Merge branch 'pnfs-submit' of git://git.open-osd.org/linux-open-osd

* 'pnfs-submit' of git://git.open-osd.org/linux-open-osd: (32 commits)
  pnfs-obj: pg_test check for max_io_size
  NFSv4.1: define nfs_generic_pg_test
  NFSv4.1: use pnfs_generic_pg_test directly by layout driver
  NFSv4.1: change pg_test return type to bool
  NFSv4.1: unify pnfs_pageio_init functions
  pnfs-obj: objlayout_encode_layoutcommit implementation
  pnfs: encode_layoutcommit
  pnfs-obj: report errors and .encode_layoutreturn Implementation.
  pnfs: encode_layoutreturn
  pnfs: layoutret_on_setattr
  pnfs: layoutreturn
  pnfs-obj: osd raid engine read/write implementation
  pnfs: support for non-rpc layout drivers
  pnfs-obj: define per-inode private structure
  pnfs: alloc and free layout_hdr layoutdriver methods
  pnfs-obj: objio_osd device information retrieval and caching
  pnfs-obj: decode layout, alloc/free lseg
  pnfs-obj: pnfs_osd XDR client implementation
  pnfs-obj: pnfs_osd XDR definitions
  pnfs-obj: objlayoutdriver module skeleton
  ...

fs/nfs/Kconfig

@@ -87,6 +87,16 @@ config NFS_V4_1
 config PNFS_FILE_LAYOUT
 	tristate
 
+config PNFS_OBJLAYOUT
+	tristate "Provide support for the pNFS Objects Layout Driver for NFSv4.1 pNFS (EXPERIMENTAL)"
+	depends on NFS_FS && NFS_V4_1 && SCSI_OSD_ULD
+	help
+	  Say M here if you want your pNFS client to support the Objects Layout Driver.
+	  Requires the SCSI osd initiator library (SCSI_OSD_INITIATOR) and
+	  upper level driver (SCSI_OSD_ULD).
+
+	  If unsure, say N.
+
 config ROOT_NFS
 	bool "Root file system on NFS"
 	depends on NFS_FS=y && IP_PNP

fs/nfs/Makefile

@@ -15,9 +15,11 @@ nfs-$(CONFIG_NFS_V4)	+= nfs4proc.o nfs4xdr.o nfs4state.o nfs4renewd.o \
 			   delegation.o idmap.o \
 			   callback.o callback_xdr.o callback_proc.o \
 			   nfs4namespace.o
-nfs-$(CONFIG_NFS_V4_1)	+= pnfs.o
+nfs-$(CONFIG_NFS_V4_1)	+= pnfs.o pnfs_dev.o
 nfs-$(CONFIG_SYSCTL) += sysctl.o
 nfs-$(CONFIG_NFS_FSCACHE) += fscache.o fscache-index.o
 
 obj-$(CONFIG_PNFS_FILE_LAYOUT) += nfs_layout_nfsv41_files.o
 nfs_layout_nfsv41_files-y := nfs4filelayout.o nfs4filelayoutdev.o
+
+obj-$(CONFIG_PNFS_OBJLAYOUT) += objlayout/

fs/nfs/callback.h

@@ -167,6 +167,23 @@ extern unsigned nfs4_callback_layoutrecall(
 
 extern void nfs4_check_drain_bc_complete(struct nfs4_session *ses);
 extern void nfs4_cb_take_slot(struct nfs_client *clp);
+
+struct cb_devicenotifyitem {
+	uint32_t		cbd_notify_type;
+	uint32_t		cbd_layout_type;
+	struct nfs4_deviceid	cbd_dev_id;
+	uint32_t		cbd_immediate;
+};
+
+struct cb_devicenotifyargs {
+	int				 ndevs;
+	struct cb_devicenotifyitem	 *devs;
+};
+
+extern __be32 nfs4_callback_devicenotify(
+	struct cb_devicenotifyargs *args,
+	void *dummy, struct cb_process_state *cps);
+
 #endif /* CONFIG_NFS_V4_1 */
 extern int check_gss_callback_principal(struct nfs_client *, struct svc_rqst *);
 extern __be32 nfs4_callback_getattr(struct cb_getattrargs *args,

fs/nfs/callback_proc.c

@@ -139,7 +139,7 @@ static u32 initiate_file_draining(struct nfs_client *clp,
 	spin_lock(&ino->i_lock);
 	if (test_bit(NFS_LAYOUT_BULK_RECALL, &lo->plh_flags) ||
 	    mark_matching_lsegs_invalid(lo, &free_me_list,
-					args->cbl_range.iomode))
+					&args->cbl_range))
 		rv = NFS4ERR_DELAY;
 	else
 		rv = NFS4ERR_NOMATCHING_LAYOUT;
@@ -184,7 +184,7 @@ static u32 initiate_bulk_draining(struct nfs_client *clp,
 		ino = lo->plh_inode;
 		spin_lock(&ino->i_lock);
 		set_bit(NFS_LAYOUT_BULK_RECALL, &lo->plh_flags);
-		if (mark_matching_lsegs_invalid(lo, &free_me_list, range.iomode))
+		if (mark_matching_lsegs_invalid(lo, &free_me_list, &range))
 			rv = NFS4ERR_DELAY;
 		list_del_init(&lo->plh_bulk_recall);
 		spin_unlock(&ino->i_lock);
@@ -241,6 +241,53 @@ static void pnfs_recall_all_layouts(struct nfs_client *clp)
 	do_callback_layoutrecall(clp, &args);
 }
 
+__be32 nfs4_callback_devicenotify(struct cb_devicenotifyargs *args,
+				  void *dummy, struct cb_process_state *cps)
+{
+	int i;
+	__be32 res = 0;
+	struct nfs_client *clp = cps->clp;
+	struct nfs_server *server = NULL;
+
+	dprintk("%s: -->\n", __func__);
+
+	if (!clp) {
+		res = cpu_to_be32(NFS4ERR_OP_NOT_IN_SESSION);
+		goto out;
+	}
+
+	for (i = 0; i < args->ndevs; i++) {
+		struct cb_devicenotifyitem *dev = &args->devs[i];
+
+		if (!server ||
+		    server->pnfs_curr_ld->id != dev->cbd_layout_type) {
+			rcu_read_lock();
+			list_for_each_entry_rcu(server, &clp->cl_superblocks, client_link)
+				if (server->pnfs_curr_ld &&
+				    server->pnfs_curr_ld->id == dev->cbd_layout_type) {
+					rcu_read_unlock();
+					goto found;
+				}
+			rcu_read_unlock();
+			dprintk("%s: layout type %u not found\n",
+				__func__, dev->cbd_layout_type);
+			continue;
+		}
+
+	found:
+		if (dev->cbd_notify_type == NOTIFY_DEVICEID4_CHANGE)
+			dprintk("%s: NOTIFY_DEVICEID4_CHANGE not supported, "
+				"deleting instead\n", __func__);
+		nfs4_delete_deviceid(server->pnfs_curr_ld, clp, &dev->cbd_dev_id);
+	}
+
+out:
+	kfree(args->devs);
+	dprintk("%s: exit with status = %u\n",
+		__func__, be32_to_cpu(res));
+	return res;
+}
+
 int nfs41_validate_delegation_stateid(struct nfs_delegation *delegation, const nfs4_stateid *stateid)
 {
 	if (delegation == NULL)

fs/nfs/callback_xdr.c

@@ -25,6 +25,7 @@
 
 #if defined(CONFIG_NFS_V4_1)
 #define CB_OP_LAYOUTRECALL_RES_MAXSZ	(CB_OP_HDR_RES_MAXSZ)
+#define CB_OP_DEVICENOTIFY_RES_MAXSZ	(CB_OP_HDR_RES_MAXSZ)
 #define CB_OP_SEQUENCE_RES_MAXSZ	(CB_OP_HDR_RES_MAXSZ + \
 					4 + 1 + 3)
 #define CB_OP_RECALLANY_RES_MAXSZ	(CB_OP_HDR_RES_MAXSZ)
@@ -284,6 +285,93 @@ out:
 	return status;
 }
 
+static
+__be32 decode_devicenotify_args(struct svc_rqst *rqstp,
+				struct xdr_stream *xdr,
+				struct cb_devicenotifyargs *args)
+{
+	__be32 *p;
+	__be32 status = 0;
+	u32 tmp;
+	int n, i;
+	args->ndevs = 0;
+
+	/* Num of device notifications */
+	p = read_buf(xdr, sizeof(uint32_t));
+	if (unlikely(p == NULL)) {
+		status = htonl(NFS4ERR_BADXDR);
+		goto out;
+	}
+	n = ntohl(*p++);
+	if (n <= 0)
+		goto out;
+
+	args->devs = kmalloc(n * sizeof(*args->devs), GFP_KERNEL);
+	if (!args->devs) {
+		status = htonl(NFS4ERR_DELAY);
+		goto out;
+	}
+
+	/* Decode each dev notification */
+	for (i = 0; i < n; i++) {
+		struct cb_devicenotifyitem *dev = &args->devs[i];
+
+		p = read_buf(xdr, (4 * sizeof(uint32_t)) + NFS4_DEVICEID4_SIZE);
+		if (unlikely(p == NULL)) {
+			status = htonl(NFS4ERR_BADXDR);
+			goto err;
+		}
+
+		tmp = ntohl(*p++);	/* bitmap size */
+		if (tmp != 1) {
+			status = htonl(NFS4ERR_INVAL);
+			goto err;
+		}
+		dev->cbd_notify_type = ntohl(*p++);
+		if (dev->cbd_notify_type != NOTIFY_DEVICEID4_CHANGE &&
+		    dev->cbd_notify_type != NOTIFY_DEVICEID4_DELETE) {
+			status = htonl(NFS4ERR_INVAL);
+			goto err;
+		}
+
+		tmp = ntohl(*p++);	/* opaque size */
+		if (((dev->cbd_notify_type == NOTIFY_DEVICEID4_CHANGE) &&
+		     (tmp != NFS4_DEVICEID4_SIZE + 8)) ||
+		    ((dev->cbd_notify_type == NOTIFY_DEVICEID4_DELETE) &&
+		     (tmp != NFS4_DEVICEID4_SIZE + 4))) {
+			status = htonl(NFS4ERR_INVAL);
+			goto err;
+		}
+		dev->cbd_layout_type = ntohl(*p++);
+		memcpy(dev->cbd_dev_id.data, p, NFS4_DEVICEID4_SIZE);
+		p += XDR_QUADLEN(NFS4_DEVICEID4_SIZE);
+
+		if (dev->cbd_layout_type == NOTIFY_DEVICEID4_CHANGE) {
+			p = read_buf(xdr, sizeof(uint32_t));
+			if (unlikely(p == NULL)) {
+				status = htonl(NFS4ERR_BADXDR);
+				goto err;
+			}
+			dev->cbd_immediate = ntohl(*p++);
+		} else {
+			dev->cbd_immediate = 0;
+		}
+
+		args->ndevs++;
+
+		dprintk("%s: type %d layout 0x%x immediate %d\n",
+			__func__, dev->cbd_notify_type, dev->cbd_layout_type,
+			dev->cbd_immediate);
+	}
+out:
+	dprintk("%s: status %d ndevs %d\n",
+		__func__, ntohl(status), args->ndevs);
+	return status;
+err:
+	kfree(args->devs);
+	goto out;
+}
+
 static __be32 decode_sessionid(struct xdr_stream *xdr,
 				 struct nfs4_sessionid *sid)
 {
@@ -639,10 +727,10 @@ preprocess_nfs41_op(int nop, unsigned int op_nr, struct callback_op **op)
 	case OP_CB_RECALL_ANY:
 	case OP_CB_RECALL_SLOT:
 	case OP_CB_LAYOUTRECALL:
+	case OP_CB_NOTIFY_DEVICEID:
 		*op = &callback_ops[op_nr];
 		break;
 
-	case OP_CB_NOTIFY_DEVICEID:
 	case OP_CB_NOTIFY:
 	case OP_CB_PUSH_DELEG:
 	case OP_CB_RECALLABLE_OBJ_AVAIL:
@@ -849,6 +937,12 @@ static struct callback_op callback_ops[] = {
 			(callback_decode_arg_t)decode_layoutrecall_args,
 		.res_maxsize = CB_OP_LAYOUTRECALL_RES_MAXSZ,
 	},
+	[OP_CB_NOTIFY_DEVICEID] = {
+		.process_op = (callback_process_op_t)nfs4_callback_devicenotify,
+		.decode_args =
+			(callback_decode_arg_t)decode_devicenotify_args,
+		.res_maxsize = CB_OP_DEVICENOTIFY_RES_MAXSZ,
+	},
 	[OP_CB_SEQUENCE] = {
 		.process_op = (callback_process_op_t)nfs4_callback_sequence,
 		.decode_args = (callback_decode_arg_t)decode_cb_sequence_args,

fs/nfs/client.c

@@ -290,6 +290,8 @@ static void nfs_free_client(struct nfs_client *clp)
 	if (clp->cl_machine_cred != NULL)
 		put_rpccred(clp->cl_machine_cred);
 
+	nfs4_deviceid_purge_client(clp);
+
 	kfree(clp->cl_hostname);
 	kfree(clp);
 

fs/nfs/dir.c

@@ -512,12 +512,7 @@ int nfs_readdir_page_filler(nfs_readdir_descriptor_t *desc, struct nfs_entry *en
 				struct page **xdr_pages, struct page *page, unsigned int buflen)
 {
 	struct xdr_stream stream;
-	struct xdr_buf buf = {
+	struct xdr_buf buf;
-		.pages = xdr_pages,
-		.page_len = buflen,
-		.buflen = buflen,
-		.len = buflen,
-	};
 	struct page *scratch;
 	struct nfs_cache_array *array;
 	unsigned int count = 0;
@@ -527,7 +522,7 @@ int nfs_readdir_page_filler(nfs_readdir_descriptor_t *desc, struct nfs_entry *en
 	if (scratch == NULL)
 		return -ENOMEM;
 
-	xdr_init_decode(&stream, &buf, NULL);
+	xdr_init_decode_pages(&stream, &buf, xdr_pages, buflen);
 	xdr_set_scratch_buffer(&stream, page_address(scratch), PAGE_SIZE);
 
 	do {

fs/nfs/inode.c

@@ -1428,9 +1428,10 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
  */
 void nfs4_evict_inode(struct inode *inode)
 {
-	pnfs_destroy_layout(NFS_I(inode));
 	truncate_inode_pages(&inode->i_data, 0);
 	end_writeback(inode);
+	pnfs_return_layout(inode);
+	pnfs_destroy_layout(NFS_I(inode));
 	/* If we are holding a delegation, return it! */
 	nfs_inode_return_delegation_noreclaim(inode);
 	/* First call standard NFS clear_inode() code */

fs/nfs/internal.h

@@ -310,6 +310,7 @@ extern int nfs_migrate_page(struct address_space *,
 #endif
 
 /* nfs4proc.c */
+extern void __nfs4_read_done_cb(struct nfs_read_data *);
 extern void nfs4_reset_read(struct rpc_task *task, struct nfs_read_data *data);
 extern int nfs4_init_client(struct nfs_client *clp,
 			    const struct rpc_timeout *timeparms,

fs/nfs/nfs4filelayout.c

@@ -421,6 +421,7 @@ filelayout_check_layout(struct pnfs_layout_hdr *lo,
 			struct nfs4_deviceid *id,
 			gfp_t gfp_flags)
 {
+	struct nfs4_deviceid_node *d;
 	struct nfs4_file_layout_dsaddr *dsaddr;
 	int status = -EINVAL;
 	struct nfs_server *nfss = NFS_SERVER(lo->plh_inode);
@@ -428,7 +429,7 @@ filelayout_check_layout(struct pnfs_layout_hdr *lo,
 	dprintk("--> %s\n", __func__);
 
 	if (fl->pattern_offset > lgr->range.offset) {
-		dprintk("%s pattern_offset %lld to large\n",
+		dprintk("%s pattern_offset %lld too large\n",
 				__func__, fl->pattern_offset);
 		goto out;
 	}
@@ -440,12 +441,14 @@ filelayout_check_layout(struct pnfs_layout_hdr *lo,
 	}
 
 	/* find and reference the deviceid */
-	dsaddr = nfs4_fl_find_get_deviceid(id);
+	d = nfs4_find_get_deviceid(NFS_SERVER(lo->plh_inode)->pnfs_curr_ld,
-	if (dsaddr == NULL) {
+				   NFS_SERVER(lo->plh_inode)->nfs_client, id);
+	if (d == NULL) {
 		dsaddr = get_device_info(lo->plh_inode, id, gfp_flags);
 		if (dsaddr == NULL)
 			goto out;
-	}
+	} else
+		dsaddr = container_of(d, struct nfs4_file_layout_dsaddr, id_node);
 	fl->dsaddr = dsaddr;
 
 	if (fl->first_stripe_index < 0 ||
@@ -507,12 +510,7 @@ filelayout_decode_layout(struct pnfs_layout_hdr *flo,
 			 gfp_t gfp_flags)
 {
 	struct xdr_stream stream;
-	struct xdr_buf buf = {
+	struct xdr_buf buf;
-		.pages =  lgr->layoutp->pages,
-		.page_len =  lgr->layoutp->len,
-		.buflen =  lgr->layoutp->len,
-		.len = lgr->layoutp->len,
-	};
 	struct page *scratch;
 	__be32 *p;
 	uint32_t nfl_util;
@@ -524,7 +522,7 @@ filelayout_decode_layout(struct pnfs_layout_hdr *flo,
 	if (!scratch)
 		return -ENOMEM;
 
-	xdr_init_decode(&stream, &buf, NULL);
+	xdr_init_decode_pages(&stream, &buf, lgr->layoutp->pages, lgr->layoutp->len);
 	xdr_set_scratch_buffer(&stream, page_address(scratch), PAGE_SIZE);
 
 	/* 20 = ufl_util (4), first_stripe_index (4), pattern_offset (8),
@@ -535,7 +533,7 @@ filelayout_decode_layout(struct pnfs_layout_hdr *flo,
 
 	memcpy(id, p, sizeof(*id));
 	p += XDR_QUADLEN(NFS4_DEVICEID4_SIZE);
-	print_deviceid(id);
+	nfs4_print_deviceid(id);
 
 	nfl_util = be32_to_cpup(p++);
 	if (nfl_util & NFL4_UFLG_COMMIT_THRU_MDS)
@@ -653,16 +651,19 @@ filelayout_alloc_lseg(struct pnfs_layout_hdr *layoutid,
 /*
  * filelayout_pg_test(). Called by nfs_can_coalesce_requests()
  *
- * return 1 :  coalesce page
+ * return true  : coalesce page
- * return 0 :  don't coalesce page
+ * return false : don't coalesce page
  */
-int
+bool
 filelayout_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev,
 		   struct nfs_page *req)
 {
 	u64 p_stripe, r_stripe;
 	u32 stripe_unit;
 
+	if (!pnfs_generic_pg_test(pgio, prev, req))
+		return 0;
+
 	if (!pgio->pg_lseg)
 		return 1;
 	p_stripe = (u64)prev->wb_index << PAGE_CACHE_SHIFT;
@@ -860,6 +861,12 @@ filelayout_commit_pagelist(struct inode *inode, struct list_head *mds_pages,
 	return -ENOMEM;
 }
 
+static void
+filelayout_free_deveiceid_node(struct nfs4_deviceid_node *d)
+{
+	nfs4_fl_free_deviceid(container_of(d, struct nfs4_file_layout_dsaddr, id_node));
+}
+
 static struct pnfs_layoutdriver_type filelayout_type = {
 	.id			= LAYOUT_NFSV4_1_FILES,
 	.name			= "LAYOUT_NFSV4_1_FILES",
@@ -872,6 +879,7 @@ static struct pnfs_layoutdriver_type filelayout_type = {
 	.commit_pagelist	= filelayout_commit_pagelist,
 	.read_pagelist		= filelayout_read_pagelist,
 	.write_pagelist		= filelayout_write_pagelist,
+	.free_deviceid_node	= filelayout_free_deveiceid_node,
 };
 
 static int __init nfs4filelayout_init(void)

fs/nfs/nfs4filelayout.h

@@ -59,9 +59,7 @@ struct nfs4_pnfs_ds {
 #define NFS4_DEVICE_ID_NEG_ENTRY	0x00000001
 
 struct nfs4_file_layout_dsaddr {
-	struct hlist_node		node;
+	struct nfs4_deviceid_node	id_node;
-	struct nfs4_deviceid		deviceid;
-	atomic_t			ref;
 	unsigned long			flags;
 	u32				stripe_count;
 	u8				*stripe_indices;
@@ -95,14 +93,12 @@ extern struct nfs_fh *
 nfs4_fl_select_ds_fh(struct pnfs_layout_segment *lseg, u32 j);
 
 extern void print_ds(struct nfs4_pnfs_ds *ds);
-extern void print_deviceid(struct nfs4_deviceid *dev_id);
 u32 nfs4_fl_calc_j_index(struct pnfs_layout_segment *lseg, loff_t offset);
 u32 nfs4_fl_calc_ds_index(struct pnfs_layout_segment *lseg, u32 j);
 struct nfs4_pnfs_ds *nfs4_fl_prepare_ds(struct pnfs_layout_segment *lseg,
 					u32 ds_idx);
-extern struct nfs4_file_layout_dsaddr *
-nfs4_fl_find_get_deviceid(struct nfs4_deviceid *dev_id);
 extern void nfs4_fl_put_deviceid(struct nfs4_file_layout_dsaddr *dsaddr);
+extern void nfs4_fl_free_deviceid(struct nfs4_file_layout_dsaddr *dsaddr);
 struct nfs4_file_layout_dsaddr *
 get_device_info(struct inode *inode, struct nfs4_deviceid *dev_id, gfp_t gfp_flags);
 

fs/nfs/nfs4filelayoutdev.c

@@ -36,30 +36,6 @@
 
 #define NFSDBG_FACILITY		NFSDBG_PNFS_LD
 
-/*
- * Device ID RCU cache. A device ID is unique per client ID and layout type.
- */
-#define NFS4_FL_DEVICE_ID_HASH_BITS	5
-#define NFS4_FL_DEVICE_ID_HASH_SIZE	(1 << NFS4_FL_DEVICE_ID_HASH_BITS)
-#define NFS4_FL_DEVICE_ID_HASH_MASK	(NFS4_FL_DEVICE_ID_HASH_SIZE - 1)
-
-static inline u32
-nfs4_fl_deviceid_hash(struct nfs4_deviceid *id)
-{
-	unsigned char *cptr = (unsigned char *)id->data;
-	unsigned int nbytes = NFS4_DEVICEID4_SIZE;
-	u32 x = 0;
-
-	while (nbytes--) {
-		x *= 37;
-		x += *cptr++;
-	}
-	return x & NFS4_FL_DEVICE_ID_HASH_MASK;
-}
-
-static struct hlist_head filelayout_deviceid_cache[NFS4_FL_DEVICE_ID_HASH_SIZE];
-static DEFINE_SPINLOCK(filelayout_deviceid_lock);
-
 /*
  * Data server cache
  *
@@ -89,27 +65,6 @@ print_ds(struct nfs4_pnfs_ds *ds)
 		ds->ds_clp ? ds->ds_clp->cl_exchange_flags : 0);
 }
 
-void
-print_ds_list(struct nfs4_file_layout_dsaddr *dsaddr)
-{
-	int i;
-
-	ifdebug(FACILITY) {
-		printk("%s dsaddr->ds_num %d\n", __func__,
-		       dsaddr->ds_num);
-		for (i = 0; i < dsaddr->ds_num; i++)
-			print_ds(dsaddr->ds_list[i]);
-	}
-}
-
-void print_deviceid(struct nfs4_deviceid *id)
-{
-	u32 *p = (u32 *)id;
-
-	dprintk("%s: device id= [%x%x%x%x]\n", __func__,
-		p[0], p[1], p[2], p[3]);
-}
-
 /* nfs4_ds_cache_lock is held */
 static struct nfs4_pnfs_ds *
 _data_server_lookup_locked(u32 ip_addr, u32 port)
@@ -201,13 +156,13 @@ destroy_ds(struct nfs4_pnfs_ds *ds)
 	kfree(ds);
 }
 
-static void
+void
 nfs4_fl_free_deviceid(struct nfs4_file_layout_dsaddr *dsaddr)
 {
 	struct nfs4_pnfs_ds *ds;
 	int i;
 
-	print_deviceid(&dsaddr->deviceid);
+	nfs4_print_deviceid(&dsaddr->id_node.deviceid);
 
 	for (i = 0; i < dsaddr->ds_num; i++) {
 		ds = dsaddr->ds_list[i];
@@ -353,12 +308,7 @@ decode_device(struct inode *ino, struct pnfs_device *pdev, gfp_t gfp_flags)
 	u8 max_stripe_index;
 	struct nfs4_file_layout_dsaddr *dsaddr = NULL;
 	struct xdr_stream stream;
-	struct xdr_buf buf = {
+	struct xdr_buf buf;
-		.pages = pdev->pages,
-		.page_len = pdev->pglen,
-		.buflen = pdev->pglen,
-		.len = pdev->pglen,
-	};
 	struct page *scratch;
 
 	/* set up xdr stream */
@@ -366,7 +316,7 @@ decode_device(struct inode *ino, struct pnfs_device *pdev, gfp_t gfp_flags)
 	if (!scratch)
 		goto out_err;
 
-	xdr_init_decode(&stream, &buf, NULL);
+	xdr_init_decode_pages(&stream, &buf, pdev->pages, pdev->pglen);
 	xdr_set_scratch_buffer(&stream, page_address(scratch), PAGE_SIZE);
 
 	/* Get the stripe count (number of stripe index) */
@@ -431,8 +381,10 @@ decode_device(struct inode *ino, struct pnfs_device *pdev, gfp_t gfp_flags)
 	dsaddr->stripe_indices = stripe_indices;
 	stripe_indices = NULL;
 	dsaddr->ds_num = num;
-
+	nfs4_init_deviceid_node(&dsaddr->id_node,
-	memcpy(&dsaddr->deviceid, &pdev->dev_id, sizeof(pdev->dev_id));
+				NFS_SERVER(ino)->pnfs_curr_ld,
+				NFS_SERVER(ino)->nfs_client,
+				&pdev->dev_id);
 
 	for (i = 0; i < dsaddr->ds_num; i++) {
 		int j;
@@ -505,8 +457,8 @@ out_err:
 static struct nfs4_file_layout_dsaddr *
 decode_and_add_device(struct inode *inode, struct pnfs_device *dev, gfp_t gfp_flags)
 {
-	struct nfs4_file_layout_dsaddr *d, *new;
+	struct nfs4_deviceid_node *d;
-	long hash;
+	struct nfs4_file_layout_dsaddr *n, *new;
 
 	new = decode_device(inode, dev, gfp_flags);
 	if (!new) {
@@ -515,20 +467,13 @@ decode_and_add_device(struct inode *inode, struct pnfs_device *dev, gfp_t gfp_fl
 		return NULL;
 	}
 
-	spin_lock(&filelayout_deviceid_lock);
+	d = nfs4_insert_deviceid_node(&new->id_node);
-	d = nfs4_fl_find_get_deviceid(&new->deviceid);
+	n = container_of(d, struct nfs4_file_layout_dsaddr, id_node);
-	if (d) {
+	if (n != new) {
-		spin_unlock(&filelayout_deviceid_lock);
 		nfs4_fl_free_deviceid(new);
-		return d;
+		return n;
 	}
 
-	INIT_HLIST_NODE(&new->node);
-	atomic_set(&new->ref, 1);
-	hash = nfs4_fl_deviceid_hash(&new->deviceid);
-	hlist_add_head_rcu(&new->node, &filelayout_deviceid_cache[hash]);
-	spin_unlock(&filelayout_deviceid_lock);
-
 	return new;
 }
 
@@ -600,35 +545,7 @@ out_free:
 void
 nfs4_fl_put_deviceid(struct nfs4_file_layout_dsaddr *dsaddr)
 {
-	if (atomic_dec_and_lock(&dsaddr->ref, &filelayout_deviceid_lock)) {
+	nfs4_put_deviceid_node(&dsaddr->id_node);
-		hlist_del_rcu(&dsaddr->node);
-		spin_unlock(&filelayout_deviceid_lock);
-
-		synchronize_rcu();
-		nfs4_fl_free_deviceid(dsaddr);
-	}
-}
-
-struct nfs4_file_layout_dsaddr *
-nfs4_fl_find_get_deviceid(struct nfs4_deviceid *id)
-{
-	struct nfs4_file_layout_dsaddr *d;
-	struct hlist_node *n;
-	long hash = nfs4_fl_deviceid_hash(id);
-
-
-	rcu_read_lock();
-	hlist_for_each_entry_rcu(d, n, &filelayout_deviceid_cache[hash], node) {
-		if (!memcmp(&d->deviceid, id, sizeof(*id))) {
-			if (!atomic_inc_not_zero(&d->ref))
-				goto fail;
-			rcu_read_unlock();
-			return d;
-		}
-	}
-fail:
-	rcu_read_unlock();
-	return NULL;
 }
 
 /*
@@ -676,15 +593,15 @@ static void
 filelayout_mark_devid_negative(struct nfs4_file_layout_dsaddr *dsaddr,
 			       int err, u32 ds_addr)
 {
-	u32 *p = (u32 *)&dsaddr->deviceid;
+	u32 *p = (u32 *)&dsaddr->id_node.deviceid;
 
 	printk(KERN_ERR "NFS: data server %x connection error %d."
 		" Deviceid [%x%x%x%x] marked out of use.\n",
 		ds_addr, err, p[0], p[1], p[2], p[3]);
 
-	spin_lock(&filelayout_deviceid_lock);
+	spin_lock(&nfs4_ds_cache_lock);
 	dsaddr->flags |= NFS4_DEVICE_ID_NEG_ENTRY;
-	spin_unlock(&filelayout_deviceid_lock);
+	spin_unlock(&nfs4_ds_cache_lock);
 }
 
 struct nfs4_pnfs_ds *

fs/nfs/nfs4proc.c

@@ -2363,6 +2363,9 @@ nfs4_proc_setattr(struct dentry *dentry, struct nfs_fattr *fattr,
 	struct nfs4_state *state = NULL;
 	int status;
 
+	if (pnfs_ld_layoutret_on_setattr(inode))
+		pnfs_return_layout(inode);
+
 	nfs_fattr_init(fattr);
 	
 	/* Search for an existing open(O_WRITE) file */
@@ -3177,6 +3180,11 @@ static int nfs4_proc_pathconf(struct nfs_server *server, struct nfs_fh *fhandle,
 	return err;
 }
 
+void __nfs4_read_done_cb(struct nfs_read_data *data)
+{
+	nfs_invalidate_atime(data->inode);
+}
+
 static int nfs4_read_done_cb(struct rpc_task *task, struct nfs_read_data *data)
 {
 	struct nfs_server *server = NFS_SERVER(data->inode);
@@ -3186,7 +3194,7 @@ static int nfs4_read_done_cb(struct rpc_task *task, struct nfs_read_data *data)
 		return -EAGAIN;
 	}
 
-	nfs_invalidate_atime(data->inode);
+	__nfs4_read_done_cb(data);
 	if (task->tk_status > 0)
 		renew_lease(server, data->timestamp);
 	return 0;
@@ -3200,7 +3208,8 @@ static int nfs4_read_done(struct rpc_task *task, struct nfs_read_data *data)
 	if (!nfs4_sequence_done(task, &data->res.seq_res))
 		return -EAGAIN;
 
-	return data->read_done_cb(task, data);
+	return data->read_done_cb ? data->read_done_cb(task, data) :
+				    nfs4_read_done_cb(task, data);
 }
 
 static void nfs4_proc_read_setup(struct nfs_read_data *data, struct rpc_message *msg)
@@ -3245,7 +3254,8 @@ static int nfs4_write_done(struct rpc_task *task, struct nfs_write_data *data)
 {
 	if (!nfs4_sequence_done(task, &data->res.seq_res))
 		return -EAGAIN;
-	return data->write_done_cb(task, data);
+	return data->write_done_cb ? data->write_done_cb(task, data) :
+		nfs4_write_done_cb(task, data);
 }
 
 /* Reset the the nfs_write_data to send the write to the MDS. */
@@ -5671,6 +5681,88 @@ int nfs4_proc_layoutget(struct nfs4_layoutget *lgp)
 	return status;
 }
 
+static void
+nfs4_layoutreturn_prepare(struct rpc_task *task, void *calldata)
+{
+	struct nfs4_layoutreturn *lrp = calldata;
+
+	dprintk("--> %s\n", __func__);
+	if (nfs41_setup_sequence(lrp->clp->cl_session, &lrp->args.seq_args,
+				&lrp->res.seq_res, 0, task))
+		return;
+	rpc_call_start(task);
+}
+
+static void nfs4_layoutreturn_done(struct rpc_task *task, void *calldata)
+{
+	struct nfs4_layoutreturn *lrp = calldata;
+	struct nfs_server *server;
+
+	dprintk("--> %s\n", __func__);
+
+	if (!nfs4_sequence_done(task, &lrp->res.seq_res))
+		return;
+
+	server = NFS_SERVER(lrp->args.inode);
+	if (nfs4_async_handle_error(task, server, NULL) == -EAGAIN) {
+		nfs_restart_rpc(task, lrp->clp);
+		return;
+	}
+	if (task->tk_status == 0) {
+		struct pnfs_layout_hdr *lo = NFS_I(lrp->args.inode)->layout;
+
+		if (lrp->res.lrs_present) {
+			spin_lock(&lo->plh_inode->i_lock);
+			pnfs_set_layout_stateid(lo, &lrp->res.stateid, true);
+			spin_unlock(&lo->plh_inode->i_lock);
+		} else
+			BUG_ON(!list_empty(&lo->plh_segs));
+	}
+	dprintk("<-- %s\n", __func__);
+}
+
+static void nfs4_layoutreturn_release(void *calldata)
+{
+	struct nfs4_layoutreturn *lrp = calldata;
+
+	dprintk("--> %s\n", __func__);
+	put_layout_hdr(NFS_I(lrp->args.inode)->layout);
+	kfree(calldata);
+	dprintk("<-- %s\n", __func__);
+}
+
+static const struct rpc_call_ops nfs4_layoutreturn_call_ops = {
+	.rpc_call_prepare = nfs4_layoutreturn_prepare,
+	.rpc_call_done = nfs4_layoutreturn_done,
+	.rpc_release = nfs4_layoutreturn_release,
+};
+
+int nfs4_proc_layoutreturn(struct nfs4_layoutreturn *lrp)
+{
+	struct rpc_task *task;
+	struct rpc_message msg = {
+		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LAYOUTRETURN],
+		.rpc_argp = &lrp->args,
+		.rpc_resp = &lrp->res,
+	};
+	struct rpc_task_setup task_setup_data = {
+		.rpc_client = lrp->clp->cl_rpcclient,
+		.rpc_message = &msg,
+		.callback_ops = &nfs4_layoutreturn_call_ops,
+		.callback_data = lrp,
+	};
+	int status;
+
+	dprintk("--> %s\n", __func__);
+	task = rpc_run_task(&task_setup_data);
+	if (IS_ERR(task))
+		return PTR_ERR(task);
+	status = task->tk_status;
+	dprintk("<-- %s status=%d\n", __func__, status);
+	rpc_put_task(task);
+	return status;
+}
+
 static int
 _nfs4_proc_getdeviceinfo(struct nfs_server *server, struct pnfs_device *pdev)
 {

fs/nfs/nfs4xdr.c

@@ -338,7 +338,11 @@ static int nfs4_stat_to_errno(int);
 				1 /* layoutupdate4 layout type */ + \
 				1 /* NULL filelayout layoutupdate4 payload */)
 #define decode_layoutcommit_maxsz (op_decode_hdr_maxsz + 3)
-
+#define encode_layoutreturn_maxsz (8 + op_encode_hdr_maxsz + \
+				encode_stateid_maxsz + \
+				1 /* FIXME: opaque lrf_body always empty at the moment */)
+#define decode_layoutreturn_maxsz (op_decode_hdr_maxsz + \
+				1 + decode_stateid_maxsz)
 #else /* CONFIG_NFS_V4_1 */
 #define encode_sequence_maxsz	0
 #define decode_sequence_maxsz	0
@@ -760,7 +764,14 @@ static int nfs4_stat_to_errno(int);
 				decode_putfh_maxsz + \
 				decode_layoutcommit_maxsz + \
 				decode_getattr_maxsz)
-
+#define NFS4_enc_layoutreturn_sz (compound_encode_hdr_maxsz + \
+				encode_sequence_maxsz + \
+				encode_putfh_maxsz + \
+				encode_layoutreturn_maxsz)
+#define NFS4_dec_layoutreturn_sz (compound_decode_hdr_maxsz + \
+				decode_sequence_maxsz + \
+				decode_putfh_maxsz + \
+				decode_layoutreturn_maxsz)
 
 const u32 nfs41_maxwrite_overhead = ((RPC_MAX_HEADER_WITH_AUTH +
 				      compound_encode_hdr_maxsz +
@@ -1864,6 +1875,7 @@ encode_layoutget(struct xdr_stream *xdr,
 
 static int
 encode_layoutcommit(struct xdr_stream *xdr,
+		    struct inode *inode,
 		    const struct nfs4_layoutcommit_args *args,
 		    struct compound_hdr *hdr)
 {
@@ -1872,7 +1884,7 @@ encode_layoutcommit(struct xdr_stream *xdr,
 	dprintk("%s: lbw: %llu type: %d\n", __func__, args->lastbytewritten,
 		NFS_SERVER(args->inode)->pnfs_curr_ld->id);
 
-	p = reserve_space(xdr, 48 + NFS4_STATEID_SIZE);
+	p = reserve_space(xdr, 44 + NFS4_STATEID_SIZE);
 	*p++ = cpu_to_be32(OP_LAYOUTCOMMIT);
 	/* Only whole file layouts */
 	p = xdr_encode_hyper(p, 0); /* offset */
@@ -1883,12 +1895,49 @@ encode_layoutcommit(struct xdr_stream *xdr,
 	p = xdr_encode_hyper(p, args->lastbytewritten);
 	*p++ = cpu_to_be32(0); /* Never send time_modify_changed */
 	*p++ = cpu_to_be32(NFS_SERVER(args->inode)->pnfs_curr_ld->id);/* type */
-	*p++ = cpu_to_be32(0); /* no file layout payload */
+
+	if (NFS_SERVER(inode)->pnfs_curr_ld->encode_layoutcommit)
+		NFS_SERVER(inode)->pnfs_curr_ld->encode_layoutcommit(
+			NFS_I(inode)->layout, xdr, args);
+	else {
+		p = reserve_space(xdr, 4);
+		*p = cpu_to_be32(0); /* no layout-type payload */
+	}
 
 	hdr->nops++;
 	hdr->replen += decode_layoutcommit_maxsz;
 	return 0;
 }
+
+static void
+encode_layoutreturn(struct xdr_stream *xdr,
+		    const struct nfs4_layoutreturn_args *args,
+		    struct compound_hdr *hdr)
+{
+	__be32 *p;
+
+	p = reserve_space(xdr, 20);
+	*p++ = cpu_to_be32(OP_LAYOUTRETURN);
+	*p++ = cpu_to_be32(0);		/* reclaim. always 0 for now */
+	*p++ = cpu_to_be32(args->layout_type);
+	*p++ = cpu_to_be32(IOMODE_ANY);
+	*p = cpu_to_be32(RETURN_FILE);
+	p = reserve_space(xdr, 16 + NFS4_STATEID_SIZE);
+	p = xdr_encode_hyper(p, 0);
+	p = xdr_encode_hyper(p, NFS4_MAX_UINT64);
+	spin_lock(&args->inode->i_lock);
+	xdr_encode_opaque_fixed(p, &args->stateid.data, NFS4_STATEID_SIZE);
+	spin_unlock(&args->inode->i_lock);
+	if (NFS_SERVER(args->inode)->pnfs_curr_ld->encode_layoutreturn) {
+		NFS_SERVER(args->inode)->pnfs_curr_ld->encode_layoutreturn(
+			NFS_I(args->inode)->layout, xdr, args);
+	} else {
+		p = reserve_space(xdr, 4);
+		*p = cpu_to_be32(0);
+	}
+	hdr->nops++;
+	hdr->replen += decode_layoutreturn_maxsz;
+}
 #endif /* CONFIG_NFS_V4_1 */
 
 /*
@@ -2706,9 +2755,30 @@ static void nfs4_xdr_enc_layoutget(struct rpc_rqst *req,
 /*
  *  Encode LAYOUTCOMMIT request
  */
-static int nfs4_xdr_enc_layoutcommit(struct rpc_rqst *req,
+static void nfs4_xdr_enc_layoutcommit(struct rpc_rqst *req,
-				     struct xdr_stream *xdr,
+				      struct xdr_stream *xdr,
-				     struct nfs4_layoutcommit_args *args)
+				      struct nfs4_layoutcommit_args *args)
+{
+	struct nfs4_layoutcommit_data *data =
+		container_of(args, struct nfs4_layoutcommit_data, args);
+	struct compound_hdr hdr = {
+		.minorversion = nfs4_xdr_minorversion(&args->seq_args),
+	};
+
+	encode_compound_hdr(xdr, req, &hdr);
+	encode_sequence(xdr, &args->seq_args, &hdr);
+	encode_putfh(xdr, NFS_FH(args->inode), &hdr);
+	encode_layoutcommit(xdr, data->args.inode, args, &hdr);
+	encode_getfattr(xdr, args->bitmask, &hdr);
+	encode_nops(&hdr);
+}
+
+/*
+ * Encode LAYOUTRETURN request
+ */
+static void nfs4_xdr_enc_layoutreturn(struct rpc_rqst *req,
+				      struct xdr_stream *xdr,
+				      struct nfs4_layoutreturn_args *args)
 {
 	struct compound_hdr hdr = {
 		.minorversion = nfs4_xdr_minorversion(&args->seq_args),
@@ -2717,10 +2787,8 @@ static int nfs4_xdr_enc_layoutcommit(struct rpc_rqst *req,
 	encode_compound_hdr(xdr, req, &hdr);
 	encode_sequence(xdr, &args->seq_args, &hdr);
 	encode_putfh(xdr, NFS_FH(args->inode), &hdr);
-	encode_layoutcommit(xdr, args, &hdr);
+	encode_layoutreturn(xdr, args, &hdr);
-	encode_getfattr(xdr, args->bitmask, &hdr);
 	encode_nops(&hdr);
-	return 0;
 }
 #endif /* CONFIG_NFS_V4_1 */
 
@@ -5203,6 +5271,27 @@ out_overflow:
 	return -EIO;
 }
 
+static int decode_layoutreturn(struct xdr_stream *xdr,
+			       struct nfs4_layoutreturn_res *res)
+{
+	__be32 *p;
+	int status;
+
+	status = decode_op_hdr(xdr, OP_LAYOUTRETURN);
+	if (status)
+		return status;
+	p = xdr_inline_decode(xdr, 4);
+	if (unlikely(!p))
+		goto out_overflow;
+	res->lrs_present = be32_to_cpup(p);
+	if (res->lrs_present)
+		status = decode_stateid(xdr, &res->stateid);
+	return status;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
+}
+
 static int decode_layoutcommit(struct xdr_stream *xdr,
 			       struct rpc_rqst *req,
 			       struct nfs4_layoutcommit_res *res)
@@ -6319,6 +6408,30 @@ out:
 	return status;
 }
 
+/*
+ * Decode LAYOUTRETURN response
+ */
+static int nfs4_xdr_dec_layoutreturn(struct rpc_rqst *rqstp,
+				     struct xdr_stream *xdr,
+				     struct nfs4_layoutreturn_res *res)
+{
+	struct compound_hdr hdr;
+	int status;
+
+	status = decode_compound_hdr(xdr, &hdr);
+	if (status)
+		goto out;
+	status = decode_sequence(xdr, &res->seq_res, rqstp);
+	if (status)
+		goto out;
+	status = decode_putfh(xdr);
+	if (status)
+		goto out;
+	status = decode_layoutreturn(xdr, res);
+out:
+	return status;
+}
+
 /*
  * Decode LAYOUTCOMMIT response
  */
@@ -6547,6 +6660,7 @@ struct rpc_procinfo	nfs4_procedures[] = {
 	PROC(GETDEVICEINFO,	enc_getdeviceinfo,	dec_getdeviceinfo),
 	PROC(LAYOUTGET,		enc_layoutget,		dec_layoutget),
 	PROC(LAYOUTCOMMIT,	enc_layoutcommit,	dec_layoutcommit),
+	PROC(LAYOUTRETURN,	enc_layoutreturn,	dec_layoutreturn),
 #endif /* CONFIG_NFS_V4_1 */
 };
 

fs/nfs/objlayout/Kbuild

@@ -0,0 +1,5 @@
+#
+# Makefile for the pNFS Objects Layout Driver kernel module
+#
+objlayoutdriver-y := objio_osd.o pnfs_osd_xdr_cli.o objlayout.o
+obj-$(CONFIG_PNFS_OBJLAYOUT) += objlayoutdriver.o

fs/nfs/objlayout/objlayout.c

@@ -0,0 +1,712 @@
+/*
+ *  pNFS Objects layout driver high level definitions
+ *
+ *  Copyright (C) 2007 Panasas Inc. [year of first publication]
+ *  All rights reserved.
+ *
+ *  Benny Halevy <bhalevy@panasas.com>
+ *  Boaz Harrosh <bharrosh@panasas.com>
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License version 2
+ *  See the file COPYING included with this distribution for more details.
+ *
+ *  Redistribution and use in source and binary forms, with or without
+ *  modification, are permitted provided that the following conditions
+ *  are met:
+ *
+ *  1. Redistributions of source code must retain the above copyright
+ *     notice, this list of conditions and the following disclaimer.
+ *  2. Redistributions in binary form must reproduce the above copyright
+ *     notice, this list of conditions and the following disclaimer in the
+ *     documentation and/or other materials provided with the distribution.
+ *  3. Neither the name of the Panasas company nor the names of its
+ *     contributors may be used to endorse or promote products derived
+ *     from this software without specific prior written permission.
+ *
+ *  THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED
+ *  WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+ *  MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ *  DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
+ *  FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ *  CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ *  SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
+ *  BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+ *  LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+ *  NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ *  SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <scsi/osd_initiator.h>
+#include "objlayout.h"
+
+#define NFSDBG_FACILITY         NFSDBG_PNFS_LD
+/*
+ * Create a objlayout layout structure for the given inode and return it.
+ */
+struct pnfs_layout_hdr *
+objlayout_alloc_layout_hdr(struct inode *inode, gfp_t gfp_flags)
+{
+	struct objlayout *objlay;
+
+	objlay = kzalloc(sizeof(struct objlayout), gfp_flags);
+	if (objlay) {
+		spin_lock_init(&objlay->lock);
+		INIT_LIST_HEAD(&objlay->err_list);
+	}
+	dprintk("%s: Return %p\n", __func__, objlay);
+	return &objlay->pnfs_layout;
+}
+
+/*
+ * Free an objlayout layout structure
+ */
+void
+objlayout_free_layout_hdr(struct pnfs_layout_hdr *lo)
+{
+	struct objlayout *objlay = OBJLAYOUT(lo);
+
+	dprintk("%s: objlay %p\n", __func__, objlay);
+
+	WARN_ON(!list_empty(&objlay->err_list));
+	kfree(objlay);
+}
+
+/*
+ * Unmarshall layout and store it in pnfslay.
+ */
+struct pnfs_layout_segment *
+objlayout_alloc_lseg(struct pnfs_layout_hdr *pnfslay,
+		     struct nfs4_layoutget_res *lgr,
+		     gfp_t gfp_flags)
+{
+	int status = -ENOMEM;
+	struct xdr_stream stream;
+	struct xdr_buf buf = {
+		.pages =  lgr->layoutp->pages,
+		.page_len =  lgr->layoutp->len,
+		.buflen =  lgr->layoutp->len,
+		.len = lgr->layoutp->len,
+	};
+	struct page *scratch;
+	struct pnfs_layout_segment *lseg;
+
+	dprintk("%s: Begin pnfslay %p\n", __func__, pnfslay);
+
+	scratch = alloc_page(gfp_flags);
+	if (!scratch)
+		goto err_nofree;
+
+	xdr_init_decode(&stream, &buf, NULL);
+	xdr_set_scratch_buffer(&stream, page_address(scratch), PAGE_SIZE);
+
+	status = objio_alloc_lseg(&lseg, pnfslay, &lgr->range, &stream, gfp_flags);
+	if (unlikely(status)) {
+		dprintk("%s: objio_alloc_lseg Return err %d\n", __func__,
+			status);
+		goto err;
+	}
+
+	__free_page(scratch);
+
+	dprintk("%s: Return %p\n", __func__, lseg);
+	return lseg;
+
+err:
+	__free_page(scratch);
+err_nofree:
+	dprintk("%s: Err Return=>%d\n", __func__, status);
+	return ERR_PTR(status);
+}
+
+/*
+ * Free a layout segement
+ */
+void
+objlayout_free_lseg(struct pnfs_layout_segment *lseg)
+{
+	dprintk("%s: freeing layout segment %p\n", __func__, lseg);
+
+	if (unlikely(!lseg))
+		return;
+
+	objio_free_lseg(lseg);
+}
+
+/*
+ * I/O Operations
+ */
+static inline u64
+end_offset(u64 start, u64 len)
+{
+	u64 end;
+
+	end = start + len;
+	return end >= start ? end : NFS4_MAX_UINT64;
+}
+
+/* last octet in a range */
+static inline u64
+last_byte_offset(u64 start, u64 len)
+{
+	u64 end;
+
+	BUG_ON(!len);
+	end = start + len;
+	return end > start ? end - 1 : NFS4_MAX_UINT64;
+}
+
+static struct objlayout_io_state *
+objlayout_alloc_io_state(struct pnfs_layout_hdr *pnfs_layout_type,
+			struct page **pages,
+			unsigned pgbase,
+			loff_t offset,
+			size_t count,
+			struct pnfs_layout_segment *lseg,
+			void *rpcdata,
+			gfp_t gfp_flags)
+{
+	struct objlayout_io_state *state;
+	u64 lseg_end_offset;
+
+	dprintk("%s: allocating io_state\n", __func__);
+	if (objio_alloc_io_state(lseg, &state, gfp_flags))
+		return NULL;
+
+	BUG_ON(offset < lseg->pls_range.offset);
+	lseg_end_offset = end_offset(lseg->pls_range.offset,
+				     lseg->pls_range.length);
+	BUG_ON(offset >= lseg_end_offset);
+	if (offset + count > lseg_end_offset) {
+		count = lseg->pls_range.length -
+				(offset - lseg->pls_range.offset);
+		dprintk("%s: truncated count %Zd\n", __func__, count);
+	}
+
+	if (pgbase > PAGE_SIZE) {
+		pages += pgbase >> PAGE_SHIFT;
+		pgbase &= ~PAGE_MASK;
+	}
+
+	INIT_LIST_HEAD(&state->err_list);
+	state->lseg = lseg;
+	state->rpcdata = rpcdata;
+	state->pages = pages;
+	state->pgbase = pgbase;
+	state->nr_pages = (pgbase + count + PAGE_SIZE - 1) >> PAGE_SHIFT;
+	state->offset = offset;
+	state->count = count;
+	state->sync = 0;
+
+	return state;
+}
+
+static void
+objlayout_free_io_state(struct objlayout_io_state *state)
+{
+	dprintk("%s: freeing io_state\n", __func__);
+	if (unlikely(!state))
+		return;
+
+	objio_free_io_state(state);
+}
+
+/*
+ * I/O done common code
+ */
+static void
+objlayout_iodone(struct objlayout_io_state *state)
+{
+	dprintk("%s: state %p status\n", __func__, state);
+
+	if (likely(state->status >= 0)) {
+		objlayout_free_io_state(state);
+	} else {
+		struct objlayout *objlay = OBJLAYOUT(state->lseg->pls_layout);
+
+		spin_lock(&objlay->lock);
+		objlay->delta_space_valid = OBJ_DSU_INVALID;
+		list_add(&objlay->err_list, &state->err_list);
+		spin_unlock(&objlay->lock);
+	}
+}
+
+/*
+ * objlayout_io_set_result - Set an osd_error code on a specific osd comp.
+ *
+ * The @index component IO failed (error returned from target). Register
+ * the error for later reporting at layout-return.
+ */
+void
+objlayout_io_set_result(struct objlayout_io_state *state, unsigned index,
+			struct pnfs_osd_objid *pooid, int osd_error,
+			u64 offset, u64 length, bool is_write)
+{
+	struct pnfs_osd_ioerr *ioerr = &state->ioerrs[index];
+
+	BUG_ON(index >= state->num_comps);
+	if (osd_error) {
+		ioerr->oer_component = *pooid;
+		ioerr->oer_comp_offset = offset;
+		ioerr->oer_comp_length = length;
+		ioerr->oer_iswrite = is_write;
+		ioerr->oer_errno = osd_error;
+
+		dprintk("%s: err[%d]: errno=%d is_write=%d dev(%llx:%llx) "
+			"par=0x%llx obj=0x%llx offset=0x%llx length=0x%llx\n",
+			__func__, index, ioerr->oer_errno,
+			ioerr->oer_iswrite,
+			_DEVID_LO(&ioerr->oer_component.oid_device_id),
+			_DEVID_HI(&ioerr->oer_component.oid_device_id),
+			ioerr->oer_component.oid_partition_id,
+			ioerr->oer_component.oid_object_id,
+			ioerr->oer_comp_offset,
+			ioerr->oer_comp_length);
+	} else {
+		/* User need not call if no error is reported */
+		ioerr->oer_errno = 0;
+	}
+}
+
+/* Function scheduled on rpc workqueue to call ->nfs_readlist_complete().
+ * This is because the osd completion is called with ints-off from
+ * the block layer
+ */
+static void _rpc_read_complete(struct work_struct *work)
+{
+	struct rpc_task *task;
+	struct nfs_read_data *rdata;
+
+	dprintk("%s enter\n", __func__);
+	task = container_of(work, struct rpc_task, u.tk_work);
+	rdata = container_of(task, struct nfs_read_data, task);
+
+	pnfs_ld_read_done(rdata);
+}
+
+void
+objlayout_read_done(struct objlayout_io_state *state, ssize_t status, bool sync)
+{
+	int eof = state->eof;
+	struct nfs_read_data *rdata;
+
+	state->status = status;
+	dprintk("%s: Begin status=%ld eof=%d\n", __func__, status, eof);
+	rdata = state->rpcdata;
+	rdata->task.tk_status = status;
+	if (status >= 0) {
+		rdata->res.count = status;
+		rdata->res.eof = eof;
+	}
+	objlayout_iodone(state);
+	/* must not use state after this point */
+
+	if (sync)
+		pnfs_ld_read_done(rdata);
+	else {
+		INIT_WORK(&rdata->task.u.tk_work, _rpc_read_complete);
+		schedule_work(&rdata->task.u.tk_work);
+	}
+}
+
+/*
+ * Perform sync or async reads.
+ */
+enum pnfs_try_status
+objlayout_read_pagelist(struct nfs_read_data *rdata)
+{
+	loff_t offset = rdata->args.offset;
+	size_t count = rdata->args.count;
+	struct objlayout_io_state *state;
+	ssize_t status = 0;
+	loff_t eof;
+
+	dprintk("%s: Begin inode %p offset %llu count %d\n",
+		__func__, rdata->inode, offset, (int)count);
+
+	eof = i_size_read(rdata->inode);
+	if (unlikely(offset + count > eof)) {
+		if (offset >= eof) {
+			status = 0;
+			rdata->res.count = 0;
+			rdata->res.eof = 1;
+			goto out;
+		}
+		count = eof - offset;
+	}
+
+	state = objlayout_alloc_io_state(NFS_I(rdata->inode)->layout,
+					 rdata->args.pages, rdata->args.pgbase,
+					 offset, count,
+					 rdata->lseg, rdata,
+					 GFP_KERNEL);
+	if (unlikely(!state)) {
+		status = -ENOMEM;
+		goto out;
+	}
+
+	state->eof = state->offset + state->count >= eof;
+
+	status = objio_read_pagelist(state);
+ out:
+	dprintk("%s: Return status %Zd\n", __func__, status);
+	rdata->pnfs_error = status;
+	return PNFS_ATTEMPTED;
+}
+
+/* Function scheduled on rpc workqueue to call ->nfs_writelist_complete().
+ * This is because the osd completion is called with ints-off from
+ * the block layer
+ */
+static void _rpc_write_complete(struct work_struct *work)
+{
+	struct rpc_task *task;
+	struct nfs_write_data *wdata;
+
+	dprintk("%s enter\n", __func__);
+	task = container_of(work, struct rpc_task, u.tk_work);
+	wdata = container_of(task, struct nfs_write_data, task);
+
+	pnfs_ld_write_done(wdata);
+}
+
+void
+objlayout_write_done(struct objlayout_io_state *state, ssize_t status,
+		     bool sync)
+{
+	struct nfs_write_data *wdata;
+
+	dprintk("%s: Begin\n", __func__);
+	wdata = state->rpcdata;
+	state->status = status;
+	wdata->task.tk_status = status;
+	if (status >= 0) {
+		wdata->res.count = status;
+		wdata->verf.committed = state->committed;
+		dprintk("%s: Return status %d committed %d\n",
+			__func__, wdata->task.tk_status,
+			wdata->verf.committed);
+	} else
+		dprintk("%s: Return status %d\n",
+			__func__, wdata->task.tk_status);
+	objlayout_iodone(state);
+	/* must not use state after this point */
+
+	if (sync)
+		pnfs_ld_write_done(wdata);
+	else {
+		INIT_WORK(&wdata->task.u.tk_work, _rpc_write_complete);
+		schedule_work(&wdata->task.u.tk_work);
+	}
+}
+
+/*
+ * Perform sync or async writes.
+ */
+enum pnfs_try_status
+objlayout_write_pagelist(struct nfs_write_data *wdata,
+			 int how)
+{
+	struct objlayout_io_state *state;
+	ssize_t status;
+
+	dprintk("%s: Begin inode %p offset %llu count %u\n",
+		__func__, wdata->inode, wdata->args.offset, wdata->args.count);
+
+	state = objlayout_alloc_io_state(NFS_I(wdata->inode)->layout,
+					 wdata->args.pages,
+					 wdata->args.pgbase,
+					 wdata->args.offset,
+					 wdata->args.count,
+					 wdata->lseg, wdata,
+					 GFP_NOFS);
+	if (unlikely(!state)) {
+		status = -ENOMEM;
+		goto out;
+	}
+
+	state->sync = how & FLUSH_SYNC;
+
+	status = objio_write_pagelist(state, how & FLUSH_STABLE);
+ out:
+	dprintk("%s: Return status %Zd\n", __func__, status);
+	wdata->pnfs_error = status;
+	return PNFS_ATTEMPTED;
+}
+
+void
+objlayout_encode_layoutcommit(struct pnfs_layout_hdr *pnfslay,
+			      struct xdr_stream *xdr,
+			      const struct nfs4_layoutcommit_args *args)
+{
+	struct objlayout *objlay = OBJLAYOUT(pnfslay);
+	struct pnfs_osd_layoutupdate lou;
+	__be32 *start;
+
+	dprintk("%s: Begin\n", __func__);
+
+	spin_lock(&objlay->lock);
+	lou.dsu_valid = (objlay->delta_space_valid == OBJ_DSU_VALID);
+	lou.dsu_delta = objlay->delta_space_used;
+	objlay->delta_space_used = 0;
+	objlay->delta_space_valid = OBJ_DSU_INIT;
+	lou.olu_ioerr_flag = !list_empty(&objlay->err_list);
+	spin_unlock(&objlay->lock);
+
+	start = xdr_reserve_space(xdr, 4);
+
+	BUG_ON(pnfs_osd_xdr_encode_layoutupdate(xdr, &lou));
+
+	*start = cpu_to_be32((xdr->p - start - 1) * 4);
+
+	dprintk("%s: Return delta_space_used %lld err %d\n", __func__,
+		lou.dsu_delta, lou.olu_ioerr_flag);
+}
+
+static int
+err_prio(u32 oer_errno)
+{
+	switch (oer_errno) {
+	case 0:
+		return 0;
+
+	case PNFS_OSD_ERR_RESOURCE:
+		return OSD_ERR_PRI_RESOURCE;
+	case PNFS_OSD_ERR_BAD_CRED:
+		return OSD_ERR_PRI_BAD_CRED;
+	case PNFS_OSD_ERR_NO_ACCESS:
+		return OSD_ERR_PRI_NO_ACCESS;
+	case PNFS_OSD_ERR_UNREACHABLE:
+		return OSD_ERR_PRI_UNREACHABLE;
+	case PNFS_OSD_ERR_NOT_FOUND:
+		return OSD_ERR_PRI_NOT_FOUND;
+	case PNFS_OSD_ERR_NO_SPACE:
+		return OSD_ERR_PRI_NO_SPACE;
+	default:
+		WARN_ON(1);
+		/* fallthrough */
+	case PNFS_OSD_ERR_EIO:
+		return OSD_ERR_PRI_EIO;
+	}
+}
+
+static void
+merge_ioerr(struct pnfs_osd_ioerr *dest_err,
+	    const struct pnfs_osd_ioerr *src_err)
+{
+	u64 dest_end, src_end;
+
+	if (!dest_err->oer_errno) {
+		*dest_err = *src_err;
+		/* accumulated device must be blank */
+		memset(&dest_err->oer_component.oid_device_id, 0,
+			sizeof(dest_err->oer_component.oid_device_id));
+
+		return;
+	}
+
+	if (dest_err->oer_component.oid_partition_id !=
+				src_err->oer_component.oid_partition_id)
+		dest_err->oer_component.oid_partition_id = 0;
+
+	if (dest_err->oer_component.oid_object_id !=
+				src_err->oer_component.oid_object_id)
+		dest_err->oer_component.oid_object_id = 0;
+
+	if (dest_err->oer_comp_offset > src_err->oer_comp_offset)
+		dest_err->oer_comp_offset = src_err->oer_comp_offset;
+
+	dest_end = end_offset(dest_err->oer_comp_offset,
+			      dest_err->oer_comp_length);
+	src_end =  end_offset(src_err->oer_comp_offset,
+			      src_err->oer_comp_length);
+	if (dest_end < src_end)
+		dest_end = src_end;
+
+	dest_err->oer_comp_length = dest_end - dest_err->oer_comp_offset;
+
+	if ((src_err->oer_iswrite == dest_err->oer_iswrite) &&
+	    (err_prio(src_err->oer_errno) > err_prio(dest_err->oer_errno))) {
+			dest_err->oer_errno = src_err->oer_errno;
+	} else if (src_err->oer_iswrite) {
+		dest_err->oer_iswrite = true;
+		dest_err->oer_errno = src_err->oer_errno;
+	}
+}
+
+static void
+encode_accumulated_error(struct objlayout *objlay, __be32 *p)
+{
+	struct objlayout_io_state *state, *tmp;
+	struct pnfs_osd_ioerr accumulated_err = {.oer_errno = 0};
+
+	list_for_each_entry_safe(state, tmp, &objlay->err_list, err_list) {
+		unsigned i;
+
+		for (i = 0; i < state->num_comps; i++) {
+			struct pnfs_osd_ioerr *ioerr = &state->ioerrs[i];
+
+			if (!ioerr->oer_errno)
+				continue;
+
+			printk(KERN_ERR "%s: err[%d]: errno=%d is_write=%d "
+				"dev(%llx:%llx) par=0x%llx obj=0x%llx "
+				"offset=0x%llx length=0x%llx\n",
+				__func__, i, ioerr->oer_errno,
+				ioerr->oer_iswrite,
+				_DEVID_LO(&ioerr->oer_component.oid_device_id),
+				_DEVID_HI(&ioerr->oer_component.oid_device_id),
+				ioerr->oer_component.oid_partition_id,
+				ioerr->oer_component.oid_object_id,
+				ioerr->oer_comp_offset,
+				ioerr->oer_comp_length);
+
+			merge_ioerr(&accumulated_err, ioerr);
+		}
+		list_del(&state->err_list);
+		objlayout_free_io_state(state);
+	}
+
+	pnfs_osd_xdr_encode_ioerr(p, &accumulated_err);
+}
+
+void
+objlayout_encode_layoutreturn(struct pnfs_layout_hdr *pnfslay,
+			      struct xdr_stream *xdr,
+			      const struct nfs4_layoutreturn_args *args)
+{
+	struct objlayout *objlay = OBJLAYOUT(pnfslay);
+	struct objlayout_io_state *state, *tmp;
+	__be32 *start;
+
+	dprintk("%s: Begin\n", __func__);
+	start = xdr_reserve_space(xdr, 4);
+	BUG_ON(!start);
+
+	spin_lock(&objlay->lock);
+
+	list_for_each_entry_safe(state, tmp, &objlay->err_list, err_list) {
+		__be32 *last_xdr = NULL, *p;
+		unsigned i;
+		int res = 0;
+
+		for (i = 0; i < state->num_comps; i++) {
+			struct pnfs_osd_ioerr *ioerr = &state->ioerrs[i];
+
+			if (!ioerr->oer_errno)
+				continue;
+
+			dprintk("%s: err[%d]: errno=%d is_write=%d "
+				"dev(%llx:%llx) par=0x%llx obj=0x%llx "
+				"offset=0x%llx length=0x%llx\n",
+				__func__, i, ioerr->oer_errno,
+				ioerr->oer_iswrite,
+				_DEVID_LO(&ioerr->oer_component.oid_device_id),
+				_DEVID_HI(&ioerr->oer_component.oid_device_id),
+				ioerr->oer_component.oid_partition_id,
+				ioerr->oer_component.oid_object_id,
+				ioerr->oer_comp_offset,
+				ioerr->oer_comp_length);
+
+			p = pnfs_osd_xdr_ioerr_reserve_space(xdr);
+			if (unlikely(!p)) {
+				res = -E2BIG;
+				break; /* accumulated_error */
+			}
+
+			last_xdr = p;
+			pnfs_osd_xdr_encode_ioerr(p, &state->ioerrs[i]);
+		}
+
+		/* TODO: use xdr_write_pages */
+		if (unlikely(res)) {
+			/* no space for even one error descriptor */
+			BUG_ON(!last_xdr);
+
+			/* we've encountered a situation with lots and lots of
+			 * errors and no space to encode them all. Use the last
+			 * available slot to report the union of all the
+			 * remaining errors.
+			 */
+			encode_accumulated_error(objlay, last_xdr);
+			goto loop_done;
+		}
+		list_del(&state->err_list);
+		objlayout_free_io_state(state);
+	}
+loop_done:
+	spin_unlock(&objlay->lock);
+
+	*start = cpu_to_be32((xdr->p - start - 1) * 4);
+	dprintk("%s: Return\n", __func__);
+}
+
+
+/*
+ * Get Device Info API for io engines
+ */
+struct objlayout_deviceinfo {
+	struct page *page;
+	struct pnfs_osd_deviceaddr da; /* This must be last */
+};
+
+/* Initialize and call nfs_getdeviceinfo, then decode and return a
+ * "struct pnfs_osd_deviceaddr *" Eventually objlayout_put_deviceinfo()
+ * should be called.
+ */
+int objlayout_get_deviceinfo(struct pnfs_layout_hdr *pnfslay,
+	struct nfs4_deviceid *d_id, struct pnfs_osd_deviceaddr **deviceaddr,
+	gfp_t gfp_flags)
+{
+	struct objlayout_deviceinfo *odi;
+	struct pnfs_device pd;
+	struct super_block *sb;
+	struct page *page, **pages;
+	u32 *p;
+	int err;
+
+	page = alloc_page(gfp_flags);
+	if (!page)
+		return -ENOMEM;
+
+	pages = &page;
+	pd.pages = pages;
+
+	memcpy(&pd.dev_id, d_id, sizeof(*d_id));
+	pd.layout_type = LAYOUT_OSD2_OBJECTS;
+	pd.pages = &page;
+	pd.pgbase = 0;
+	pd.pglen = PAGE_SIZE;
+	pd.mincount = 0;
+
+	sb = pnfslay->plh_inode->i_sb;
+	err = nfs4_proc_getdeviceinfo(NFS_SERVER(pnfslay->plh_inode), &pd);
+	dprintk("%s nfs_getdeviceinfo returned %d\n", __func__, err);
+	if (err)
+		goto err_out;
+
+	p = page_address(page);
+	odi = kzalloc(sizeof(*odi), gfp_flags);
+	if (!odi) {
+		err = -ENOMEM;
+		goto err_out;
+	}
+	pnfs_osd_xdr_decode_deviceaddr(&odi->da, p);
+	odi->page = page;
+	*deviceaddr = &odi->da;
+	return 0;
+
+err_out:
+	__free_page(page);
+	return err;
+}
+
+void objlayout_put_deviceinfo(struct pnfs_osd_deviceaddr *deviceaddr)
+{
+	struct objlayout_deviceinfo *odi = container_of(deviceaddr,
+						struct objlayout_deviceinfo,
+						da);
+
+	__free_page(odi->page);
+	kfree(odi);
+}

fs/nfs/objlayout/objlayout.h

@@ -0,0 +1,187 @@
+/*
+ *  Data types and function declerations for interfacing with the
+ *  pNFS standard object layout driver.
+ *
+ *  Copyright (C) 2007 Panasas Inc. [year of first publication]
+ *  All rights reserved.
+ *
+ *  Benny Halevy <bhalevy@panasas.com>
+ *  Boaz Harrosh <bharrosh@panasas.com>
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License version 2
+ *  See the file COPYING included with this distribution for more details.
+ *
+ *  Redistribution and use in source and binary forms, with or without
+ *  modification, are permitted provided that the following conditions
+ *  are met:
+ *
+ *  1. Redistributions of source code must retain the above copyright
+ *     notice, this list of conditions and the following disclaimer.
+ *  2. Redistributions in binary form must reproduce the above copyright
+ *     notice, this list of conditions and the following disclaimer in the
+ *     documentation and/or other materials provided with the distribution.
+ *  3. Neither the name of the Panasas company nor the names of its
+ *     contributors may be used to endorse or promote products derived
+ *     from this software without specific prior written permission.
+ *
+ *  THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED
+ *  WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+ *  MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ *  DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
+ *  FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ *  CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ *  SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
+ *  BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+ *  LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+ *  NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ *  SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef _OBJLAYOUT_H
+#define _OBJLAYOUT_H
+
+#include <linux/nfs_fs.h>
+#include <linux/pnfs_osd_xdr.h>
+#include "../pnfs.h"
+
+/*
+ * per-inode layout
+ */
+struct objlayout {
+	struct pnfs_layout_hdr pnfs_layout;
+
+	 /* for layout_commit */
+	enum osd_delta_space_valid_enum {
+		OBJ_DSU_INIT = 0,
+		OBJ_DSU_VALID,
+		OBJ_DSU_INVALID,
+	} delta_space_valid;
+	s64 delta_space_used;  /* consumed by write ops */
+
+	 /* for layout_return */
+	spinlock_t lock;
+	struct list_head err_list;
+};
+
+static inline struct objlayout *
+OBJLAYOUT(struct pnfs_layout_hdr *lo)
+{
+	return container_of(lo, struct objlayout, pnfs_layout);
+}
+
+/*
+ * per-I/O operation state
+ * embedded in objects provider io_state data structure
+ */
+struct objlayout_io_state {
+	struct pnfs_layout_segment *lseg;
+
+	struct page **pages;
+	unsigned pgbase;
+	unsigned nr_pages;
+	unsigned long count;
+	loff_t offset;
+	bool sync;
+
+	void *rpcdata;
+	int status;             /* res */
+	int eof;                /* res */
+	int committed;          /* res */
+
+	/* Error reporting (layout_return) */
+	struct list_head err_list;
+	unsigned num_comps;
+	/* Pointer to array of error descriptors of size num_comps.
+	 * It should contain as many entries as devices in the osd_layout
+	 * that participate in the I/O. It is up to the io_engine to allocate
+	 * needed space and set num_comps.
+	 */
+	struct pnfs_osd_ioerr *ioerrs;
+};
+
+/*
+ * Raid engine I/O API
+ */
+extern int objio_alloc_lseg(struct pnfs_layout_segment **outp,
+	struct pnfs_layout_hdr *pnfslay,
+	struct pnfs_layout_range *range,
+	struct xdr_stream *xdr,
+	gfp_t gfp_flags);
+extern void objio_free_lseg(struct pnfs_layout_segment *lseg);
+
+extern int objio_alloc_io_state(
+	struct pnfs_layout_segment *lseg,
+	struct objlayout_io_state **outp,
+	gfp_t gfp_flags);
+extern void objio_free_io_state(struct objlayout_io_state *state);
+
+extern ssize_t objio_read_pagelist(struct objlayout_io_state *ol_state);
+extern ssize_t objio_write_pagelist(struct objlayout_io_state *ol_state,
+				    bool stable);
+
+/*
+ * callback API
+ */
+extern void objlayout_io_set_result(struct objlayout_io_state *state,
+			unsigned index, struct pnfs_osd_objid *pooid,
+			int osd_error, u64 offset, u64 length, bool is_write);
+
+static inline void
+objlayout_add_delta_space_used(struct objlayout_io_state *state, s64 space_used)
+{
+	struct objlayout *objlay = OBJLAYOUT(state->lseg->pls_layout);
+
+	/* If one of the I/Os errored out and the delta_space_used was
+	 * invalid we render the complete report as invalid. Protocol mandate
+	 * the DSU be accurate or not reported.
+	 */
+	spin_lock(&objlay->lock);
+	if (objlay->delta_space_valid != OBJ_DSU_INVALID) {
+		objlay->delta_space_valid = OBJ_DSU_VALID;
+		objlay->delta_space_used += space_used;
+	}
+	spin_unlock(&objlay->lock);
+}
+
+extern void objlayout_read_done(struct objlayout_io_state *state,
+				ssize_t status, bool sync);
+extern void objlayout_write_done(struct objlayout_io_state *state,
+				 ssize_t status, bool sync);
+
+extern int objlayout_get_deviceinfo(struct pnfs_layout_hdr *pnfslay,
+	struct nfs4_deviceid *d_id, struct pnfs_osd_deviceaddr **deviceaddr,
+	gfp_t gfp_flags);
+extern void objlayout_put_deviceinfo(struct pnfs_osd_deviceaddr *deviceaddr);
+
+/*
+ * exported generic objects function vectors
+ */
+
+extern struct pnfs_layout_hdr *objlayout_alloc_layout_hdr(struct inode *, gfp_t gfp_flags);
+extern void objlayout_free_layout_hdr(struct pnfs_layout_hdr *);
+
+extern struct pnfs_layout_segment *objlayout_alloc_lseg(
+	struct pnfs_layout_hdr *,
+	struct nfs4_layoutget_res *,
+	gfp_t gfp_flags);
+extern void objlayout_free_lseg(struct pnfs_layout_segment *);
+
+extern enum pnfs_try_status objlayout_read_pagelist(
+	struct nfs_read_data *);
+
+extern enum pnfs_try_status objlayout_write_pagelist(
+	struct nfs_write_data *,
+	int how);
+
+extern void objlayout_encode_layoutcommit(
+	struct pnfs_layout_hdr *,
+	struct xdr_stream *,
+	const struct nfs4_layoutcommit_args *);
+
+extern void objlayout_encode_layoutreturn(
+	struct pnfs_layout_hdr *,
+	struct xdr_stream *,
+	const struct nfs4_layoutreturn_args *);
+
+#endif /* _OBJLAYOUT_H */

fs/nfs/objlayout/pnfs_osd_xdr_cli.c

@@ -0,0 +1,412 @@
+/*
+ *  Object-Based pNFS Layout XDR layer
+ *
+ *  Copyright (C) 2007 Panasas Inc. [year of first publication]
+ *  All rights reserved.
+ *
+ *  Benny Halevy <bhalevy@panasas.com>
+ *  Boaz Harrosh <bharrosh@panasas.com>
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License version 2
+ *  See the file COPYING included with this distribution for more details.
+ *
+ *  Redistribution and use in source and binary forms, with or without
+ *  modification, are permitted provided that the following conditions
+ *  are met:
+ *
+ *  1. Redistributions of source code must retain the above copyright
+ *     notice, this list of conditions and the following disclaimer.
+ *  2. Redistributions in binary form must reproduce the above copyright
+ *     notice, this list of conditions and the following disclaimer in the
+ *     documentation and/or other materials provided with the distribution.
+ *  3. Neither the name of the Panasas company nor the names of its
+ *     contributors may be used to endorse or promote products derived
+ *     from this software without specific prior written permission.
+ *
+ *  THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED
+ *  WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+ *  MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ *  DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
+ *  FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ *  CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ *  SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
+ *  BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+ *  LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+ *  NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ *  SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <linux/pnfs_osd_xdr.h>
+
+#define NFSDBG_FACILITY         NFSDBG_PNFS_LD
+
+/*
+ * The following implementation is based on RFC5664
+ */
+
+/*
+ * struct pnfs_osd_objid {
+ *	struct nfs4_deviceid	oid_device_id;
+ *	u64			oid_partition_id;
+ *	u64			oid_object_id;
+ * }; // xdr size 32 bytes
+ */
+static __be32 *
+_osd_xdr_decode_objid(__be32 *p, struct pnfs_osd_objid *objid)
+{
+	p = xdr_decode_opaque_fixed(p, objid->oid_device_id.data,
+				    sizeof(objid->oid_device_id.data));
+
+	p = xdr_decode_hyper(p, &objid->oid_partition_id);
+	p = xdr_decode_hyper(p, &objid->oid_object_id);
+	return p;
+}
+/*
+ * struct pnfs_osd_opaque_cred {
+ *	u32 cred_len;
+ *	void *cred;
+ * }; // xdr size [variable]
+ * The return pointers are from the xdr buffer
+ */
+static int
+_osd_xdr_decode_opaque_cred(struct pnfs_osd_opaque_cred *opaque_cred,
+			    struct xdr_stream *xdr)
+{
+	__be32 *p = xdr_inline_decode(xdr, 1);
+
+	if (!p)
+		return -EINVAL;
+
+	opaque_cred->cred_len = be32_to_cpu(*p++);
+
+	p = xdr_inline_decode(xdr, opaque_cred->cred_len);
+	if (!p)
+		return -EINVAL;
+
+	opaque_cred->cred = p;
+	return 0;
+}
+
+/*
+ * struct pnfs_osd_object_cred {
+ *	struct pnfs_osd_objid		oc_object_id;
+ *	u32				oc_osd_version;
+ *	u32				oc_cap_key_sec;
+ *	struct pnfs_osd_opaque_cred	oc_cap_key
+ *	struct pnfs_osd_opaque_cred	oc_cap;
+ * }; // xdr size 32 + 4 + 4 + [variable] + [variable]
+ */
+static int
+_osd_xdr_decode_object_cred(struct pnfs_osd_object_cred *comp,
+			    struct xdr_stream *xdr)
+{
+	__be32 *p = xdr_inline_decode(xdr, 32 + 4 + 4);
+	int ret;
+
+	if (!p)
+		return -EIO;
+
+	p = _osd_xdr_decode_objid(p, &comp->oc_object_id);
+	comp->oc_osd_version = be32_to_cpup(p++);
+	comp->oc_cap_key_sec = be32_to_cpup(p);
+
+	ret = _osd_xdr_decode_opaque_cred(&comp->oc_cap_key, xdr);
+	if (unlikely(ret))
+		return ret;
+
+	ret = _osd_xdr_decode_opaque_cred(&comp->oc_cap, xdr);
+	return ret;
+}
+
+/*
+ * struct pnfs_osd_data_map {
+ *	u32	odm_num_comps;
+ *	u64	odm_stripe_unit;
+ *	u32	odm_group_width;
+ *	u32	odm_group_depth;
+ *	u32	odm_mirror_cnt;
+ *	u32	odm_raid_algorithm;
+ * }; // xdr size 4 + 8 + 4 + 4 + 4 + 4
+ */
+static inline int
+_osd_data_map_xdr_sz(void)
+{
+	return 4 + 8 + 4 + 4 + 4 + 4;
+}
+
+static __be32 *
+_osd_xdr_decode_data_map(__be32 *p, struct pnfs_osd_data_map *data_map)
+{
+	data_map->odm_num_comps = be32_to_cpup(p++);
+	p = xdr_decode_hyper(p, &data_map->odm_stripe_unit);
+	data_map->odm_group_width = be32_to_cpup(p++);
+	data_map->odm_group_depth = be32_to_cpup(p++);
+	data_map->odm_mirror_cnt = be32_to_cpup(p++);
+	data_map->odm_raid_algorithm = be32_to_cpup(p++);
+	dprintk("%s: odm_num_comps=%u odm_stripe_unit=%llu odm_group_width=%u "
+		"odm_group_depth=%u odm_mirror_cnt=%u odm_raid_algorithm=%u\n",
+		__func__,
+		data_map->odm_num_comps,
+		(unsigned long long)data_map->odm_stripe_unit,
+		data_map->odm_group_width,
+		data_map->odm_group_depth,
+		data_map->odm_mirror_cnt,
+		data_map->odm_raid_algorithm);
+	return p;
+}
+
+int pnfs_osd_xdr_decode_layout_map(struct pnfs_osd_layout *layout,
+	struct pnfs_osd_xdr_decode_layout_iter *iter, struct xdr_stream *xdr)
+{
+	__be32 *p;
+
+	memset(iter, 0, sizeof(*iter));
+
+	p = xdr_inline_decode(xdr, _osd_data_map_xdr_sz() + 4 + 4);
+	if (unlikely(!p))
+		return -EINVAL;
+
+	p = _osd_xdr_decode_data_map(p, &layout->olo_map);
+	layout->olo_comps_index = be32_to_cpup(p++);
+	layout->olo_num_comps = be32_to_cpup(p++);
+	iter->total_comps = layout->olo_num_comps;
+	return 0;
+}
+
+bool pnfs_osd_xdr_decode_layout_comp(struct pnfs_osd_object_cred *comp,
+	struct pnfs_osd_xdr_decode_layout_iter *iter, struct xdr_stream *xdr,
+	int *err)
+{
+	BUG_ON(iter->decoded_comps > iter->total_comps);
+	if (iter->decoded_comps == iter->total_comps)
+		return false;
+
+	*err = _osd_xdr_decode_object_cred(comp, xdr);
+	if (unlikely(*err)) {
+		dprintk("%s: _osd_xdr_decode_object_cred=>%d decoded_comps=%d "
+			"total_comps=%d\n", __func__, *err,
+			iter->decoded_comps, iter->total_comps);
+		return false; /* stop the loop */
+	}
+	dprintk("%s: dev(%llx:%llx) par=0x%llx obj=0x%llx "
+		"key_len=%u cap_len=%u\n",
+		__func__,
+		_DEVID_LO(&comp->oc_object_id.oid_device_id),
+		_DEVID_HI(&comp->oc_object_id.oid_device_id),
+		comp->oc_object_id.oid_partition_id,
+		comp->oc_object_id.oid_object_id,
+		comp->oc_cap_key.cred_len, comp->oc_cap.cred_len);
+
+	iter->decoded_comps++;
+	return true;
+}
+
+/*
+ * Get Device Information Decoding
+ *
+ * Note: since Device Information is currently done synchronously, all
+ *       variable strings fields are left inside the rpc buffer and are only
+ *       pointed to by the pnfs_osd_deviceaddr members. So the read buffer
+ *       should not be freed while the returned information is in use.
+ */
+/*
+ *struct nfs4_string {
+ *	unsigned int len;
+ *	char *data;
+ *}; // size [variable]
+ * NOTE: Returned string points to inside the XDR buffer
+ */
+static __be32 *
+__read_u8_opaque(__be32 *p, struct nfs4_string *str)
+{
+	str->len = be32_to_cpup(p++);
+	str->data = (char *)p;
+
+	p += XDR_QUADLEN(str->len);
+	return p;
+}
+
+/*
+ * struct pnfs_osd_targetid {
+ *	u32			oti_type;
+ *	struct nfs4_string	oti_scsi_device_id;
+ * };// size 4 + [variable]
+ */
+static __be32 *
+__read_targetid(__be32 *p, struct pnfs_osd_targetid* targetid)
+{
+	u32 oti_type;
+
+	oti_type = be32_to_cpup(p++);
+	targetid->oti_type = oti_type;
+
+	switch (oti_type) {
+	case OBJ_TARGET_SCSI_NAME:
+	case OBJ_TARGET_SCSI_DEVICE_ID:
+		p = __read_u8_opaque(p, &targetid->oti_scsi_device_id);
+	}
+
+	return p;
+}
+
+/*
+ * struct pnfs_osd_net_addr {
+ *	struct nfs4_string	r_netid;
+ *	struct nfs4_string	r_addr;
+ * };
+ */
+static __be32 *
+__read_net_addr(__be32 *p, struct pnfs_osd_net_addr* netaddr)
+{
+	p = __read_u8_opaque(p, &netaddr->r_netid);
+	p = __read_u8_opaque(p, &netaddr->r_addr);
+
+	return p;
+}
+
+/*
+ * struct pnfs_osd_targetaddr {
+ *	u32				ota_available;
+ *	struct pnfs_osd_net_addr	ota_netaddr;
+ * };
+ */
+static __be32 *
+__read_targetaddr(__be32 *p, struct pnfs_osd_targetaddr *targetaddr)
+{
+	u32 ota_available;
+
+	ota_available = be32_to_cpup(p++);
+	targetaddr->ota_available = ota_available;
+
+	if (ota_available)
+		p = __read_net_addr(p, &targetaddr->ota_netaddr);
+
+
+	return p;
+}
+
+/*
+ * struct pnfs_osd_deviceaddr {
+ *	struct pnfs_osd_targetid	oda_targetid;
+ *	struct pnfs_osd_targetaddr	oda_targetaddr;
+ *	u8				oda_lun[8];
+ *	struct nfs4_string		oda_systemid;
+ *	struct pnfs_osd_object_cred	oda_root_obj_cred;
+ *	struct nfs4_string		oda_osdname;
+ * };
+ */
+
+/* We need this version for the pnfs_osd_xdr_decode_deviceaddr which does
+ * not have an xdr_stream
+ */
+static __be32 *
+__read_opaque_cred(__be32 *p,
+			      struct pnfs_osd_opaque_cred *opaque_cred)
+{
+	opaque_cred->cred_len = be32_to_cpu(*p++);
+	opaque_cred->cred = p;
+	return p + XDR_QUADLEN(opaque_cred->cred_len);
+}
+
+static __be32 *
+__read_object_cred(__be32 *p, struct pnfs_osd_object_cred *comp)
+{
+	p = _osd_xdr_decode_objid(p, &comp->oc_object_id);
+	comp->oc_osd_version = be32_to_cpup(p++);
+	comp->oc_cap_key_sec = be32_to_cpup(p++);
+
+	p = __read_opaque_cred(p, &comp->oc_cap_key);
+	p = __read_opaque_cred(p, &comp->oc_cap);
+	return p;
+}
+
+void pnfs_osd_xdr_decode_deviceaddr(
+	struct pnfs_osd_deviceaddr *deviceaddr, __be32 *p)
+{
+	p = __read_targetid(p, &deviceaddr->oda_targetid);
+
+	p = __read_targetaddr(p, &deviceaddr->oda_targetaddr);
+
+	p = xdr_decode_opaque_fixed(p, deviceaddr->oda_lun,
+				    sizeof(deviceaddr->oda_lun));
+
+	p = __read_u8_opaque(p, &deviceaddr->oda_systemid);
+
+	p = __read_object_cred(p, &deviceaddr->oda_root_obj_cred);
+
+	p = __read_u8_opaque(p, &deviceaddr->oda_osdname);
+
+	/* libosd likes this terminated in dbg. It's last, so no problems */
+	deviceaddr->oda_osdname.data[deviceaddr->oda_osdname.len] = 0;
+}
+
+/*
+ * struct pnfs_osd_layoutupdate {
+ *	u32	dsu_valid;
+ *	s64	dsu_delta;
+ *	u32	olu_ioerr_flag;
+ * }; xdr size 4 + 8 + 4
+ */
+int
+pnfs_osd_xdr_encode_layoutupdate(struct xdr_stream *xdr,
+				 struct pnfs_osd_layoutupdate *lou)
+{
+	__be32 *p = xdr_reserve_space(xdr,  4 + 8 + 4);
+
+	if (!p)
+		return -E2BIG;
+
+	*p++ = cpu_to_be32(lou->dsu_valid);
+	if (lou->dsu_valid)
+		p = xdr_encode_hyper(p, lou->dsu_delta);
+	*p++ = cpu_to_be32(lou->olu_ioerr_flag);
+	return 0;
+}
+
+/*
+ * struct pnfs_osd_objid {
+ *	struct nfs4_deviceid	oid_device_id;
+ *	u64			oid_partition_id;
+ *	u64			oid_object_id;
+ * }; // xdr size 32 bytes
+ */
+static inline __be32 *
+pnfs_osd_xdr_encode_objid(__be32 *p, struct pnfs_osd_objid *object_id)
+{
+	p = xdr_encode_opaque_fixed(p, &object_id->oid_device_id.data,
+				    sizeof(object_id->oid_device_id.data));
+	p = xdr_encode_hyper(p, object_id->oid_partition_id);
+	p = xdr_encode_hyper(p, object_id->oid_object_id);
+
+	return p;
+}
+
+/*
+ * struct pnfs_osd_ioerr {
+ *	struct pnfs_osd_objid	oer_component;
+ *	u64			oer_comp_offset;
+ *	u64			oer_comp_length;
+ *	u32			oer_iswrite;
+ *	u32			oer_errno;
+ * }; // xdr size 32 + 24 bytes
+ */
+void pnfs_osd_xdr_encode_ioerr(__be32 *p, struct pnfs_osd_ioerr *ioerr)
+{
+	p = pnfs_osd_xdr_encode_objid(p, &ioerr->oer_component);
+	p = xdr_encode_hyper(p, ioerr->oer_comp_offset);
+	p = xdr_encode_hyper(p, ioerr->oer_comp_length);
+	*p++ = cpu_to_be32(ioerr->oer_iswrite);
+	*p   = cpu_to_be32(ioerr->oer_errno);
+}
+
+__be32 *pnfs_osd_xdr_ioerr_reserve_space(struct xdr_stream *xdr)
+{
+	__be32 *p;
+
+	p = xdr_reserve_space(xdr, 32 + 24);
+	if (unlikely(!p))
+		dprintk("%s: out of xdr space\n", __func__);
+
+	return p;
+}

fs/nfs/pagelist.c

@@ -204,6 +204,21 @@ nfs_wait_on_request(struct nfs_page *req)
 			TASK_UNINTERRUPTIBLE);
 }
 
+static bool nfs_generic_pg_test(struct nfs_pageio_descriptor *desc, struct nfs_page *prev, struct nfs_page *req)
+{
+	/*
+	 * FIXME: ideally we should be able to coalesce all requests
+	 * that are not block boundary aligned, but currently this
+	 * is problematic for the case of bsize < PAGE_CACHE_SIZE,
+	 * since nfs_flush_multi and nfs_pagein_multi assume you
+	 * can have only one struct nfs_page.
+	 */
+	if (desc->pg_bsize < PAGE_SIZE)
+		return 0;
+
+	return desc->pg_count + req->wb_bytes <= desc->pg_bsize;
+}
+
 /**
  * nfs_pageio_init - initialise a page io descriptor
  * @desc: pointer to descriptor
@@ -229,6 +244,8 @@ void nfs_pageio_init(struct nfs_pageio_descriptor *desc,
 	desc->pg_ioflags = io_flags;
 	desc->pg_error = 0;
 	desc->pg_lseg = NULL;
+	desc->pg_test = nfs_generic_pg_test;
+	pnfs_pageio_init(desc, inode);
 }
 
 /**
@@ -242,29 +259,23 @@ void nfs_pageio_init(struct nfs_pageio_descriptor *desc,
  *
  * Return 'true' if this is the case, else return 'false'.
  */
-static int nfs_can_coalesce_requests(struct nfs_page *prev,
+static bool nfs_can_coalesce_requests(struct nfs_page *prev,
-				     struct nfs_page *req,
+				      struct nfs_page *req,
-				     struct nfs_pageio_descriptor *pgio)
+				      struct nfs_pageio_descriptor *pgio)
 {
 	if (req->wb_context->cred != prev->wb_context->cred)
-		return 0;
+		return false;
 	if (req->wb_lock_context->lockowner != prev->wb_lock_context->lockowner)
-		return 0;
+		return false;
 	if (req->wb_context->state != prev->wb_context->state)
-		return 0;
+		return false;
 	if (req->wb_index != (prev->wb_index + 1))
-		return 0;
+		return false;
 	if (req->wb_pgbase != 0)
-		return 0;
+		return false;
 	if (prev->wb_pgbase + prev->wb_bytes != PAGE_CACHE_SIZE)
-		return 0;
+		return false;
-	/*
+	return pgio->pg_test(pgio, prev, req);
-	 * Non-whole file layouts need to check that req is inside of
-	 * pgio->pg_lseg.
-	 */
-	if (pgio->pg_test && !pgio->pg_test(pgio, prev, req))
-		return 0;
-	return 1;
 }
 
 /**
@@ -278,31 +289,18 @@ static int nfs_can_coalesce_requests(struct nfs_page *prev,
 static int nfs_pageio_do_add_request(struct nfs_pageio_descriptor *desc,
 				     struct nfs_page *req)
 {
-	size_t newlen = req->wb_bytes;
-
 	if (desc->pg_count != 0) {
 		struct nfs_page *prev;
 
-		/*
-		 * FIXME: ideally we should be able to coalesce all requests
-		 * that are not block boundary aligned, but currently this
-		 * is problematic for the case of bsize < PAGE_CACHE_SIZE,
-		 * since nfs_flush_multi and nfs_pagein_multi assume you
-		 * can have only one struct nfs_page.
-		 */
-		if (desc->pg_bsize < PAGE_SIZE)
-			return 0;
-		newlen += desc->pg_count;
-		if (newlen > desc->pg_bsize)
-			return 0;
 		prev = nfs_list_entry(desc->pg_list.prev);
 		if (!nfs_can_coalesce_requests(prev, req, desc))
 			return 0;
-	} else
+	} else {
 		desc->pg_base = req->wb_pgbase;
+	}
 	nfs_list_remove_request(req);
 	nfs_list_add_request(req, &desc->pg_list);
-	desc->pg_count = newlen;
+	desc->pg_count += req->wb_bytes;
 	return 1;
 }
 

fs/nfs/pnfs.c

@@ -177,13 +177,28 @@ get_layout_hdr(struct pnfs_layout_hdr *lo)
 	atomic_inc(&lo->plh_refcount);
 }
 
+static struct pnfs_layout_hdr *
+pnfs_alloc_layout_hdr(struct inode *ino, gfp_t gfp_flags)
+{
+	struct pnfs_layoutdriver_type *ld = NFS_SERVER(ino)->pnfs_curr_ld;
+	return ld->alloc_layout_hdr ? ld->alloc_layout_hdr(ino, gfp_flags) :
+		kzalloc(sizeof(struct pnfs_layout_hdr), gfp_flags);
+}
+
+static void
+pnfs_free_layout_hdr(struct pnfs_layout_hdr *lo)
+{
+	struct pnfs_layoutdriver_type *ld = NFS_SERVER(lo->plh_inode)->pnfs_curr_ld;
+	return ld->alloc_layout_hdr ? ld->free_layout_hdr(lo) : kfree(lo);
+}
+
 static void
 destroy_layout_hdr(struct pnfs_layout_hdr *lo)
 {
 	dprintk("%s: freeing layout cache %p\n", __func__, lo);
 	BUG_ON(!list_empty(&lo->plh_layouts));
 	NFS_I(lo->plh_inode)->layout = NULL;
-	kfree(lo);
+	pnfs_free_layout_hdr(lo);
 }
 
 static void
@@ -228,7 +243,7 @@ put_lseg_common(struct pnfs_layout_segment *lseg)
 {
 	struct inode *inode = lseg->pls_layout->plh_inode;
 
-	BUG_ON(test_bit(NFS_LSEG_VALID, &lseg->pls_flags));
+	WARN_ON(test_bit(NFS_LSEG_VALID, &lseg->pls_flags));
 	list_del_init(&lseg->pls_list);
 	if (list_empty(&lseg->pls_layout->plh_segs)) {
 		set_bit(NFS_LAYOUT_DESTROYED, &lseg->pls_layout->plh_flags);
@@ -261,11 +276,72 @@ put_lseg(struct pnfs_layout_segment *lseg)
 }
 EXPORT_SYMBOL_GPL(put_lseg);
 
-static bool
+static inline u64
-should_free_lseg(u32 lseg_iomode, u32 recall_iomode)
+end_offset(u64 start, u64 len)
 {
-	return (recall_iomode == IOMODE_ANY ||
+	u64 end;
-		lseg_iomode == recall_iomode);
+
+	end = start + len;
+	return end >= start ? end : NFS4_MAX_UINT64;
+}
+
+/* last octet in a range */
+static inline u64
+last_byte_offset(u64 start, u64 len)
+{
+	u64 end;
+
+	BUG_ON(!len);
+	end = start + len;
+	return end > start ? end - 1 : NFS4_MAX_UINT64;
+}
+
+/*
+ * is l2 fully contained in l1?
+ *   start1                             end1
+ *   [----------------------------------)
+ *           start2           end2
+ *           [----------------)
+ */
+static inline int
+lo_seg_contained(struct pnfs_layout_range *l1,
+		 struct pnfs_layout_range *l2)
+{
+	u64 start1 = l1->offset;
+	u64 end1 = end_offset(start1, l1->length);
+	u64 start2 = l2->offset;
+	u64 end2 = end_offset(start2, l2->length);
+
+	return (start1 <= start2) && (end1 >= end2);
+}
+
+/*
+ * is l1 and l2 intersecting?
+ *   start1                             end1
+ *   [----------------------------------)
+ *                              start2           end2
+ *                              [----------------)
+ */
+static inline int
+lo_seg_intersecting(struct pnfs_layout_range *l1,
+		    struct pnfs_layout_range *l2)
+{
+	u64 start1 = l1->offset;
+	u64 end1 = end_offset(start1, l1->length);
+	u64 start2 = l2->offset;
+	u64 end2 = end_offset(start2, l2->length);
+
+	return (end1 == NFS4_MAX_UINT64 || end1 > start2) &&
+	       (end2 == NFS4_MAX_UINT64 || end2 > start1);
+}
+
+static bool
+should_free_lseg(struct pnfs_layout_range *lseg_range,
+		 struct pnfs_layout_range *recall_range)
+{
+	return (recall_range->iomode == IOMODE_ANY ||
+		lseg_range->iomode == recall_range->iomode) &&
+	       lo_seg_intersecting(lseg_range, recall_range);
 }
 
 /* Returns 1 if lseg is removed from list, 0 otherwise */
@@ -296,7 +372,7 @@ static int mark_lseg_invalid(struct pnfs_layout_segment *lseg,
 int
 mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo,
 			    struct list_head *tmp_list,
-			    u32 iomode)
+			    struct pnfs_layout_range *recall_range)
 {
 	struct pnfs_layout_segment *lseg, *next;
 	int invalid = 0, removed = 0;
@@ -309,7 +385,8 @@ mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo,
 		return 0;
 	}
 	list_for_each_entry_safe(lseg, next, &lo->plh_segs, pls_list)
-		if (should_free_lseg(lseg->pls_range.iomode, iomode)) {
+		if (!recall_range ||
+		    should_free_lseg(&lseg->pls_range, recall_range)) {
 			dprintk("%s: freeing lseg %p iomode %d "
 				"offset %llu length %llu\n", __func__,
 				lseg, lseg->pls_range.iomode, lseg->pls_range.offset,
@@ -358,7 +435,7 @@ pnfs_destroy_layout(struct nfs_inode *nfsi)
 	lo = nfsi->layout;
 	if (lo) {
 		lo->plh_block_lgets++; /* permanently block new LAYOUTGETs */
-		mark_matching_lsegs_invalid(lo, &tmp_list, IOMODE_ANY);
+		mark_matching_lsegs_invalid(lo, &tmp_list, NULL);
 	}
 	spin_unlock(&nfsi->vfs_inode.i_lock);
 	pnfs_free_lseg_list(&tmp_list);
@@ -467,7 +544,7 @@ pnfs_choose_layoutget_stateid(nfs4_stateid *dst, struct pnfs_layout_hdr *lo,
 static struct pnfs_layout_segment *
 send_layoutget(struct pnfs_layout_hdr *lo,
 	   struct nfs_open_context *ctx,
-	   u32 iomode,
+	   struct pnfs_layout_range *range,
 	   gfp_t gfp_flags)
 {
 	struct inode *ino = lo->plh_inode;
@@ -499,11 +576,11 @@ send_layoutget(struct pnfs_layout_hdr *lo,
 			goto out_err_free;
 	}
 
-	lgp->args.minlength = NFS4_MAX_UINT64;
+	lgp->args.minlength = PAGE_CACHE_SIZE;
+	if (lgp->args.minlength > range->length)
+		lgp->args.minlength = range->length;
 	lgp->args.maxcount = PNFS_LAYOUT_MAXSIZE;
-	lgp->args.range.iomode = iomode;
+	lgp->args.range = *range;
-	lgp->args.range.offset = 0;
-	lgp->args.range.length = NFS4_MAX_UINT64;
 	lgp->args.type = server->pnfs_curr_ld->id;
 	lgp->args.inode = ino;
 	lgp->args.ctx = get_nfs_open_context(ctx);
@@ -518,7 +595,7 @@ send_layoutget(struct pnfs_layout_hdr *lo,
 	nfs4_proc_layoutget(lgp);
 	if (!lseg) {
 		/* remember that LAYOUTGET failed and suspend trying */
-		set_bit(lo_fail_bit(iomode), &lo->plh_flags);
+		set_bit(lo_fail_bit(range->iomode), &lo->plh_flags);
 	}
 
 	/* free xdr pages */
@@ -542,6 +619,51 @@ out_err_free:
 	return NULL;
 }
 
+/* Initiates a LAYOUTRETURN(FILE) */
+int
+_pnfs_return_layout(struct inode *ino)
+{
+	struct pnfs_layout_hdr *lo = NULL;
+	struct nfs_inode *nfsi = NFS_I(ino);
+	LIST_HEAD(tmp_list);
+	struct nfs4_layoutreturn *lrp;
+	nfs4_stateid stateid;
+	int status = 0;
+
+	dprintk("--> %s\n", __func__);
+
+	spin_lock(&ino->i_lock);
+	lo = nfsi->layout;
+	if (!lo || !mark_matching_lsegs_invalid(lo, &tmp_list, NULL)) {
+		spin_unlock(&ino->i_lock);
+		dprintk("%s: no layout segments to return\n", __func__);
+		goto out;
+	}
+	stateid = nfsi->layout->plh_stateid;
+	/* Reference matched in nfs4_layoutreturn_release */
+	get_layout_hdr(lo);
+	spin_unlock(&ino->i_lock);
+	pnfs_free_lseg_list(&tmp_list);
+
+	WARN_ON(test_bit(NFS_INO_LAYOUTCOMMIT, &nfsi->flags));
+
+	lrp = kzalloc(sizeof(*lrp), GFP_KERNEL);
+	if (unlikely(lrp == NULL)) {
+		status = -ENOMEM;
+		goto out;
+	}
+
+	lrp->args.stateid = stateid;
+	lrp->args.layout_type = NFS_SERVER(ino)->pnfs_curr_ld->id;
+	lrp->args.inode = ino;
+	lrp->clp = NFS_SERVER(ino)->nfs_client;
+
+	status = nfs4_proc_layoutreturn(lrp);
+out:
+	dprintk("<-- %s status: %d\n", __func__, status);
+	return status;
+}
+
 bool pnfs_roc(struct inode *ino)
 {
 	struct pnfs_layout_hdr *lo;
@@ -625,10 +747,23 @@ bool pnfs_roc_drain(struct inode *ino, u32 *barrier)
  * are seen first.
  */
 static s64
-cmp_layout(u32 iomode1, u32 iomode2)
+cmp_layout(struct pnfs_layout_range *l1,
+	   struct pnfs_layout_range *l2)
 {
+	s64 d;
+
+	/* high offset > low offset */
+	d = l1->offset - l2->offset;
+	if (d)
+		return d;
+
+	/* short length > long length */
+	d = l2->length - l1->length;
+	if (d)
+		return d;
+
 	/* read > read/write */
-	return (int)(iomode2 == IOMODE_READ) - (int)(iomode1 == IOMODE_READ);
+	return (int)(l1->iomode == IOMODE_READ) - (int)(l2->iomode == IOMODE_READ);
 }
 
 static void
@@ -636,13 +771,12 @@ pnfs_insert_layout(struct pnfs_layout_hdr *lo,
 		   struct pnfs_layout_segment *lseg)
 {
 	struct pnfs_layout_segment *lp;
-	int found = 0;
 
 	dprintk("%s:Begin\n", __func__);
 
 	assert_spin_locked(&lo->plh_inode->i_lock);
 	list_for_each_entry(lp, &lo->plh_segs, pls_list) {
-		if (cmp_layout(lp->pls_range.iomode, lseg->pls_range.iomode) > 0)
+		if (cmp_layout(&lseg->pls_range, &lp->pls_range) > 0)
 			continue;
 		list_add_tail(&lseg->pls_list, &lp->pls_list);
 		dprintk("%s: inserted lseg %p "
@@ -652,16 +786,14 @@ pnfs_insert_layout(struct pnfs_layout_hdr *lo,
 			lseg->pls_range.offset, lseg->pls_range.length,
 			lp, lp->pls_range.iomode, lp->pls_range.offset,
 			lp->pls_range.length);
-		found = 1;
+		goto out;
-		break;
-	}
-	if (!found) {
-		list_add_tail(&lseg->pls_list, &lo->plh_segs);
-		dprintk("%s: inserted lseg %p "
-			"iomode %d offset %llu length %llu at tail\n",
-			__func__, lseg, lseg->pls_range.iomode,
-			lseg->pls_range.offset, lseg->pls_range.length);
 	}
+	list_add_tail(&lseg->pls_list, &lo->plh_segs);
+	dprintk("%s: inserted lseg %p "
+		"iomode %d offset %llu length %llu at tail\n",
+		__func__, lseg, lseg->pls_range.iomode,
+		lseg->pls_range.offset, lseg->pls_range.length);
+out:
 	get_layout_hdr(lo);
 
 	dprintk("%s:Return\n", __func__);
@@ -672,7 +804,7 @@ alloc_init_layout_hdr(struct inode *ino, gfp_t gfp_flags)
 {
 	struct pnfs_layout_hdr *lo;
 
-	lo = kzalloc(sizeof(struct pnfs_layout_hdr), gfp_flags);
+	lo = pnfs_alloc_layout_hdr(ino, gfp_flags);
 	if (!lo)
 		return NULL;
 	atomic_set(&lo->plh_refcount, 1);
@@ -705,7 +837,7 @@ pnfs_find_alloc_layout(struct inode *ino, gfp_t gfp_flags)
 	if (likely(nfsi->layout == NULL))	/* Won the race? */
 		nfsi->layout = new;
 	else
-		kfree(new);
+		pnfs_free_layout_hdr(new);
 	return nfsi->layout;
 }
 
@@ -721,16 +853,28 @@ pnfs_find_alloc_layout(struct inode *ino, gfp_t gfp_flags)
  * READ		RW	true
  */
 static int
-is_matching_lseg(struct pnfs_layout_segment *lseg, u32 iomode)
+is_matching_lseg(struct pnfs_layout_range *ls_range,
+		 struct pnfs_layout_range *range)
 {
-	return (iomode != IOMODE_RW || lseg->pls_range.iomode == IOMODE_RW);
+	struct pnfs_layout_range range1;
+
+	if ((range->iomode == IOMODE_RW &&
+	     ls_range->iomode != IOMODE_RW) ||
+	    !lo_seg_intersecting(ls_range, range))
+		return 0;
+
+	/* range1 covers only the first byte in the range */
+	range1 = *range;
+	range1.length = 1;
+	return lo_seg_contained(ls_range, &range1);
 }
 
 /*
  * lookup range in layout
  */
 static struct pnfs_layout_segment *
-pnfs_find_lseg(struct pnfs_layout_hdr *lo, u32 iomode)
+pnfs_find_lseg(struct pnfs_layout_hdr *lo,
+		struct pnfs_layout_range *range)
 {
 	struct pnfs_layout_segment *lseg, *ret = NULL;
 
@@ -739,11 +883,11 @@ pnfs_find_lseg(struct pnfs_layout_hdr *lo, u32 iomode)
 	assert_spin_locked(&lo->plh_inode->i_lock);
 	list_for_each_entry(lseg, &lo->plh_segs, pls_list) {
 		if (test_bit(NFS_LSEG_VALID, &lseg->pls_flags) &&
-		    is_matching_lseg(lseg, iomode)) {
+		    is_matching_lseg(&lseg->pls_range, range)) {
 			ret = get_lseg(lseg);
 			break;
 		}
-		if (cmp_layout(iomode, lseg->pls_range.iomode) > 0)
+		if (cmp_layout(range, &lseg->pls_range) > 0)
 			break;
 	}
 
@@ -759,9 +903,17 @@ pnfs_find_lseg(struct pnfs_layout_hdr *lo, u32 iomode)
 struct pnfs_layout_segment *
 pnfs_update_layout(struct inode *ino,
 		   struct nfs_open_context *ctx,
+		   loff_t pos,
+		   u64 count,
 		   enum pnfs_iomode iomode,
 		   gfp_t gfp_flags)
 {
+	struct pnfs_layout_range arg = {
+		.iomode = iomode,
+		.offset = pos,
+		.length = count,
+	};
+	unsigned pg_offset;
 	struct nfs_inode *nfsi = NFS_I(ino);
 	struct nfs_client *clp = NFS_SERVER(ino)->nfs_client;
 	struct pnfs_layout_hdr *lo;
@@ -789,7 +941,7 @@ pnfs_update_layout(struct inode *ino,
 		goto out_unlock;
 
 	/* Check to see if the layout for the given range already exists */
-	lseg = pnfs_find_lseg(lo, iomode);
+	lseg = pnfs_find_lseg(lo, &arg);
 	if (lseg)
 		goto out_unlock;
 
@@ -811,7 +963,14 @@ pnfs_update_layout(struct inode *ino,
 		spin_unlock(&clp->cl_lock);
 	}
 
-	lseg = send_layoutget(lo, ctx, iomode, gfp_flags);
+	pg_offset = arg.offset & ~PAGE_CACHE_MASK;
+	if (pg_offset) {
+		arg.offset -= pg_offset;
+		arg.length += pg_offset;
+	}
+	arg.length = PAGE_CACHE_ALIGN(arg.length);
+
+	lseg = send_layoutget(lo, ctx, &arg, gfp_flags);
 	if (!lseg && first) {
 		spin_lock(&clp->cl_lock);
 		list_del_init(&lo->plh_layouts);
@@ -838,17 +997,6 @@ pnfs_layout_process(struct nfs4_layoutget *lgp)
 	struct nfs_client *clp = NFS_SERVER(ino)->nfs_client;
 	int status = 0;
 
-	/* Verify we got what we asked for.
-	 * Note that because the xdr parsing only accepts a single
-	 * element array, this can fail even if the server is behaving
-	 * correctly.
-	 */
-	if (lgp->args.range.iomode > res->range.iomode ||
-	    res->range.offset != 0 ||
-	    res->range.length != NFS4_MAX_UINT64) {
-		status = -EINVAL;
-		goto out;
-	}
 	/* Inject layout blob into I/O device driver */
 	lseg = NFS_SERVER(ino)->pnfs_curr_ld->alloc_lseg(lo, res, lgp->gfp_flags);
 	if (!lseg || IS_ERR(lseg)) {
@@ -895,51 +1043,64 @@ out_forget_reply:
 	goto out;
 }
 
-static int pnfs_read_pg_test(struct nfs_pageio_descriptor *pgio,
+bool
-			     struct nfs_page *prev,
+pnfs_generic_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev,
-			     struct nfs_page *req)
+		     struct nfs_page *req)
 {
+	enum pnfs_iomode access_type;
+	gfp_t gfp_flags;
+
+	/* We assume that pg_ioflags == 0 iff we're reading a page */
+	if (pgio->pg_ioflags == 0) {
+		access_type = IOMODE_READ;
+		gfp_flags = GFP_KERNEL;
+	} else {
+		access_type = IOMODE_RW;
+		gfp_flags = GFP_NOFS;
+	}
+
 	if (pgio->pg_count == prev->wb_bytes) {
 		/* This is first coelesce call for a series of nfs_pages */
 		pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
 						   prev->wb_context,
-						   IOMODE_READ,
+						   req_offset(req),
-						   GFP_KERNEL);
+						   pgio->pg_count,
+						   access_type,
+						   gfp_flags);
+		return true;
 	}
-	return NFS_SERVER(pgio->pg_inode)->pnfs_curr_ld->pg_test(pgio, prev, req);
+
+	if (pgio->pg_lseg &&
+	    req_offset(req) > end_offset(pgio->pg_lseg->pls_range.offset,
+					 pgio->pg_lseg->pls_range.length))
+		return false;
+
+	return true;
 }
+EXPORT_SYMBOL_GPL(pnfs_generic_pg_test);
 
-void
+/*
-pnfs_pageio_init_read(struct nfs_pageio_descriptor *pgio, struct inode *inode)
+ * Called by non rpc-based layout drivers
+ */
+int
+pnfs_ld_write_done(struct nfs_write_data *data)
 {
-	struct pnfs_layoutdriver_type *ld;
+	int status;
 
-	ld = NFS_SERVER(inode)->pnfs_curr_ld;
+	if (!data->pnfs_error) {
-	pgio->pg_test = (ld && ld->pg_test) ? pnfs_read_pg_test : NULL;
+		pnfs_set_layoutcommit(data);
-}
+		data->mds_ops->rpc_call_done(&data->task, data);
-
+		data->mds_ops->rpc_release(data);
-static int pnfs_write_pg_test(struct nfs_pageio_descriptor *pgio,
+		return 0;
-			      struct nfs_page *prev,
-			      struct nfs_page *req)
-{
-	if (pgio->pg_count == prev->wb_bytes) {
-		/* This is first coelesce call for a series of nfs_pages */
-		pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
-						   prev->wb_context,
-						   IOMODE_RW,
-						   GFP_NOFS);
 	}
-	return NFS_SERVER(pgio->pg_inode)->pnfs_curr_ld->pg_test(pgio, prev, req);
-}
 
-void
+	dprintk("%s: pnfs_error=%d, retry via MDS\n", __func__,
-pnfs_pageio_init_write(struct nfs_pageio_descriptor *pgio, struct inode *inode)
+		data->pnfs_error);
-{
+	status = nfs_initiate_write(data, NFS_CLIENT(data->inode),
-	struct pnfs_layoutdriver_type *ld;
+				    data->mds_ops, NFS_FILE_SYNC);
-
+	return status ? : -EAGAIN;
-	ld = NFS_SERVER(inode)->pnfs_curr_ld;
-	pgio->pg_test = (ld && ld->pg_test) ? pnfs_write_pg_test : NULL;
 }
+EXPORT_SYMBOL_GPL(pnfs_ld_write_done);
 
 enum pnfs_try_status
 pnfs_try_to_write_data(struct nfs_write_data *wdata,
@@ -965,6 +1126,29 @@ pnfs_try_to_write_data(struct nfs_write_data *wdata,
 	return trypnfs;
 }
 
+/*
+ * Called by non rpc-based layout drivers
+ */
+int
+pnfs_ld_read_done(struct nfs_read_data *data)
+{
+	int status;
+
+	if (!data->pnfs_error) {
+		__nfs4_read_done_cb(data);
+		data->mds_ops->rpc_call_done(&data->task, data);
+		data->mds_ops->rpc_release(data);
+		return 0;
+	}
+
+	dprintk("%s: pnfs_error=%d, retry via MDS\n", __func__,
+		data->pnfs_error);
+	status = nfs_initiate_read(data, NFS_CLIENT(data->inode),
+				   data->mds_ops);
+	return status ? : -EAGAIN;
+}
+EXPORT_SYMBOL_GPL(pnfs_ld_read_done);
+
 /*
  * Call the appropriate parallel I/O subsystem read function.
  */

fs/nfs/pnfs.h

@@ -30,6 +30,7 @@
 #ifndef FS_NFS_PNFS_H
 #define FS_NFS_PNFS_H
 
+#include <linux/nfs_fs.h>
 #include <linux/nfs_page.h>
 
 enum {
@@ -64,17 +65,29 @@ enum {
 	NFS_LAYOUT_DESTROYED,		/* no new use of layout allowed */
 };
 
+enum layoutdriver_policy_flags {
+	/* Should the pNFS client commit and return the layout upon a setattr */
+	PNFS_LAYOUTRET_ON_SETATTR	= 1 << 0,
+};
+
+struct nfs4_deviceid_node;
+
 /* Per-layout driver specific registration structure */
 struct pnfs_layoutdriver_type {
 	struct list_head pnfs_tblid;
 	const u32 id;
 	const char *name;
 	struct module *owner;
+	unsigned flags;
+
+	struct pnfs_layout_hdr * (*alloc_layout_hdr) (struct inode *inode, gfp_t gfp_flags);
+	void (*free_layout_hdr) (struct pnfs_layout_hdr *);
+
 	struct pnfs_layout_segment * (*alloc_lseg) (struct pnfs_layout_hdr *layoutid, struct nfs4_layoutget_res *lgr, gfp_t gfp_flags);
 	void (*free_lseg) (struct pnfs_layout_segment *lseg);
 
 	/* test for nfs page cache coalescing */
-	int (*pg_test)(struct nfs_pageio_descriptor *, struct nfs_page *, struct nfs_page *);
+	bool (*pg_test)(struct nfs_pageio_descriptor *, struct nfs_page *, struct nfs_page *);
 
 	/* Returns true if layoutdriver wants to divert this request to
 	 * driver's commit routine.
@@ -89,6 +102,16 @@ struct pnfs_layoutdriver_type {
 	 */
 	enum pnfs_try_status (*read_pagelist) (struct nfs_read_data *nfs_data);
 	enum pnfs_try_status (*write_pagelist) (struct nfs_write_data *nfs_data, int how);
+
+	void (*free_deviceid_node) (struct nfs4_deviceid_node *);
+
+	void (*encode_layoutreturn) (struct pnfs_layout_hdr *layoutid,
+				     struct xdr_stream *xdr,
+				     const struct nfs4_layoutreturn_args *args);
+
+	void (*encode_layoutcommit) (struct pnfs_layout_hdr *layoutid,
+				     struct xdr_stream *xdr,
+				     const struct nfs4_layoutcommit_args *args);
 };
 
 struct pnfs_layout_hdr {
@@ -120,21 +143,22 @@ extern void pnfs_unregister_layoutdriver(struct pnfs_layoutdriver_type *);
 extern int nfs4_proc_getdeviceinfo(struct nfs_server *server,
 				   struct pnfs_device *dev);
 extern int nfs4_proc_layoutget(struct nfs4_layoutget *lgp);
+extern int nfs4_proc_layoutreturn(struct nfs4_layoutreturn *lrp);
 
 /* pnfs.c */
 void get_layout_hdr(struct pnfs_layout_hdr *lo);
 void put_lseg(struct pnfs_layout_segment *lseg);
 struct pnfs_layout_segment *
 pnfs_update_layout(struct inode *ino, struct nfs_open_context *ctx,
-		   enum pnfs_iomode access_type, gfp_t gfp_flags);
+		   loff_t pos, u64 count, enum pnfs_iomode access_type,
+		   gfp_t gfp_flags);
 void set_pnfs_layoutdriver(struct nfs_server *, u32 id);
 void unset_pnfs_layoutdriver(struct nfs_server *);
 enum pnfs_try_status pnfs_try_to_write_data(struct nfs_write_data *,
 					     const struct rpc_call_ops *, int);
 enum pnfs_try_status pnfs_try_to_read_data(struct nfs_read_data *,
 					    const struct rpc_call_ops *);
-void pnfs_pageio_init_read(struct nfs_pageio_descriptor *, struct inode *);
+bool pnfs_generic_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev, struct nfs_page *req);
-void pnfs_pageio_init_write(struct nfs_pageio_descriptor *, struct inode *);
 int pnfs_layout_process(struct nfs4_layoutget *lgp);
 void pnfs_free_lseg_list(struct list_head *tmp_list);
 void pnfs_destroy_layout(struct nfs_inode *);
@@ -148,13 +172,37 @@ int pnfs_choose_layoutget_stateid(nfs4_stateid *dst,
 				  struct nfs4_state *open_state);
 int mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo,
 				struct list_head *tmp_list,
-				u32 iomode);
+				struct pnfs_layout_range *recall_range);
 bool pnfs_roc(struct inode *ino);
 void pnfs_roc_release(struct inode *ino);
 void pnfs_roc_set_barrier(struct inode *ino, u32 barrier);
 bool pnfs_roc_drain(struct inode *ino, u32 *barrier);
 void pnfs_set_layoutcommit(struct nfs_write_data *wdata);
 int pnfs_layoutcommit_inode(struct inode *inode, bool sync);
+int _pnfs_return_layout(struct inode *);
+int pnfs_ld_write_done(struct nfs_write_data *);
+int pnfs_ld_read_done(struct nfs_read_data *);
+
+/* pnfs_dev.c */
+struct nfs4_deviceid_node {
+	struct hlist_node		node;
+	const struct pnfs_layoutdriver_type *ld;
+	const struct nfs_client		*nfs_client;
+	struct nfs4_deviceid		deviceid;
+	atomic_t			ref;
+};
+
+void nfs4_print_deviceid(const struct nfs4_deviceid *dev_id);
+struct nfs4_deviceid_node *nfs4_find_get_deviceid(const struct pnfs_layoutdriver_type *, const struct nfs_client *, const struct nfs4_deviceid *);
+struct nfs4_deviceid_node *nfs4_unhash_put_deviceid(const struct pnfs_layoutdriver_type *, const struct nfs_client *, const struct nfs4_deviceid *);
+void nfs4_delete_deviceid(const struct pnfs_layoutdriver_type *, const struct nfs_client *, const struct nfs4_deviceid *);
+void nfs4_init_deviceid_node(struct nfs4_deviceid_node *,
+			     const struct pnfs_layoutdriver_type *,
+			     const struct nfs_client *,
+			     const struct nfs4_deviceid *);
+struct nfs4_deviceid_node *nfs4_insert_deviceid_node(struct nfs4_deviceid_node *);
+bool nfs4_put_deviceid_node(struct nfs4_deviceid_node *);
+void nfs4_deviceid_purge_client(const struct nfs_client *);
 
 static inline int lo_fail_bit(u32 iomode)
 {
@@ -223,6 +271,36 @@ static inline void pnfs_clear_request_commit(struct nfs_page *req)
 		put_lseg(req->wb_commit_lseg);
 }
 
+/* Should the pNFS client commit and return the layout upon a setattr */
+static inline bool
+pnfs_ld_layoutret_on_setattr(struct inode *inode)
+{
+	if (!pnfs_enabled_sb(NFS_SERVER(inode)))
+		return false;
+	return NFS_SERVER(inode)->pnfs_curr_ld->flags &
+		PNFS_LAYOUTRET_ON_SETATTR;
+}
+
+static inline int pnfs_return_layout(struct inode *ino)
+{
+	struct nfs_inode *nfsi = NFS_I(ino);
+	struct nfs_server *nfss = NFS_SERVER(ino);
+
+	if (pnfs_enabled_sb(nfss) && nfsi->layout)
+		return _pnfs_return_layout(ino);
+
+	return 0;
+}
+
+static inline void pnfs_pageio_init(struct nfs_pageio_descriptor *pgio,
+				    struct inode *inode)
+{
+	struct pnfs_layoutdriver_type *ld = NFS_SERVER(inode)->pnfs_curr_ld;
+
+	if (ld)
+		pgio->pg_test = ld->pg_test;
+}
+
 #else  /* CONFIG_NFS_V4_1 */
 
 static inline void pnfs_destroy_all_layouts(struct nfs_client *clp)
@@ -245,7 +323,8 @@ static inline void put_lseg(struct pnfs_layout_segment *lseg)
 
 static inline struct pnfs_layout_segment *
 pnfs_update_layout(struct inode *ino, struct nfs_open_context *ctx,
-		   enum pnfs_iomode access_type, gfp_t gfp_flags)
+		   loff_t pos, u64 count, enum pnfs_iomode access_type,
+		   gfp_t gfp_flags)
 {
 	return NULL;
 }
@@ -264,6 +343,17 @@ pnfs_try_to_write_data(struct nfs_write_data *data,
 	return PNFS_NOT_ATTEMPTED;
 }
 
+static inline int pnfs_return_layout(struct inode *ino)
+{
+	return 0;
+}
+
+static inline bool
+pnfs_ld_layoutret_on_setattr(struct inode *inode)
+{
+	return false;
+}
+
 static inline bool
 pnfs_roc(struct inode *ino)
 {
@@ -294,16 +384,9 @@ static inline void unset_pnfs_layoutdriver(struct nfs_server *s)
 {
 }
 
-static inline void
+static inline void pnfs_pageio_init(struct nfs_pageio_descriptor *pgio,
-pnfs_pageio_init_read(struct nfs_pageio_descriptor *pgio, struct inode *ino)
+				    struct inode *inode)
 {
-	pgio->pg_test = NULL;
-}
-
-static inline void
-pnfs_pageio_init_write(struct nfs_pageio_descriptor *pgio, struct inode *ino)
-{
-	pgio->pg_test = NULL;
 }
 
 static inline void
@@ -331,6 +414,10 @@ static inline int pnfs_layoutcommit_inode(struct inode *inode, bool sync)
 {
 	return 0;
 }
+
+static inline void nfs4_deviceid_purge_client(struct nfs_client *ncl)
+{
+}
 #endif /* CONFIG_NFS_V4_1 */
 
 #endif /* FS_NFS_PNFS_H */

fs/nfs/pnfs_dev.c

@@ -0,0 +1,270 @@
+/*
+ *  Device operations for the pnfs client.
+ *
+ *  Copyright (c) 2002
+ *  The Regents of the University of Michigan
+ *  All Rights Reserved
+ *
+ *  Dean Hildebrand <dhildebz@umich.edu>
+ *  Garth Goodson   <Garth.Goodson@netapp.com>
+ *
+ *  Permission is granted to use, copy, create derivative works, and
+ *  redistribute this software and such derivative works for any purpose,
+ *  so long as the name of the University of Michigan is not used in
+ *  any advertising or publicity pertaining to the use or distribution
+ *  of this software without specific, written prior authorization. If
+ *  the above copyright notice or any other identification of the
+ *  University of Michigan is included in any copy of any portion of
+ *  this software, then the disclaimer below must also be included.
+ *
+ *  This software is provided as is, without representation or warranty
+ *  of any kind either express or implied, including without limitation
+ *  the implied warranties of merchantability, fitness for a particular
+ *  purpose, or noninfringement.  The Regents of the University of
+ *  Michigan shall not be liable for any damages, including special,
+ *  indirect, incidental, or consequential damages, with respect to any
+ *  claim arising out of or in connection with the use of the software,
+ *  even if it has been or is hereafter advised of the possibility of
+ *  such damages.
+ */
+
+#include "pnfs.h"
+
+#define NFSDBG_FACILITY		NFSDBG_PNFS
+
+/*
+ * Device ID RCU cache. A device ID is unique per server and layout type.
+ */
+#define NFS4_DEVICE_ID_HASH_BITS	5
+#define NFS4_DEVICE_ID_HASH_SIZE	(1 << NFS4_DEVICE_ID_HASH_BITS)
+#define NFS4_DEVICE_ID_HASH_MASK	(NFS4_DEVICE_ID_HASH_SIZE - 1)
+
+static struct hlist_head nfs4_deviceid_cache[NFS4_DEVICE_ID_HASH_SIZE];
+static DEFINE_SPINLOCK(nfs4_deviceid_lock);
+
+void
+nfs4_print_deviceid(const struct nfs4_deviceid *id)
+{
+	u32 *p = (u32 *)id;
+
+	dprintk("%s: device id= [%x%x%x%x]\n", __func__,
+		p[0], p[1], p[2], p[3]);
+}
+EXPORT_SYMBOL_GPL(nfs4_print_deviceid);
+
+static inline u32
+nfs4_deviceid_hash(const struct nfs4_deviceid *id)
+{
+	unsigned char *cptr = (unsigned char *)id->data;
+	unsigned int nbytes = NFS4_DEVICEID4_SIZE;
+	u32 x = 0;
+
+	while (nbytes--) {
+		x *= 37;
+		x += *cptr++;
+	}
+	return x & NFS4_DEVICE_ID_HASH_MASK;
+}
+
+static struct nfs4_deviceid_node *
+_lookup_deviceid(const struct pnfs_layoutdriver_type *ld,
+		 const struct nfs_client *clp, const struct nfs4_deviceid *id,
+		 long hash)
+{
+	struct nfs4_deviceid_node *d;
+	struct hlist_node *n;
+
+	hlist_for_each_entry_rcu(d, n, &nfs4_deviceid_cache[hash], node)
+		if (d->ld == ld && d->nfs_client == clp &&
+		    !memcmp(&d->deviceid, id, sizeof(*id))) {
+			if (atomic_read(&d->ref))
+				return d;
+			else
+				continue;
+		}
+	return NULL;
+}
+
+/*
+ * Lookup a deviceid in cache and get a reference count on it if found
+ *
+ * @clp nfs_client associated with deviceid
+ * @id deviceid to look up
+ */
+struct nfs4_deviceid_node *
+_find_get_deviceid(const struct pnfs_layoutdriver_type *ld,
+		   const struct nfs_client *clp, const struct nfs4_deviceid *id,
+		   long hash)
+{
+	struct nfs4_deviceid_node *d;
+
+	rcu_read_lock();
+	d = _lookup_deviceid(ld, clp, id, hash);
+	if (d && !atomic_inc_not_zero(&d->ref))
+		d = NULL;
+	rcu_read_unlock();
+	return d;
+}
+
+struct nfs4_deviceid_node *
+nfs4_find_get_deviceid(const struct pnfs_layoutdriver_type *ld,
+		       const struct nfs_client *clp, const struct nfs4_deviceid *id)
+{
+	return _find_get_deviceid(ld, clp, id, nfs4_deviceid_hash(id));
+}
+EXPORT_SYMBOL_GPL(nfs4_find_get_deviceid);
+
+/*
+ * Unhash and put deviceid
+ *
+ * @clp nfs_client associated with deviceid
+ * @id the deviceid to unhash
+ *
+ * @ret the unhashed node, if found and dereferenced to zero, NULL otherwise.
+ */
+struct nfs4_deviceid_node *
+nfs4_unhash_put_deviceid(const struct pnfs_layoutdriver_type *ld,
+			 const struct nfs_client *clp, const struct nfs4_deviceid *id)
+{
+	struct nfs4_deviceid_node *d;
+
+	spin_lock(&nfs4_deviceid_lock);
+	rcu_read_lock();
+	d = _lookup_deviceid(ld, clp, id, nfs4_deviceid_hash(id));
+	rcu_read_unlock();
+	if (!d) {
+		spin_unlock(&nfs4_deviceid_lock);
+		return NULL;
+	}
+	hlist_del_init_rcu(&d->node);
+	spin_unlock(&nfs4_deviceid_lock);
+	synchronize_rcu();
+
+	/* balance the initial ref set in pnfs_insert_deviceid */
+	if (atomic_dec_and_test(&d->ref))
+		return d;
+
+	return NULL;
+}
+EXPORT_SYMBOL_GPL(nfs4_unhash_put_deviceid);
+
+/*
+ * Delete a deviceid from cache
+ *
+ * @clp struct nfs_client qualifying the deviceid
+ * @id deviceid to delete
+ */
+void
+nfs4_delete_deviceid(const struct pnfs_layoutdriver_type *ld,
+		     const struct nfs_client *clp, const struct nfs4_deviceid *id)
+{
+	struct nfs4_deviceid_node *d;
+
+	d = nfs4_unhash_put_deviceid(ld, clp, id);
+	if (!d)
+		return;
+	d->ld->free_deviceid_node(d);
+}
+EXPORT_SYMBOL_GPL(nfs4_delete_deviceid);
+
+void
+nfs4_init_deviceid_node(struct nfs4_deviceid_node *d,
+			const struct pnfs_layoutdriver_type *ld,
+			const struct nfs_client *nfs_client,
+			const struct nfs4_deviceid *id)
+{
+	INIT_HLIST_NODE(&d->node);
+	d->ld = ld;
+	d->nfs_client = nfs_client;
+	d->deviceid = *id;
+	atomic_set(&d->ref, 1);
+}
+EXPORT_SYMBOL_GPL(nfs4_init_deviceid_node);
+
+/*
+ * Uniquely initialize and insert a deviceid node into cache
+ *
+ * @new new deviceid node
+ *      Note that the caller must set up the following members:
+ *        new->ld
+ *        new->nfs_client
+ *        new->deviceid
+ *
+ * @ret the inserted node, if none found, otherwise, the found entry.
+ */
+struct nfs4_deviceid_node *
+nfs4_insert_deviceid_node(struct nfs4_deviceid_node *new)
+{
+	struct nfs4_deviceid_node *d;
+	long hash;
+
+	spin_lock(&nfs4_deviceid_lock);
+	hash = nfs4_deviceid_hash(&new->deviceid);
+	d = _find_get_deviceid(new->ld, new->nfs_client, &new->deviceid, hash);
+	if (d) {
+		spin_unlock(&nfs4_deviceid_lock);
+		return d;
+	}
+
+	hlist_add_head_rcu(&new->node, &nfs4_deviceid_cache[hash]);
+	spin_unlock(&nfs4_deviceid_lock);
+
+	return new;
+}
+EXPORT_SYMBOL_GPL(nfs4_insert_deviceid_node);
+
+/*
+ * Dereference a deviceid node and delete it when its reference count drops
+ * to zero.
+ *
+ * @d deviceid node to put
+ *
+ * @ret true iff the node was deleted
+ */
+bool
+nfs4_put_deviceid_node(struct nfs4_deviceid_node *d)
+{
+	if (!atomic_dec_and_lock(&d->ref, &nfs4_deviceid_lock))
+		return false;
+	hlist_del_init_rcu(&d->node);
+	spin_unlock(&nfs4_deviceid_lock);
+	synchronize_rcu();
+	d->ld->free_deviceid_node(d);
+	return true;
+}
+EXPORT_SYMBOL_GPL(nfs4_put_deviceid_node);
+
+static void
+_deviceid_purge_client(const struct nfs_client *clp, long hash)
+{
+	struct nfs4_deviceid_node *d;
+	struct hlist_node *n, *next;
+	HLIST_HEAD(tmp);
+
+	rcu_read_lock();
+	hlist_for_each_entry_rcu(d, n, &nfs4_deviceid_cache[hash], node)
+		if (d->nfs_client == clp && atomic_read(&d->ref)) {
+			hlist_del_init_rcu(&d->node);
+			hlist_add_head(&d->node, &tmp);
+		}
+	rcu_read_unlock();
+
+	if (hlist_empty(&tmp))
+		return;
+
+	synchronize_rcu();
+	hlist_for_each_entry_safe(d, n, next, &tmp, node)
+		if (atomic_dec_and_test(&d->ref))
+			d->ld->free_deviceid_node(d);
+}
+
+void
+nfs4_deviceid_purge_client(const struct nfs_client *clp)
+{
+	long h;
+
+	spin_lock(&nfs4_deviceid_lock);
+	for (h = 0; h < NFS4_DEVICE_ID_HASH_SIZE; h++)
+		_deviceid_purge_client(clp, h);
+	spin_unlock(&nfs4_deviceid_lock);
+}

fs/nfs/read.c

@@ -288,7 +288,9 @@ static int nfs_pagein_multi(struct nfs_pageio_descriptor *desc)
 	atomic_set(&req->wb_complete, requests);
 
 	BUG_ON(desc->pg_lseg != NULL);
-	lseg = pnfs_update_layout(desc->pg_inode, req->wb_context, IOMODE_READ, GFP_KERNEL);
+	lseg = pnfs_update_layout(desc->pg_inode, req->wb_context,
+				  req_offset(req), desc->pg_count,
+				  IOMODE_READ, GFP_KERNEL);
 	ClearPageError(page);
 	offset = 0;
 	nbytes = desc->pg_count;
@@ -351,7 +353,9 @@ static int nfs_pagein_one(struct nfs_pageio_descriptor *desc)
 	}
 	req = nfs_list_entry(data->pages.next);
 	if ((!lseg) && list_is_singular(&data->pages))
-		lseg = pnfs_update_layout(desc->pg_inode, req->wb_context, IOMODE_READ, GFP_KERNEL);
+		lseg = pnfs_update_layout(desc->pg_inode, req->wb_context,
+					  req_offset(req), desc->pg_count,
+					  IOMODE_READ, GFP_KERNEL);
 
 	ret = nfs_read_rpcsetup(req, data, &nfs_read_full_ops, desc->pg_count,
 				0, lseg);
@@ -660,7 +664,6 @@ int nfs_readpages(struct file *filp, struct address_space *mapping,
 	if (ret == 0)
 		goto read_complete; /* all pages were read */
 
-	pnfs_pageio_init_read(&pgio, inode);
 	if (rsize < PAGE_CACHE_SIZE)
 		nfs_pageio_init(&pgio, inode, nfs_pagein_multi, rsize, 0);
 	else

fs/nfs/super.c

@@ -63,6 +63,7 @@
 #include "iostat.h"
 #include "internal.h"
 #include "fscache.h"
+#include "pnfs.h"
 
 #define NFSDBG_FACILITY		NFSDBG_VFS
 
@@ -732,6 +733,28 @@ static int nfs_show_options(struct seq_file *m, struct vfsmount *mnt)
 
 	return 0;
 }
+#ifdef CONFIG_NFS_V4_1
+void show_sessions(struct seq_file *m, struct nfs_server *server)
+{
+	if (nfs4_has_session(server->nfs_client))
+		seq_printf(m, ",sessions");
+}
+#else
+void show_sessions(struct seq_file *m, struct nfs_server *server) {}
+#endif
+
+#ifdef CONFIG_NFS_V4_1
+void show_pnfs(struct seq_file *m, struct nfs_server *server)
+{
+	seq_printf(m, ",pnfs=");
+	if (server->pnfs_curr_ld)
+		seq_printf(m, "%s", server->pnfs_curr_ld->name);
+	else
+		seq_printf(m, "not configured");
+}
+#else  /* CONFIG_NFS_V4_1 */
+void show_pnfs(struct seq_file *m, struct nfs_server *server) {}
+#endif /* CONFIG_NFS_V4_1 */
 
 static int nfs_show_devname(struct seq_file *m, struct vfsmount *mnt)
 {
@@ -792,6 +815,8 @@ static int nfs_show_stats(struct seq_file *m, struct vfsmount *mnt)
 		seq_printf(m, "bm0=0x%x", nfss->attr_bitmask[0]);
 		seq_printf(m, ",bm1=0x%x", nfss->attr_bitmask[1]);
 		seq_printf(m, ",acl=0x%x", nfss->acl_bitmask);
+		show_sessions(m, nfss);
+		show_pnfs(m, nfss);
 	}
 #endif
 

fs/nfs/write.c

@@ -939,7 +939,9 @@ static int nfs_flush_multi(struct nfs_pageio_descriptor *desc)
 	atomic_set(&req->wb_complete, requests);
 
 	BUG_ON(desc->pg_lseg);
-	lseg = pnfs_update_layout(desc->pg_inode, req->wb_context, IOMODE_RW, GFP_NOFS);
+	lseg = pnfs_update_layout(desc->pg_inode, req->wb_context,
+				  req_offset(req), desc->pg_count,
+				  IOMODE_RW, GFP_NOFS);
 	ClearPageError(page);
 	offset = 0;
 	nbytes = desc->pg_count;
@@ -1013,7 +1015,9 @@ static int nfs_flush_one(struct nfs_pageio_descriptor *desc)
 	}
 	req = nfs_list_entry(data->pages.next);
 	if ((!lseg) && list_is_singular(&data->pages))
-		lseg = pnfs_update_layout(desc->pg_inode, req->wb_context, IOMODE_RW, GFP_NOFS);
+		lseg = pnfs_update_layout(desc->pg_inode, req->wb_context,
+					  req_offset(req), desc->pg_count,
+					  IOMODE_RW, GFP_NOFS);
 
 	if ((desc->pg_ioflags & FLUSH_COND_STABLE) &&
 	    (desc->pg_moreio || NFS_I(desc->pg_inode)->ncommit))
@@ -1032,8 +1036,6 @@ static void nfs_pageio_init_write(struct nfs_pageio_descriptor *pgio,
 {
 	size_t wsize = NFS_SERVER(inode)->wsize;
 
-	pnfs_pageio_init_write(pgio, inode);
-
 	if (wsize < PAGE_CACHE_SIZE)
 		nfs_pageio_init(pgio, inode, nfs_flush_multi, wsize, ioflags);
 	else

include/linux/nfs4.h

@@ -562,6 +562,7 @@ enum {
 	NFSPROC4_CLNT_LAYOUTGET,
 	NFSPROC4_CLNT_GETDEVICEINFO,
 	NFSPROC4_CLNT_LAYOUTCOMMIT,
+	NFSPROC4_CLNT_LAYOUTRETURN,
 };
 
 /* nfs41 types */

include/linux/nfs_page.h

@@ -68,7 +68,7 @@ struct nfs_pageio_descriptor {
 	int 			pg_ioflags;
 	int			pg_error;
 	struct pnfs_layout_segment *pg_lseg;
-	int			(*pg_test)(struct nfs_pageio_descriptor *, struct nfs_page *, struct nfs_page *);
+	bool			(*pg_test)(struct nfs_pageio_descriptor *, struct nfs_page *, struct nfs_page *);
 };
 
 #define NFS_WBACK_BUSY(req)	(test_bit(PG_BUSY,&(req)->wb_flags))

include/linux/nfs_xdr.h

@@ -269,6 +269,27 @@ struct nfs4_layoutcommit_data {
 	struct nfs4_layoutcommit_res res;
 };
 
+struct nfs4_layoutreturn_args {
+	__u32   layout_type;
+	struct inode *inode;
+	nfs4_stateid stateid;
+	struct nfs4_sequence_args seq_args;
+};
+
+struct nfs4_layoutreturn_res {
+	struct nfs4_sequence_res seq_res;
+	u32 lrs_present;
+	nfs4_stateid stateid;
+};
+
+struct nfs4_layoutreturn {
+	struct nfs4_layoutreturn_args args;
+	struct nfs4_layoutreturn_res res;
+	struct rpc_cred *cred;
+	struct nfs_client *clp;
+	int rpc_status;
+};
+
 /*
  * Arguments to the open call.
  */
@@ -1087,6 +1108,7 @@ struct nfs_read_data {
 	const struct rpc_call_ops *mds_ops;
 	int (*read_done_cb) (struct rpc_task *task, struct nfs_read_data *data);
 	__u64			mds_offset;
+	int			pnfs_error;
 	struct page		*page_array[NFS_PAGEVEC_SIZE];
 };
 
@@ -1112,6 +1134,7 @@ struct nfs_write_data {
 	unsigned long		timestamp;	/* For lease renewal */
 #endif
 	__u64			mds_offset;	/* Filelayout dense stripe */
+	int			pnfs_error;
 	struct page		*page_array[NFS_PAGEVEC_SIZE];
 };
 

include/linux/pnfs_osd_xdr.h

@@ -0,0 +1,345 @@
+/*
+ *  pNFS-osd on-the-wire data structures
+ *
+ *  Copyright (C) 2007 Panasas Inc. [year of first publication]
+ *  All rights reserved.
+ *
+ *  Benny Halevy <bhalevy@panasas.com>
+ *  Boaz Harrosh <bharrosh@panasas.com>
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License version 2
+ *  See the file COPYING included with this distribution for more details.
+ *
+ *  Redistribution and use in source and binary forms, with or without
+ *  modification, are permitted provided that the following conditions
+ *  are met:
+ *
+ *  1. Redistributions of source code must retain the above copyright
+ *     notice, this list of conditions and the following disclaimer.
+ *  2. Redistributions in binary form must reproduce the above copyright
+ *     notice, this list of conditions and the following disclaimer in the
+ *     documentation and/or other materials provided with the distribution.
+ *  3. Neither the name of the Panasas company nor the names of its
+ *     contributors may be used to endorse or promote products derived
+ *     from this software without specific prior written permission.
+ *
+ *  THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED
+ *  WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+ *  MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ *  DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
+ *  FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ *  CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ *  SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
+ *  BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+ *  LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+ *  NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ *  SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+#ifndef __PNFS_OSD_XDR_H__
+#define __PNFS_OSD_XDR_H__
+
+#include <linux/nfs_fs.h>
+#include <linux/nfs_page.h>
+#include <scsi/osd_protocol.h>
+
+#define PNFS_OSD_OSDNAME_MAXSIZE 256
+
+/*
+ * draft-ietf-nfsv4-minorversion-22
+ * draft-ietf-nfsv4-pnfs-obj-12
+ */
+
+/* Layout Structure */
+
+enum pnfs_osd_raid_algorithm4 {
+	PNFS_OSD_RAID_0		= 1,
+	PNFS_OSD_RAID_4		= 2,
+	PNFS_OSD_RAID_5		= 3,
+	PNFS_OSD_RAID_PQ	= 4     /* Reed-Solomon P+Q */
+};
+
+/*   struct pnfs_osd_data_map4 {
+ *       uint32_t                    odm_num_comps;
+ *       length4                     odm_stripe_unit;
+ *       uint32_t                    odm_group_width;
+ *       uint32_t                    odm_group_depth;
+ *       uint32_t                    odm_mirror_cnt;
+ *       pnfs_osd_raid_algorithm4    odm_raid_algorithm;
+ *   };
+ */
+struct pnfs_osd_data_map {
+	u32	odm_num_comps;
+	u64	odm_stripe_unit;
+	u32	odm_group_width;
+	u32	odm_group_depth;
+	u32	odm_mirror_cnt;
+	u32	odm_raid_algorithm;
+};
+
+/*   struct pnfs_osd_objid4 {
+ *       deviceid4       oid_device_id;
+ *       uint64_t        oid_partition_id;
+ *       uint64_t        oid_object_id;
+ *   };
+ */
+struct pnfs_osd_objid {
+	struct nfs4_deviceid	oid_device_id;
+	u64			oid_partition_id;
+	u64			oid_object_id;
+};
+
+/* For printout. I use:
+ * kprint("dev(%llx:%llx)", _DEVID_LO(pointer), _DEVID_HI(pointer));
+ * BE style
+ */
+#define _DEVID_LO(oid_device_id) \
+	(unsigned long long)be64_to_cpup((__be64 *)(oid_device_id)->data)
+
+#define _DEVID_HI(oid_device_id) \
+	(unsigned long long)be64_to_cpup(((__be64 *)(oid_device_id)->data) + 1)
+
+static inline int
+pnfs_osd_objid_xdr_sz(void)
+{
+	return (NFS4_DEVICEID4_SIZE / 4) + 2 + 2;
+}
+
+enum pnfs_osd_version {
+	PNFS_OSD_MISSING              = 0,
+	PNFS_OSD_VERSION_1            = 1,
+	PNFS_OSD_VERSION_2            = 2
+};
+
+struct pnfs_osd_opaque_cred {
+	u32 cred_len;
+	void *cred;
+};
+
+enum pnfs_osd_cap_key_sec {
+	PNFS_OSD_CAP_KEY_SEC_NONE     = 0,
+	PNFS_OSD_CAP_KEY_SEC_SSV      = 1,
+};
+
+/*   struct pnfs_osd_object_cred4 {
+ *       pnfs_osd_objid4         oc_object_id;
+ *       pnfs_osd_version4       oc_osd_version;
+ *       pnfs_osd_cap_key_sec4   oc_cap_key_sec;
+ *       opaque                  oc_capability_key<>;
+ *       opaque                  oc_capability<>;
+ *   };
+ */
+struct pnfs_osd_object_cred {
+	struct pnfs_osd_objid		oc_object_id;
+	u32				oc_osd_version;
+	u32				oc_cap_key_sec;
+	struct pnfs_osd_opaque_cred	oc_cap_key;
+	struct pnfs_osd_opaque_cred	oc_cap;
+};
+
+/*   struct pnfs_osd_layout4 {
+ *       pnfs_osd_data_map4      olo_map;
+ *       uint32_t                olo_comps_index;
+ *       pnfs_osd_object_cred4   olo_components<>;
+ *   };
+ */
+struct pnfs_osd_layout {
+	struct pnfs_osd_data_map	olo_map;
+	u32				olo_comps_index;
+	u32				olo_num_comps;
+	struct pnfs_osd_object_cred	*olo_comps;
+};
+
+/* Device Address */
+enum pnfs_osd_targetid_type {
+	OBJ_TARGET_ANON = 1,
+	OBJ_TARGET_SCSI_NAME = 2,
+	OBJ_TARGET_SCSI_DEVICE_ID = 3,
+};
+
+/*   union pnfs_osd_targetid4 switch (pnfs_osd_targetid_type4 oti_type) {
+ *       case OBJ_TARGET_SCSI_NAME:
+ *           string              oti_scsi_name<>;
+ *
+ *       case OBJ_TARGET_SCSI_DEVICE_ID:
+ *           opaque              oti_scsi_device_id<>;
+ *
+ *       default:
+ *           void;
+ *   };
+ *
+ *   union pnfs_osd_targetaddr4 switch (bool ota_available) {
+ *       case TRUE:
+ *           netaddr4            ota_netaddr;
+ *       case FALSE:
+ *           void;
+ *   };
+ *
+ *   struct pnfs_osd_deviceaddr4 {
+ *       pnfs_osd_targetid4      oda_targetid;
+ *       pnfs_osd_targetaddr4    oda_targetaddr;
+ *       uint64_t                oda_lun;
+ *       opaque                  oda_systemid<>;
+ *       pnfs_osd_object_cred4   oda_root_obj_cred;
+ *       opaque                  oda_osdname<>;
+ *   };
+ */
+struct pnfs_osd_targetid {
+	u32				oti_type;
+	struct nfs4_string		oti_scsi_device_id;
+};
+
+enum { PNFS_OSD_TARGETID_MAX = 1 + PNFS_OSD_OSDNAME_MAXSIZE / 4 };
+
+/*   struct netaddr4 {
+ *       // see struct rpcb in RFC1833
+ *       string r_netid<>;    // network id
+ *       string r_addr<>;     // universal address
+ *   };
+ */
+struct pnfs_osd_net_addr {
+	struct nfs4_string	r_netid;
+	struct nfs4_string	r_addr;
+};
+
+struct pnfs_osd_targetaddr {
+	u32				ota_available;
+	struct pnfs_osd_net_addr	ota_netaddr;
+};
+
+enum {
+	NETWORK_ID_MAX = 16 / 4,
+	UNIVERSAL_ADDRESS_MAX = 64 / 4,
+	PNFS_OSD_TARGETADDR_MAX = 3 +  NETWORK_ID_MAX + UNIVERSAL_ADDRESS_MAX,
+};
+
+struct pnfs_osd_deviceaddr {
+	struct pnfs_osd_targetid	oda_targetid;
+	struct pnfs_osd_targetaddr	oda_targetaddr;
+	u8				oda_lun[8];
+	struct nfs4_string		oda_systemid;
+	struct pnfs_osd_object_cred	oda_root_obj_cred;
+	struct nfs4_string		oda_osdname;
+};
+
+enum {
+	ODA_OSDNAME_MAX = PNFS_OSD_OSDNAME_MAXSIZE / 4,
+	PNFS_OSD_DEVICEADDR_MAX =
+		PNFS_OSD_TARGETID_MAX + PNFS_OSD_TARGETADDR_MAX +
+		2 /*oda_lun*/ +
+		1 + OSD_SYSTEMID_LEN +
+		1 + ODA_OSDNAME_MAX,
+};
+
+/* LAYOUTCOMMIT: layoutupdate */
+
+/*   union pnfs_osd_deltaspaceused4 switch (bool dsu_valid) {
+ *       case TRUE:
+ *           int64_t     dsu_delta;
+ *       case FALSE:
+ *           void;
+ *   };
+ *
+ *   struct pnfs_osd_layoutupdate4 {
+ *       pnfs_osd_deltaspaceused4    olu_delta_space_used;
+ *       bool                        olu_ioerr_flag;
+ *   };
+ */
+struct pnfs_osd_layoutupdate {
+	u32	dsu_valid;
+	s64	dsu_delta;
+	u32	olu_ioerr_flag;
+};
+
+/* LAYOUTRETURN: I/O Rrror Report */
+
+enum pnfs_osd_errno {
+	PNFS_OSD_ERR_EIO		= 1,
+	PNFS_OSD_ERR_NOT_FOUND		= 2,
+	PNFS_OSD_ERR_NO_SPACE		= 3,
+	PNFS_OSD_ERR_BAD_CRED		= 4,
+	PNFS_OSD_ERR_NO_ACCESS		= 5,
+	PNFS_OSD_ERR_UNREACHABLE	= 6,
+	PNFS_OSD_ERR_RESOURCE		= 7
+};
+
+/*   struct pnfs_osd_ioerr4 {
+ *       pnfs_osd_objid4     oer_component;
+ *       length4             oer_comp_offset;
+ *       length4             oer_comp_length;
+ *       bool                oer_iswrite;
+ *       pnfs_osd_errno4     oer_errno;
+ *   };
+ */
+struct pnfs_osd_ioerr {
+	struct pnfs_osd_objid	oer_component;
+	u64			oer_comp_offset;
+	u64			oer_comp_length;
+	u32			oer_iswrite;
+	u32			oer_errno;
+};
+
+/* OSD XDR API */
+/* Layout helpers */
+/* Layout decoding is done in two parts:
+ * 1. First Call pnfs_osd_xdr_decode_layout_map to read in only the header part
+ *    of the layout. @iter members need not be initialized.
+ *    Returned:
+ *             @layout members are set. (@layout->olo_comps set to NULL).
+ *
+ *             Zero on success, or negative error if passed xdr is broken.
+ *
+ * 2. 2nd Call pnfs_osd_xdr_decode_layout_comp() in a loop until it returns
+ *    false, to decode the next component.
+ *    Returned:
+ *       true if there is more to decode or false if we are done or error.
+ *
+ * Example:
+ *	struct pnfs_osd_xdr_decode_layout_iter iter;
+ *	struct pnfs_osd_layout layout;
+ *	struct pnfs_osd_object_cred comp;
+ *	int status;
+ *
+ *	status = pnfs_osd_xdr_decode_layout_map(&layout, &iter, xdr);
+ *	if (unlikely(status))
+ *		goto err;
+ *	while(pnfs_osd_xdr_decode_layout_comp(&comp, &iter, xdr, &status)) {
+ *		// All of @comp strings point to inside the xdr_buffer
+ *		// or scrach buffer. Copy them out to user memory eg.
+ *		copy_single_comp(dest_comp++, &comp);
+ *	}
+ *	if (unlikely(status))
+ *		goto err;
+ */
+
+struct pnfs_osd_xdr_decode_layout_iter {
+	unsigned total_comps;
+	unsigned decoded_comps;
+};
+
+extern int pnfs_osd_xdr_decode_layout_map(struct pnfs_osd_layout *layout,
+	struct pnfs_osd_xdr_decode_layout_iter *iter, struct xdr_stream *xdr);
+
+extern bool pnfs_osd_xdr_decode_layout_comp(struct pnfs_osd_object_cred *comp,
+	struct pnfs_osd_xdr_decode_layout_iter *iter, struct xdr_stream *xdr,
+	int *err);
+
+/* Device Info helpers */
+
+/* Note: All strings inside @deviceaddr point to space inside @p.
+ * @p should stay valid while @deviceaddr is in use.
+ */
+extern void pnfs_osd_xdr_decode_deviceaddr(
+	struct pnfs_osd_deviceaddr *deviceaddr, __be32 *p);
+
+/* layoutupdate (layout_commit) xdr helpers */
+extern int
+pnfs_osd_xdr_encode_layoutupdate(struct xdr_stream *xdr,
+				 struct pnfs_osd_layoutupdate *lou);
+
+/* osd_ioerror encoding/decoding (layout_return) */
+/* Client */
+extern __be32 *pnfs_osd_xdr_ioerr_reserve_space(struct xdr_stream *xdr);
+extern void pnfs_osd_xdr_encode_ioerr(__be32 *p, struct pnfs_osd_ioerr *ioerr);
+
+#endif /* __PNFS_OSD_XDR_H__ */

include/linux/sunrpc/xdr.h

@@ -216,6 +216,8 @@ extern __be32 *xdr_reserve_space(struct xdr_stream *xdr, size_t nbytes);
 extern void xdr_write_pages(struct xdr_stream *xdr, struct page **pages,
 		unsigned int base, unsigned int len);
 extern void xdr_init_decode(struct xdr_stream *xdr, struct xdr_buf *buf, __be32 *p);
+extern void xdr_init_decode_pages(struct xdr_stream *xdr, struct xdr_buf *buf,
+		struct page **pages, unsigned int len);
 extern void xdr_set_scratch_buffer(struct xdr_stream *xdr, void *buf, size_t buflen);
 extern __be32 *xdr_inline_decode(struct xdr_stream *xdr, size_t nbytes);
 extern void xdr_read_pages(struct xdr_stream *xdr, unsigned int len);

net/sunrpc/xdr.c

@@ -638,6 +638,25 @@ void xdr_init_decode(struct xdr_stream *xdr, struct xdr_buf *buf, __be32 *p)
 }
 EXPORT_SYMBOL_GPL(xdr_init_decode);
 
+/**
+ * xdr_init_decode - Initialize an xdr_stream for decoding data.
+ * @xdr: pointer to xdr_stream struct
+ * @buf: pointer to XDR buffer from which to decode data
+ * @pages: list of pages to decode into
+ * @len: length in bytes of buffer in pages
+ */
+void xdr_init_decode_pages(struct xdr_stream *xdr, struct xdr_buf *buf,
+			   struct page **pages, unsigned int len)
+{
+	memset(buf, 0, sizeof(*buf));
+	buf->pages =  pages;
+	buf->page_len =  len;
+	buf->buflen =  len;
+	buf->len = len;
+	xdr_init_decode(xdr, buf, NULL);
+}
+EXPORT_SYMBOL_GPL(xdr_init_decode_pages);
+
 static __be32 * __xdr_inline_decode(struct xdr_stream *xdr, size_t nbytes)
 {
 	__be32 *p = xdr->p;