Merge branch 'for-2.6.40/drivers' of git://git.kernel.dk/linux-2.6-block

* 'for-2.6.40/drivers' of git://git.kernel.dk/linux-2.6-block: (110 commits)
  loop: handle on-demand devices correctly
  loop: limit 'max_part' module param to DISK_MAX_PARTS
  drbd: fix warning
  drbd: fix warning
  drbd: Fix spelling
  drbd: fix schedule in atomic
  drbd: Take a more conservative approach when deciding max_bio_size
  drbd: Fixed state transitions after async outdate-peer-handler returned
  drbd: Disallow the peer_disk_state to be D_OUTDATED while connected
  drbd: Fix for the connection problems on high latency links
  drbd: fix potential activity log refcount imbalance in error path
  drbd: Only downgrade the disk state in case of disk failures
  drbd: fix disconnect/reconnect loop, if ping-timeout == ping-int
  drbd: fix potential distributed deadlock
  lru_cache.h: fix comments referring to ts_ instead of lc_
  drbd: Fix for application IO with the on-io-error=pass-on policy
  xen/p2m: Add EXPORT_SYMBOL_GPL to the M2P override functions.
  xen/p2m/m2p/gnttab: Support GNTMAP_host_map in the M2P override.
  xen/blkback: don't fail empty barrier requests
  xen/blkback: fix xenbus_transaction_start() hang caused by double xenbus_transaction_end()
  ...

Documentation/blockdev/cciss.txt

@@ -169,3 +169,18 @@ is issued which positions the tape to a known position.  Typically you
 must rewind the tape (by issuing "mt -f /dev/st0 rewind" for example)
 before i/o can proceed again to a tape drive which was reset.
 
+There is a cciss_tape_cmds module parameter which can be used to make cciss
+allocate more commands for use by tape drives.  Ordinarily only a few commands
+(6) are allocated for tape drives because tape drives are slow and
+infrequently used and the primary purpose of Smart Array controllers is to
+act as a RAID controller for disk drives, so the vast majority of commands
+are allocated for disk devices.  However, if you have more than a few tape
+drives attached to a smart array, the default number of commands may not be
+enought (for example, if you have 8 tape drives, you could only rewind 6
+at one time with the default number of commands.)  The cciss_tape_cmds module
+parameter allows more commands (up to 16 more) to be allocated for use by
+tape drives.  For example:
+
+        insmod cciss.ko cciss_tape_cmds=16
+
+Or, as a kernel boot parameter passed in via grub:  cciss.cciss_tape_cmds=8

drivers/block/Kconfig

@@ -470,6 +470,27 @@ config XEN_BLKDEV_FRONTEND
 	  block device driver.  It communicates with a back-end driver
 	  in another domain which drives the actual block device.
 
+config XEN_BLKDEV_BACKEND
+	tristate "Block-device backend driver"
+	depends on XEN_BACKEND
+	help
+	  The block-device backend driver allows the kernel to export its
+	  block devices to other guests via a high-performance shared-memory
+	  interface.
+
+	  The corresponding Linux frontend driver is enabled by the
+	  CONFIG_XEN_BLKDEV_FRONTEND configuration option.
+
+	  The backend driver attaches itself to a any block device specified
+	  in the XenBus configuration. There are no limits to what the block
+	  device as long as it has a major and minor.
+
+	  If you are compiling a kernel to run in a Xen block backend driver
+	  domain (often this is domain 0) you should say Y here. To
+	  compile this driver as a module, chose M here: the module
+	  will be called xen-blkback.
+
+
 config VIRTIO_BLK
 	tristate "Virtio block driver (EXPERIMENTAL)"
 	depends on EXPERIMENTAL && VIRTIO

drivers/block/Makefile

@@ -36,6 +36,7 @@ obj-$(CONFIG_BLK_DEV_UB)	+= ub.o
 obj-$(CONFIG_BLK_DEV_HD)	+= hd.o
 
 obj-$(CONFIG_XEN_BLKDEV_FRONTEND)	+= xen-blkfront.o
+obj-$(CONFIG_XEN_BLKDEV_BACKEND)	+= xen-blkback/
 obj-$(CONFIG_BLK_DEV_DRBD)     += drbd/
 obj-$(CONFIG_BLK_DEV_RBD)     += rbd.o
 

drivers/block/cciss.h

@@ -200,7 +200,7 @@ struct ctlr_info
  * the above.
  */
 #define CCISS_BOARD_READY_WAIT_SECS (120)
-#define CCISS_BOARD_NOT_READY_WAIT_SECS (10)
+#define CCISS_BOARD_NOT_READY_WAIT_SECS (100)
 #define CCISS_BOARD_READY_POLL_INTERVAL_MSECS (100)
 #define CCISS_BOARD_READY_ITERATIONS \
 	((CCISS_BOARD_READY_WAIT_SECS * 1000) / \
@@ -209,8 +209,9 @@ struct ctlr_info
 	((CCISS_BOARD_NOT_READY_WAIT_SECS * 1000) / \
 		CCISS_BOARD_READY_POLL_INTERVAL_MSECS)
 #define CCISS_POST_RESET_PAUSE_MSECS (3000)
-#define CCISS_POST_RESET_NOOP_INTERVAL_MSECS (1000)
+#define CCISS_POST_RESET_NOOP_INTERVAL_MSECS (4000)
 #define CCISS_POST_RESET_NOOP_RETRIES (12)
+#define CCISS_POST_RESET_NOOP_TIMEOUT_MSECS (10000)
 
 /* 
 	Send the command to the hardware 
@@ -239,11 +240,13 @@ static void SA5_intr_mask(ctlr_info_t *h, unsigned long val)
 	{ /* Turn interrupts on */
 		h->interrupts_enabled = 1;
 		writel(0, h->vaddr + SA5_REPLY_INTR_MASK_OFFSET);
+		(void) readl(h->vaddr + SA5_REPLY_INTR_MASK_OFFSET);
 	} else /* Turn them off */
 	{
 		h->interrupts_enabled = 0;
         	writel( SA5_INTR_OFF, 
 			h->vaddr + SA5_REPLY_INTR_MASK_OFFSET);
+		(void) readl(h->vaddr + SA5_REPLY_INTR_MASK_OFFSET);
 	}
 }
 /*
@@ -257,11 +260,13 @@ static void SA5B_intr_mask(ctlr_info_t *h, unsigned long val)
         { /* Turn interrupts on */
 		h->interrupts_enabled = 1;
                 writel(0, h->vaddr + SA5_REPLY_INTR_MASK_OFFSET);
+		(void) readl(h->vaddr + SA5_REPLY_INTR_MASK_OFFSET);
         } else /* Turn them off */
         {
 		h->interrupts_enabled = 0;
                 writel( SA5B_INTR_OFF,
                         h->vaddr + SA5_REPLY_INTR_MASK_OFFSET);
+		(void) readl(h->vaddr + SA5_REPLY_INTR_MASK_OFFSET);
         }
 }
 
@@ -271,10 +276,12 @@ static void SA5_performant_intr_mask(ctlr_info_t *h, unsigned long val)
 	if (val) { /* turn on interrupts */
 		h->interrupts_enabled = 1;
 		writel(0, h->vaddr + SA5_REPLY_INTR_MASK_OFFSET);
+		(void) readl(h->vaddr + SA5_REPLY_INTR_MASK_OFFSET);
 	} else {
 		h->interrupts_enabled = 0;
 		writel(SA5_PERF_INTR_OFF,
 				h->vaddr + SA5_REPLY_INTR_MASK_OFFSET);
+		(void) readl(h->vaddr + SA5_REPLY_INTR_MASK_OFFSET);
 	}
 }
 

drivers/block/cciss_cmd.h

@@ -53,6 +53,7 @@
 #define CFGTBL_ChangeReq        0x00000001l
 #define CFGTBL_AccCmds          0x00000001l
 #define DOORBELL_CTLR_RESET     0x00000004l
+#define DOORBELL_CTLR_RESET2    0x00000020l
 
 #define CFGTBL_Trans_Simple     0x00000002l
 #define CFGTBL_Trans_Performant 0x00000004l
@@ -142,6 +143,14 @@ typedef struct _ReadCapdata_struct_16
 #define BMIC_CACHE_FLUSH 0xc2
 #define CCISS_CACHE_FLUSH 0x01	/* C2 was already being used by CCISS */
 
+#define CCISS_ABORT_MSG 0x00
+#define CCISS_RESET_MSG 0x01
+#define CCISS_RESET_TYPE_CONTROLLER 0x00
+#define CCISS_RESET_TYPE_BUS 0x01
+#define CCISS_RESET_TYPE_TARGET 0x03
+#define CCISS_RESET_TYPE_LUN 0x04
+#define CCISS_NOOP_MSG 0x03
+
 /* Command List Structure */
 #define CTLR_LUNID "\0\0\0\0\0\0\0\0"
 
@@ -235,6 +244,8 @@ typedef struct _CfgTable_struct {
   u8		   reserved[0x78 - 0x58];
   u32		   misc_fw_support; /* offset 0x78 */
 #define MISC_FW_DOORBELL_RESET (0x02)
+#define MISC_FW_DOORBELL_RESET2 (0x10)
+	u8	   driver_version[32];
 } CfgTable_struct;
 
 struct TransTable_struct {

drivers/block/cciss_scsi.c

@@ -84,7 +84,6 @@ static struct scsi_host_template cciss_driver_template = {
 	.proc_name		= "cciss",
 	.proc_info		= cciss_scsi_proc_info,
 	.queuecommand		= cciss_scsi_queue_command,
-	.can_queue		= SCSI_CCISS_CAN_QUEUE,
 	.this_id		= 7,
 	.cmd_per_lun		= 1,
 	.use_clustering		= DISABLE_CLUSTERING,
@@ -108,16 +107,13 @@ struct cciss_scsi_cmd_stack_elem_t {
 
 #pragma pack()
 
-#define CMD_STACK_SIZE (SCSI_CCISS_CAN_QUEUE * \
-		CCISS_MAX_SCSI_DEVS_PER_HBA + 2)
-			// plus two for init time usage
-
 #pragma pack(1)
 struct cciss_scsi_cmd_stack_t {
 	struct cciss_scsi_cmd_stack_elem_t *pool;
-	struct cciss_scsi_cmd_stack_elem_t *elem[CMD_STACK_SIZE];
+	struct cciss_scsi_cmd_stack_elem_t **elem;
 	dma_addr_t cmd_pool_handle;
 	int top;
+	int nelems;
 };
 #pragma pack()
 
@@ -191,7 +187,7 @@ scsi_cmd_free(ctlr_info_t *h, CommandList_struct *c)
 	sa = h->scsi_ctlr;
 	stk = &sa->cmd_stack; 
 	stk->top++;
-	if (stk->top >= CMD_STACK_SIZE) {
+	if (stk->top >= stk->nelems) {
 		dev_err(&h->pdev->dev,
 			"scsi_cmd_free called too many times.\n");
 		BUG();
@@ -206,13 +202,14 @@ scsi_cmd_stack_setup(ctlr_info_t *h, struct cciss_scsi_adapter_data_t *sa)
 	struct cciss_scsi_cmd_stack_t *stk;
 	size_t size;
 
+	stk = &sa->cmd_stack;
+	stk->nelems = cciss_tape_cmds + 2;
 	sa->cmd_sg_list = cciss_allocate_sg_chain_blocks(h,
-		h->chainsize, CMD_STACK_SIZE);
+		h->chainsize, stk->nelems);
 	if (!sa->cmd_sg_list && h->chainsize > 0)
 		return -ENOMEM;
 
-	stk = &sa->cmd_stack; 
-	size = sizeof(struct cciss_scsi_cmd_stack_elem_t) * CMD_STACK_SIZE;
+	size = sizeof(struct cciss_scsi_cmd_stack_elem_t) * stk->nelems;
 
 	/* Check alignment, see cciss_cmd.h near CommandList_struct def. */
 	BUILD_BUG_ON((sizeof(*stk->pool) % COMMANDLIST_ALIGNMENT) != 0);
@@ -221,18 +218,23 @@ scsi_cmd_stack_setup(ctlr_info_t *h, struct cciss_scsi_adapter_data_t *sa)
 		pci_alloc_consistent(h->pdev, size, &stk->cmd_pool_handle);
 
 	if (stk->pool == NULL) {
-		cciss_free_sg_chain_blocks(sa->cmd_sg_list, CMD_STACK_SIZE);
+		cciss_free_sg_chain_blocks(sa->cmd_sg_list, stk->nelems);
 		sa->cmd_sg_list = NULL;
 		return -ENOMEM;
 	}
-
-	for (i=0; i<CMD_STACK_SIZE; i++) {
+	stk->elem = kmalloc(sizeof(stk->elem[0]) * stk->nelems, GFP_KERNEL);
+	if (!stk->elem) {
+		pci_free_consistent(h->pdev, size, stk->pool,
+		stk->cmd_pool_handle);
+		return -1;
+	}
+	for (i = 0; i < stk->nelems; i++) {
 		stk->elem[i] = &stk->pool[i];
 		stk->elem[i]->busaddr = (__u32) (stk->cmd_pool_handle + 
 			(sizeof(struct cciss_scsi_cmd_stack_elem_t) * i));
 		stk->elem[i]->cmdindex = i;
 	}
-	stk->top = CMD_STACK_SIZE-1;
+	stk->top = stk->nelems-1;
 	return 0;
 }
 
@@ -245,16 +247,18 @@ scsi_cmd_stack_free(ctlr_info_t *h)
 
 	sa = h->scsi_ctlr;
 	stk = &sa->cmd_stack; 
-	if (stk->top != CMD_STACK_SIZE-1) {
+	if (stk->top != stk->nelems-1) {
 		dev_warn(&h->pdev->dev,
 			"bug: %d scsi commands are still outstanding.\n",
-			CMD_STACK_SIZE - stk->top);
+			stk->nelems - stk->top);
 	}
-	size = sizeof(struct cciss_scsi_cmd_stack_elem_t) * CMD_STACK_SIZE;
+	size = sizeof(struct cciss_scsi_cmd_stack_elem_t) * stk->nelems;
 
 	pci_free_consistent(h->pdev, size, stk->pool, stk->cmd_pool_handle);
 	stk->pool = NULL;
-	cciss_free_sg_chain_blocks(sa->cmd_sg_list, CMD_STACK_SIZE);
+	cciss_free_sg_chain_blocks(sa->cmd_sg_list, stk->nelems);
+	kfree(stk->elem);
+	stk->elem = NULL;
 }
 
 #if 0
@@ -859,6 +863,7 @@ cciss_scsi_detect(ctlr_info_t *h)
 	sh->io_port = 0;	// good enough?  FIXME, 
 	sh->n_io_port = 0;	// I don't think we use these two...
 	sh->this_id = SELF_SCSI_ID;  
+	sh->can_queue = cciss_tape_cmds;
 	sh->sg_tablesize = h->maxsgentries;
 	sh->max_cmd_len = MAX_COMMAND_SIZE;
 

drivers/block/cciss_scsi.h

@@ -36,13 +36,9 @@
 		   addressible natively, and may in fact turn
 		   out to be not scsi at all. */
 
-#define SCSI_CCISS_CAN_QUEUE 2
 
 /* 
 
-Note, cmd_per_lun could give us some trouble, so I'm setting it very low.
-Likewise, SCSI_CCISS_CAN_QUEUE is set very conservatively.
-
 If the upper scsi layer tries to track how many commands we have 
 outstanding, it will be operating under the misapprehension that it is
 the only one sending us requests.  We also have the block interface,

drivers/block/drbd/drbd_actlog.c

@@ -28,7 +28,7 @@
 #include "drbd_int.h"
 #include "drbd_wrappers.h"
 
-/* We maintain a trivial check sum in our on disk activity log.
+/* We maintain a trivial checksum in our on disk activity log.
  * With that we can ensure correct operation even when the storage
  * device might do a partial (last) sector write while losing power.
  */

drivers/block/drbd/drbd_bitmap.c

@@ -74,7 +74,7 @@
  *	as we are "attached" to a local disk, which at 32 GiB for 1PiB storage
  *	seems excessive.
  *
- *	We plan to reduce the amount of in-core bitmap pages by pageing them in
+ *	We plan to reduce the amount of in-core bitmap pages by paging them in
  *	and out against their on-disk location as necessary, but need to make
  *	sure we don't cause too much meta data IO, and must not deadlock in
  *	tight memory situations. This needs some more work.
@@ -200,7 +200,7 @@ void drbd_bm_unlock(struct drbd_conf *mdev)
  * we if bits have been cleared since last IO. */
 #define BM_PAGE_LAZY_WRITEOUT	28
 
-/* store_page_idx uses non-atomic assingment. It is only used directly after
+/* store_page_idx uses non-atomic assignment. It is only used directly after
  * allocating the page.  All other bm_set_page_* and bm_clear_page_* need to
  * use atomic bit manipulation, as set_out_of_sync (and therefore bitmap
  * changes) may happen from various contexts, and wait_on_bit/wake_up_bit
@@ -318,7 +318,7 @@ static void bm_unmap(unsigned long *p_addr)
 /* word offset from start of bitmap to word number _in_page_
  * modulo longs per page
 #define MLPP(X) ((X) % (PAGE_SIZE/sizeof(long))
- hm, well, Philipp thinks gcc might not optimze the % into & (... - 1)
+ hm, well, Philipp thinks gcc might not optimize the % into & (... - 1)
  so do it explicitly:
  */
 #define MLPP(X) ((X) & ((PAGE_SIZE/sizeof(long))-1))

drivers/block/drbd/drbd_int.h

@@ -700,7 +700,7 @@ struct drbd_request {
 	 * see drbd_endio_pri(). */
 	struct bio *private_bio;
 
-	struct hlist_node colision;
+	struct hlist_node collision;
 	sector_t sector;
 	unsigned int size;
 	unsigned int epoch; /* barrier_nr */
@@ -766,7 +766,7 @@ struct digest_info {
 
 struct drbd_epoch_entry {
 	struct drbd_work w;
-	struct hlist_node colision;
+	struct hlist_node collision;
 	struct drbd_epoch *epoch; /* for writes */
 	struct drbd_conf *mdev;
 	struct page *pages;
@@ -1129,6 +1129,8 @@ struct drbd_conf {
 	int rs_in_flight; /* resync sectors in flight (to proxy, in proxy and from proxy) */
 	int rs_planed;    /* resync sectors already planned */
 	atomic_t ap_in_flight; /* App sectors in flight (waiting for ack) */
+	int peer_max_bio_size;
+	int local_max_bio_size;
 };
 
 static inline struct drbd_conf *minor_to_mdev(unsigned int minor)
@@ -1218,8 +1220,6 @@ extern void drbd_free_resources(struct drbd_conf *mdev);
 extern void tl_release(struct drbd_conf *mdev, unsigned int barrier_nr,
 		       unsigned int set_size);
 extern void tl_clear(struct drbd_conf *mdev);
-enum drbd_req_event;
-extern void tl_restart(struct drbd_conf *mdev, enum drbd_req_event what);
 extern void _tl_add_barrier(struct drbd_conf *, struct drbd_tl_epoch *);
 extern void drbd_free_sock(struct drbd_conf *mdev);
 extern int drbd_send(struct drbd_conf *mdev, struct socket *sock,
@@ -1434,6 +1434,7 @@ struct bm_extent {
  * hash table. */
 #define HT_SHIFT 8
 #define DRBD_MAX_BIO_SIZE (1U<<(9+HT_SHIFT))
+#define DRBD_MAX_BIO_SIZE_SAFE (1 << 12)       /* Works always = 4k */
 
 #define DRBD_MAX_SIZE_H80_PACKET (1 << 15) /* The old header only allows packets up to 32Kib data */
 
@@ -1518,9 +1519,9 @@ extern void drbd_resume_io(struct drbd_conf *mdev);
 extern char *ppsize(char *buf, unsigned long long size);
 extern sector_t drbd_new_dev_size(struct drbd_conf *, struct drbd_backing_dev *, int);
 enum determine_dev_size { dev_size_error = -1, unchanged = 0, shrunk = 1, grew = 2 };
-extern enum determine_dev_size drbd_determin_dev_size(struct drbd_conf *, enum dds_flags) __must_hold(local);
+extern enum determine_dev_size drbd_determine_dev_size(struct drbd_conf *, enum dds_flags) __must_hold(local);
 extern void resync_after_online_grow(struct drbd_conf *);
-extern void drbd_setup_queue_param(struct drbd_conf *mdev, unsigned int) __must_hold(local);
+extern void drbd_reconsider_max_bio_size(struct drbd_conf *mdev);
 extern enum drbd_state_rv drbd_set_role(struct drbd_conf *mdev,
 					enum drbd_role new_role,
 					int force);
@@ -1828,6 +1829,8 @@ static inline void __drbd_chk_io_error_(struct drbd_conf *mdev, int forcedetach,
 		if (!forcedetach) {
 			if (__ratelimit(&drbd_ratelimit_state))
 				dev_err(DEV, "Local IO failed in %s.\n", where);
+			if (mdev->state.disk > D_INCONSISTENT)
+				_drbd_set_state(_NS(mdev, disk, D_INCONSISTENT), CS_HARD, NULL);
 			break;
 		}
 		/* NOTE fall through to detach case if forcedetach set */
@@ -2153,6 +2156,10 @@ static inline int get_net_conf(struct drbd_conf *mdev)
 static inline void put_ldev(struct drbd_conf *mdev)
 {
 	int i = atomic_dec_return(&mdev->local_cnt);
+
+	/* This may be called from some endio handler,
+	 * so we must not sleep here. */
+
 	__release(local);
 	D_ASSERT(i >= 0);
 	if (i == 0) {

drivers/block/drbd/drbd_main.c

@@ -745,6 +745,9 @@ is_valid_state(struct drbd_conf *mdev, union drbd_state ns)
 		  mdev->agreed_pro_version < 88)
 		rv = SS_NOT_SUPPORTED;
 
+	else if (ns.conn >= C_CONNECTED && ns.pdsk == D_UNKNOWN)
+		rv = SS_CONNECTED_OUTDATES;
+
 	return rv;
 }
 
@@ -1565,6 +1568,10 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
 		put_ldev(mdev);
 	}
 
+	/* Notify peer that I had a local IO error, and did not detached.. */
+	if (os.disk == D_UP_TO_DATE && ns.disk == D_INCONSISTENT)
+		drbd_send_state(mdev);
+
 	/* Disks got bigger while they were detached */
 	if (ns.disk > D_NEGOTIATING && ns.pdsk > D_NEGOTIATING &&
 	    test_and_clear_bit(RESYNC_AFTER_NEG, &mdev->flags)) {
@@ -2064,7 +2071,7 @@ int drbd_send_sizes(struct drbd_conf *mdev, int trigger_reply, enum dds_flags fl
 {
 	struct p_sizes p;
 	sector_t d_size, u_size;
-	int q_order_type;
+	int q_order_type, max_bio_size;
 	int ok;
 
 	if (get_ldev_if_state(mdev, D_NEGOTIATING)) {
@@ -2072,17 +2079,20 @@ int drbd_send_sizes(struct drbd_conf *mdev, int trigger_reply, enum dds_flags fl
 		d_size = drbd_get_max_capacity(mdev->ldev);
 		u_size = mdev->ldev->dc.disk_size;
 		q_order_type = drbd_queue_order_type(mdev);
+		max_bio_size = queue_max_hw_sectors(mdev->ldev->backing_bdev->bd_disk->queue) << 9;
+		max_bio_size = min_t(int, max_bio_size, DRBD_MAX_BIO_SIZE);
 		put_ldev(mdev);
 	} else {
 		d_size = 0;
 		u_size = 0;
 		q_order_type = QUEUE_ORDERED_NONE;
+		max_bio_size = DRBD_MAX_BIO_SIZE; /* ... multiple BIOs per peer_request */
 	}
 
 	p.d_size = cpu_to_be64(d_size);
 	p.u_size = cpu_to_be64(u_size);
 	p.c_size = cpu_to_be64(trigger_reply ? 0 : drbd_get_capacity(mdev->this_bdev));
-	p.max_bio_size = cpu_to_be32(queue_max_hw_sectors(mdev->rq_queue) << 9);
+	p.max_bio_size = cpu_to_be32(max_bio_size);
 	p.queue_order_type = cpu_to_be16(q_order_type);
 	p.dds_flags = cpu_to_be16(flags);
 
@@ -2722,7 +2732,7 @@ int drbd_send_dblock(struct drbd_conf *mdev, struct drbd_request *req)
 
 		/* double check digest, sometimes buffers have been modified in flight. */
 		if (dgs > 0 && dgs <= 64) {
-			/* 64 byte, 512 bit, is the larges digest size
+			/* 64 byte, 512 bit, is the largest digest size
 			 * currently supported in kernel crypto. */
 			unsigned char digest[64];
 			drbd_csum_bio(mdev, mdev->integrity_w_tfm, req->master_bio, digest);
@@ -3041,6 +3051,8 @@ void drbd_init_set_defaults(struct drbd_conf *mdev)
 	mdev->agreed_pro_version = PRO_VERSION_MAX;
 	mdev->write_ordering = WO_bdev_flush;
 	mdev->resync_wenr = LC_FREE;
+	mdev->peer_max_bio_size = DRBD_MAX_BIO_SIZE_SAFE;
+	mdev->local_max_bio_size = DRBD_MAX_BIO_SIZE_SAFE;
 }
 
 void drbd_mdev_cleanup(struct drbd_conf *mdev)
@@ -3275,7 +3287,7 @@ static void drbd_delete_device(unsigned int minor)
 
 	drbd_release_ee_lists(mdev);
 
-	/* should be free'd on disconnect? */
+	/* should be freed on disconnect? */
 	kfree(mdev->ee_hash);
 	/*
 	mdev->ee_hash_s = 0;
@@ -3415,7 +3427,9 @@ struct drbd_conf *drbd_new_device(unsigned int minor)
 	q->backing_dev_info.congested_data = mdev;
 
 	blk_queue_make_request(q, drbd_make_request);
-	blk_queue_max_hw_sectors(q, DRBD_MAX_BIO_SIZE >> 9);
+	/* Setting the max_hw_sectors to an odd value of 8kibyte here
+	   This triggers a max_bio_size message upon first attach or connect */
+	blk_queue_max_hw_sectors(q, DRBD_MAX_BIO_SIZE_SAFE >> 8);
 	blk_queue_bounce_limit(q, BLK_BOUNCE_ANY);
 	blk_queue_merge_bvec(q, drbd_merge_bvec);
 	q->queue_lock = &mdev->req_lock;
@@ -3627,7 +3641,8 @@ struct meta_data_on_disk {
 	      /* `-- act_log->nr_elements <-- sync_conf.al_extents */
 	u32 bm_offset;         /* offset to the bitmap, from here */
 	u32 bm_bytes_per_bit;  /* BM_BLOCK_SIZE */
-	u32 reserved_u32[4];
+	u32 la_peer_max_bio_size;   /* last peer max_bio_size */
+	u32 reserved_u32[3];
 
 } __packed;
 
@@ -3668,6 +3683,7 @@ void drbd_md_sync(struct drbd_conf *mdev)
 	buffer->device_uuid = cpu_to_be64(mdev->ldev->md.device_uuid);
 
 	buffer->bm_offset = cpu_to_be32(mdev->ldev->md.bm_offset);
+	buffer->la_peer_max_bio_size = cpu_to_be32(mdev->peer_max_bio_size);
 
 	D_ASSERT(drbd_md_ss__(mdev, mdev->ldev) == mdev->ldev->md.md_offset);
 	sector = mdev->ldev->md.md_offset;
@@ -3751,6 +3767,15 @@ int drbd_md_read(struct drbd_conf *mdev, struct drbd_backing_dev *bdev)
 	mdev->sync_conf.al_extents = be32_to_cpu(buffer->al_nr_extents);
 	bdev->md.device_uuid = be64_to_cpu(buffer->device_uuid);
 
+	spin_lock_irq(&mdev->req_lock);
+	if (mdev->state.conn < C_CONNECTED) {
+		int peer;
+		peer = be32_to_cpu(buffer->la_peer_max_bio_size);
+		peer = max_t(int, peer, DRBD_MAX_BIO_SIZE_SAFE);
+		mdev->peer_max_bio_size = peer;
+	}
+	spin_unlock_irq(&mdev->req_lock);
+
 	if (mdev->sync_conf.al_extents < 7)
 		mdev->sync_conf.al_extents = 127;
 

drivers/block/drbd/drbd_nl.c

@@ -272,9 +272,28 @@ static int _try_outdate_peer_async(void *data)
 {
 	struct drbd_conf *mdev = (struct drbd_conf *)data;
 	enum drbd_disk_state nps;
+	union drbd_state ns;
 
 	nps = drbd_try_outdate_peer(mdev);
-	drbd_request_state(mdev, NS(pdsk, nps));
+
+	/* Not using
+	   drbd_request_state(mdev, NS(pdsk, nps));
+	   here, because we might were able to re-establish the connection
+	   in the meantime. This can only partially be solved in the state's
+	   engine is_valid_state() and is_valid_state_transition()
+	   functions.
+
+	   nps can be D_INCONSISTENT, D_OUTDATED or D_UNKNOWN.
+	   pdsk == D_INCONSISTENT while conn >= C_CONNECTED is valid,
+	   therefore we have to have the pre state change check here.
+	*/
+	spin_lock_irq(&mdev->req_lock);
+	ns = mdev->state;
+	if (ns.conn < C_WF_REPORT_PARAMS) {
+		ns.pdsk = nps;
+		_drbd_set_state(mdev, ns, CS_VERBOSE, NULL);
+	}
+	spin_unlock_irq(&mdev->req_lock);
 
 	return 0;
 }
@@ -577,7 +596,7 @@ void drbd_resume_io(struct drbd_conf *mdev)
  * Returns 0 on success, negative return values indicate errors.
  * You should call drbd_md_sync() after calling this function.
  */
-enum determine_dev_size drbd_determin_dev_size(struct drbd_conf *mdev, enum dds_flags flags) __must_hold(local)
+enum determine_dev_size drbd_determine_dev_size(struct drbd_conf *mdev, enum dds_flags flags) __must_hold(local)
 {
 	sector_t prev_first_sect, prev_size; /* previous meta location */
 	sector_t la_size;
@@ -773,30 +792,78 @@ static int drbd_check_al_size(struct drbd_conf *mdev)
 	return 0;
 }
 
-void drbd_setup_queue_param(struct drbd_conf *mdev, unsigned int max_bio_size) __must_hold(local)
+static void drbd_setup_queue_param(struct drbd_conf *mdev, unsigned int max_bio_size)
 {
 	struct request_queue * const q = mdev->rq_queue;
-	struct request_queue * const b = mdev->ldev->backing_bdev->bd_disk->queue;
-	int max_segments = mdev->ldev->dc.max_bio_bvecs;
-	int max_hw_sectors = min(queue_max_hw_sectors(b), max_bio_size >> 9);
+	int max_hw_sectors = max_bio_size >> 9;
+	int max_segments = 0;
+
+	if (get_ldev_if_state(mdev, D_ATTACHING)) {
+		struct request_queue * const b = mdev->ldev->backing_bdev->bd_disk->queue;
+
+		max_hw_sectors = min(queue_max_hw_sectors(b), max_bio_size >> 9);
+		max_segments = mdev->ldev->dc.max_bio_bvecs;
+		put_ldev(mdev);
+	}
 
 	blk_queue_logical_block_size(q, 512);
 	blk_queue_max_hw_sectors(q, max_hw_sectors);
 	/* This is the workaround for "bio would need to, but cannot, be split" */
 	blk_queue_max_segments(q, max_segments ? max_segments : BLK_MAX_SEGMENTS);
 	blk_queue_segment_boundary(q, PAGE_CACHE_SIZE-1);
-	blk_queue_stack_limits(q, b);
 
-	dev_info(DEV, "max BIO size = %u\n", queue_max_hw_sectors(q) << 9);
+	if (get_ldev_if_state(mdev, D_ATTACHING)) {
+		struct request_queue * const b = mdev->ldev->backing_bdev->bd_disk->queue;
+
+		blk_queue_stack_limits(q, b);
 
-	if (q->backing_dev_info.ra_pages != b->backing_dev_info.ra_pages) {
-		dev_info(DEV, "Adjusting my ra_pages to backing device's (%lu -> %lu)\n",
-		     q->backing_dev_info.ra_pages,
-		     b->backing_dev_info.ra_pages);
-		q->backing_dev_info.ra_pages = b->backing_dev_info.ra_pages;
+		if (q->backing_dev_info.ra_pages != b->backing_dev_info.ra_pages) {
+			dev_info(DEV, "Adjusting my ra_pages to backing device's (%lu -> %lu)\n",
+				 q->backing_dev_info.ra_pages,
+				 b->backing_dev_info.ra_pages);
+			q->backing_dev_info.ra_pages = b->backing_dev_info.ra_pages;
+		}
+		put_ldev(mdev);
 	}
 }
 
+void drbd_reconsider_max_bio_size(struct drbd_conf *mdev)
+{
+	int now, new, local, peer;
+
+	now = queue_max_hw_sectors(mdev->rq_queue) << 9;
+	local = mdev->local_max_bio_size; /* Eventually last known value, from volatile memory */
+	peer = mdev->peer_max_bio_size; /* Eventually last known value, from meta data */
+
+	if (get_ldev_if_state(mdev, D_ATTACHING)) {
+		local = queue_max_hw_sectors(mdev->ldev->backing_bdev->bd_disk->queue) << 9;
+		mdev->local_max_bio_size = local;
+		put_ldev(mdev);
+	}
+
+	/* We may ignore peer limits if the peer is modern enough.
+	   Because new from 8.3.8 onwards the peer can use multiple
+	   BIOs for a single peer_request */
+	if (mdev->state.conn >= C_CONNECTED) {
+		if (mdev->agreed_pro_version < 94)
+			peer = mdev->peer_max_bio_size;
+		else if (mdev->agreed_pro_version == 94)
+			peer = DRBD_MAX_SIZE_H80_PACKET;
+		else /* drbd 8.3.8 onwards */
+			peer = DRBD_MAX_BIO_SIZE;
+	}
+
+	new = min_t(int, local, peer);
+
+	if (mdev->state.role == R_PRIMARY && new < now)
+		dev_err(DEV, "ASSERT FAILED new < now; (%d < %d)\n", new, now);
+
+	if (new != now)
+		dev_info(DEV, "max BIO size = %u\n", new);
+
+	drbd_setup_queue_param(mdev, new);
+}
+
 /* serialize deconfig (worker exiting, doing cleanup)
  * and reconfig (drbdsetup disk, drbdsetup net)
  *
@@ -865,7 +932,6 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
 	struct block_device *bdev;
 	struct lru_cache *resync_lru = NULL;
 	union drbd_state ns, os;
-	unsigned int max_bio_size;
 	enum drbd_state_rv rv;
 	int cp_discovered = 0;
 	int logical_block_size;
@@ -1117,20 +1183,7 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
 	mdev->read_cnt = 0;
 	mdev->writ_cnt = 0;
 
-	max_bio_size = DRBD_MAX_BIO_SIZE;
-	if (mdev->state.conn == C_CONNECTED) {
-		/* We are Primary, Connected, and now attach a new local
-		 * backing store. We must not increase the user visible maximum
-		 * bio size on this device to something the peer may not be
-		 * able to handle. */
-		if (mdev->agreed_pro_version < 94)
-			max_bio_size = queue_max_hw_sectors(mdev->rq_queue) << 9;
-		else if (mdev->agreed_pro_version == 94)
-			max_bio_size = DRBD_MAX_SIZE_H80_PACKET;
-		/* else: drbd 8.3.9 and later, stay with default */
-	}
-
-	drbd_setup_queue_param(mdev, max_bio_size);
+	drbd_reconsider_max_bio_size(mdev);
 
 	/* If I am currently not R_PRIMARY,
 	 * but meta data primary indicator is set,
@@ -1152,7 +1205,7 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
 	    !drbd_md_test_flag(mdev->ldev, MDF_CONNECTED_IND))
 		set_bit(USE_DEGR_WFC_T, &mdev->flags);
 
-	dd = drbd_determin_dev_size(mdev, 0);
+	dd = drbd_determine_dev_size(mdev, 0);
 	if (dd == dev_size_error) {
 		retcode = ERR_NOMEM_BITMAP;
 		goto force_diskless_dec;
@@ -1281,11 +1334,19 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
 static int drbd_nl_detach(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
 			  struct drbd_nl_cfg_reply *reply)
 {
+	enum drbd_ret_code retcode;
+	int ret;
 	drbd_suspend_io(mdev); /* so no-one is stuck in drbd_al_begin_io */
-	reply->ret_code = drbd_request_state(mdev, NS(disk, D_DISKLESS));
-	if (mdev->state.disk == D_DISKLESS)
-		wait_event(mdev->misc_wait, !atomic_read(&mdev->local_cnt));
+	retcode = drbd_request_state(mdev, NS(disk, D_FAILED));
+	/* D_FAILED will transition to DISKLESS. */
+	ret = wait_event_interruptible(mdev->misc_wait,
+			mdev->state.disk != D_FAILED);
 	drbd_resume_io(mdev);
+	if ((int)retcode == (int)SS_IS_DISKLESS)
+		retcode = SS_NOTHING_TO_DO;
+	if (ret)
+		retcode = ERR_INTR;
+	reply->ret_code = retcode;
 	return 0;
 }
 
@@ -1658,7 +1719,7 @@ static int drbd_nl_resize(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
 
 	mdev->ldev->dc.disk_size = (sector_t)rs.resize_size;
 	ddsf = (rs.resize_force ? DDSF_FORCED : 0) | (rs.no_resync ? DDSF_NO_RESYNC : 0);
-	dd = drbd_determin_dev_size(mdev, ddsf);
+	dd = drbd_determine_dev_size(mdev, ddsf);
 	drbd_md_sync(mdev);
 	put_ldev(mdev);
 	if (dd == dev_size_error) {

drivers/block/drbd/drbd_receiver.c

@@ -333,7 +333,7 @@ struct drbd_epoch_entry *drbd_alloc_ee(struct drbd_conf *mdev,
 	if (!page)
 		goto fail;
 
-	INIT_HLIST_NODE(&e->colision);
+	INIT_HLIST_NODE(&e->collision);
 	e->epoch = NULL;
 	e->mdev = mdev;
 	e->pages = page;
@@ -356,7 +356,7 @@ void drbd_free_some_ee(struct drbd_conf *mdev, struct drbd_epoch_entry *e, int i
 		kfree(e->digest);
 	drbd_pp_free(mdev, e->pages, is_net);
 	D_ASSERT(atomic_read(&e->pending_bios) == 0);
-	D_ASSERT(hlist_unhashed(&e->colision));
+	D_ASSERT(hlist_unhashed(&e->collision));
 	mempool_free(e, drbd_ee_mempool);
 }
 
@@ -787,7 +787,7 @@ static int drbd_connect(struct drbd_conf *mdev)
 		}
 
 		if (sock && msock) {
-			schedule_timeout_interruptible(HZ / 10);
+			schedule_timeout_interruptible(mdev->net_conf->ping_timeo*HZ/10);
 			ok = drbd_socket_okay(mdev, &sock);
 			ok = drbd_socket_okay(mdev, &msock) && ok;
 			if (ok)
@@ -899,11 +899,6 @@ static int drbd_connect(struct drbd_conf *mdev)
 
 	drbd_thread_start(&mdev->asender);
 
-	if (mdev->agreed_pro_version < 95 && get_ldev(mdev)) {
-		drbd_setup_queue_param(mdev, DRBD_MAX_SIZE_H80_PACKET);
-		put_ldev(mdev);
-	}
-
 	if (drbd_send_protocol(mdev) == -1)
 		return -1;
 	drbd_send_sync_param(mdev, &mdev->sync_conf);
@@ -1418,7 +1413,7 @@ static int e_end_resync_block(struct drbd_conf *mdev, struct drbd_work *w, int u
 	sector_t sector = e->sector;
 	int ok;
 
-	D_ASSERT(hlist_unhashed(&e->colision));
+	D_ASSERT(hlist_unhashed(&e->collision));
 
 	if (likely((e->flags & EE_WAS_ERROR) == 0)) {
 		drbd_set_in_sync(mdev, sector, e->size);
@@ -1487,7 +1482,7 @@ static int receive_DataReply(struct drbd_conf *mdev, enum drbd_packets cmd, unsi
 		return false;
 	}
 
-	/* hlist_del(&req->colision) is done in _req_may_be_done, to avoid
+	/* hlist_del(&req->collision) is done in _req_may_be_done, to avoid
 	 * special casing it there for the various failure cases.
 	 * still no race with drbd_fail_pending_reads */
 	ok = recv_dless_read(mdev, req, sector, data_size);
@@ -1558,11 +1553,11 @@ static int e_end_block(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
 	 * P_WRITE_ACK / P_NEG_ACK, to get the sequence number right.  */
 	if (mdev->net_conf->two_primaries) {
 		spin_lock_irq(&mdev->req_lock);
-		D_ASSERT(!hlist_unhashed(&e->colision));
-		hlist_del_init(&e->colision);
+		D_ASSERT(!hlist_unhashed(&e->collision));
+		hlist_del_init(&e->collision);
 		spin_unlock_irq(&mdev->req_lock);
 	} else {
-		D_ASSERT(hlist_unhashed(&e->colision));
+		D_ASSERT(hlist_unhashed(&e->collision));
 	}
 
 	drbd_may_finish_epoch(mdev, e->epoch, EV_PUT + (cancel ? EV_CLEANUP : 0));
@@ -1579,8 +1574,8 @@ static int e_send_discard_ack(struct drbd_conf *mdev, struct drbd_work *w, int u
 	ok = drbd_send_ack(mdev, P_DISCARD_ACK, e);
 
 	spin_lock_irq(&mdev->req_lock);
-	D_ASSERT(!hlist_unhashed(&e->colision));
-	hlist_del_init(&e->colision);
+	D_ASSERT(!hlist_unhashed(&e->collision));
+	hlist_del_init(&e->collision);
 	spin_unlock_irq(&mdev->req_lock);
 
 	dec_unacked(mdev);
@@ -1755,7 +1750,7 @@ static int receive_Data(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned
 
 		spin_lock_irq(&mdev->req_lock);
 
-		hlist_add_head(&e->colision, ee_hash_slot(mdev, sector));
+		hlist_add_head(&e->collision, ee_hash_slot(mdev, sector));
 
 #define OVERLAPS overlaps(i->sector, i->size, sector, size)
 		slot = tl_hash_slot(mdev, sector);
@@ -1765,7 +1760,7 @@ static int receive_Data(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned
 			int have_conflict = 0;
 			prepare_to_wait(&mdev->misc_wait, &wait,
 				TASK_INTERRUPTIBLE);
-			hlist_for_each_entry(i, n, slot, colision) {
+			hlist_for_each_entry(i, n, slot, collision) {
 				if (OVERLAPS) {
 					/* only ALERT on first iteration,
 					 * we may be woken up early... */
@@ -1804,7 +1799,7 @@ static int receive_Data(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned
 			}
 
 			if (signal_pending(current)) {
-				hlist_del_init(&e->colision);
+				hlist_del_init(&e->collision);
 
 				spin_unlock_irq(&mdev->req_lock);
 
@@ -1862,7 +1857,7 @@ static int receive_Data(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned
 	dev_err(DEV, "submit failed, triggering re-connect\n");
 	spin_lock_irq(&mdev->req_lock);
 	list_del(&e->w.list);
-	hlist_del_init(&e->colision);
+	hlist_del_init(&e->collision);
 	spin_unlock_irq(&mdev->req_lock);
 	if (e->flags & EE_CALL_AL_COMPLETE_IO)
 		drbd_al_complete_io(mdev, e->sector);
@@ -2916,12 +2911,6 @@ static int receive_SyncParam(struct drbd_conf *mdev, enum drbd_packets cmd, unsi
 	return false;
 }
 
-static void drbd_setup_order_type(struct drbd_conf *mdev, int peer)
-{
-	/* sorry, we currently have no working implementation
-	 * of distributed TCQ */
-}
-
 /* warn if the arguments differ by more than 12.5% */
 static void warn_if_differ_considerably(struct drbd_conf *mdev,
 	const char *s, sector_t a, sector_t b)
@@ -2939,7 +2928,6 @@ static int receive_sizes(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned
 {
 	struct p_sizes *p = &mdev->data.rbuf.sizes;
 	enum determine_dev_size dd = unchanged;
-	unsigned int max_bio_size;
 	sector_t p_size, p_usize, my_usize;
 	int ldsc = 0; /* local disk size changed */
 	enum dds_flags ddsf;
@@ -2994,7 +2982,7 @@ static int receive_sizes(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned
 
 	ddsf = be16_to_cpu(p->dds_flags);
 	if (get_ldev(mdev)) {
-		dd = drbd_determin_dev_size(mdev, ddsf);
+		dd = drbd_determine_dev_size(mdev, ddsf);
 		put_ldev(mdev);
 		if (dd == dev_size_error)
 			return false;
@@ -3004,23 +2992,15 @@ static int receive_sizes(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned
 		drbd_set_my_capacity(mdev, p_size);
 	}
 
+	mdev->peer_max_bio_size = be32_to_cpu(p->max_bio_size);
+	drbd_reconsider_max_bio_size(mdev);
+
 	if (get_ldev(mdev)) {
 		if (mdev->ldev->known_size != drbd_get_capacity(mdev->ldev->backing_bdev)) {
 			mdev->ldev->known_size = drbd_get_capacity(mdev->ldev->backing_bdev);
 			ldsc = 1;
 		}
 
-		if (mdev->agreed_pro_version < 94)
-			max_bio_size = be32_to_cpu(p->max_bio_size);
-		else if (mdev->agreed_pro_version == 94)
-			max_bio_size = DRBD_MAX_SIZE_H80_PACKET;
-		else /* drbd 8.3.8 onwards */
-			max_bio_size = DRBD_MAX_BIO_SIZE;
-
-		if (max_bio_size != queue_max_hw_sectors(mdev->rq_queue) << 9)
-			drbd_setup_queue_param(mdev, max_bio_size);
-
-		drbd_setup_order_type(mdev, be16_to_cpu(p->queue_order_type));
 		put_ldev(mdev);
 	}
 
@@ -4275,7 +4255,7 @@ static struct drbd_request *_ack_id_to_req(struct drbd_conf *mdev,
 	struct hlist_node *n;
 	struct drbd_request *req;
 
-	hlist_for_each_entry(req, n, slot, colision) {
+	hlist_for_each_entry(req, n, slot, collision) {
 		if ((unsigned long)req == (unsigned long)id) {
 			if (req->sector != sector) {
 				dev_err(DEV, "_ack_id_to_req: found req %p but it has "
@@ -4554,6 +4534,7 @@ int drbd_asender(struct drbd_thread *thi)
 	int received = 0;
 	int expect   = sizeof(struct p_header80);
 	int empty;
+	int ping_timeout_active = 0;
 
 	sprintf(current->comm, "drbd%d_asender", mdev_to_minor(mdev));
 
@@ -4566,6 +4547,7 @@ int drbd_asender(struct drbd_thread *thi)
 			ERR_IF(!drbd_send_ping(mdev)) goto reconnect;
 			mdev->meta.socket->sk->sk_rcvtimeo =
 				mdev->net_conf->ping_timeo*HZ/10;
+			ping_timeout_active = 1;
 		}
 
 		/* conditionally cork;
@@ -4620,8 +4602,7 @@ int drbd_asender(struct drbd_thread *thi)
 			dev_err(DEV, "meta connection shut down by peer.\n");
 			goto reconnect;
 		} else if (rv == -EAGAIN) {
-			if (mdev->meta.socket->sk->sk_rcvtimeo ==
-			    mdev->net_conf->ping_timeo*HZ/10) {
+			if (ping_timeout_active) {
 				dev_err(DEV, "PingAck did not arrive in time.\n");
 				goto reconnect;
 			}
@@ -4660,6 +4641,11 @@ int drbd_asender(struct drbd_thread *thi)
 			if (!cmd->process(mdev, h))
 				goto reconnect;
 
+			/* the idle_timeout (ping-int)
+			 * has been restored in got_PingAck() */
+			if (cmd == get_asender_cmd(P_PING_ACK))
+				ping_timeout_active = 0;
+
 			buf	 = h;
 			received = 0;
 			expect	 = sizeof(struct p_header80);

drivers/block/drbd/drbd_req.c

@@ -163,7 +163,7 @@ static void _about_to_complete_local_write(struct drbd_conf *mdev,
 		 * they must have been failed on the spot */
 #define OVERLAPS overlaps(sector, size, i->sector, i->size)
 		slot = tl_hash_slot(mdev, sector);
-		hlist_for_each_entry(i, n, slot, colision) {
+		hlist_for_each_entry(i, n, slot, collision) {
 			if (OVERLAPS) {
 				dev_alert(DEV, "LOGIC BUG: completed: %p %llus +%u; "
 				      "other: %p %llus +%u\n",
@@ -187,7 +187,7 @@ static void _about_to_complete_local_write(struct drbd_conf *mdev,
 #undef OVERLAPS
 #define OVERLAPS overlaps(sector, size, e->sector, e->size)
 		slot = ee_hash_slot(mdev, req->sector);
-		hlist_for_each_entry(e, n, slot, colision) {
+		hlist_for_each_entry(e, n, slot, collision) {
 			if (OVERLAPS) {
 				wake_up(&mdev->misc_wait);
 				break;
@@ -260,8 +260,8 @@ void _req_may_be_done(struct drbd_request *req, struct bio_and_error *m)
 
 		/* remove the request from the conflict detection
 		 * respective block_id verification hash */
-		if (!hlist_unhashed(&req->colision))
-			hlist_del(&req->colision);
+		if (!hlist_unhashed(&req->collision))
+			hlist_del(&req->collision);
 		else
 			D_ASSERT((s & (RQ_NET_MASK & ~RQ_NET_DONE)) == 0);
 
@@ -329,7 +329,7 @@ static int _req_conflicts(struct drbd_request *req)
 	struct hlist_node *n;
 	struct hlist_head *slot;
 
-	D_ASSERT(hlist_unhashed(&req->colision));
+	D_ASSERT(hlist_unhashed(&req->collision));
 
 	if (!get_net_conf(mdev))
 		return 0;
@@ -341,7 +341,7 @@ static int _req_conflicts(struct drbd_request *req)
 
 #define OVERLAPS overlaps(i->sector, i->size, sector, size)
 	slot = tl_hash_slot(mdev, sector);
-	hlist_for_each_entry(i, n, slot, colision) {
+	hlist_for_each_entry(i, n, slot, collision) {
 		if (OVERLAPS) {
 			dev_alert(DEV, "%s[%u] Concurrent local write detected! "
 			      "[DISCARD L] new: %llus +%u; "
@@ -359,7 +359,7 @@ static int _req_conflicts(struct drbd_request *req)
 #undef OVERLAPS
 #define OVERLAPS overlaps(e->sector, e->size, sector, size)
 		slot = ee_hash_slot(mdev, sector);
-		hlist_for_each_entry(e, n, slot, colision) {
+		hlist_for_each_entry(e, n, slot, collision) {
 			if (OVERLAPS) {
 				dev_alert(DEV, "%s[%u] Concurrent remote write detected!"
 				      " [DISCARD L] new: %llus +%u; "
@@ -491,7 +491,7 @@ int __req_mod(struct drbd_request *req, enum drbd_req_event what,
 
 		/* so we can verify the handle in the answer packet
 		 * corresponding hlist_del is in _req_may_be_done() */
-		hlist_add_head(&req->colision, ar_hash_slot(mdev, req->sector));
+		hlist_add_head(&req->collision, ar_hash_slot(mdev, req->sector));
 
 		set_bit(UNPLUG_REMOTE, &mdev->flags);
 
@@ -507,7 +507,7 @@ int __req_mod(struct drbd_request *req, enum drbd_req_event what,
 		/* assert something? */
 		/* from drbd_make_request_common only */
 
-		hlist_add_head(&req->colision, tl_hash_slot(mdev, req->sector));
+		hlist_add_head(&req->collision, tl_hash_slot(mdev, req->sector));
 		/* corresponding hlist_del is in _req_may_be_done() */
 
 		/* NOTE
@@ -1033,7 +1033,7 @@ static int drbd_make_request_common(struct drbd_conf *mdev, struct bio *bio, uns
 	err = 0;
 
 fail_free_complete:
-	if (rw == WRITE && local)
+	if (req->rq_state & RQ_IN_ACT_LOG)
 		drbd_al_complete_io(mdev, sector);
 fail_and_free_req:
 	if (local) {

drivers/block/drbd/drbd_req.h

@@ -256,7 +256,7 @@ static inline struct drbd_request *_ar_id_to_req(struct drbd_conf *mdev,
 	struct hlist_node *n;
 	struct drbd_request *req;
 
-	hlist_for_each_entry(req, n, slot, colision) {
+	hlist_for_each_entry(req, n, slot, collision) {
 		if ((unsigned long)req == (unsigned long)id) {
 			D_ASSERT(req->sector == sector);
 			return req;
@@ -291,7 +291,7 @@ static inline struct drbd_request *drbd_req_new(struct drbd_conf *mdev,
 		req->epoch       = 0;
 		req->sector      = bio_src->bi_sector;
 		req->size        = bio_src->bi_size;
-		INIT_HLIST_NODE(&req->colision);
+		INIT_HLIST_NODE(&req->collision);
 		INIT_LIST_HEAD(&req->tl_requests);
 		INIT_LIST_HEAD(&req->w.list);
 	}
@@ -323,6 +323,7 @@ extern int __req_mod(struct drbd_request *req, enum drbd_req_event what,
 extern void complete_master_bio(struct drbd_conf *mdev,
 		struct bio_and_error *m);
 extern void request_timer_fn(unsigned long data);
+extern void tl_restart(struct drbd_conf *mdev, enum drbd_req_event what);
 
 /* use this if you don't want to deal with calling complete_master_bio()
  * outside the spinlock, e.g. when walking some list on cleanup. */

drivers/block/drbd/drbd_worker.c

@@ -126,7 +126,7 @@ static void drbd_endio_write_sec_final(struct drbd_epoch_entry *e) __releases(lo
 	list_del(&e->w.list); /* has been on active_ee or sync_ee */
 	list_add_tail(&e->w.list, &mdev->done_ee);
 
-	/* No hlist_del_init(&e->colision) here, we did not send the Ack yet,
+	/* No hlist_del_init(&e->collision) here, we did not send the Ack yet,
 	 * neither did we wake possibly waiting conflicting requests.
 	 * done from "drbd_process_done_ee" within the appropriate w.cb
 	 * (e_end_block/e_end_resync_block) or from _drbd_clear_done_ee */
@@ -297,42 +297,48 @@ void drbd_csum_bio(struct drbd_conf *mdev, struct crypto_hash *tfm, struct bio *
 	crypto_hash_final(&desc, digest);
 }
 
-static int w_e_send_csum(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
+/* TODO merge common code with w_e_end_ov_req */
+int w_e_send_csum(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
 {
 	struct drbd_epoch_entry *e = container_of(w, struct drbd_epoch_entry, w);
 	int digest_size;
 	void *digest;
-	int ok;
+	int ok = 1;
 
 	D_ASSERT(e->block_id == DRBD_MAGIC + 0xbeef);
 
-	if (unlikely(cancel)) {
-		drbd_free_ee(mdev, e);
-		return 1;
-	}
+	if (unlikely(cancel))
+		goto out;
 
-	if (likely((e->flags & EE_WAS_ERROR) == 0)) {
-		digest_size = crypto_hash_digestsize(mdev->csums_tfm);
-		digest = kmalloc(digest_size, GFP_NOIO);
-		if (digest) {
-			drbd_csum_ee(mdev, mdev->csums_tfm, e, digest);
+	if (likely((e->flags & EE_WAS_ERROR) != 0))
+		goto out;
 
-			inc_rs_pending(mdev);
-			ok = drbd_send_drequest_csum(mdev,
-						     e->sector,
-						     e->size,
-						     digest,
-						     digest_size,
-						     P_CSUM_RS_REQUEST);
-			kfree(digest);
-		} else {
-			dev_err(DEV, "kmalloc() of digest failed.\n");
-			ok = 0;
-		}
-	} else
-		ok = 1;
+	digest_size = crypto_hash_digestsize(mdev->csums_tfm);
+	digest = kmalloc(digest_size, GFP_NOIO);
+	if (digest) {
+		sector_t sector = e->sector;
+		unsigned int size = e->size;
+		drbd_csum_ee(mdev, mdev->csums_tfm, e, digest);
+		/* Free e and pages before send.
+		 * In case we block on congestion, we could otherwise run into
+		 * some distributed deadlock, if the other side blocks on
+		 * congestion as well, because our receiver blocks in
+		 * drbd_pp_alloc due to pp_in_use > max_buffers. */
+		drbd_free_ee(mdev, e);
+		e = NULL;
+		inc_rs_pending(mdev);
+		ok = drbd_send_drequest_csum(mdev, sector, size,
+					     digest, digest_size,
+					     P_CSUM_RS_REQUEST);
+		kfree(digest);
+	} else {
+		dev_err(DEV, "kmalloc() of digest failed.\n");
+		ok = 0;
+	}
 
-	drbd_free_ee(mdev, e);
+out:
+	if (e)
+		drbd_free_ee(mdev, e);
 
 	if (unlikely(!ok))
 		dev_err(DEV, "drbd_send_drequest(..., csum) failed\n");
@@ -834,7 +840,7 @@ int drbd_resync_finished(struct drbd_conf *mdev)
 			const int ratio =
 				(t == 0)     ? 0 :
 			(t < 100000) ? ((s*100)/t) : (s/(t/100));
-			dev_info(DEV, "%u %% had equal check sums, eliminated: %luK; "
+			dev_info(DEV, "%u %% had equal checksums, eliminated: %luK; "
 			     "transferred %luK total %luK\n",
 			     ratio,
 			     Bit2KB(mdev->rs_same_csum),
@@ -1071,9 +1077,12 @@ int w_e_end_csum_rs_req(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
 	return ok;
 }
 
+/* TODO merge common code with w_e_send_csum */
 int w_e_end_ov_req(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
 {
 	struct drbd_epoch_entry *e = container_of(w, struct drbd_epoch_entry, w);
+	sector_t sector = e->sector;
+	unsigned int size = e->size;
 	int digest_size;
 	void *digest;
 	int ok = 1;
@@ -1093,17 +1102,25 @@ int w_e_end_ov_req(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
 	else
 		memset(digest, 0, digest_size);
 
+	/* Free e and pages before send.
+	 * In case we block on congestion, we could otherwise run into
+	 * some distributed deadlock, if the other side blocks on
+	 * congestion as well, because our receiver blocks in
+	 * drbd_pp_alloc due to pp_in_use > max_buffers. */
+	drbd_free_ee(mdev, e);
+	e = NULL;
 	inc_rs_pending(mdev);
-	ok = drbd_send_drequest_csum(mdev, e->sector, e->size,
-				     digest, digest_size, P_OV_REPLY);
+	ok = drbd_send_drequest_csum(mdev, sector, size,
+				     digest, digest_size,
+				     P_OV_REPLY);
 	if (!ok)
 		dec_rs_pending(mdev);
 	kfree(digest);
 
 out:
-	drbd_free_ee(mdev, e);
+	if (e)
+		drbd_free_ee(mdev, e);
 	dec_unacked(mdev);
-
 	return ok;
 }
 
@@ -1122,8 +1139,10 @@ int w_e_end_ov_reply(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
 {
 	struct drbd_epoch_entry *e = container_of(w, struct drbd_epoch_entry, w);
 	struct digest_info *di;
-	int digest_size;
 	void *digest;
+	sector_t sector = e->sector;
+	unsigned int size = e->size;
+	int digest_size;
 	int ok, eq = 0;
 
 	if (unlikely(cancel)) {
@@ -1153,16 +1172,21 @@ int w_e_end_ov_reply(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
 		}
 	}
 
-	dec_unacked(mdev);
+		/* Free e and pages before send.
+		 * In case we block on congestion, we could otherwise run into
+		 * some distributed deadlock, if the other side blocks on
+		 * congestion as well, because our receiver blocks in
+		 * drbd_pp_alloc due to pp_in_use > max_buffers. */
+	drbd_free_ee(mdev, e);
 	if (!eq)
-		drbd_ov_oos_found(mdev, e->sector, e->size);
+		drbd_ov_oos_found(mdev, sector, size);
 	else
 		ov_oos_print(mdev);
 
-	ok = drbd_send_ack_ex(mdev, P_OV_RESULT, e->sector, e->size,
+	ok = drbd_send_ack_ex(mdev, P_OV_RESULT, sector, size,
 			      eq ? ID_IN_SYNC : ID_OUT_OF_SYNC);
 
-	drbd_free_ee(mdev, e);
+	dec_unacked(mdev);
 
 	--mdev->ov_left;
 

drivers/block/loop.c

@@ -1658,7 +1658,7 @@ static struct kobject *loop_probe(dev_t dev, int *part, void *data)
 	struct kobject *kobj;
 
 	mutex_lock(&loop_devices_mutex);
-	lo = loop_init_one(dev & MINORMASK);
+	lo = loop_init_one(MINOR(dev) >> part_shift);
 	kobj = lo ? get_disk(lo->lo_disk) : ERR_PTR(-ENOMEM);
 	mutex_unlock(&loop_devices_mutex);
 
@@ -1691,15 +1691,18 @@ static int __init loop_init(void)
 	if (max_part > 0)
 		part_shift = fls(max_part);
 
+	if ((1UL << part_shift) > DISK_MAX_PARTS)
+		return -EINVAL;
+
 	if (max_loop > 1UL << (MINORBITS - part_shift))
 		return -EINVAL;
 
 	if (max_loop) {
 		nr = max_loop;
-		range = max_loop;
+		range = max_loop << part_shift;
 	} else {
 		nr = 8;
-		range = 1UL << (MINORBITS - part_shift);
+		range = 1UL << MINORBITS;
 	}
 
 	if (register_blkdev(LOOP_MAJOR, "loop"))
@@ -1738,7 +1741,7 @@ static void __exit loop_exit(void)
 	unsigned long range;
 	struct loop_device *lo, *next;
 
-	range = max_loop ? max_loop :  1UL << (MINORBITS - part_shift);
+	range = max_loop ? max_loop << part_shift : 1UL << MINORBITS;
 
 	list_for_each_entry_safe(lo, next, &loop_devices, lo_list)
 		loop_del_one(lo);

drivers/block/xen-blkback/Makefile

+obj-$(CONFIG_XEN_BLKDEV_BACKEND) := xen-blkback.o
+
+xen-blkback-y	:= blkback.o xenbus.o

drivers/block/xen-blkback/common.h

+/*
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __XEN_BLKIF__BACKEND__COMMON_H__
+#define __XEN_BLKIF__BACKEND__COMMON_H__
+
+#include <linux/version.h>
+#include <linux/module.h>
+#include <linux/interrupt.h>
+#include <linux/slab.h>
+#include <linux/blkdev.h>
+#include <linux/vmalloc.h>
+#include <linux/wait.h>
+#include <linux/io.h>
+#include <asm/setup.h>
+#include <asm/pgalloc.h>
+#include <asm/hypervisor.h>
+#include <xen/grant_table.h>
+#include <xen/xenbus.h>
+#include <xen/interface/io/ring.h>
+#include <xen/interface/io/blkif.h>
+#include <xen/interface/io/protocols.h>
+
+#define DRV_PFX "xen-blkback:"
+#define DPRINTK(fmt, args...)				\
+	pr_debug(DRV_PFX "(%s:%d) " fmt ".\n",	\
+		 __func__, __LINE__, ##args)
+
+
+/* Not a real protocol.  Used to generate ring structs which contain
+ * the elements common to all protocols only.  This way we get a
+ * compiler-checkable way to use common struct elements, so we can
+ * avoid using switch(protocol) in a number of places.  */
+struct blkif_common_request {
+	char dummy;
+};
+struct blkif_common_response {
+	char dummy;
+};
+
+/* i386 protocol version */
+#pragma pack(push, 4)
+struct blkif_x86_32_request {
+	uint8_t        operation;    /* BLKIF_OP_???                         */
+	uint8_t        nr_segments;  /* number of segments                   */
+	blkif_vdev_t   handle;       /* only for read/write requests         */
+	uint64_t       id;           /* private guest value, echoed in resp  */
+	blkif_sector_t sector_number;/* start sector idx on disk (r/w only)  */
+	struct blkif_request_segment seg[BLKIF_MAX_SEGMENTS_PER_REQUEST];
+};
+struct blkif_x86_32_response {
+	uint64_t        id;              /* copied from request */
+	uint8_t         operation;       /* copied from request */
+	int16_t         status;          /* BLKIF_RSP_???       */
+};
+#pragma pack(pop)
+
+/* x86_64 protocol version */
+struct blkif_x86_64_request {
+	uint8_t        operation;    /* BLKIF_OP_???                         */
+	uint8_t        nr_segments;  /* number of segments                   */
+	blkif_vdev_t   handle;       /* only for read/write requests         */
+	uint64_t       __attribute__((__aligned__(8))) id;
+	blkif_sector_t sector_number;/* start sector idx on disk (r/w only)  */
+	struct blkif_request_segment seg[BLKIF_MAX_SEGMENTS_PER_REQUEST];
+};
+struct blkif_x86_64_response {
+	uint64_t       __attribute__((__aligned__(8))) id;
+	uint8_t         operation;       /* copied from request */
+	int16_t         status;          /* BLKIF_RSP_???       */
+};
+
+DEFINE_RING_TYPES(blkif_common, struct blkif_common_request,
+		  struct blkif_common_response);
+DEFINE_RING_TYPES(blkif_x86_32, struct blkif_x86_32_request,
+		  struct blkif_x86_32_response);
+DEFINE_RING_TYPES(blkif_x86_64, struct blkif_x86_64_request,
+		  struct blkif_x86_64_response);
+
+union blkif_back_rings {
+	struct blkif_back_ring        native;
+	struct blkif_common_back_ring common;
+	struct blkif_x86_32_back_ring x86_32;
+	struct blkif_x86_64_back_ring x86_64;
+};
+
+enum blkif_protocol {
+	BLKIF_PROTOCOL_NATIVE = 1,
+	BLKIF_PROTOCOL_X86_32 = 2,
+	BLKIF_PROTOCOL_X86_64 = 3,
+};
+
+struct xen_vbd {
+	/* What the domain refers to this vbd as. */
+	blkif_vdev_t		handle;
+	/* Non-zero -> read-only */
+	unsigned char		readonly;
+	/* VDISK_xxx */
+	unsigned char		type;
+	/* phys device that this vbd maps to. */
+	u32			pdevice;
+	struct block_device	*bdev;
+	/* Cached size parameter. */
+	sector_t		size;
+	bool			flush_support;
+};
+
+struct backend_info;
+
+struct xen_blkif {
+	/* Unique identifier for this interface. */
+	domid_t			domid;
+	unsigned int		handle;
+	/* Physical parameters of the comms window. */
+	unsigned int		irq;
+	/* Comms information. */
+	enum blkif_protocol	blk_protocol;
+	union blkif_back_rings	blk_rings;
+	struct vm_struct	*blk_ring_area;
+	/* 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;
+	/* One thread per one blkif. */
+	struct task_struct	*xenblkd;
+	unsigned int		waiting_reqs;
+
+	/* statistics */
+	unsigned long		st_print;
+	int			st_rd_req;
+	int			st_wr_req;
+	int			st_oo_req;
+	int			st_f_req;
+	int			st_rd_sect;
+	int			st_wr_sect;
+
+	wait_queue_head_t	waiting_to_free;
+
+	grant_handle_t		shmem_handle;
+	grant_ref_t		shmem_ref;
+};
+
+
+#define vbd_sz(_v)	((_v)->bdev->bd_part ? \
+			 (_v)->bdev->bd_part->nr_sects : \
+			  get_capacity((_v)->bdev->bd_disk))
+
+#define xen_blkif_get(_b) (atomic_inc(&(_b)->refcnt))
+#define xen_blkif_put(_b)				\
+	do {						\
+		if (atomic_dec_and_test(&(_b)->refcnt))	\
+			wake_up(&(_b)->waiting_to_free);\
+	} while (0)
+
+struct phys_req {
+	unsigned short		dev;
+	unsigned short		nr_sects;
+	struct block_device	*bdev;
+	blkif_sector_t		sector_number;
+};
+int xen_blkif_interface_init(void);
+
+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_blkbk_flush_diskcache(struct xenbus_transaction xbt,
+			      struct backend_info *be, int state);
+
+struct xenbus_device *xen_blkbk_xenbus(struct backend_info *be);
+
+static inline void blkif_get_x86_32_req(struct blkif_request *dst,
+					struct blkif_x86_32_request *src)
+{
+	int i, n = BLKIF_MAX_SEGMENTS_PER_REQUEST;
+	dst->operation = src->operation;
+	dst->nr_segments = src->nr_segments;
+	dst->handle = src->handle;
+	dst->id = src->id;
+	dst->u.rw.sector_number = src->sector_number;
+	barrier();
+	if (n > dst->nr_segments)
+		n = dst->nr_segments;
+	for (i = 0; i < n; i++)
+		dst->u.rw.seg[i] = src->seg[i];
+}
+
+static inline void blkif_get_x86_64_req(struct blkif_request *dst,
+					struct blkif_x86_64_request *src)
+{
+	int i, n = BLKIF_MAX_SEGMENTS_PER_REQUEST;
+	dst->operation = src->operation;
+	dst->nr_segments = src->nr_segments;
+	dst->handle = src->handle;
+	dst->id = src->id;
+	dst->u.rw.sector_number = src->sector_number;
+	barrier();
+	if (n > dst->nr_segments)
+		n = dst->nr_segments;
+	for (i = 0; i < n; i++)
+		dst->u.rw.seg[i] = src->seg[i];
+}
+
+#endif /* __XEN_BLKIF__BACKEND__COMMON_H__ */

drivers/block/xen-blkback/xenbus.c

+/*  Xenbus code for blkif backend
+    Copyright (C) 2005 Rusty Russell <rusty@rustcorp.com.au>
+    Copyright (C) 2005 XenSource Ltd
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+*/
+
+#include <stdarg.h>
+#include <linux/module.h>
+#include <linux/kthread.h>
+#include <xen/events.h>
+#include <xen/grant_table.h>
+#include "common.h"
+
+struct backend_info {
+	struct xenbus_device	*dev;
+	struct xen_blkif	*blkif;
+	struct xenbus_watch	backend_watch;
+	unsigned		major;
+	unsigned		minor;
+	char			*mode;
+};
+
+static struct kmem_cache *xen_blkif_cachep;
+static void connect(struct backend_info *);
+static int connect_ring(struct backend_info *);
+static void backend_changed(struct xenbus_watch *, const char **,
+			    unsigned int);
+
+struct xenbus_device *xen_blkbk_xenbus(struct backend_info *be)
+{
+	return be->dev;
+}
+
+static int blkback_name(struct xen_blkif *blkif, char *buf)
+{
+	char *devpath, *devname;
+	struct xenbus_device *dev = blkif->be->dev;
+
+	devpath = xenbus_read(XBT_NIL, dev->nodename, "dev", NULL);
+	if (IS_ERR(devpath))
+		return PTR_ERR(devpath);
+
+	devname = strstr(devpath, "/dev/");
+	if (devname != NULL)
+		devname += strlen("/dev/");
+	else
+		devname  = devpath;
+
+	snprintf(buf, TASK_COMM_LEN, "blkback.%d.%s", blkif->domid, devname);
+	kfree(devpath);
+
+	return 0;
+}
+
+static void xen_update_blkif_status(struct xen_blkif *blkif)
+{
+	int err;
+	char name[TASK_COMM_LEN];
+
+	/* Not ready to connect? */
+	if (!blkif->irq || !blkif->vbd.bdev)
+		return;
+
+	/* Already connected? */
+	if (blkif->be->dev->state == XenbusStateConnected)
+		return;
+
+	/* Attempt to connect: exit if we fail to. */
+	connect(blkif->be);
+	if (blkif->be->dev->state != XenbusStateConnected)
+		return;
+
+	err = blkback_name(blkif, name);
+	if (err) {
+		xenbus_dev_error(blkif->be->dev, err, "get blkback dev name");
+		return;
+	}
+
+	err = filemap_write_and_wait(blkif->vbd.bdev->bd_inode->i_mapping);
+	if (err) {
+		xenbus_dev_error(blkif->be->dev, err, "block flush");
+		return;
+	}
+	invalidate_inode_pages2(blkif->vbd.bdev->bd_inode->i_mapping);
+
+	blkif->xenblkd = kthread_run(xen_blkif_schedule, blkif, name);
+	if (IS_ERR(blkif->xenblkd)) {
+		err = PTR_ERR(blkif->xenblkd);
+		blkif->xenblkd = NULL;
+		xenbus_dev_error(blkif->be->dev, err, "start xenblkd");
+	}
+}
+
+static struct xen_blkif *xen_blkif_alloc(domid_t domid)
+{
+	struct xen_blkif *blkif;
+
+	blkif = kmem_cache_alloc(xen_blkif_cachep, GFP_KERNEL);
+	if (!blkif)
+		return ERR_PTR(-ENOMEM);
+
+	memset(blkif, 0, sizeof(*blkif));
+	blkif->domid = domid;
+	spin_lock_init(&blkif->blk_ring_lock);
+	atomic_set(&blkif->refcnt, 1);
+	init_waitqueue_head(&blkif->wq);
+	blkif->st_print = jiffies;
+	init_waitqueue_head(&blkif->waiting_to_free);
+
+	return blkif;
+}
+
+static int map_frontend_page(struct xen_blkif *blkif, unsigned long shared_page)
+{
+	struct gnttab_map_grant_ref op;
+
+	gnttab_set_map_op(&op, (unsigned long)blkif->blk_ring_area->addr,
+			  GNTMAP_host_map, shared_page, blkif->domid);
+
+	if (HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, &op, 1))
+		BUG();
+
+	if (op.status) {
+		DPRINTK("Grant table operation failure !\n");
+		return op.status;
+	}
+
+	blkif->shmem_ref = shared_page;
+	blkif->shmem_handle = op.handle;
+
+	return 0;
+}
+
+static void unmap_frontend_page(struct xen_blkif *blkif)
+{
+	struct gnttab_unmap_grant_ref op;
+
+	gnttab_set_unmap_op(&op, (unsigned long)blkif->blk_ring_area->addr,
+			    GNTMAP_host_map, blkif->shmem_handle);
+
+	if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, &op, 1))
+		BUG();
+}
+
+static int xen_blkif_map(struct xen_blkif *blkif, unsigned long shared_page,
+			 unsigned int evtchn)
+{
+	int err;
+
+	/* Already connected through? */
+	if (blkif->irq)
+		return 0;
+
+	blkif->blk_ring_area = alloc_vm_area(PAGE_SIZE);
+	if (!blkif->blk_ring_area)
+		return -ENOMEM;
+
+	err = map_frontend_page(blkif, shared_page);
+	if (err) {
+		free_vm_area(blkif->blk_ring_area);
+		return err;
+	}
+
+	switch (blkif->blk_protocol) {
+	case BLKIF_PROTOCOL_NATIVE:
+	{
+		struct blkif_sring *sring;
+		sring = (struct blkif_sring *)blkif->blk_ring_area->addr;
+		BACK_RING_INIT(&blkif->blk_rings.native, sring, PAGE_SIZE);
+		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_area->addr;
+		BACK_RING_INIT(&blkif->blk_rings.x86_32, sring_x86_32, PAGE_SIZE);
+		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_area->addr;
+		BACK_RING_INIT(&blkif->blk_rings.x86_64, sring_x86_64, PAGE_SIZE);
+		break;
+	}
+	default:
+		BUG();
+	}
+
+	err = bind_interdomain_evtchn_to_irqhandler(blkif->domid, evtchn,
+						    xen_blkif_be_int, 0,
+						    "blkif-backend", blkif);
+	if (err < 0) {
+		unmap_frontend_page(blkif);
+		free_vm_area(blkif->blk_ring_area);
+		blkif->blk_rings.common.sring = NULL;
+		return err;
+	}
+	blkif->irq = err;
+
+	return 0;
+}
+
+static void xen_blkif_disconnect(struct xen_blkif *blkif)
+{
+	if (blkif->xenblkd) {
+		kthread_stop(blkif->xenblkd);
+		blkif->xenblkd = NULL;
+	}
+
+	atomic_dec(&blkif->refcnt);
+	wait_event(blkif->waiting_to_free, atomic_read(&blkif->refcnt) == 0);
+	atomic_inc(&blkif->refcnt);
+
+	if (blkif->irq) {
+		unbind_from_irqhandler(blkif->irq, blkif);
+		blkif->irq = 0;
+	}
+
+	if (blkif->blk_rings.common.sring) {
+		unmap_frontend_page(blkif);
+		free_vm_area(blkif->blk_ring_area);
+		blkif->blk_rings.common.sring = NULL;
+	}
+}
+
+void xen_blkif_free(struct xen_blkif *blkif)
+{
+	if (!atomic_dec_and_test(&blkif->refcnt))
+		BUG();
+	kmem_cache_free(xen_blkif_cachep, blkif);
+}
+
+int __init xen_blkif_interface_init(void)
+{
+	xen_blkif_cachep = kmem_cache_create("blkif_cache",
+					     sizeof(struct xen_blkif),
+					     0, 0, NULL);
+	if (!xen_blkif_cachep)
+		return -ENOMEM;
+
+	return 0;
+}
+
+/*
+ *  sysfs interface for VBD I/O requests
+ */
+
+#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(oo_req,  "%d\n", be->blkif->st_oo_req);
+VBD_SHOW(rd_req,  "%d\n", be->blkif->st_rd_req);
+VBD_SHOW(wr_req,  "%d\n", be->blkif->st_wr_req);
+VBD_SHOW(f_req,  "%d\n", be->blkif->st_f_req);
+VBD_SHOW(rd_sect, "%d\n", be->blkif->st_rd_sect);
+VBD_SHOW(wr_sect, "%d\n", be->blkif->st_wr_sect);
+
+static struct attribute *xen_vbdstat_attrs[] = {
+	&dev_attr_oo_req.attr,
+	&dev_attr_rd_req.attr,
+	&dev_attr_wr_req.attr,
+	&dev_attr_f_req.attr,
+	&dev_attr_rd_sect.attr,
+	&dev_attr_wr_sect.attr,
+	NULL
+};
+
+static struct attribute_group xen_vbdstat_group = {
+	.name = "statistics",
+	.attrs = xen_vbdstat_attrs,
+};
+
+VBD_SHOW(physical_device, "%x:%x\n", be->major, be->minor);
+VBD_SHOW(mode, "%s\n", be->mode);
+
+int xenvbd_sysfs_addif(struct xenbus_device *dev)
+{
+	int error;
+
+	error = device_create_file(&dev->dev, &dev_attr_physical_device);
+	if (error)
+		goto fail1;
+
+	error = device_create_file(&dev->dev, &dev_attr_mode);
+	if (error)
+		goto fail2;
+
+	error = sysfs_create_group(&dev->dev.kobj, &xen_vbdstat_group);
+	if (error)
+		goto fail3;
+
+	return 0;
+
+fail3:	sysfs_remove_group(&dev->dev.kobj, &xen_vbdstat_group);
+fail2:	device_remove_file(&dev->dev, &dev_attr_mode);
+fail1:	device_remove_file(&dev->dev, &dev_attr_physical_device);
+	return error;
+}
+
+void xenvbd_sysfs_delif(struct xenbus_device *dev)
+{
+	sysfs_remove_group(&dev->dev.kobj, &xen_vbdstat_group);
+	device_remove_file(&dev->dev, &dev_attr_mode);
+	device_remove_file(&dev->dev, &dev_attr_physical_device);
+}
+
+
+static void xen_vbd_free(struct xen_vbd *vbd)
+{
+	if (vbd->bdev)
+		blkdev_put(vbd->bdev, vbd->readonly ? FMODE_READ : FMODE_WRITE);
+	vbd->bdev = NULL;
+}
+
+static int xen_vbd_create(struct xen_blkif *blkif, blkif_vdev_t handle,
+			  unsigned major, unsigned minor, int readonly,
+			  int cdrom)
+{
+	struct xen_vbd *vbd;
+	struct block_device *bdev;
+	struct request_queue *q;
+
+	vbd = &blkif->vbd;
+	vbd->handle   = handle;
+	vbd->readonly = readonly;
+	vbd->type     = 0;
+
+	vbd->pdevice  = MKDEV(major, minor);
+
+	bdev = blkdev_get_by_dev(vbd->pdevice, vbd->readonly ?
+				 FMODE_READ : FMODE_WRITE, NULL);
+
+	if (IS_ERR(bdev)) {
+		DPRINTK("xen_vbd_create: device %08x could not be opened.\n",
+			vbd->pdevice);
+		return -ENOENT;
+	}
+
+	vbd->bdev = bdev;
+	vbd->size = vbd_sz(vbd);
+
+	if (vbd->bdev->bd_disk == NULL) {
+		DPRINTK("xen_vbd_create: device %08x doesn't exist.\n",
+			vbd->pdevice);
+		xen_vbd_free(vbd);
+		return -ENOENT;
+	}
+
+	if (vbd->bdev->bd_disk->flags & GENHD_FL_CD || cdrom)
+		vbd->type |= VDISK_CDROM;
+	if (vbd->bdev->bd_disk->flags & GENHD_FL_REMOVABLE)
+		vbd->type |= VDISK_REMOVABLE;
+
+	q = bdev_get_queue(bdev);
+	if (q && q->flush_flags)
+		vbd->flush_support = true;
+
+	DPRINTK("Successful creation of handle=%04x (dom=%u)\n",
+		handle, blkif->domid);
+	return 0;
+}
+static int xen_blkbk_remove(struct xenbus_device *dev)
+{
+	struct backend_info *be = dev_get_drvdata(&dev->dev);
+
+	DPRINTK("");
+
+	if (be->major || be->minor)
+		xenvbd_sysfs_delif(dev);
+
+	if (be->backend_watch.node) {
+		unregister_xenbus_watch(&be->backend_watch);
+		kfree(be->backend_watch.node);
+		be->backend_watch.node = NULL;
+	}
+
+	if (be->blkif) {
+		xen_blkif_disconnect(be->blkif);
+		xen_vbd_free(&be->blkif->vbd);
+		xen_blkif_free(be->blkif);
+		be->blkif = NULL;
+	}
+
+	kfree(be);
+	dev_set_drvdata(&dev->dev, NULL);
+	return 0;
+}
+
+int xen_blkbk_flush_diskcache(struct xenbus_transaction xbt,
+			      struct backend_info *be, int state)
+{
+	struct xenbus_device *dev = be->dev;
+	int err;
+
+	err = xenbus_printf(xbt, dev->nodename, "feature-flush-cache",
+			    "%d", state);
+	if (err)
+		xenbus_dev_fatal(dev, err, "writing feature-flush-cache");
+
+	return err;
+}
+
+/*
+ * Entry point to this code when a new device is created.  Allocate the basic
+ * structures, and watch the store waiting for the hotplug scripts to tell us
+ * the device's physical major and minor numbers.  Switch to InitWait.
+ */
+static int xen_blkbk_probe(struct xenbus_device *dev,
+			   const struct xenbus_device_id *id)
+{
+	int err;
+	struct backend_info *be = kzalloc(sizeof(struct backend_info),
+					  GFP_KERNEL);
+	if (!be) {
+		xenbus_dev_fatal(dev, -ENOMEM,
+				 "allocating backend structure");
+		return -ENOMEM;
+	}
+	be->dev = dev;
+	dev_set_drvdata(&dev->dev, be);
+
+	be->blkif = xen_blkif_alloc(dev->otherend_id);
+	if (IS_ERR(be->blkif)) {
+		err = PTR_ERR(be->blkif);
+		be->blkif = NULL;
+		xenbus_dev_fatal(dev, err, "creating block interface");
+		goto fail;
+	}
+
+	/* setup back pointer */
+	be->blkif->be = be;
+
+	err = xenbus_watch_pathfmt(dev, &be->backend_watch, backend_changed,
+				   "%s/%s", dev->nodename, "physical-device");
+	if (err)
+		goto fail;
+
+	err = xenbus_switch_state(dev, XenbusStateInitWait);
+	if (err)
+		goto fail;
+
+	return 0;
+
+fail:
+	DPRINTK("failed");
+	xen_blkbk_remove(dev);
+	return err;
+}
+
+
+/*
+ * Callback received when the hotplug scripts have placed the physical-device
+ * node.  Read it and the mode node, and create a vbd.  If the frontend is
+ * ready, connect.
+ */
+static void backend_changed(struct xenbus_watch *watch,
+			    const char **vec, unsigned int len)
+{
+	int err;
+	unsigned major;
+	unsigned minor;
+	struct backend_info *be
+		= container_of(watch, struct backend_info, backend_watch);
+	struct xenbus_device *dev = be->dev;
+	int cdrom = 0;
+	char *device_type;
+
+	DPRINTK("");
+
+	err = xenbus_scanf(XBT_NIL, dev->nodename, "physical-device", "%x:%x",
+			   &major, &minor);
+	if (XENBUS_EXIST_ERR(err)) {
+		/*
+		 * Since this watch will fire once immediately after it is
+		 * registered, we expect this.  Ignore it, and wait for the
+		 * hotplug scripts.
+		 */
+		return;
+	}
+	if (err != 2) {
+		xenbus_dev_fatal(dev, err, "reading physical-device");
+		return;
+	}
+
+	if ((be->major || be->minor) &&
+	    ((be->major != major) || (be->minor != minor))) {
+		pr_warn(DRV_PFX "changing physical device (from %x:%x to %x:%x) not supported.\n",
+			be->major, be->minor, major, minor);
+		return;
+	}
+
+	be->mode = xenbus_read(XBT_NIL, dev->nodename, "mode", NULL);
+	if (IS_ERR(be->mode)) {
+		err = PTR_ERR(be->mode);
+		be->mode = NULL;
+		xenbus_dev_fatal(dev, err, "reading mode");
+		return;
+	}
+
+	device_type = xenbus_read(XBT_NIL, dev->otherend, "device-type", NULL);
+	if (!IS_ERR(device_type)) {
+		cdrom = strcmp(device_type, "cdrom") == 0;
+		kfree(device_type);
+	}
+
+	if (be->major == 0 && be->minor == 0) {
+		/* Front end dir is a number, which is used as the handle. */
+
+		char *p = strrchr(dev->otherend, '/') + 1;
+		long handle;
+		err = strict_strtoul(p, 0, &handle);
+		if (err)
+			return;
+
+		be->major = major;
+		be->minor = minor;
+
+		err = xen_vbd_create(be->blkif, handle, major, minor,
+				 (NULL == strchr(be->mode, 'w')), cdrom);
+		if (err) {
+			be->major = 0;
+			be->minor = 0;
+			xenbus_dev_fatal(dev, err, "creating vbd structure");
+			return;
+		}
+
+		err = xenvbd_sysfs_addif(dev);
+		if (err) {
+			xen_vbd_free(&be->blkif->vbd);
+			be->major = 0;
+			be->minor = 0;
+			xenbus_dev_fatal(dev, err, "creating sysfs entries");
+			return;
+		}
+
+		/* We're potentially connected now */
+		xen_update_blkif_status(be->blkif);
+	}
+}
+
+
+/*
+ * Callback received when the frontend's state changes.
+ */
+static void frontend_changed(struct xenbus_device *dev,
+			     enum xenbus_state frontend_state)
+{
+	struct backend_info *be = dev_get_drvdata(&dev->dev);
+	int err;
+
+	DPRINTK("%s", xenbus_strstate(frontend_state));
+
+	switch (frontend_state) {
+	case XenbusStateInitialising:
+		if (dev->state == XenbusStateClosed) {
+			pr_info(DRV_PFX "%s: prepare for reconnect\n",
+				dev->nodename);
+			xenbus_switch_state(dev, XenbusStateInitWait);
+		}
+		break;
+
+	case XenbusStateInitialised:
+	case XenbusStateConnected:
+		/*
+		 * Ensure we connect even when two watches fire in
+		 * close successsion and we miss the intermediate value
+		 * of frontend_state.
+		 */
+		if (dev->state == XenbusStateConnected)
+			break;
+
+		/*
+		 * Enforce precondition before potential leak point.
+		 * blkif_disconnect() is idempotent.
+		 */
+		xen_blkif_disconnect(be->blkif);
+
+		err = connect_ring(be);
+		if (err)
+			break;
+		xen_update_blkif_status(be->blkif);
+		break;
+
+	case XenbusStateClosing:
+		xen_blkif_disconnect(be->blkif);
+		xenbus_switch_state(dev, XenbusStateClosing);
+		break;
+
+	case XenbusStateClosed:
+		xenbus_switch_state(dev, XenbusStateClosed);
+		if (xenbus_dev_is_online(dev))
+			break;
+		/* fall through if not online */
+	case XenbusStateUnknown:
+		/* implies blkif_disconnect() via blkback_remove() */
+		device_unregister(&dev->dev);
+		break;
+
+	default:
+		xenbus_dev_fatal(dev, -EINVAL, "saw state %d at frontend",
+				 frontend_state);
+		break;
+	}
+}
+
+
+/* ** Connection ** */
+
+
+/*
+ * Write the physical details regarding the block device to the store, and
+ * switch to Connected state.
+ */
+static void connect(struct backend_info *be)
+{
+	struct xenbus_transaction xbt;
+	int err;
+	struct xenbus_device *dev = be->dev;
+
+	DPRINTK("%s", dev->otherend);
+
+	/* Supply the information about the device the frontend needs */
+again:
+	err = xenbus_transaction_start(&xbt);
+	if (err) {
+		xenbus_dev_fatal(dev, err, "starting transaction");
+		return;
+	}
+
+	err = xen_blkbk_flush_diskcache(xbt, be, be->blkif->vbd.flush_support);
+	if (err)
+		goto abort;
+
+	err = xenbus_printf(xbt, dev->nodename, "sectors", "%llu",
+			    (unsigned long long)vbd_sz(&be->blkif->vbd));
+	if (err) {
+		xenbus_dev_fatal(dev, err, "writing %s/sectors",
+				 dev->nodename);
+		goto abort;
+	}
+
+	/* FIXME: use a typename instead */
+	err = xenbus_printf(xbt, dev->nodename, "info", "%u",
+			    be->blkif->vbd.type |
+			    (be->blkif->vbd.readonly ? VDISK_READONLY : 0));
+	if (err) {
+		xenbus_dev_fatal(dev, err, "writing %s/info",
+				 dev->nodename);
+		goto abort;
+	}
+	err = xenbus_printf(xbt, dev->nodename, "sector-size", "%lu",
+			    (unsigned long)
+			    bdev_logical_block_size(be->blkif->vbd.bdev));
+	if (err) {
+		xenbus_dev_fatal(dev, err, "writing %s/sector-size",
+				 dev->nodename);
+		goto abort;
+	}
+
+	err = xenbus_transaction_end(xbt, 0);
+	if (err == -EAGAIN)
+		goto again;
+	if (err)
+		xenbus_dev_fatal(dev, err, "ending transaction");
+
+	err = xenbus_switch_state(dev, XenbusStateConnected);
+	if (err)
+		xenbus_dev_fatal(dev, err, "switching to Connected state",
+				 dev->nodename);
+
+	return;
+ abort:
+	xenbus_transaction_end(xbt, 1);
+}
+
+
+static int connect_ring(struct backend_info *be)
+{
+	struct xenbus_device *dev = be->dev;
+	unsigned long ring_ref;
+	unsigned int evtchn;
+	char protocol[64] = "";
+	int err;
+
+	DPRINTK("%s", dev->otherend);
+
+	err = xenbus_gather(XBT_NIL, dev->otherend, "ring-ref", "%lu",
+			    &ring_ref, "event-channel", "%u", &evtchn, NULL);
+	if (err) {
+		xenbus_dev_fatal(dev, err,
+				 "reading %s/ring-ref and event-channel",
+				 dev->otherend);
+		return err;
+	}
+
+	be->blkif->blk_protocol = BLKIF_PROTOCOL_NATIVE;
+	err = xenbus_gather(XBT_NIL, dev->otherend, "protocol",
+			    "%63s", protocol, NULL);
+	if (err)
+		strcpy(protocol, "unspecified, assuming native");
+	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;
+	}
+	pr_info(DRV_PFX "ring-ref %ld, event-channel %d, protocol %d (%s)\n",
+		ring_ref, evtchn, be->blkif->blk_protocol, protocol);
+
+	/* Map the shared frame, irq etc. */
+	err = xen_blkif_map(be->blkif, ring_ref, evtchn);
+	if (err) {
+		xenbus_dev_fatal(dev, err, "mapping ring-ref %lu port %u",
+				 ring_ref, evtchn);
+		return err;
+	}
+
+	return 0;
+}
+
+
+/* ** Driver Registration ** */
+
+
+static const struct xenbus_device_id xen_blkbk_ids[] = {
+	{ "vbd" },
+	{ "" }
+};
+
+
+static struct xenbus_driver xen_blkbk = {
+	.name = "vbd",
+	.owner = THIS_MODULE,
+	.ids = xen_blkbk_ids,
+	.probe = xen_blkbk_probe,
+	.remove = xen_blkbk_remove,
+	.otherend_changed = frontend_changed
+};
+
+
+int xen_blkif_xenbus_init(void)
+{
+	return xenbus_register_backend(&xen_blkbk);
+}

drivers/block/xen-blkfront.c

@@ -97,6 +97,7 @@ struct blkfront_info
 	struct blk_shadow shadow[BLK_RING_SIZE];
 	unsigned long shadow_free;
 	unsigned int feature_flush;
+	unsigned int flush_op;
 	int is_ready;
 };
 
@@ -250,8 +251,7 @@ static int blkif_ioctl(struct block_device *bdev, fmode_t mode,
 
 /*
  * Generate a Xen blkfront IO request from a blk layer request.  Reads
- * and writes are handled as expected.  Since we lack a loose flush
- * request, we map flushes into a full ordered barrier.
+ * and writes are handled as expected.
  *
  * @req: a request struct
  */
@@ -293,14 +293,13 @@ static int blkif_queue_request(struct request *req)
 
 	if (req->cmd_flags & (REQ_FLUSH | REQ_FUA)) {
 		/*
-		 * Ideally we could just do an unordered
-		 * flush-to-disk, but all we have is a full write
-		 * barrier at the moment.  However, a barrier write is
+		 * Ideally we can do an unordered flush-to-disk. In case the
+		 * backend onlysupports barriers, use that. A barrier request
 		 * a superset of FUA, so we can implement it the same
 		 * way.  (It's also a FLUSH+FUA, since it is
 		 * guaranteed ordered WRT previous writes.)
 		 */
-		ring_req->operation = BLKIF_OP_WRITE_BARRIER;
+		ring_req->operation = info->flush_op;
 	}
 
 	ring_req->nr_segments = blk_rq_map_sg(req->q, req, info->sg);
@@ -433,8 +432,11 @@ static int xlvbd_init_blk_queue(struct gendisk *gd, u16 sector_size)
 static void xlvbd_flush(struct blkfront_info *info)
 {
 	blk_queue_flush(info->rq, info->feature_flush);
-	printk(KERN_INFO "blkfront: %s: barriers %s\n",
+	printk(KERN_INFO "blkfront: %s: %s: %s\n",
 	       info->gd->disk_name,
+	       info->flush_op == BLKIF_OP_WRITE_BARRIER ?
+		"barrier" : (info->flush_op == BLKIF_OP_FLUSH_DISKCACHE ?
+		"flush diskcache" : "barrier or flush"),
 	       info->feature_flush ? "enabled" : "disabled");
 }
 
@@ -720,15 +722,20 @@ static irqreturn_t blkif_interrupt(int irq, void *dev_id)
 
 		error = (bret->status == BLKIF_RSP_OKAY) ? 0 : -EIO;
 		switch (bret->operation) {
+		case BLKIF_OP_FLUSH_DISKCACHE:
 		case BLKIF_OP_WRITE_BARRIER:
 			if (unlikely(bret->status == BLKIF_RSP_EOPNOTSUPP)) {
-				printk(KERN_WARNING "blkfront: %s: write barrier op failed\n",
+				printk(KERN_WARNING "blkfront: %s: write %s op failed\n",
+				       info->flush_op == BLKIF_OP_WRITE_BARRIER ?
+				       "barrier" :  "flush disk cache",
 				       info->gd->disk_name);
 				error = -EOPNOTSUPP;
 			}
 			if (unlikely(bret->status == BLKIF_RSP_ERROR &&
 				     info->shadow[id].req.nr_segments == 0)) {
-				printk(KERN_WARNING "blkfront: %s: empty write barrier op failed\n",
+				printk(KERN_WARNING "blkfront: %s: empty write %s op failed\n",
+				       info->flush_op == BLKIF_OP_WRITE_BARRIER ?
+				       "barrier" :  "flush disk cache",
 				       info->gd->disk_name);
 				error = -EOPNOTSUPP;
 			}
@@ -736,6 +743,7 @@ static irqreturn_t blkif_interrupt(int irq, void *dev_id)
 				if (error == -EOPNOTSUPP)
 					error = 0;
 				info->feature_flush = 0;
+				info->flush_op = 0;
 				xlvbd_flush(info);
 			}
 			/* fall through */
@@ -1100,7 +1108,7 @@ static void blkfront_connect(struct blkfront_info *info)
 	unsigned long sector_size;
 	unsigned int binfo;
 	int err;
-	int barrier;
+	int barrier, flush;
 
 	switch (info->connected) {
 	case BLKIF_STATE_CONNECTED:
@@ -1140,8 +1148,11 @@ static void blkfront_connect(struct blkfront_info *info)
 		return;
 	}
 
+	info->feature_flush = 0;
+	info->flush_op = 0;
+
 	err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
-			    "feature-barrier", "%lu", &barrier,
+			    "feature-barrier", "%d", &barrier,
 			    NULL);
 
 	/*
@@ -1151,11 +1162,23 @@ static void blkfront_connect(struct blkfront_info *info)
 	 *
 	 * If there are barriers, then we use flush.
 	 */
-	info->feature_flush = 0;
-
-	if (!err && barrier)
+	if (!err && barrier) {
 		info->feature_flush = REQ_FLUSH | REQ_FUA;
+		info->flush_op = BLKIF_OP_WRITE_BARRIER;
+	}
+	/*
+	 * And if there is "feature-flush-cache" use that above
+	 * barriers.
+	 */
+	err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
+			    "feature-flush-cache", "%d", &flush,
+			    NULL);
 
+	if (!err && flush) {
+		info->feature_flush = REQ_FLUSH;
+		info->flush_op = BLKIF_OP_FLUSH_DISKCACHE;
+	}
+		
 	err = xlvbd_alloc_gendisk(sectors, info, binfo, sector_size);
 	if (err) {
 		xenbus_dev_fatal(info->xbdev, err, "xlvbd_add at %s",

include/linux/drbd.h

@@ -38,7 +38,7 @@
 
 /* Although the Linux source code makes a difference between
    generic endianness and the bitfields' endianness, there is no
-   architecture as of Linux-2.6.24-rc4 where the bitfileds' endianness
+   architecture as of Linux-2.6.24-rc4 where the bitfields' endianness
    does not match the generic endianness. */
 
 #if __BYTE_ORDER == __LITTLE_ENDIAN
@@ -53,7 +53,7 @@
 
 
 extern const char *drbd_buildtag(void);
-#define REL_VERSION "8.3.10"
+#define REL_VERSION "8.3.11"
 #define API_VERSION 88
 #define PRO_VERSION_MIN 86
 #define PRO_VERSION_MAX 96
@@ -195,7 +195,7 @@ enum drbd_conns {
 	C_WF_REPORT_PARAMS, /* we have a socket */
 	C_CONNECTED,      /* we have introduced each other */
 	C_STARTING_SYNC_S,  /* starting full sync by admin request. */
-	C_STARTING_SYNC_T,  /* stariing full sync by admin request. */
+	C_STARTING_SYNC_T,  /* starting full sync by admin request. */
 	C_WF_BITMAP_S,
 	C_WF_BITMAP_T,
 	C_WF_SYNC_UUID,
@@ -236,7 +236,7 @@ union drbd_state {
  * pointed out by Maxim Uvarov q<muvarov@ru.mvista.com>
  * even though we transmit as "cpu_to_be32(state)",
  * the offsets of the bitfields still need to be swapped
- * on different endianess.
+ * on different endianness.
  */
 	struct {
 #if defined(__LITTLE_ENDIAN_BITFIELD)
@@ -266,7 +266,7 @@ union drbd_state {
 		unsigned peer:2 ;   /* 3/4	 primary/secondary/unknown */
 		unsigned role:2 ;   /* 3/4	 primary/secondary/unknown */
 #else
-# error "this endianess is not supported"
+# error "this endianness is not supported"
 #endif
 	};
 	unsigned int i;

include/linux/drbd_tag_magic.h

@@ -30,7 +30,7 @@ enum packet_types {
 	int tag_and_len ## member;
 #include "linux/drbd_nl.h"
 
-/* declate tag-list-sizes */
+/* declare tag-list-sizes */
 static const int tag_list_sizes[] = {
 #define NL_PACKET(name, number, fields) 2 fields ,
 #define NL_INTEGER(pn, pr, member)      + 4 + 4

include/linux/lru_cache.h

@@ -139,9 +139,9 @@ write intent log information, three of which are mentioned here.
  * .list is on one of three lists:
  *  in_use: currently in use (refcnt > 0, lc_number != LC_FREE)
  *     lru: unused but ready to be reused or recycled
- *          (ts_refcnt == 0, lc_number != LC_FREE),
+ *          (lc_refcnt == 0, lc_number != LC_FREE),
  *    free: unused but ready to be recycled
- *          (ts_refcnt == 0, lc_number == LC_FREE),
+ *          (lc_refcnt == 0, lc_number == LC_FREE),
  *
  * an element is said to be "in the active set",
  * if either on "in_use" or "lru", i.e. lc_number != LC_FREE.
@@ -160,8 +160,8 @@ struct lc_element {
 	struct hlist_node colision;
 	struct list_head list;		 /* LRU list or free list */
 	unsigned refcnt;
-	/* back "pointer" into ts_cache->element[index],
-	 * for paranoia, and for "ts_element_to_index" */
+	/* back "pointer" into lc_cache->element[index],
+	 * for paranoia, and for "lc_element_to_index" */
 	unsigned lc_index;
 	/* if we want to track a larger set of objects,
 	 * it needs to become arch independend u64 */
@@ -190,8 +190,8 @@ struct lru_cache {
 	/* Arbitrary limit on maximum tracked objects. Practical limit is much
 	 * lower due to allocation failures, probably. For typical use cases,
 	 * nr_elements should be a few thousand at most.
-	 * This also limits the maximum value of ts_element.ts_index, allowing the
-	 * 8 high bits of .ts_index to be overloaded with flags in the future. */
+	 * This also limits the maximum value of lc_element.lc_index, allowing the
+	 * 8 high bits of .lc_index to be overloaded with flags in the future. */
 #define LC_MAX_ACTIVE	(1<<24)
 
 	/* statistics */

include/xen/interface/io/blkif.h

@@ -44,6 +44,19 @@ typedef uint64_t blkif_sector_t;
  */
 #define BLKIF_OP_WRITE_BARRIER     2
 
+/*
+ * Recognised if "feature-flush-cache" is present in backend xenbus
+ * info.  A flush will ask the underlying storage hardware to flush its
+ * non-volatile caches as appropriate.  The "feature-flush-cache" node
+ * contains a boolean indicating whether flush requests are likely to
+ * succeed or fail. Either way, a flush request may fail at any time
+ * with BLKIF_RSP_EOPNOTSUPP if it is unsupported by the underlying
+ * block-device hardware. The boolean simply indicates whether or not it
+ * is worthwhile for the frontend to attempt flushes.  If a backend does
+ * not recognise BLKIF_OP_WRITE_FLUSH_CACHE, it should *not* create the
+ * "feature-flush-cache" node!
+ */
+#define BLKIF_OP_FLUSH_DISKCACHE   3
 /*
  * Maximum scatter/gather segments per request.
  * This is carefully chosen so that sizeof(struct blkif_ring) <= PAGE_SIZE.