out;
+	for (i = 0; i < info->nr_rings; i++) {
+		struct blkfront_ring_info *rinfo = &info->rinfo[i];
+
+		/* Create shared ring, alloc event channel. */
+		err = setup_blkring(dev, rinfo);
+		if (err)
+			goto destroy_blkring;
+	}
 
 again:
 	err = xenbus_transaction_start(&xbt);
@@ -1487,38 +1784,49 @@ static int talk_to_blkback(struct xenbus_device *dev,
 		goto destroy_blkring;
 	}
 
-	if (info->nr_ring_pages == 1) {
-		err = xenbus_printf(xbt, dev->nodename,
-				    "ring-ref", "%u", info->ring_ref[0]);
+	if (info->nr_ring_pages > 1) {
+		err = xenbus_printf(xbt, dev->nodename, "ring-page-order", "%u",
+				    ring_page_order);
 		if (err) {
-			message = "writing ring-ref";
+			message = "writing ring-page-order";
 			goto abort_transaction;
 		}
+	}
+
+	/* We already got the number of queues/rings in _probe */
+	if (info->nr_rings == 1) {
+		err = write_per_ring_nodes(xbt, &info->rinfo[0], dev->nodename);
+		if (err)
+			goto destroy_blkring;
 	} else {
-		err = xenbus_printf(xbt, dev->nodename,
-				    "ring-page-order", "%u", ring_page_order);
+		char *path;
+		size_t pathsize;
+
+		err = xenbus_printf(xbt, dev->nodename, "multi-queue-num-queues", "%u",
+				    info->nr_rings);
 		if (err) {
-			message = "writing ring-page-order";
+			message = "writing multi-queue-num-queues";
 			goto abort_transaction;
 		}
 
-		for (i = 0; i < info->nr_ring_pages; i++) {
-			char ring_ref_name[RINGREF_NAME_LEN];
+		pathsize = strlen(dev->nodename) + QUEUE_NAME_LEN;
+		path = kmalloc(pathsize, GFP_KERNEL);
+		if (!path) {
+			err = -ENOMEM;
+			message = "ENOMEM while writing ring references";
+			goto abort_transaction;
+		}
 
-			snprintf(ring_ref_name, RINGREF_NAME_LEN, "ring-ref%u", i);
-			err = xenbus_printf(xbt, dev->nodename, ring_ref_name,
-					    "%u", info->ring_ref[i]);
+		for (i = 0; i < info->nr_rings; i++) {
+			memset(path, 0, pathsize);
+			snprintf(path, pathsize, "%s/queue-%u", dev->nodename, i);
+			err = write_per_ring_nodes(xbt, &info->rinfo[i], path);
 			if (err) {
-				message = "writing ring-ref";
-				goto abort_transaction;
+				kfree(path);
+				goto destroy_blkring;
 			}
 		}
-	}
-	err = xenbus_printf(xbt, dev->nodename,
-			    "event-channel", "%u", info->evtchn);
-	if (err) {
-		message = "writing event-channel";
-		goto abort_transaction;
+		kfree(path);
 	}
 	err = xenbus_printf(xbt, dev->nodename, "protocol", "%s",
 			    XEN_IO_PROTO_ABI_NATIVE);
@@ -1540,9 +1848,14 @@ static int talk_to_blkback(struct xenbus_device *dev,
 		goto destroy_blkring;
 	}
 
-	for (i = 0; i < BLK_RING_SIZE(info); i++)
-		info->shadow[i].req.u.rw.id = i+1;
-	info->shadow[BLK_RING_SIZE(info)-1].req.u.rw.id = 0x0fffffff;
+	for (i = 0; i < info->nr_rings; i++) {
+		unsigned int j;
+		struct blkfront_ring_info *rinfo = &info->rinfo[i];
+
+		for (j = 0; j < BLK_RING_SIZE(info); j++)
+			rinfo->shadow[j].req.u.rw.id = j + 1;
+		rinfo->shadow[BLK_RING_SIZE(info)-1].req.u.rw.id = 0x0fffffff;
+	}
 	xenbus_switch_state(dev, XenbusStateInitialised);
 
 	return 0;
@@ -1553,7 +1866,10 @@ static int talk_to_blkback(struct xenbus_device *dev,
 		xenbus_dev_fatal(dev, err, "%s", message);
  destroy_blkring:
 	blkif_free(info, 0);
- out:
+
+	kfree(info);
+	dev_set_drvdata(&dev->dev, NULL);
+
 	return err;
 }
 
@@ -1567,7 +1883,9 @@ static int blkfront_probe(struct xenbus_device *dev,
 			  const struct xenbus_device_id *id)
 {
 	int err, vdevice;
+	unsigned int r_index;
 	struct blkfront_info *info;
+	unsigned int backend_max_queues = 0;
 
 	/* FIXME: Use dynamic device id if this is not set. */
 	err = xenbus_scanf(XBT_NIL, dev->nodename,
@@ -1617,15 +1935,39 @@ static int blkfront_probe(struct xenbus_device *dev,
 		return -ENOMEM;
 	}
 
-	mutex_init(&info->mutex);
-	spin_lock_init(&info->io_lock);
 	info->xbdev = dev;
+	/* Check if backend supports multiple queues. */
+	err = xenbus_scanf(XBT_NIL, info->xbdev->otherend,
+			   "multi-queue-max-queues", "%u", &backend_max_queues);
+	if (err < 0)
+		backend_max_queues = 1;
+
+	info->nr_rings = min(backend_max_queues, xen_blkif_max_queues);
+	/* We need at least one ring. */
+	if (!info->nr_rings)
+		info->nr_rings = 1;
+
+	info->rinfo = kzalloc(sizeof(struct blkfront_ring_info) * info->nr_rings, GFP_KERNEL);
+	if (!info->rinfo) {
+		xenbus_dev_fatal(dev, -ENOMEM, "allocating ring_info structure");
+		kfree(info);
+		return -ENOMEM;
+	}
+
+	for (r_index = 0; r_index < info->nr_rings; r_index++) {
+		struct blkfront_ring_info *rinfo;
+
+		rinfo = &info->rinfo[r_index];
+		INIT_LIST_HEAD(&rinfo->indirect_pages);
+		INIT_LIST_HEAD(&rinfo->grants);
+		rinfo->dev_info = info;
+		INIT_WORK(&rinfo->work, blkif_restart_queue);
+		spin_lock_init(&rinfo->ring_lock);
+	}
+
+	mutex_init(&info->mutex);
 	info->vdevice = vdevice;
-	INIT_LIST_HEAD(&info->grants);
-	INIT_LIST_HEAD(&info->indirect_pages);
-	info->persistent_gnts_c = 0;
 	info->connected = BLKIF_STATE_DISCONNECTED;
-	INIT_WORK(&info->work, blkif_restart_queue);
 
 	/* Front end dir is a number, which is used as the id. */
 	info->handle = simple_strtoul(strrchr(dev->nodename, '/')+1, NULL, 0);
@@ -1649,7 +1991,7 @@ static void split_bio_end(struct bio *bio)
 
 static int blkif_recover(struct blkfront_info *info)
 {
-	int i;
+	unsigned int i, r_index;
 	struct request *req, *n;
 	struct blk_shadow *copy;
 	int rc;
@@ -1660,64 +2002,73 @@ static int blkif_recover(struct blkfront_info *info)
 	struct split_bio *split_bio;
 	struct list_head requests;
 
-	/* Stage 1: Make a safe copy of the shadow state. */
-	copy = kmemdup(info->shadow, sizeof(info->shadow),
-		       GFP_NOIO | __GFP_REPEAT | __GFP_HIGH);
-	if (!copy)
-		return -ENOMEM;
-
-	/* Stage 2: Set up free list. */
-	memset(&info->shadow, 0, sizeof(info->shadow));
-	for (i = 0; i < BLK_RING_SIZE(info); i++)
-		info->shadow[i].req.u.rw.id = i+1;
-	info->shadow_free = info->ring.req_prod_pvt;
-	info->shadow[BLK_RING_SIZE(info)-1].req.u.rw.id = 0x0fffffff;
-
-	rc = blkfront_gather_backend_features(info);
-	if (rc) {
-		kfree(copy);
-		return rc;
-	}
-
+	blkfront_gather_backend_features(info);
 	segs = info->max_indirect_segments ? : BLKIF_MAX_SEGMENTS_PER_REQUEST;
 	blk_queue_max_segments(info->rq, segs);
 	bio_list_init(&bio_list);
 	INIT_LIST_HEAD(&requests);
-	for (i = 0; i < BLK_RING_SIZE(info); i++) {
-		/* Not in use? */
-		if (!copy[i].request)
-			continue;
 
-		/*
-		 * Get the bios in the request so we can re-queue them.
-		 */
-		if (copy[i].request->cmd_flags &
-		    (REQ_FLUSH | REQ_FUA | REQ_DISCARD | REQ_SECURE)) {
+	for (r_index = 0; r_index < info->nr_rings; r_index++) {
+		struct blkfront_ring_info *rinfo;
+
+		rinfo = &info->rinfo[r_index];
+		/* Stage 1: Make a safe copy of the shadow state. */
+		copy = kmemdup(rinfo->shadow, sizeof(rinfo->shadow),
+			       GFP_NOIO | __GFP_REPEAT | __GFP_HIGH);
+		if (!copy)
+			return -ENOMEM;
+
+		/* Stage 2: Set up free list. */
+		memset(&rinfo->shadow, 0, sizeof(rinfo->shadow));
+		for (i = 0; i < BLK_RING_SIZE(info); i++)
+			rinfo->shadow[i].req.u.rw.id = i+1;
+		rinfo->shadow_free = rinfo->ring.req_prod_pvt;
+		rinfo->shadow[BLK_RING_SIZE(info)-1].req.u.rw.id = 0x0fffffff;
+
+		rc = blkfront_setup_indirect(rinfo);
+		if (rc) {
+			kfree(copy);
+			return rc;
+		}
+
+		for (i = 0; i < BLK_RING_SIZE(info); i++) {
+			/* Not in use? */
+			if (!copy[i].request)
+				continue;
+
 			/*
-			 * Flush operations don't contain bios, so
-			 * we need to requeue the whole request
+			 * Get the bios in the request so we can re-queue them.
 			 */
-			list_add(&copy[i].request->queuelist, &requests);
-			continue;
+			if (copy[i].request->cmd_flags &
+			    (REQ_FLUSH | REQ_FUA | REQ_DISCARD | REQ_SECURE)) {
+				/*
+				 * Flush operations don't contain bios, so
+				 * we need to requeue the whole request
+				 */
+				list_add(&copy[i].request->queuelist, &requests);
+				continue;
+			}
+			merge_bio.head = copy[i].request->bio;
+			merge_bio.tail = copy[i].request->biotail;
+			bio_list_merge(&bio_list, &merge_bio);
+			copy[i].request->bio = NULL;
+			blk_end_request_all(copy[i].request, 0);
 		}
-		merge_bio.head = copy[i].request->bio;
-		merge_bio.tail = copy[i].request->biotail;
-		bio_list_merge(&bio_list, &merge_bio);
-		copy[i].request->bio = NULL;
-		blk_end_request_all(copy[i].request, 0);
-	}
-
-	kfree(copy);
 
+		kfree(copy);
+	}
 	xenbus_switch_state(info->xbdev, XenbusStateConnected);
 
-	spin_lock_irq(&info->io_lock);
-
 	/* Now safe for us to use the shared ring */
 	info->connected = BLKIF_STATE_CONNECTED;
 
-	/* Kick any other new requests queued since we resumed */
-	kick_pending_request_queues(info);
+	for (r_index = 0; r_index < info->nr_rings; r_index++) {
+		struct blkfront_ring_info *rinfo;
+
+		rinfo = &info->rinfo[r_index];
+		/* Kick any other new requests queued since we resumed */
+		kick_pending_request_queues(rinfo);
+	}
 
 	list_for_each_entry_safe(req, n, &requests, queuelist) {
 		/* Requeue pending requests (flush or discard) */
@@ -1725,7 +2076,6 @@ static int blkif_recover(struct blkfront_info *info)
 		BUG_ON(req->nr_phys_segments > segs);
 		blk_mq_requeue_request(req);
 	}
-	spin_unlock_irq(&info->io_lock);
 	blk_mq_kick_requeue_list(info->rq);
 
 	while ((bio = bio_list_pop(&bio_list)) != NULL) {
@@ -1790,8 +2140,7 @@ static int blkfront_resume(struct xenbus_device *dev)
 	return err;
 }
 
-static void
-blkfront_closing(struct blkfront_info *info)
+static void blkfront_closing(struct blkfront_info *info)
 {
 	struct xenbus_device *xbdev = info->xbdev;
 	struct block_device *bdev = NULL;
@@ -1851,18 +2200,29 @@ static void blkfront_setup_discard(struct blkfront_info *info)
 		info->feature_secdiscard = !!discard_secure;
 }
 
-static int blkfront_setup_indirect(struct blkfront_info *info)
+static int blkfront_setup_indirect(struct blkfront_ring_info *rinfo)
 {
 	unsigned int psegs, grants;
 	int err, i;
+	struct blkfront_info *info = rinfo->dev_info;
 
-	if (info->max_indirect_segments == 0)
-		grants = BLKIF_MAX_SEGMENTS_PER_REQUEST;
+	if (info->max_indirect_segments == 0) {
+		if (!HAS_EXTRA_REQ)
+			grants = BLKIF_MAX_SEGMENTS_PER_REQUEST;
+		else {
+			/*
+			 * When an extra req is required, the maximum
+			 * grants supported is related to the size of the
+			 * Linux block segment.
+			 */
+			grants = GRANTS_PER_PSEG;
+		}
+	}
 	else
 		grants = info->max_indirect_segments;
 	psegs = grants / GRANTS_PER_PSEG;
 
-	err = fill_grant_buffer(info,
+	err = fill_grant_buffer(rinfo,
 				(grants + INDIRECT_GREFS(grants)) * BLK_RING_SIZE(info));
 	if (err)
 		goto out_of_memory;
@@ -1875,31 +2235,31 @@ static int blkfront_setup_indirect(struct blkfront_info *info)
 		 */
 		int num = INDIRECT_GREFS(grants) * BLK_RING_SIZE(info);
 
-		BUG_ON(!list_empty(&info->indirect_pages));
+		BUG_ON(!list_empty(&rinfo->indirect_pages));
 		for (i = 0; i < num; i++) {
 			struct page *indirect_page = alloc_page(GFP_NOIO);
 			if (!indirect_page)
 				goto out_of_memory;
-			list_add(&indirect_page->lru, &info->indirect_pages);
+			list_add(&indirect_page->lru, &rinfo->indirect_pages);
 		}
 	}
 
 	for (i = 0; i < BLK_RING_SIZE(info); i++) {
-		info->shadow[i].grants_used = kzalloc(
-			sizeof(info->shadow[i].grants_used[0]) * grants,
+		rinfo->shadow[i].grants_used = kzalloc(
+			sizeof(rinfo->shadow[i].grants_used[0]) * grants,
 			GFP_NOIO);
-		info->shadow[i].sg = kzalloc(sizeof(info->shadow[i].sg[0]) * psegs, GFP_NOIO);
+		rinfo->shadow[i].sg = kzalloc(sizeof(rinfo->shadow[i].sg[0]) * psegs, GFP_NOIO);
 		if (info->max_indirect_segments)
-			info->shadow[i].indirect_grants = kzalloc(
-				sizeof(info->shadow[i].indirect_grants[0]) *
+			rinfo->shadow[i].indirect_grants = kzalloc(
+				sizeof(rinfo->shadow[i].indirect_grants[0]) *
 				INDIRECT_GREFS(grants),
 				GFP_NOIO);
-		if ((info->shadow[i].grants_used == NULL) ||
-			(info->shadow[i].sg == NULL) ||
+		if ((rinfo->shadow[i].grants_used == NULL) ||
+			(rinfo->shadow[i].sg == NULL) ||
 		     (info->max_indirect_segments &&
-		     (info->shadow[i].indirect_grants == NULL)))
+		     (rinfo->shadow[i].indirect_grants == NULL)))
 			goto out_of_memory;
-		sg_init_table(info->shadow[i].sg, psegs);
+		sg_init_table(rinfo->shadow[i].sg, psegs);
 	}
 
 
@@ -1907,16 +2267,16 @@ static int blkfront_setup_indirect(struct blkfront_info *info)
 
 out_of_memory:
 	for (i = 0; i < BLK_RING_SIZE(info); i++) {
-		kfree(info->shadow[i].grants_used);
-		info->shadow[i].grants_used = NULL;
-		kfree(info->shadow[i].sg);
-		info->shadow[i].sg = NULL;
-		kfree(info->shadow[i].indirect_grants);
-		info->shadow[i].indirect_grants = NULL;
-	}
-	if (!list_empty(&info->indirect_pages)) {
+		kfree(rinfo->shadow[i].grants_used);
+		rinfo->shadow[i].grants_used = NULL;
+		kfree(rinfo->shadow[i].sg);
+		rinfo->shadow[i].sg = NULL;
+		kfree(rinfo->shadow[i].indirect_grants);
+		rinfo->shadow[i].indirect_grants = NULL;
+	}
+	if (!list_empty(&rinfo->indirect_pages)) {
 		struct page *indirect_page, *n;
-		list_for_each_entry_safe(indirect_page, n, &info->indirect_pages, lru) {
+		list_for_each_entry_safe(indirect_page, n, &rinfo->indirect_pages, lru) {
 			list_del(&indirect_page->lru);
 			__free_page(indirect_page);
 		}
@@ -1927,7 +2287,7 @@ static int blkfront_setup_indirect(struct blkfront_info *info)
 /*
  * Gather all backend feature-*
  */
-static int blkfront_gather_backend_features(struct blkfront_info *info)
+static void blkfront_gather_backend_features(struct blkfront_info *info)
 {
 	int err;
 	int barrier, flush, discard, persistent;
@@ -1982,8 +2342,6 @@ static int blkfront_gather_backend_features(struct blkfront_info *info)
 	else
 		info->max_indirect_segments = min(indirect_segments,
 						  xen_blkif_max_segments);
-
-	return blkfront_setup_indirect(info);
 }
 
 /*
@@ -1996,7 +2354,7 @@ static void blkfront_connect(struct blkfront_info *info)
 	unsigned long sector_size;
 	unsigned int physical_sector_size;
 	unsigned int binfo;
-	int err;
+	int err, i;
 
 	switch (info->connected) {
 	case BLKIF_STATE_CONNECTED:
@@ -2053,11 +2411,15 @@ static void blkfront_connect(struct blkfront_info *info)
 	if (err != 1)
 		physical_sector_size = sector_size;
 
-	err = blkfront_gather_backend_features(info);
-	if (err) {
-		xenbus_dev_fatal(info->xbdev, err, "setup_indirect at %s",
-				 info->xbdev->otherend);
-		return;
+	blkfront_gather_backend_features(info);
+	for (i = 0; i < info->nr_rings; i++) {
+		err = blkfront_setup_indirect(&info->rinfo[i]);
+		if (err) {
+			xenbus_dev_fatal(info->xbdev, err, "setup_indirect at %s",
+					 info->xbdev->otherend);
+			blkif_free(info, 0);
+			break;
+		}
 	}
 
 	err = xlvbd_alloc_gendisk(sectors, info, binfo, sector_size,
@@ -2071,10 +2433,9 @@ static void blkfront_connect(struct blkfront_info *info)
 	xenbus_switch_state(info->xbdev, XenbusStateConnected);
 
 	/* Kick pending requests. */
-	spin_lock_irq(&info->io_lock);
 	info->connected = BLKIF_STATE_CONNECTED;
-	kick_pending_request_queues(info);
-	spin_unlock_irq(&info->io_lock);
+	for (i = 0; i < info->nr_rings; i++)
+		kick_pending_request_queues(&info->rinfo[i]);
 
 	add_disk(info->gd);
 
@@ -2095,11 +2456,8 @@ static void blkback_changed(struct xenbus_device *dev,
 	case XenbusStateInitWait:
 		if (dev->state != XenbusStateInitialising)
 			break;
-		if (talk_to_blkback(dev, info)) {
-			kfree(info);
-			dev_set_drvdata(&dev->dev, NULL);
+		if (talk_to_blkback(dev, info))
 			break;
-		}
 	case XenbusStateInitialising:
 	case XenbusStateInitialised:
 	case XenbusStateReconfiguring:
@@ -2108,6 +2466,10 @@ static void blkback_changed(struct xenbus_device *dev,
 		break;
 
 	case XenbusStateConnected:
+		if (dev->state != XenbusStateInitialised) {
+			if (talk_to_blkback(dev, info))
+				break;
+		}
 		blkfront_connect(info);
 		break;
 
@@ -2281,6 +2643,7 @@ static struct xenbus_driver blkfront_driver = {
 static int __init xlblk_init(void)
 {
 	int ret;
+	int nr_cpus = num_online_cpus();
 
 	if (!xen_domain())
 		return -ENODEV;
@@ -2288,7 +2651,13 @@ static int __init xlblk_init(void)
 	if (xen_blkif_max_ring_order > XENBUS_MAX_RING_GRANT_ORDER) {
 		pr_info("Invalid max_ring_order (%d), will use default max: %d.\n",
 			xen_blkif_max_ring_order, XENBUS_MAX_RING_GRANT_ORDER);
-		xen_blkif_max_ring_order = 0;
+		xen_blkif_max_ring_order = XENBUS_MAX_RING_GRANT_ORDER;
+	}
+
+	if (xen_blkif_max_queues > nr_cpus) {
+		pr_info("Invalid max_queues (%d), will use default max: %d.\n",
+			xen_blkif_max_queues, nr_cpus);
+		xen_blkif_max_queues = nr_cpus;
 	}
 
 	if (!xen_has_pv_disk_devices())

drivers/md/bcache/btree.c

@@ -1741,6 +1741,7 @@ static void bch_btree_gc(struct cache_set *c)
 	do {
 		ret = btree_root(gc_root, c, &op, &writes, &stats);
 		closure_sync(&writes);
+		cond_resched();
 
 		if (ret && ret != -EAGAIN)
 			pr_warn("gc failed!");
@@ -2162,8 +2163,10 @@ int bch_btree_insert_check_key(struct btree *b, struct btree_op *op,
 		rw_lock(true, b, b->level);
 
 		if (b->key.ptr[0] != btree_ptr ||
-		    b->seq != seq + 1)
+                   b->seq != seq + 1) {
+                       op->lock = b->level;
 			goto out;
+               }
 	}
 
 	SET_KEY_PTRS(check_key, 1);

drivers/md/bcache/super.c

@@ -685,6 +685,8 @@ static void bcache_device_link(struct bcache_device *d, struct cache_set *c,
 	WARN(sysfs_create_link(&d->kobj, &c->kobj, "cache") ||
 	     sysfs_create_link(&c->kobj, &d->kobj, d->name),
 	     "Couldn't create device <-> cache set symlinks");
+
+	clear_bit(BCACHE_DEV_UNLINK_DONE, &d->flags);
 }
 
 static void bcache_device_detach(struct bcache_device *d)
@@ -847,8 +849,11 @@ void bch_cached_dev_run(struct cached_dev *dc)
 	buf[SB_LABEL_SIZE] = '\0';
 	env[2] = kasprintf(GFP_KERNEL, "CACHED_LABEL=%s", buf);
 
-	if (atomic_xchg(&dc->running, 1))
+	if (atomic_xchg(&dc->running, 1)) {
+		kfree(env[1]);
+		kfree(env[2]);
 		return;
+	}
 
 	if (!d->c &&
 	    BDEV_STATE(&dc->sb) != BDEV_STATE_NONE) {
@@ -1933,6 +1938,8 @@ static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr,
 			else
 				err = "device busy";
 			mutex_unlock(&bch_register_lock);
+			if (attr == &ksysfs_register_quiet)
+				goto out;
 		}
 		goto err;
 	}
@@ -1971,8 +1978,7 @@ static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr,
 err_close:
 	blkdev_put(bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
 err:
-	if (attr != &ksysfs_register_quiet)
-		pr_info("error opening %s: %s", path, err);
+	pr_info("error opening %s: %s", path, err);
 	ret = -EINVAL;
 	goto out;
 }
@@ -2066,8 +2072,10 @@ static int __init bcache_init(void)
 	closure_debug_init();
 
 	bcache_major = register_blkdev(0, "bcache");
-	if (bcache_major < 0)
+	if (bcache_major < 0) {
+		unregister_reboot_notifier(&reboot);
 		return bcache_major;
+	}
 
 	if (!(bcache_wq = create_workqueue("bcache")) ||
 	    !(bcache_kobj = kobject_create_and_add("bcache", fs_kobj)) ||

drivers/md/bcache/writeback.c

@@ -323,6 +323,10 @@ void bcache_dev_sectors_dirty_add(struct cache_set *c, unsigned inode,
 
 static bool dirty_pred(struct keybuf *buf, struct bkey *k)
 {
+	struct cached_dev *dc = container_of(buf, struct cached_dev, writeback_keys);
+
+	BUG_ON(KEY_INODE(k) != dc->disk.id);
+
 	return KEY_DIRTY(k);
 }
 
@@ -372,11 +376,24 @@ static void refill_full_stripes(struct cached_dev *dc)
 	}
 }
 
+/*
+ * Returns true if we scanned the entire disk
+ */
 static bool refill_dirty(struct cached_dev *dc)
 {
 	struct keybuf *buf = &dc->writeback_keys;
+	struct bkey start = KEY(dc->disk.id, 0, 0);
 	struct bkey end = KEY(dc->disk.id, MAX_KEY_OFFSET, 0);
-	bool searched_from_start = false;
+	struct bkey start_pos;
+
+	/*
+	 * make sure keybuf pos is inside the range for this disk - at bringup
+	 * we might not be attached yet so this disk's inode nr isn't
+	 * initialized then
+	 */
+	if (bkey_cmp(&buf->last_scanned, &start) < 0 ||
+	    bkey_cmp(&buf->last_scanned, &end) > 0)
+		buf->last_scanned = start;
 
 	if (dc->partial_stripes_expensive) {
 		refill_full_stripes(dc);
@@ -384,14 +401,20 @@ static bool refill_dirty(struct cached_dev *dc)
 			return false;
 	}
 
-	if (bkey_cmp(&buf->last_scanned, &end) >= 0) {
-		buf->last_scanned = KEY(dc->disk.id, 0, 0);
-		searched_from_start = true;
-	}
-
+	start_pos = buf->last_scanned;
 	bch_refill_keybuf(dc->disk.c, buf, &end, dirty_pred);
 
-	return bkey_cmp(&buf->last_scanned, &end) >= 0 && searched_from_start;
+	if (bkey_cmp(&buf->last_scanned, &end) < 0)
+		return false;
+
+	/*
+	 * If we get to the end start scanning again from the beginning, and
+	 * only scan up to where we initially started scanning from:
+	 */
+	buf->last_scanned = start;
+	bch_refill_keybuf(dc->disk.c, buf, &start_pos, dirty_pred);
+
+	return bkey_cmp(&buf->last_scanned, &start_pos) >= 0;
 }
 
 static int bch_writeback_thread(void *arg)

drivers/md/bcache/writeback.h

@@ -63,7 +63,8 @@ static inline bool should_writeback(struct cached_dev *dc, struct bio *bio,
 
 static inline void bch_writeback_queue(struct cached_dev *dc)
 {
-	wake_up_process(dc->writeback_thread);
+	if (!IS_ERR_OR_NULL(dc->writeback_thread))
+		wake_up_process(dc->writeback_thread);
 }
 
 static inline void bch_writeback_add(struct cached_dev *dc)

include/linux/drbd.h

@@ -25,7 +25,6 @@
 */
 #ifndef DRBD_H
 #define DRBD_H
-#include <linux/connector.h>
 #include <asm/types.h>
 
 #ifdef __KERNEL__
@@ -52,7 +51,7 @@
 #endif
 
 extern const char *drbd_buildtag(void);
-#define REL_VERSION "8.4.5"
+#define REL_VERSION "8.4.6"
 #define API_VERSION 1
 #define PRO_VERSION_MIN 86
 #define PRO_VERSION_MAX 101
@@ -339,6 +338,8 @@ enum drbd_state_rv {
 #define MDF_AL_CLEAN		(1 << 7)
 #define MDF_AL_DISABLED		(1 << 8)
 
+#define MAX_PEERS 32
+
 enum drbd_uuid_index {
 	UI_CURRENT,
 	UI_BITMAP,
@@ -349,14 +350,35 @@ enum drbd_uuid_index {
 	UI_EXTENDED_SIZE   /* Everything. */
 };
 
+#define HISTORY_UUIDS MAX_PEERS
+
 enum drbd_timeout_flag {
 	UT_DEFAULT      = 0,
 	UT_DEGRADED     = 1,
 	UT_PEER_OUTDATED = 2,
 };
 
+enum drbd_notification_type {
+	NOTIFY_EXISTS,
+	NOTIFY_CREATE,
+	NOTIFY_CHANGE,
+	NOTIFY_DESTROY,
+	NOTIFY_CALL,
+	NOTIFY_RESPONSE,
+
+	NOTIFY_CONTINUES = 0x8000,
+	NOTIFY_FLAGS = NOTIFY_CONTINUES,
+};
+
 #define UUID_JUST_CREATED ((__u64)4)
 
+enum write_ordering_e {
+	WO_NONE,
+	WO_DRAIN_IO,
+	WO_BDEV_FLUSH,
+	WO_BIO_BARRIER
+};
+
 /* magic numbers used in meta data and network packets */
 #define DRBD_MAGIC 0x83740267
 #define DRBD_MAGIC_BIG 0x835a

include/linux/drbd_genl.h

@@ -250,6 +250,76 @@ GENL_struct(DRBD_NLA_DETACH_PARMS, 13, detach_parms,
 	__flg_field(1, DRBD_GENLA_F_MANDATORY,	force_detach)
 )
 
+GENL_struct(DRBD_NLA_RESOURCE_INFO, 15, resource_info,
+	__u32_field(1, 0, res_role)
+	__flg_field(2, 0, res_susp)
+	__flg_field(3, 0, res_susp_nod)
+	__flg_field(4, 0, res_susp_fen)
+	/* __flg_field(5, 0, res_weak) */
+)
+
+GENL_struct(DRBD_NLA_DEVICE_INFO, 16, device_info,
+	__u32_field(1, 0, dev_disk_state)
+)
+
+GENL_struct(DRBD_NLA_CONNECTION_INFO, 17, connection_info,
+	__u32_field(1, 0, conn_connection_state)
+	__u32_field(2, 0, conn_role)
+)
+
+GENL_struct(DRBD_NLA_PEER_DEVICE_INFO, 18, peer_device_info,
+	__u32_field(1, 0, peer_repl_state)
+	__u32_field(2, 0, peer_disk_state)
+	__u32_field(3, 0, peer_resync_susp_user)
+	__u32_field(4, 0, peer_resync_susp_peer)
+	__u32_field(5, 0, peer_resync_susp_dependency)
+)
+
+GENL_struct(DRBD_NLA_RESOURCE_STATISTICS, 19, resource_statistics,
+	__u32_field(1, 0, res_stat_write_ordering)
+)
+
+GENL_struct(DRBD_NLA_DEVICE_STATISTICS, 20, device_statistics,
+	__u64_field(1, 0, dev_size)  /* (sectors) */
+	__u64_field(2, 0, dev_read)  /* (sectors) */
+	__u64_field(3, 0, dev_write)  /* (sectors) */
+	__u64_field(4, 0, dev_al_writes)  /* activity log writes (count) */
+	__u64_field(5, 0, dev_bm_writes)  /*  bitmap writes  (count) */
+	__u32_field(6, 0, dev_upper_pending)  /* application requests in progress */
+	__u32_field(7, 0, dev_lower_pending)  /* backing device requests in progress */
+	__flg_field(8, 0, dev_upper_blocked)
+	__flg_field(9, 0, dev_lower_blocked)
+	__flg_field(10, 0, dev_al_suspended)  /* activity log suspended */
+	__u64_field(11, 0, dev_exposed_data_uuid)
+	__u64_field(12, 0, dev_current_uuid)
+	__u32_field(13, 0, dev_disk_flags)
+	__bin_field(14, 0, history_uuids, HISTORY_UUIDS * sizeof(__u64))
+)
+
+GENL_struct(DRBD_NLA_CONNECTION_STATISTICS, 21, connection_statistics,
+	__flg_field(1, 0, conn_congested)
+)
+
+GENL_struct(DRBD_NLA_PEER_DEVICE_STATISTICS, 22, peer_device_statistics,
+	__u64_field(1, 0, peer_dev_received)  /* sectors */
+	__u64_field(2, 0, peer_dev_sent)  /* sectors */
+	__u32_field(3, 0, peer_dev_pending)  /* number of requests */
+	__u32_field(4, 0, peer_dev_unacked)  /* number of requests */
+	__u64_field(5, 0, peer_dev_out_of_sync)  /* sectors */
+	__u64_field(6, 0, peer_dev_resync_failed)  /* sectors */
+	__u64_field(7, 0, peer_dev_bitmap_uuid)
+	__u32_field(9, 0, peer_dev_flags)
+)
+
+GENL_struct(DRBD_NLA_NOTIFICATION_HEADER, 23, drbd_notification_header,
+	__u32_field(1, DRBD_GENLA_F_MANDATORY, nh_type)
+)
+
+GENL_struct(DRBD_NLA_HELPER, 24, drbd_helper_info,
+	__str_field(1, DRBD_GENLA_F_MANDATORY, helper_name, 32)
+	__u32_field(2, DRBD_GENLA_F_MANDATORY, helper_status)
+)
+
 /*
  * Notifications and commands (genlmsghdr->cmd)
  */
@@ -382,3 +452,82 @@ GENL_op(DRBD_ADM_GET_TIMEOUT_TYPE, 26, GENL_doit(drbd_adm_get_timeout_type),
 	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED))
 GENL_op(DRBD_ADM_DOWN,		27, GENL_doit(drbd_adm_down),
 	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED))
+
+GENL_op(DRBD_ADM_GET_RESOURCES, 30,
+	 GENL_op_init(
+		 .dumpit = drbd_adm_dump_resources,
+	 ),
+	 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_GENLA_F_MANDATORY)
+	 GENL_tla_expected(DRBD_NLA_RESOURCE_INFO, DRBD_GENLA_F_MANDATORY)
+	 GENL_tla_expected(DRBD_NLA_RESOURCE_STATISTICS, DRBD_GENLA_F_MANDATORY))
+
+GENL_op(DRBD_ADM_GET_DEVICES, 31,
+	 GENL_op_init(
+		 .dumpit = drbd_adm_dump_devices,
+		 .done = drbd_adm_dump_devices_done,
+	 ),
+	 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_GENLA_F_MANDATORY)
+	 GENL_tla_expected(DRBD_NLA_DEVICE_INFO, DRBD_GENLA_F_MANDATORY)
+	 GENL_tla_expected(DRBD_NLA_DEVICE_STATISTICS, DRBD_GENLA_F_MANDATORY))
+
+GENL_op(DRBD_ADM_GET_CONNECTIONS, 32,
+	 GENL_op_init(
+		 .dumpit = drbd_adm_dump_connections,
+		 .done = drbd_adm_dump_connections_done,
+	 ),
+	 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_GENLA_F_MANDATORY)
+	 GENL_tla_expected(DRBD_NLA_CONNECTION_INFO, DRBD_GENLA_F_MANDATORY)
+	 GENL_tla_expected(DRBD_NLA_CONNECTION_STATISTICS, DRBD_GENLA_F_MANDATORY))
+
+GENL_op(DRBD_ADM_GET_PEER_DEVICES, 33,
+	 GENL_op_init(
+		 .dumpit = drbd_adm_dump_peer_devices,
+		 .done = drbd_adm_dump_peer_devices_done,
+	 ),
+	 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_GENLA_F_MANDATORY)
+	 GENL_tla_expected(DRBD_NLA_PEER_DEVICE_INFO, DRBD_GENLA_F_MANDATORY)
+	 GENL_tla_expected(DRBD_NLA_PEER_DEVICE_STATISTICS, DRBD_GENLA_F_MANDATORY))
+
+GENL_notification(
+	DRBD_RESOURCE_STATE, 34, events,
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_NOTIFICATION_HEADER, DRBD_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_RESOURCE_INFO, DRBD_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_RESOURCE_STATISTICS, DRBD_F_REQUIRED))
+
+GENL_notification(
+	DRBD_DEVICE_STATE, 35, events,
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_NOTIFICATION_HEADER, DRBD_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_DEVICE_INFO, DRBD_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_DEVICE_STATISTICS, DRBD_F_REQUIRED))
+
+GENL_notification(
+	DRBD_CONNECTION_STATE, 36, events,
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_NOTIFICATION_HEADER, DRBD_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_CONNECTION_INFO, DRBD_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_CONNECTION_STATISTICS, DRBD_F_REQUIRED))
+
+GENL_notification(
+	DRBD_PEER_DEVICE_STATE, 37, events,
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_NOTIFICATION_HEADER, DRBD_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_PEER_DEVICE_INFO, DRBD_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_PEER_DEVICE_STATISTICS, DRBD_F_REQUIRED))
+
+GENL_op(
+	DRBD_ADM_GET_INITIAL_STATE, 38,
+	GENL_op_init(
+	        .dumpit = drbd_adm_get_initial_state,
+	),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_GENLA_F_MANDATORY))
+
+GENL_notification(
+	DRBD_HELPER, 40, events,
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_HELPER, DRBD_F_REQUIRED))
+
+GENL_notification(
+	DRBD_INITIAL_STATE_DONE, 41, events,
+	GENL_tla_expected(DRBD_NLA_NOTIFICATION_HEADER, DRBD_F_REQUIRED))

include/linux/idr.h

@@ -135,6 +135,20 @@ static inline void *idr_find(struct idr *idr, int id)
 #define idr_for_each_entry(idp, entry, id)			\
 	for (id = 0; ((entry) = idr_get_next(idp, &(id))) != NULL; ++id)
 
+/**
+ * idr_for_each_entry - continue iteration over an idr's elements of a given type
+ * @idp:     idr handle
+ * @entry:   the type * to use as cursor
+ * @id:      id entry's key
+ *
+ * Continue to iterate over list of given type, continuing after
+ * the current position.
+ */
+#define idr_for_each_entry_continue(idp, entry, id)			\
+	for ((entry) = idr_get_next((idp), &(id));			\
+	     entry;							\
+	     ++id, (entry) = idr_get_next((idp), &(id)))
+
 /*
  * IDA - IDR based id allocator, use when translation from id to
  * pointer isn't necessary.

include/linux/lru_cache.h

@@ -264,7 +264,7 @@ extern unsigned int lc_put(struct lru_cache *lc, struct lc_element *e);
 extern void lc_committed(struct lru_cache *lc);
 
 struct seq_file;
-extern size_t lc_seq_printf_stats(struct seq_file *seq, struct lru_cache *lc);
+extern void lc_seq_printf_stats(struct seq_file *seq, struct lru_cache *lc);
 
 extern void lc_seq_dump_details(struct seq_file *seq, struct lru_cache *lc, char *utext,
 				void (*detail) (struct seq_file *, struct lc_element *));

include/xen/interface/io/blkif.h

@@ -27,6 +27,54 @@
 typedef uint16_t blkif_vdev_t;
 typedef uint64_t blkif_sector_t;
 
+/*
+ * Multiple hardware queues/rings:
+ * If supported, the backend will write the key "multi-queue-max-queues" to
+ * the directory for that vbd, and set its value to the maximum supported
+ * number of queues.
+ * Frontends that are aware of this feature and wish to use it can write the
+ * key "multi-queue-num-queues" with the number they wish to use, which must be
+ * greater than zero, and no more than the value reported by the backend in
+ * "multi-queue-max-queues".
+ *
+ * For frontends requesting just one queue, the usual event-channel and
+ * ring-ref keys are written as before, simplifying the backend processing
+ * to avoid distinguishing between a frontend that doesn't understand the
+ * multi-queue feature, and one that does, but requested only one queue.
+ *
+ * Frontends requesting two or more queues must not write the toplevel
+ * event-channel and ring-ref keys, instead writing those keys under sub-keys
+ * having the name "queue-N" where N is the integer ID of the queue/ring for
+ * which those keys belong. Queues are indexed from zero.
+ * For example, a frontend with two queues must write the following set of
+ * queue-related keys:
+ *
+ * /local/domain/1/device/vbd/0/multi-queue-num-queues = "2"
+ * /local/domain/1/device/vbd/0/queue-0 = ""
+ * /local/domain/1/device/vbd/0/queue-0/ring-ref = "<ring-ref#0>"
+ * /local/domain/1/device/vbd/0/queue-0/event-channel = "<evtchn#0>"
+ * /local/domain/1/device/vbd/0/queue-1 = ""
+ * /local/domain/1/device/vbd/0/queue-1/ring-ref = "<ring-ref#1>"
+ * /local/domain/1/device/vbd/0/queue-1/event-channel = "<evtchn#1>"
+ *
+ * It is also possible to use multiple queues/rings together with
+ * feature multi-page ring buffer.
+ * For example, a frontend requests two queues/rings and the size of each ring
+ * buffer is two pages must write the following set of related keys:
+ *
+ * /local/domain/1/device/vbd/0/multi-queue-num-queues = "2"
+ * /local/domain/1/device/vbd/0/ring-page-order = "1"
+ * /local/domain/1/device/vbd/0/queue-0 = ""
+ * /local/domain/1/device/vbd/0/queue-0/ring-ref0 = "<ring-ref#0>"
+ * /local/domain/1/device/vbd/0/queue-0/ring-ref1 = "<ring-ref#1>"
+ * /local/domain/1/device/vbd/0/queue-0/event-channel = "<evtchn#0>"
+ * /local/domain/1/device/vbd/0/queue-1 = ""
+ * /local/domain/1/device/vbd/0/queue-1/ring-ref0 = "<ring-ref#2>"
+ * /local/domain/1/device/vbd/0/queue-1/ring-ref1 = "<ring-ref#3>"
+ * /local/domain/1/device/vbd/0/queue-1/event-channel = "<evtchn#1>"
+ *
+ */
+
 /*
  * REQUEST CODES.
  */

lib/lru_cache.c

@@ -238,7 +238,7 @@ void lc_reset(struct lru_cache *lc)
  * @seq: the seq_file to print into
  * @lc: the lru cache to print statistics of
  */
-size_t lc_seq_printf_stats(struct seq_file *seq, struct lru_cache *lc)
+void lc_seq_printf_stats(struct seq_file *seq, struct lru_cache *lc)
 {
 	/* NOTE:
 	 * total calls to lc_get are
@@ -250,8 +250,6 @@ size_t lc_seq_printf_stats(struct seq_file *seq, struct lru_cache *lc)
 	seq_printf(seq, "\t%s: used:%u/%u hits:%lu misses:%lu starving:%lu locked:%lu changed:%lu\n",
 		   lc->name, lc->used, lc->nr_elements,
 		   lc->hits, lc->misses, lc->starving, lc->locked, lc->changed);
-
-	return 0;
 }
 
 static struct hlist_head *lc_hash_slot(struct lru_cache *lc, unsigned int enr)