1 files changed, 83 insertions, 34 deletions

diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 70806cae0d36..56b2d96f9103 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -30,6 +30,7 @@
 #include <linux/nfs_fs.h>
 #include <linux/nfs_page.h>
 #include <linux/module.h>
+#include <linux/sort.h>
 #include "internal.h"
 #include "pnfs.h"
 #include "iostat.h"
@@ -99,35 +100,79 @@ unset_pnfs_layoutdriver(struct nfs_server *nfss)
 }
 
 /*
+ * When the server sends a list of layout types, we choose one in the order
+ * given in the list below.
+ *
+ * FIXME: should this list be configurable in some fashion? module param?
+ * 	  mount option? something else?
+ */
+static const u32 ld_prefs[] = {
+	LAYOUT_SCSI,
+	LAYOUT_BLOCK_VOLUME,
+	LAYOUT_OSD2_OBJECTS,
+	LAYOUT_FLEX_FILES,
+	LAYOUT_NFSV4_1_FILES,
+	0
+};
+
+static int
+ld_cmp(const void *e1, const void *e2)
+{
+	u32 ld1 = *((u32 *)e1);
+	u32 ld2 = *((u32 *)e2);
+	int i;
+
+	for (i = 0; ld_prefs[i] != 0; i++) {
+		if (ld1 == ld_prefs[i])
+			return -1;
+
+		if (ld2 == ld_prefs[i])
+			return 1;
+	}
+	return 0;
+}
+
+/*
  * Try to set the server's pnfs module to the pnfs layout type specified by id.
  * Currently only one pNFS layout driver per filesystem is supported.
  *
- * @id layout type. Zero (illegal layout type) indicates pNFS not in use.
+ * @ids array of layout types supported by MDS.
  */
 void
 set_pnfs_layoutdriver(struct nfs_server *server, const struct nfs_fh *mntfh,
-		      u32 id)
+		      struct nfs_fsinfo *fsinfo)
 {
 	struct pnfs_layoutdriver_type *ld_type = NULL;
+	u32 id;
+	int i;
 
-	if (id == 0)
-		goto out_no_driver;
 	if (!(server->nfs_client->cl_exchange_flags &
 		 (EXCHGID4_FLAG_USE_NON_PNFS | EXCHGID4_FLAG_USE_PNFS_MDS))) {
-		printk(KERN_ERR "NFS: %s: id %u cl_exchange_flags 0x%x\n",
-			__func__, id, server->nfs_client->cl_exchange_flags);
+		printk(KERN_ERR "NFS: %s: cl_exchange_flags 0x%x\n",
+			__func__, server->nfs_client->cl_exchange_flags);
 		goto out_no_driver;
 	}
-	ld_type = find_pnfs_driver(id);
-	if (!ld_type) {
-		request_module("%s-%u", LAYOUT_NFSV4_1_MODULE_PREFIX, id);
+
+	sort(fsinfo->layouttype, fsinfo->nlayouttypes,
+		sizeof(*fsinfo->layouttype), ld_cmp, NULL);
+
+	for (i = 0; i < fsinfo->nlayouttypes; i++) {
+		id = fsinfo->layouttype[i];
 		ld_type = find_pnfs_driver(id);
 		if (!ld_type) {
-			dprintk("%s: No pNFS module found for %u.\n",
-				__func__, id);
-			goto out_no_driver;
+			request_module("%s-%u", LAYOUT_NFSV4_1_MODULE_PREFIX,
+					id);
+			ld_type = find_pnfs_driver(id);
 		}
+		if (ld_type)
+			break;
 	}
+
+	if (!ld_type) {
+		dprintk("%s: No pNFS module found!\n", __func__);
+		goto out_no_driver;
+	}
+
 	server->pnfs_curr_ld = ld_type;
 	if (ld_type->set_layoutdriver
 	    && ld_type->set_layoutdriver(server, mntfh)) {
@@ -365,7 +410,8 @@ pnfs_layout_remove_lseg(struct pnfs_layout_hdr *lo,
 	/* Matched by pnfs_get_layout_hdr in pnfs_layout_insert_lseg */
 	atomic_dec(&lo->plh_refcount);
 	if (list_empty(&lo->plh_segs)) {
-		set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags);
+		if (atomic_read(&lo->plh_outstanding) == 0)
+			set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags);
 		clear_bit(NFS_LAYOUT_BULK_RECALL, &lo->plh_flags);
 	}
 	rpc_wake_up(&NFS_SERVER(inode)->roc_rpcwaitq);
@@ -768,17 +814,32 @@ pnfs_destroy_all_layouts(struct nfs_client *clp)
 	pnfs_destroy_layouts_byclid(clp, false);
 }
 
+static void
+pnfs_clear_layoutreturn_info(struct pnfs_layout_hdr *lo)
+{
+	lo->plh_return_iomode = 0;
+	lo->plh_return_seq = 0;
+	clear_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags);
+}
+
 /* update lo->plh_stateid with new if is more recent */
 void
 pnfs_set_layout_stateid(struct pnfs_layout_hdr *lo, const nfs4_stateid *new,
 			bool update_barrier)
 {
 	u32 oldseq, newseq, new_barrier = 0;
-	bool invalid = !pnfs_layout_is_valid(lo);
 
 	oldseq = be32_to_cpu(lo->plh_stateid.seqid);
 	newseq = be32_to_cpu(new->seqid);
-	if (invalid || pnfs_seqid_is_newer(newseq, oldseq)) {
+
+	if (!pnfs_layout_is_valid(lo)) {
+		nfs4_stateid_copy(&lo->plh_stateid, new);
+		lo->plh_barrier = newseq;
+		pnfs_clear_layoutreturn_info(lo);
+		clear_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags);
+		return;
+	}
+	if (pnfs_seqid_is_newer(newseq, oldseq)) {
 		nfs4_stateid_copy(&lo->plh_stateid, new);
 		/*
 		 * Because of wraparound, we want to keep the barrier
@@ -790,7 +851,7 @@ pnfs_set_layout_stateid(struct pnfs_layout_hdr *lo, const nfs4_stateid *new,
 		new_barrier = be32_to_cpu(new->seqid);
 	else if (new_barrier == 0)
 		return;
-	if (invalid || pnfs_seqid_is_newer(new_barrier, lo->plh_barrier))
+	if (pnfs_seqid_is_newer(new_barrier, lo->plh_barrier))
 		lo->plh_barrier = new_barrier;
 }
 
@@ -886,19 +947,14 @@ void pnfs_clear_layoutreturn_waitbit(struct pnfs_layout_hdr *lo)
 	rpc_wake_up(&NFS_SERVER(lo->plh_inode)->roc_rpcwaitq);
 }
 
-static void
-pnfs_clear_layoutreturn_info(struct pnfs_layout_hdr *lo)
-{
-	lo->plh_return_iomode = 0;
-	lo->plh_return_seq = 0;
-	clear_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags);
-}
-
 static bool
 pnfs_prepare_layoutreturn(struct pnfs_layout_hdr *lo,
 		nfs4_stateid *stateid,
 		enum pnfs_iomode *iomode)
 {
+	/* Serialise LAYOUTGET/LAYOUTRETURN */
+	if (atomic_read(&lo->plh_outstanding) != 0)
+		return false;
 	if (test_and_set_bit(NFS_LAYOUT_RETURN, &lo->plh_flags))
 		return false;
 	pnfs_get_layout_hdr(lo);
@@ -1555,6 +1611,7 @@ pnfs_update_layout(struct inode *ino,
 	}
 
 lookup_again:
+	nfs4_client_recover_expired_lease(clp);
 	first = false;
 	spin_lock(&ino->i_lock);
 	lo = pnfs_find_alloc_layout(ino, ctx, gfp_flags);
@@ -1797,16 +1854,11 @@ pnfs_layout_process(struct nfs4_layoutget *lgp)
 		 */
 		pnfs_mark_layout_stateid_invalid(lo, &free_me);
 
-		nfs4_stateid_copy(&lo->plh_stateid, &res->stateid);
-		lo->plh_barrier = be32_to_cpu(res->stateid.seqid);
+		pnfs_set_layout_stateid(lo, &res->stateid, true);
 	}
 
 	pnfs_get_lseg(lseg);
 	pnfs_layout_insert_lseg(lo, lseg, &free_me);
-	if (!pnfs_layout_is_valid(lo)) {
-		pnfs_clear_layoutreturn_info(lo);
-		clear_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags);
-	}
 
 
 	if (res->return_on_close)
@@ -2178,10 +2230,8 @@ static void pnfs_ld_handle_read_error(struct nfs_pgio_header *hdr)
  */
 void pnfs_ld_read_done(struct nfs_pgio_header *hdr)
 {
-	if (likely(!hdr->pnfs_error)) {
-		__nfs4_read_done_cb(hdr);
+	if (likely(!hdr->pnfs_error))
 		hdr->mds_ops->rpc_call_done(&hdr->task, hdr);
-	}
 	trace_nfs4_pnfs_read(hdr, hdr->pnfs_error);
 	if (unlikely(hdr->pnfs_error))
 		pnfs_ld_handle_read_error(hdr);
@@ -2510,7 +2560,6 @@ pnfs_report_layoutstat(struct inode *inode, gfp_t gfp_flags)
 
 	data->args.fh = NFS_FH(inode);
 	data->args.inode = inode;
-	nfs4_stateid_copy(&data->args.stateid, &hdr->plh_stateid);
 	status = ld->prepare_layoutstats(&data->args);
 	if (status)
 		goto out_free;