From c9895cb69b07a4b17d8fdae26667f9a9fba5183b Mon Sep 17 00:00:00 2001
From: Weston Andros Adamson <dros@netapp.com>
Date: Tue, 31 May 2011 18:48:56 -0400
Subject: NFS: pnfs IPv6 support

Handle ipv6 remote addresses from GETDEVICEINFO

 - supports netid "tcp" for ipv4 and "tcp6" for ipv6 as rfc 5665 specifies
 - added ds_remotestr to avoid having to handle different AFs in every dprintk
 - tested against pynfs 4.1 server, submitting ipv6 support patch to pynfs
 - tested with IPv6 disabled, it compiles cleanly and relies on rpc_pton to
   refuse to accept IPv6 addresses

Signed-off-by: Weston Andros Adamson <dros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/nfs4filelayout.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

(limited to 'fs/nfs/nfs4filelayout.c')

diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c
index f9d03abcd04..b690bb5ee58 100644
--- a/fs/nfs/nfs4filelayout.c
+++ b/fs/nfs/nfs4filelayout.c
@@ -344,8 +344,7 @@ filelayout_read_pagelist(struct nfs_read_data *data)
 		set_bit(lo_fail_bit(IOMODE_READ), &lseg->pls_layout->plh_flags);
 		return PNFS_NOT_ATTEMPTED;
 	}
-	dprintk("%s USE DS:ip %x %hu\n", __func__,
-		ntohl(ds->ds_ip_addr), ntohs(ds->ds_port));
+	dprintk("%s USE DS: %s\n", __func__, ds->ds_remotestr);
 
 	/* No multipath support. Use first DS */
 	data->ds_clp = ds->ds_clp;
@@ -384,9 +383,9 @@ filelayout_write_pagelist(struct nfs_write_data *data, int sync)
 		set_bit(lo_fail_bit(IOMODE_READ), &lseg->pls_layout->plh_flags);
 		return PNFS_NOT_ATTEMPTED;
 	}
-	dprintk("%s ino %lu sync %d req %Zu@%llu DS:%x:%hu\n", __func__,
+	dprintk("%s ino %lu sync %d req %Zu@%llu DS: %s\n", __func__,
 		data->inode->i_ino, sync, (size_t) data->args.count, offset,
-		ntohl(ds->ds_ip_addr), ntohs(ds->ds_port));
+		ds->ds_remotestr);
 
 	data->write_done_cb = filelayout_write_done_cb;
 	data->ds_clp = ds->ds_clp;
-- 
cgit v1.2.3-70-g09d2


From 1751c3638f2a07a8c66a803a31791bab9bd3fced Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
Date: Fri, 10 Jun 2011 13:30:23 -0400
Subject: NFS: Cleanup of the nfs_pageio code in preparation for a pnfs bugfix

We need to ensure that the layouts are set up before we can decide to
coalesce requests. To do so, we want to further split up the struct
nfs_pageio_descriptor operations into an initialisation callback, a
coalescing test callback, and a 'do i/o' callback.

This patch cleans up the existing callback methods before adding the
'initialisation' callback.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/nfs4filelayout.c      | 15 +++++++++++++--
 fs/nfs/objlayout/objio_osd.c | 13 ++++++++++++-
 fs/nfs/pagelist.c            | 10 ++++------
 fs/nfs/pnfs.c                | 24 ++++++++++++++++++++++++
 fs/nfs/pnfs.h                | 25 +++++++++++++------------
 fs/nfs/read.c                | 43 ++++++++++++++++++++++++++++++++-----------
 fs/nfs/write.c               | 28 ++++++++++++++++++++++------
 include/linux/nfs_page.h     | 14 +++++++++++---
 8 files changed, 131 insertions(+), 41 deletions(-)

(limited to 'fs/nfs/nfs4filelayout.c')

diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c
index b690bb5ee58..46bb31b7968 100644
--- a/fs/nfs/nfs4filelayout.c
+++ b/fs/nfs/nfs4filelayout.c
@@ -658,7 +658,7 @@ filelayout_alloc_lseg(struct pnfs_layout_hdr *layoutid,
  * return true  : coalesce page
  * return false : don't coalesce page
  */
-bool
+static bool
 filelayout_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev,
 		   struct nfs_page *req)
 {
@@ -681,6 +681,16 @@ filelayout_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev,
 	return (p_stripe == r_stripe);
 }
 
+static const struct nfs_pageio_ops filelayout_pg_read_ops = {
+	.pg_test = filelayout_pg_test,
+	.pg_doio = nfs_generic_pg_readpages,
+};
+
+static const struct nfs_pageio_ops filelayout_pg_write_ops = {
+	.pg_test = filelayout_pg_test,
+	.pg_doio = nfs_generic_pg_writepages,
+};
+
 static bool filelayout_mark_pnfs_commit(struct pnfs_layout_segment *lseg)
 {
 	return !FILELAYOUT_LSEG(lseg)->commit_through_mds;
@@ -878,7 +888,8 @@ static struct pnfs_layoutdriver_type filelayout_type = {
 	.owner			= THIS_MODULE,
 	.alloc_lseg		= filelayout_alloc_lseg,
 	.free_lseg		= filelayout_free_lseg,
-	.pg_test		= filelayout_pg_test,
+	.pg_read_ops		= &filelayout_pg_read_ops,
+	.pg_write_ops		= &filelayout_pg_write_ops,
 	.mark_pnfs_commit	= filelayout_mark_pnfs_commit,
 	.choose_commit_list	= filelayout_choose_commit_list,
 	.commit_pagelist	= filelayout_commit_pagelist,
diff --git a/fs/nfs/objlayout/objio_osd.c b/fs/nfs/objlayout/objio_osd.c
index 8ff2ea3f10e..c203fab694a 100644
--- a/fs/nfs/objlayout/objio_osd.c
+++ b/fs/nfs/objlayout/objio_osd.c
@@ -1007,6 +1007,16 @@ static bool objio_pg_test(struct nfs_pageio_descriptor *pgio,
 			OBJIO_LSEG(pgio->pg_lseg)->max_io_size;
 }
 
+static const struct nfs_pageio_ops objio_pg_read_ops = {
+	.pg_test = objio_pg_test,
+	.pg_doio = nfs_generic_pg_readpages,
+};
+
+static const struct nfs_pageio_ops objio_pg_write_ops = {
+	.pg_test = objio_pg_test,
+	.pg_doio = nfs_generic_pg_writepages,
+};
+
 static struct pnfs_layoutdriver_type objlayout_type = {
 	.id = LAYOUT_OSD2_OBJECTS,
 	.name = "LAYOUT_OSD2_OBJECTS",
@@ -1020,7 +1030,8 @@ static struct pnfs_layoutdriver_type objlayout_type = {
 
 	.read_pagelist           = objlayout_read_pagelist,
 	.write_pagelist          = objlayout_write_pagelist,
-	.pg_test                 = objio_pg_test,
+	.pg_read_ops             = &objio_pg_read_ops,
+	.pg_write_ops            = &objio_pg_write_ops,
 
 	.free_deviceid_node	 = objio_free_deviceid_node,
 
diff --git a/fs/nfs/pagelist.c b/fs/nfs/pagelist.c
index 00985571628..9b8a4730f0b 100644
--- a/fs/nfs/pagelist.c
+++ b/fs/nfs/pagelist.c
@@ -230,7 +230,7 @@ EXPORT_SYMBOL_GPL(nfs_generic_pg_test);
  */
 void nfs_pageio_init(struct nfs_pageio_descriptor *desc,
 		     struct inode *inode,
-		     int (*doio)(struct nfs_pageio_descriptor *),
+		     const struct nfs_pageio_ops *pg_ops,
 		     size_t bsize,
 		     int io_flags)
 {
@@ -241,12 +241,10 @@ void nfs_pageio_init(struct nfs_pageio_descriptor *desc,
 	desc->pg_base = 0;
 	desc->pg_moreio = 0;
 	desc->pg_inode = inode;
-	desc->pg_doio = doio;
+	desc->pg_ops = pg_ops;
 	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);
 }
 
 /**
@@ -276,7 +274,7 @@ static bool nfs_can_coalesce_requests(struct nfs_page *prev,
 		return false;
 	if (prev->wb_pgbase + prev->wb_bytes != PAGE_CACHE_SIZE)
 		return false;
-	return pgio->pg_test(pgio, prev, req);
+	return pgio->pg_ops->pg_test(pgio, prev, req);
 }
 
 /**
@@ -311,7 +309,7 @@ static int nfs_pageio_do_add_request(struct nfs_pageio_descriptor *desc,
 static void nfs_pageio_doio(struct nfs_pageio_descriptor *desc)
 {
 	if (!list_empty(&desc->pg_list)) {
-		int error = desc->pg_doio(desc);
+		int error = desc->pg_ops->pg_doio(desc);
 		if (error < 0)
 			desc->pg_error = error;
 		else
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index ff820077237..7ec46d5f05a 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -1055,6 +1055,30 @@ out_forget_reply:
 	goto out;
 }
 
+bool
+pnfs_pageio_init_read(struct nfs_pageio_descriptor *pgio, struct inode *inode)
+{
+	struct nfs_server *server = NFS_SERVER(inode);
+	struct pnfs_layoutdriver_type *ld = server->pnfs_curr_ld;
+
+	if (ld == NULL)
+		return false;
+	nfs_pageio_init(pgio, inode, ld->pg_read_ops, server->rsize, 0);
+	return true;
+}
+
+bool
+pnfs_pageio_init_write(struct nfs_pageio_descriptor *pgio, struct inode *inode, int ioflags)
+{
+	struct nfs_server *server = NFS_SERVER(inode);
+	struct pnfs_layoutdriver_type *ld = server->pnfs_curr_ld;
+
+	if (ld == NULL)
+		return false;
+	nfs_pageio_init(pgio, inode, ld->pg_write_ops, server->wsize, ioflags);
+	return true;
+}
+
 bool
 pnfs_generic_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev,
 		     struct nfs_page *req)
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index 96bf4e6f45b..137a2bd5c8c 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -87,7 +87,8 @@ struct pnfs_layoutdriver_type {
 	void (*free_lseg) (struct pnfs_layout_segment *lseg);
 
 	/* test for nfs page cache coalescing */
-	bool (*pg_test)(struct nfs_pageio_descriptor *, struct nfs_page *, struct nfs_page *);
+	const struct nfs_pageio_ops *pg_read_ops;
+	const struct nfs_pageio_ops *pg_write_ops;
 
 	/* Returns true if layoutdriver wants to divert this request to
 	 * driver's commit routine.
@@ -152,6 +153,10 @@ struct pnfs_layout_segment *
 pnfs_update_layout(struct inode *ino, struct nfs_open_context *ctx,
 		   loff_t pos, u64 count, enum pnfs_iomode access_type,
 		   gfp_t gfp_flags);
+
+bool pnfs_pageio_init_read(struct nfs_pageio_descriptor *, struct inode *);
+bool pnfs_pageio_init_write(struct nfs_pageio_descriptor *, struct inode *, int);
+
 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 *,
@@ -293,15 +298,6 @@ static inline int pnfs_return_layout(struct inode *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)
@@ -385,9 +381,14 @@ static inline void unset_pnfs_layoutdriver(struct nfs_server *s)
 {
 }
 
-static inline void pnfs_pageio_init(struct nfs_pageio_descriptor *pgio,
-				    struct inode *inode)
+static inline bool pnfs_pageio_init_read(struct nfs_pageio_descriptor *pgio, struct inode *inode)
 {
+	return false;
+}
+
+static inline bool pnfs_pageio_init_write(struct nfs_pageio_descriptor *pgio, struct inode *inode, int ioflags)
+{
+	return false;
 }
 
 static inline void
diff --git a/fs/nfs/read.c b/fs/nfs/read.c
index 20a7f952e24..b6d9ec9a208 100644
--- a/fs/nfs/read.c
+++ b/fs/nfs/read.c
@@ -32,6 +32,7 @@
 
 static int nfs_pagein_multi(struct nfs_pageio_descriptor *desc);
 static int nfs_pagein_one(struct nfs_pageio_descriptor *desc);
+static const struct nfs_pageio_ops nfs_pageio_read_ops;
 static const struct rpc_call_ops nfs_read_partial_ops;
 static const struct rpc_call_ops nfs_read_full_ops;
 
@@ -113,6 +114,20 @@ static void nfs_readpage_truncate_uninitialised_page(struct nfs_read_data *data)
 	}
 }
 
+static void nfs_pageio_init_read_mds(struct nfs_pageio_descriptor *pgio,
+		struct inode *inode)
+{
+	nfs_pageio_init(pgio, inode, &nfs_pageio_read_ops,
+			NFS_SERVER(inode)->rsize, 0);
+}
+
+static void nfs_pageio_init_read(struct nfs_pageio_descriptor *pgio,
+		struct inode *inode)
+{
+	if (!pnfs_pageio_init_read(pgio, inode))
+		nfs_pageio_init_read_mds(pgio, inode);
+}
+
 int nfs_readpage_async(struct nfs_open_context *ctx, struct inode *inode,
 		       struct page *page)
 {
@@ -131,14 +146,11 @@ int nfs_readpage_async(struct nfs_open_context *ctx, struct inode *inode,
 	if (len < PAGE_CACHE_SIZE)
 		zero_user_segment(page, len, PAGE_CACHE_SIZE);
 
-	nfs_pageio_init(&pgio, inode, NULL, 0, 0);
+	nfs_pageio_init_read(&pgio, inode);
 	nfs_list_add_request(new, &pgio.pg_list);
 	pgio.pg_count = len;
 
-	if (NFS_SERVER(inode)->rsize < PAGE_CACHE_SIZE)
-		nfs_pagein_multi(&pgio);
-	else
-		nfs_pagein_one(&pgio);
+	nfs_pageio_complete(&pgio);
 	return 0;
 }
 
@@ -365,6 +377,20 @@ out:
 	return ret;
 }
 
+int nfs_generic_pg_readpages(struct nfs_pageio_descriptor *desc)
+{
+	if (desc->pg_bsize < PAGE_CACHE_SIZE)
+		return nfs_pagein_multi(desc);
+	return nfs_pagein_one(desc);
+}
+EXPORT_SYMBOL_GPL(nfs_generic_pg_readpages);
+
+
+static const struct nfs_pageio_ops nfs_pageio_read_ops = {
+	.pg_test = nfs_generic_pg_test,
+	.pg_doio = nfs_generic_pg_readpages,
+};
+
 /*
  * This is the callback from RPC telling us whether a reply was
  * received or some error occurred (timeout or socket shutdown).
@@ -635,8 +661,6 @@ int nfs_readpages(struct file *filp, struct address_space *mapping,
 		.pgio = &pgio,
 	};
 	struct inode *inode = mapping->host;
-	struct nfs_server *server = NFS_SERVER(inode);
-	size_t rsize = server->rsize;
 	unsigned long npages;
 	int ret = -ESTALE;
 
@@ -664,10 +688,7 @@ int nfs_readpages(struct file *filp, struct address_space *mapping,
 	if (ret == 0)
 		goto read_complete; /* all pages were read */
 
-	if (rsize < PAGE_CACHE_SIZE)
-		nfs_pageio_init(&pgio, inode, nfs_pagein_multi, rsize, 0);
-	else
-		nfs_pageio_init(&pgio, inode, nfs_pagein_one, rsize, 0);
+	nfs_pageio_init_read(&pgio, inode);
 
 	ret = read_cache_pages(mapping, pages, readpage_async_filler, &desc);
 
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index 72716805968..b144b547454 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -1033,15 +1033,31 @@ out:
 	return ret;
 }
 
-static void nfs_pageio_init_write(struct nfs_pageio_descriptor *pgio,
+int nfs_generic_pg_writepages(struct nfs_pageio_descriptor *desc)
+{
+	if (desc->pg_bsize < PAGE_CACHE_SIZE)
+		return nfs_flush_multi(desc);
+	return nfs_flush_one(desc);
+}
+EXPORT_SYMBOL_GPL(nfs_generic_pg_writepages);
+
+static const struct nfs_pageio_ops nfs_pageio_write_ops = {
+	.pg_test = nfs_generic_pg_test,
+	.pg_doio = nfs_generic_pg_writepages,
+};
+
+static void nfs_pageio_init_write_mds(struct nfs_pageio_descriptor *pgio,
 				  struct inode *inode, int ioflags)
 {
-	size_t wsize = NFS_SERVER(inode)->wsize;
+	nfs_pageio_init(pgio, inode, &nfs_pageio_write_ops,
+				NFS_SERVER(inode)->wsize, ioflags);
+}
 
-	if (wsize < PAGE_CACHE_SIZE)
-		nfs_pageio_init(pgio, inode, nfs_flush_multi, wsize, ioflags);
-	else
-		nfs_pageio_init(pgio, inode, nfs_flush_one, wsize, ioflags);
+static void nfs_pageio_init_write(struct nfs_pageio_descriptor *pgio,
+				  struct inode *inode, int ioflags)
+{
+	if (!pnfs_pageio_init_write(pgio, inode, ioflags))
+		nfs_pageio_init_write_mds(pgio, inode, ioflags);
 }
 
 /*
diff --git a/include/linux/nfs_page.h b/include/linux/nfs_page.h
index 25311b3bedf..d378f08b905 100644
--- a/include/linux/nfs_page.h
+++ b/include/linux/nfs_page.h
@@ -55,6 +55,12 @@ struct nfs_page {
 	struct nfs_writeverf	wb_verf;	/* Commit cookie */
 };
 
+struct nfs_pageio_descriptor;
+struct nfs_pageio_ops {
+	bool	(*pg_test)(struct nfs_pageio_descriptor *, struct nfs_page *, struct nfs_page *);
+	int	(*pg_doio)(struct nfs_pageio_descriptor *);
+};
+
 struct nfs_pageio_descriptor {
 	struct list_head	pg_list;
 	unsigned long		pg_bytes_written;
@@ -64,11 +70,10 @@ struct nfs_pageio_descriptor {
 	char			pg_moreio;
 
 	struct inode		*pg_inode;
-	int			(*pg_doio)(struct nfs_pageio_descriptor *);
+	const struct nfs_pageio_ops *pg_ops;
 	int 			pg_ioflags;
 	int			pg_error;
 	struct pnfs_layout_segment *pg_lseg;
-	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))
@@ -85,7 +90,7 @@ extern	int nfs_scan_list(struct nfs_inode *nfsi, struct list_head *dst,
 			  pgoff_t idx_start, unsigned int npages, int tag);
 extern	void nfs_pageio_init(struct nfs_pageio_descriptor *desc,
 			     struct inode *inode,
-			     int (*doio)(struct nfs_pageio_descriptor *desc),
+			     const struct nfs_pageio_ops *pg_ops,
 			     size_t bsize,
 			     int how);
 extern	int nfs_pageio_add_request(struct nfs_pageio_descriptor *,
@@ -100,6 +105,9 @@ extern	void nfs_unlock_request(struct nfs_page *req);
 extern	int nfs_set_page_tag_locked(struct nfs_page *req);
 extern  void nfs_clear_page_tag_locked(struct nfs_page *req);
 
+extern	int nfs_generic_pg_readpages(struct nfs_pageio_descriptor *desc);
+extern	int nfs_generic_pg_writepages(struct nfs_pageio_descriptor *desc);
+
 
 /*
  * Lock the page of an asynchronous request without getting a new reference
-- 
cgit v1.2.3-70-g09d2


From d8007d4dd6ff8749cc8a4063c3ec87442db76d82 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
Date: Fri, 10 Jun 2011 13:30:23 -0400
Subject: NFSv4.1: Add an initialisation callback for pNFS

Ensure that we always get a layout before setting up the i/o request.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/nfs4filelayout.c      |  2 ++
 fs/nfs/objlayout/objio_osd.c |  2 ++
 fs/nfs/pagelist.c            |  2 ++
 fs/nfs/pnfs.c                | 57 ++++++++++++++++++++++++--------------------
 fs/nfs/pnfs.h                | 14 ++---------
 fs/nfs/read.c                | 16 ++++---------
 fs/nfs/write.c               | 12 +++-------
 include/linux/nfs_page.h     |  1 +
 8 files changed, 47 insertions(+), 59 deletions(-)

(limited to 'fs/nfs/nfs4filelayout.c')

diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c
index 46bb31b7968..faac87f2479 100644
--- a/fs/nfs/nfs4filelayout.c
+++ b/fs/nfs/nfs4filelayout.c
@@ -682,11 +682,13 @@ filelayout_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev,
 }
 
 static const struct nfs_pageio_ops filelayout_pg_read_ops = {
+	.pg_init = pnfs_generic_pg_init_read,
 	.pg_test = filelayout_pg_test,
 	.pg_doio = nfs_generic_pg_readpages,
 };
 
 static const struct nfs_pageio_ops filelayout_pg_write_ops = {
+	.pg_init = pnfs_generic_pg_init_write,
 	.pg_test = filelayout_pg_test,
 	.pg_doio = nfs_generic_pg_writepages,
 };
diff --git a/fs/nfs/objlayout/objio_osd.c b/fs/nfs/objlayout/objio_osd.c
index c203fab694a..b759aeba57b 100644
--- a/fs/nfs/objlayout/objio_osd.c
+++ b/fs/nfs/objlayout/objio_osd.c
@@ -1008,11 +1008,13 @@ static bool objio_pg_test(struct nfs_pageio_descriptor *pgio,
 }
 
 static const struct nfs_pageio_ops objio_pg_read_ops = {
+	.pg_init = pnfs_generic_pg_init_read,
 	.pg_test = objio_pg_test,
 	.pg_doio = nfs_generic_pg_readpages,
 };
 
 static const struct nfs_pageio_ops objio_pg_write_ops = {
+	.pg_init = pnfs_generic_pg_init_write,
 	.pg_test = objio_pg_test,
 	.pg_doio = nfs_generic_pg_writepages,
 };
diff --git a/fs/nfs/pagelist.c b/fs/nfs/pagelist.c
index 9b8a4730f0b..d421e19557a 100644
--- a/fs/nfs/pagelist.c
+++ b/fs/nfs/pagelist.c
@@ -295,6 +295,8 @@ static int nfs_pageio_do_add_request(struct nfs_pageio_descriptor *desc,
 		if (!nfs_can_coalesce_requests(prev, req, desc))
 			return 0;
 	} else {
+		if (desc->pg_ops->pg_init)
+			desc->pg_ops->pg_init(desc, req);
 		desc->pg_base = req->wb_pgbase;
 	}
 	nfs_list_remove_request(req);
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 7ec46d5f05a..69a0c3f1e46 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -911,7 +911,7 @@ pnfs_find_lseg(struct pnfs_layout_hdr *lo,
  * Layout segment is retreived from the server if not cached.
  * The appropriate layout segment is referenced and returned to the caller.
  */
-struct pnfs_layout_segment *
+static struct pnfs_layout_segment *
 pnfs_update_layout(struct inode *ino,
 		   struct nfs_open_context *ctx,
 		   loff_t pos,
@@ -1055,6 +1055,34 @@ out_forget_reply:
 	goto out;
 }
 
+void
+pnfs_generic_pg_init_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *req)
+{
+	BUG_ON(pgio->pg_lseg != NULL);
+
+	pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
+					   req->wb_context,
+					   req_offset(req),
+					   req->wb_bytes,
+					   IOMODE_READ,
+					   GFP_KERNEL);
+}
+EXPORT_SYMBOL_GPL(pnfs_generic_pg_init_read);
+
+void
+pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *pgio, struct nfs_page *req)
+{
+	BUG_ON(pgio->pg_lseg != NULL);
+
+	pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
+					   req->wb_context,
+					   req_offset(req),
+					   req->wb_bytes,
+					   IOMODE_RW,
+					   GFP_NOFS);
+}
+EXPORT_SYMBOL_GPL(pnfs_generic_pg_init_write);
+
 bool
 pnfs_pageio_init_read(struct nfs_pageio_descriptor *pgio, struct inode *inode)
 {
@@ -1083,31 +1111,8 @@ bool
 pnfs_generic_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev,
 		     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_lseg == NULL) {
-		if (pgio->pg_count != prev->wb_bytes)
-			return true;
-		/* This is first coelesce call for a series of nfs_pages */
-		pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
-						   prev->wb_context,
-						   req_offset(prev),
-						   pgio->pg_count,
-						   access_type,
-						   gfp_flags);
-		if (pgio->pg_lseg == NULL)
-			return true;
-	}
+	if (pgio->pg_lseg == NULL)
+		return nfs_generic_pg_test(pgio, prev, req);
 
 	/*
 	 * Test if a nfs_page is fully contained in the pnfs_layout_range.
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index 137a2bd5c8c..dd54351ae6a 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -149,10 +149,6 @@ 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,
-		   loff_t pos, u64 count, enum pnfs_iomode access_type,
-		   gfp_t gfp_flags);
 
 bool pnfs_pageio_init_read(struct nfs_pageio_descriptor *, struct inode *);
 bool pnfs_pageio_init_write(struct nfs_pageio_descriptor *, struct inode *, int);
@@ -163,6 +159,8 @@ 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_generic_pg_init_read(struct nfs_pageio_descriptor *, struct nfs_page *);
+void pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *, struct nfs_page *);
 bool pnfs_generic_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev, struct nfs_page *req);
 int pnfs_layout_process(struct nfs4_layoutget *lgp);
 void pnfs_free_lseg_list(struct list_head *tmp_list);
@@ -318,14 +316,6 @@ 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,
-		   loff_t pos, u64 count, enum pnfs_iomode access_type,
-		   gfp_t gfp_flags)
-{
-	return NULL;
-}
-
 static inline enum pnfs_try_status
 pnfs_try_to_read_data(struct nfs_read_data *data,
 		      const struct rpc_call_ops *call_ops)
diff --git a/fs/nfs/read.c b/fs/nfs/read.c
index b6d9ec9a208..2cf728b832c 100644
--- a/fs/nfs/read.c
+++ b/fs/nfs/read.c
@@ -147,9 +147,7 @@ int nfs_readpage_async(struct nfs_open_context *ctx, struct inode *inode,
 		zero_user_segment(page, len, PAGE_CACHE_SIZE);
 
 	nfs_pageio_init_read(&pgio, inode);
-	nfs_list_add_request(new, &pgio.pg_list);
-	pgio.pg_count = len;
-
+	nfs_pageio_add_request(&pgio, new);
 	nfs_pageio_complete(&pgio);
 	return 0;
 }
@@ -281,7 +279,7 @@ static int nfs_pagein_multi(struct nfs_pageio_descriptor *desc)
 	unsigned int offset;
 	int requests = 0;
 	int ret = 0;
-	struct pnfs_layout_segment *lseg;
+	struct pnfs_layout_segment *lseg = desc->pg_lseg;
 	LIST_HEAD(list);
 
 	nfs_list_remove_request(req);
@@ -299,10 +297,6 @@ static int nfs_pagein_multi(struct nfs_pageio_descriptor *desc)
 	} while(nbytes != 0);
 	atomic_set(&req->wb_complete, requests);
 
-	BUG_ON(desc->pg_lseg != NULL);
-	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;
@@ -336,6 +330,8 @@ out_bad:
 	}
 	SetPageError(page);
 	nfs_readpage_release(req);
+	put_lseg(lseg);
+	desc->pg_lseg = NULL;
 	return -ENOMEM;
 }
 
@@ -364,10 +360,6 @@ static int nfs_pagein_one(struct nfs_pageio_descriptor *desc)
 		*pages++ = req->wb_page;
 	}
 	req = nfs_list_entry(data->pages.next);
-	if ((!lseg) && list_is_singular(&data->pages))
-		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);
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index b144b547454..d1ae7e45886 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -916,7 +916,7 @@ static int nfs_flush_multi(struct nfs_pageio_descriptor *desc)
 	unsigned int offset;
 	int requests = 0;
 	int ret = 0;
-	struct pnfs_layout_segment *lseg;
+	struct pnfs_layout_segment *lseg = desc->pg_lseg;
 	LIST_HEAD(list);
 
 	nfs_list_remove_request(req);
@@ -940,10 +940,6 @@ static int nfs_flush_multi(struct nfs_pageio_descriptor *desc)
 	} while (nbytes != 0);
 	atomic_set(&req->wb_complete, requests);
 
-	BUG_ON(desc->pg_lseg);
-	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;
@@ -976,6 +972,8 @@ out_bad:
 		nfs_writedata_free(data);
 	}
 	nfs_redirty_request(req);
+	put_lseg(lseg);
+	desc->pg_lseg = NULL;
 	return -ENOMEM;
 }
 
@@ -1016,10 +1014,6 @@ static int nfs_flush_one(struct nfs_pageio_descriptor *desc)
 		*pages++ = req->wb_page;
 	}
 	req = nfs_list_entry(data->pages.next);
-	if ((!lseg) && list_is_singular(&data->pages))
-		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))
diff --git a/include/linux/nfs_page.h b/include/linux/nfs_page.h
index d378f08b905..9ac2dd158d0 100644
--- a/include/linux/nfs_page.h
+++ b/include/linux/nfs_page.h
@@ -57,6 +57,7 @@ struct nfs_page {
 
 struct nfs_pageio_descriptor;
 struct nfs_pageio_ops {
+	void	(*pg_init)(struct nfs_pageio_descriptor *, struct nfs_page *);
 	bool	(*pg_test)(struct nfs_pageio_descriptor *, struct nfs_page *, struct nfs_page *);
 	int	(*pg_doio)(struct nfs_pageio_descriptor *);
 };
-- 
cgit v1.2.3-70-g09d2


From e885de1a5bc9f46ef8f934c5a7602c89d2d51e8d Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
Date: Fri, 10 Jun 2011 13:30:23 -0400
Subject: NFSv4.1: Fall back to ordinary i/o through the mds if we have no
 layout segment

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/internal.h            | 6 ++++++
 fs/nfs/nfs4filelayout.c      | 2 --
 fs/nfs/objlayout/objio_osd.c | 3 ---
 fs/nfs/pnfs.c                | 7 +++++++
 fs/nfs/read.c                | 2 +-
 fs/nfs/write.c               | 2 +-
 6 files changed, 15 insertions(+), 7 deletions(-)

(limited to 'fs/nfs/nfs4filelayout.c')

diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h
index fc017eadfe0..31e8b50011a 100644
--- a/fs/nfs/internal.h
+++ b/fs/nfs/internal.h
@@ -296,7 +296,13 @@ extern int nfs_initiate_read(struct nfs_read_data *data, struct rpc_clnt *clnt,
 			     const struct rpc_call_ops *call_ops);
 extern void nfs_read_prepare(struct rpc_task *task, void *calldata);
 
+struct nfs_pageio_descriptor;
+extern void nfs_pageio_init_read_mds(struct nfs_pageio_descriptor *pgio,
+		struct inode *inode);
+
 /* write.c */
+extern void nfs_pageio_init_write_mds(struct nfs_pageio_descriptor *pgio,
+				  struct inode *inode, int ioflags);
 extern void nfs_commit_free(struct nfs_write_data *p);
 extern int nfs_initiate_write(struct nfs_write_data *data,
 			      struct rpc_clnt *clnt,
diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c
index faac87f2479..d1ae5bfa668 100644
--- a/fs/nfs/nfs4filelayout.c
+++ b/fs/nfs/nfs4filelayout.c
@@ -669,8 +669,6 @@ filelayout_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev,
 	    !nfs_generic_pg_test(pgio, prev, req))
 		return false;
 
-	if (!pgio->pg_lseg)
-		return 1;
 	p_stripe = (u64)prev->wb_index << PAGE_CACHE_SHIFT;
 	r_stripe = (u64)req->wb_index << PAGE_CACHE_SHIFT;
 	stripe_unit = FILELAYOUT_LSEG(pgio->pg_lseg)->stripe_unit;
diff --git a/fs/nfs/objlayout/objio_osd.c b/fs/nfs/objlayout/objio_osd.c
index b759aeba57b..70272d5355b 100644
--- a/fs/nfs/objlayout/objio_osd.c
+++ b/fs/nfs/objlayout/objio_osd.c
@@ -1000,9 +1000,6 @@ static bool objio_pg_test(struct nfs_pageio_descriptor *pgio,
 	if (!pnfs_generic_pg_test(pgio, prev, req))
 		return false;
 
-	if (pgio->pg_lseg == NULL)
-		return true;
-
 	return pgio->pg_count + req->wb_bytes <=
 			OBJIO_LSEG(pgio->pg_lseg)->max_io_size;
 }
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 69a0c3f1e46..cc807fe4695 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -1066,6 +1066,10 @@ pnfs_generic_pg_init_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *r
 					   req->wb_bytes,
 					   IOMODE_READ,
 					   GFP_KERNEL);
+	/* If no lseg, fall back to read through mds */
+	if (pgio->pg_lseg == NULL)
+		nfs_pageio_init_read_mds(pgio, pgio->pg_inode);
+
 }
 EXPORT_SYMBOL_GPL(pnfs_generic_pg_init_read);
 
@@ -1080,6 +1084,9 @@ pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *pgio, struct nfs_page *
 					   req->wb_bytes,
 					   IOMODE_RW,
 					   GFP_NOFS);
+	/* If no lseg, fall back to write through mds */
+	if (pgio->pg_lseg == NULL)
+		nfs_pageio_init_write_mds(pgio, pgio->pg_inode, pgio->pg_ioflags);
 }
 EXPORT_SYMBOL_GPL(pnfs_generic_pg_init_write);
 
diff --git a/fs/nfs/read.c b/fs/nfs/read.c
index 2cf728b832c..9b2a1d7fb76 100644
--- a/fs/nfs/read.c
+++ b/fs/nfs/read.c
@@ -114,7 +114,7 @@ static void nfs_readpage_truncate_uninitialised_page(struct nfs_read_data *data)
 	}
 }
 
-static void nfs_pageio_init_read_mds(struct nfs_pageio_descriptor *pgio,
+void nfs_pageio_init_read_mds(struct nfs_pageio_descriptor *pgio,
 		struct inode *inode)
 {
 	nfs_pageio_init(pgio, inode, &nfs_pageio_read_ops,
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index d1ae7e45886..00940dc9ab0 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -1040,7 +1040,7 @@ static const struct nfs_pageio_ops nfs_pageio_write_ops = {
 	.pg_doio = nfs_generic_pg_writepages,
 };
 
-static void nfs_pageio_init_write_mds(struct nfs_pageio_descriptor *pgio,
+void nfs_pageio_init_write_mds(struct nfs_pageio_descriptor *pgio,
 				  struct inode *inode, int ioflags)
 {
 	nfs_pageio_init(pgio, inode, &nfs_pageio_write_ops,
-- 
cgit v1.2.3-70-g09d2


From 7c24d9489fe57d67cb56c6bdad58d89806e7fd97 Mon Sep 17 00:00:00 2001
From: Andy Adamson <andros@netapp.com>
Date: Mon, 13 Jun 2011 18:22:38 -0400
Subject: NFSv4.1: File layout only supports whole file layouts

Ask for whole file layouts. Until support for layout segments is fully
supported in the file layout code, discard non-whole file layouts.

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Fred Isaman <iisaman@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/nfs4filelayout.c | 47 +++++++++++++++++++++++++++++++++++++++++++++--
 fs/nfs/pnfs.c           |  6 ++++--
 fs/nfs/pnfs.h           |  6 ++++++
 fs/nfs/read.c           |  1 +
 fs/nfs/write.c          |  1 +
 5 files changed, 57 insertions(+), 4 deletions(-)

(limited to 'fs/nfs/nfs4filelayout.c')

diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c
index d1ae5bfa668..51c19093022 100644
--- a/fs/nfs/nfs4filelayout.c
+++ b/fs/nfs/nfs4filelayout.c
@@ -427,6 +427,14 @@ filelayout_check_layout(struct pnfs_layout_hdr *lo,
 
 	dprintk("--> %s\n", __func__);
 
+	/* FIXME: remove this check when layout segment support is added */
+	if (lgr->range.offset != 0 ||
+	    lgr->range.length != NFS4_MAX_UINT64) {
+		dprintk("%s Only whole file layouts supported. Use MDS i/o\n",
+			__func__);
+		goto out;
+	}
+
 	if (fl->pattern_offset > lgr->range.offset) {
 		dprintk("%s pattern_offset %lld too large\n",
 				__func__, fl->pattern_offset);
@@ -679,14 +687,49 @@ filelayout_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev,
 	return (p_stripe == r_stripe);
 }
 
+void
+filelayout_pg_init_read(struct nfs_pageio_descriptor *pgio,
+			struct nfs_page *req)
+{
+	BUG_ON(pgio->pg_lseg != NULL);
+
+	pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
+					   req->wb_context,
+					   0,
+					   NFS4_MAX_UINT64,
+					   IOMODE_READ,
+					   GFP_KERNEL);
+	/* If no lseg, fall back to read through mds */
+	if (pgio->pg_lseg == NULL)
+		nfs_pageio_init_read_mds(pgio, pgio->pg_inode);
+}
+
+void
+filelayout_pg_init_write(struct nfs_pageio_descriptor *pgio,
+			 struct nfs_page *req)
+{
+	BUG_ON(pgio->pg_lseg != NULL);
+
+	pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
+					   req->wb_context,
+					   0,
+					   NFS4_MAX_UINT64,
+					   IOMODE_RW,
+					   GFP_NOFS);
+	/* If no lseg, fall back to write through mds */
+	if (pgio->pg_lseg == NULL)
+		nfs_pageio_init_write_mds(pgio, pgio->pg_inode,
+					  pgio->pg_ioflags);
+}
+
 static const struct nfs_pageio_ops filelayout_pg_read_ops = {
-	.pg_init = pnfs_generic_pg_init_read,
+	.pg_init = filelayout_pg_init_read,
 	.pg_test = filelayout_pg_test,
 	.pg_doio = nfs_generic_pg_readpages,
 };
 
 static const struct nfs_pageio_ops filelayout_pg_write_ops = {
-	.pg_init = pnfs_generic_pg_init_write,
+	.pg_init = filelayout_pg_init_write,
 	.pg_test = filelayout_pg_test,
 	.pg_doio = nfs_generic_pg_writepages,
 };
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index cc807fe4695..a7dc3367a85 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -911,7 +911,7 @@ pnfs_find_lseg(struct pnfs_layout_hdr *lo,
  * Layout segment is retreived from the server if not cached.
  * The appropriate layout segment is referenced and returned to the caller.
  */
-static struct pnfs_layout_segment *
+struct pnfs_layout_segment *
 pnfs_update_layout(struct inode *ino,
 		   struct nfs_open_context *ctx,
 		   loff_t pos,
@@ -980,7 +980,8 @@ pnfs_update_layout(struct inode *ino,
 		arg.offset -= pg_offset;
 		arg.length += pg_offset;
 	}
-	arg.length = PAGE_CACHE_ALIGN(arg.length);
+	if (arg.length != NFS4_MAX_UINT64)
+		arg.length = PAGE_CACHE_ALIGN(arg.length);
 
 	lseg = send_layoutget(lo, ctx, &arg, gfp_flags);
 	if (!lseg && first) {
@@ -998,6 +999,7 @@ out_unlock:
 	spin_unlock(&ino->i_lock);
 	goto out;
 }
+EXPORT_SYMBOL_GPL(pnfs_update_layout);
 
 int
 pnfs_layout_process(struct nfs4_layoutget *lgp)
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index 1cfc96f8c45..678c4c7b14d 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -185,6 +185,12 @@ 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 *);
+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);
 
 /* pnfs_dev.c */
 struct nfs4_deviceid_node {
diff --git a/fs/nfs/read.c b/fs/nfs/read.c
index 9b2a1d7fb76..c394662c790 100644
--- a/fs/nfs/read.c
+++ b/fs/nfs/read.c
@@ -120,6 +120,7 @@ void nfs_pageio_init_read_mds(struct nfs_pageio_descriptor *pgio,
 	nfs_pageio_init(pgio, inode, &nfs_pageio_read_ops,
 			NFS_SERVER(inode)->rsize, 0);
 }
+EXPORT_SYMBOL_GPL(nfs_pageio_init_read_mds);
 
 static void nfs_pageio_init_read(struct nfs_pageio_descriptor *pgio,
 		struct inode *inode)
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index 00940dc9ab0..7f8732e3198 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -1046,6 +1046,7 @@ void nfs_pageio_init_write_mds(struct nfs_pageio_descriptor *pgio,
 	nfs_pageio_init(pgio, inode, &nfs_pageio_write_ops,
 				NFS_SERVER(inode)->wsize, ioflags);
 }
+EXPORT_SYMBOL_GPL(nfs_pageio_init_write_mds);
 
 static void nfs_pageio_init_write(struct nfs_pageio_descriptor *pgio,
 				  struct inode *inode, int ioflags)
-- 
cgit v1.2.3-70-g09d2


From c47abcf8ff4d0c56d20ce541e80d3e1c975f54b5 Mon Sep 17 00:00:00 2001
From: Andy Adamson <andros@netapp.com>
Date: Wed, 15 Jun 2011 17:52:40 -0400
Subject: NFSv4.1: do not use deviceids after MDS clientid invalidation

Mark all deviceids established under an expired MDS clientid as invalid.
Stop all new i/o through DS and send through the MDS.
Don't use any new LAYOUTGETs that use the invalid deviceid. Purge all layouts
established under the expired MDS clientid.
Remove the MDS clientid deviceid and data servers reference

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/nfs4filelayout.c | 10 ++++++++++
 fs/nfs/nfs4filelayout.h |  6 ++++++
 fs/nfs/pnfs.c           |  3 +++
 fs/nfs/pnfs.h           |  8 ++++++++
 fs/nfs/pnfs_dev.c       | 20 ++++++++++++++++++++
 5 files changed, 47 insertions(+)

(limited to 'fs/nfs/nfs4filelayout.c')

diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c
index 51c19093022..af9bf9eed4c 100644
--- a/fs/nfs/nfs4filelayout.c
+++ b/fs/nfs/nfs4filelayout.c
@@ -334,6 +334,9 @@ filelayout_read_pagelist(struct nfs_read_data *data)
 		__func__, data->inode->i_ino,
 		data->args.pgbase, (size_t)data->args.count, offset);
 
+	if (test_bit(NFS_DEVICEID_INVALID, &FILELAYOUT_DEVID_NODE(lseg)->flags))
+		return PNFS_NOT_ATTEMPTED;
+
 	/* Retrieve the correct rpc_client for the byte range */
 	j = nfs4_fl_calc_j_index(lseg, offset);
 	idx = nfs4_fl_calc_ds_index(lseg, j);
@@ -373,6 +376,9 @@ filelayout_write_pagelist(struct nfs_write_data *data, int sync)
 	struct nfs_fh *fh;
 	int status;
 
+	if (test_bit(NFS_DEVICEID_INVALID, &FILELAYOUT_DEVID_NODE(lseg)->flags))
+		return PNFS_NOT_ATTEMPTED;
+
 	/* Retrieve the correct rpc_client for the byte range */
 	j = nfs4_fl_calc_j_index(lseg, offset);
 	idx = nfs4_fl_calc_ds_index(lseg, j);
@@ -456,6 +462,10 @@ filelayout_check_layout(struct pnfs_layout_hdr *lo,
 			goto out;
 	} else
 		dsaddr = container_of(d, struct nfs4_file_layout_dsaddr, id_node);
+	/* Found deviceid is being reaped */
+	if (test_bit(NFS_DEVICEID_INVALID, &dsaddr->id_node.flags))
+			goto out_put;
+
 	fl->dsaddr = dsaddr;
 
 	if (fl->first_stripe_index < 0 ||
diff --git a/fs/nfs/nfs4filelayout.h b/fs/nfs/nfs4filelayout.h
index 68cce730b80..2e42284253f 100644
--- a/fs/nfs/nfs4filelayout.h
+++ b/fs/nfs/nfs4filelayout.h
@@ -96,6 +96,12 @@ FILELAYOUT_LSEG(struct pnfs_layout_segment *lseg)
 			    generic_hdr);
 }
 
+static inline struct nfs4_deviceid_node *
+FILELAYOUT_DEVID_NODE(struct pnfs_layout_segment *lseg)
+{
+	return &FILELAYOUT_LSEG(lseg)->dsaddr->id_node;
+}
+
 extern struct nfs_fh *
 nfs4_fl_select_ds_fh(struct pnfs_layout_segment *lseg, u32 j);
 
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 5fc2e5d755a..5b3cc3f4bb3 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -452,6 +452,9 @@ pnfs_destroy_all_layouts(struct nfs_client *clp)
 	struct pnfs_layout_hdr *lo;
 	LIST_HEAD(tmp_list);
 
+	nfs4_deviceid_mark_client_invalid(clp);
+	nfs4_deviceid_purge_client(clp);
+
 	spin_lock(&clp->cl_lock);
 	rcu_read_lock();
 	list_for_each_entry_rcu(server, &clp->cl_superblocks, client_link) {
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index 678c4c7b14d..a59736eae6e 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -192,12 +192,20 @@ struct pnfs_layout_segment *pnfs_update_layout(struct inode *ino,
 					       enum pnfs_iomode iomode,
 					       gfp_t gfp_flags);
 
+void nfs4_deviceid_mark_client_invalid(struct nfs_client *clp);
+
+/* nfs4_deviceid_flags */
+enum {
+	NFS_DEVICEID_INVALID = 0,       /* set when MDS clientid recalled */
+};
+
 /* pnfs_dev.c */
 struct nfs4_deviceid_node {
 	struct hlist_node		node;
 	struct hlist_node		tmpnode;
 	const struct pnfs_layoutdriver_type *ld;
 	const struct nfs_client		*nfs_client;
+	unsigned long 			flags;
 	struct nfs4_deviceid		deviceid;
 	atomic_t			ref;
 };
diff --git a/fs/nfs/pnfs_dev.c b/fs/nfs/pnfs_dev.c
index fb9498d91f6..6fda5228ef5 100644
--- a/fs/nfs/pnfs_dev.c
+++ b/fs/nfs/pnfs_dev.c
@@ -156,6 +156,7 @@ nfs4_init_deviceid_node(struct nfs4_deviceid_node *d,
 	INIT_HLIST_NODE(&d->tmpnode);
 	d->ld = ld;
 	d->nfs_client = nfs_client;
+	d->flags = 0;
 	d->deviceid = *id;
 	atomic_set(&d->ref, 1);
 }
@@ -253,3 +254,22 @@ nfs4_deviceid_purge_client(const struct nfs_client *clp)
 	for (h = 0; h < NFS4_DEVICE_ID_HASH_SIZE; h++)
 		_deviceid_purge_client(clp, h);
 }
+
+/*
+ * Stop use of all deviceids associated with an nfs_client
+ */
+void
+nfs4_deviceid_mark_client_invalid(struct nfs_client *clp)
+{
+	struct nfs4_deviceid_node *d;
+	struct hlist_node *n;
+	int i;
+
+	rcu_read_lock();
+	for (i = 0; i < NFS4_DEVICE_ID_HASH_SIZE; i ++){
+		hlist_for_each_entry_rcu(d, n, &nfs4_deviceid_cache[i], node)
+			if (d->nfs_client == clp)
+				set_bit(NFS_DEVICEID_INVALID, &d->flags);
+	}
+	rcu_read_unlock();
+}
-- 
cgit v1.2.3-70-g09d2


From 493292ddc78d18ee2ad2d5c24c2b7dd6a24641d2 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
Date: Wed, 13 Jul 2011 15:58:28 -0400
Subject: NFS: Move the pnfs read code into pnfs.c

...and ensure that we recoalese to take into account differences in
block sizes when falling back to read through the MDS.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/internal.h            |  6 ++++-
 fs/nfs/nfs4filelayout.c      |  2 +-
 fs/nfs/objlayout/objio_osd.c |  2 +-
 fs/nfs/pnfs.c                | 57 ++++++++++++++++++++++++++++++++++++++++++--
 fs/nfs/pnfs.h                | 10 +-------
 fs/nfs/read.c                | 46 ++++++++++++++++-------------------
 include/linux/nfs_page.h     |  1 -
 7 files changed, 83 insertions(+), 41 deletions(-)

(limited to 'fs/nfs/nfs4filelayout.c')

diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h
index 31e8b50011a..d74d7dea917 100644
--- a/fs/nfs/internal.h
+++ b/fs/nfs/internal.h
@@ -291,14 +291,18 @@ extern struct dentry *nfs4_get_root(struct super_block *, struct nfs_fh *,
 extern int nfs4_get_rootfh(struct nfs_server *server, struct nfs_fh *mntfh);
 #endif
 
+struct nfs_pageio_descriptor;
 /* read.c */
 extern int nfs_initiate_read(struct nfs_read_data *data, struct rpc_clnt *clnt,
 			     const struct rpc_call_ops *call_ops);
 extern void nfs_read_prepare(struct rpc_task *task, void *calldata);
+extern int nfs_generic_pagein(struct nfs_pageio_descriptor *desc,
+		struct list_head *head);
 
-struct nfs_pageio_descriptor;
 extern void nfs_pageio_init_read_mds(struct nfs_pageio_descriptor *pgio,
 		struct inode *inode);
+extern void nfs_pageio_reset_read_mds(struct nfs_pageio_descriptor *pgio);
+extern void nfs_readdata_release(struct nfs_read_data *rdata);
 
 /* write.c */
 extern void nfs_pageio_init_write_mds(struct nfs_pageio_descriptor *pgio,
diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c
index af9bf9eed4c..fc556d6b90f 100644
--- a/fs/nfs/nfs4filelayout.c
+++ b/fs/nfs/nfs4filelayout.c
@@ -735,7 +735,7 @@ filelayout_pg_init_write(struct nfs_pageio_descriptor *pgio,
 static const struct nfs_pageio_ops filelayout_pg_read_ops = {
 	.pg_init = filelayout_pg_init_read,
 	.pg_test = filelayout_pg_test,
-	.pg_doio = nfs_generic_pg_readpages,
+	.pg_doio = pnfs_generic_pg_readpages,
 };
 
 static const struct nfs_pageio_ops filelayout_pg_write_ops = {
diff --git a/fs/nfs/objlayout/objio_osd.c b/fs/nfs/objlayout/objio_osd.c
index 70272d5355b..add62894f36 100644
--- a/fs/nfs/objlayout/objio_osd.c
+++ b/fs/nfs/objlayout/objio_osd.c
@@ -1007,7 +1007,7 @@ static bool objio_pg_test(struct nfs_pageio_descriptor *pgio,
 static const struct nfs_pageio_ops objio_pg_read_ops = {
 	.pg_init = pnfs_generic_pg_init_read,
 	.pg_test = objio_pg_test,
-	.pg_doio = nfs_generic_pg_readpages,
+	.pg_doio = pnfs_generic_pg_readpages,
 };
 
 static const struct nfs_pageio_ops objio_pg_write_ops = {
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 5b3cc3f4bb3..9eca5a8cdbd 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -28,6 +28,7 @@
  */
 
 #include <linux/nfs_fs.h>
+#include <linux/nfs_page.h>
 #include "internal.h"
 #include "pnfs.h"
 #include "iostat.h"
@@ -1216,18 +1217,32 @@ pnfs_ld_read_done(struct nfs_read_data *data)
 }
 EXPORT_SYMBOL_GPL(pnfs_ld_read_done);
 
+static void
+pnfs_read_through_mds(struct nfs_pageio_descriptor *desc,
+		struct nfs_read_data *data)
+{
+	list_splice_tail_init(&data->pages, &desc->pg_list);
+	if (data->req && list_empty(&data->req->wb_list))
+		nfs_list_add_request(data->req, &desc->pg_list);
+	nfs_pageio_reset_read_mds(desc);
+	desc->pg_recoalesce = 1;
+	nfs_readdata_release(data);
+}
+
 /*
  * Call the appropriate parallel I/O subsystem read function.
  */
-enum pnfs_try_status
+static enum pnfs_try_status
 pnfs_try_to_read_data(struct nfs_read_data *rdata,
-		       const struct rpc_call_ops *call_ops)
+		       const struct rpc_call_ops *call_ops,
+		       struct pnfs_layout_segment *lseg)
 {
 	struct inode *inode = rdata->inode;
 	struct nfs_server *nfss = NFS_SERVER(inode);
 	enum pnfs_try_status trypnfs;
 
 	rdata->mds_ops = call_ops;
+	rdata->lseg = get_lseg(lseg);
 
 	dprintk("%s: Reading ino:%lu %u@%llu\n",
 		__func__, inode->i_ino, rdata->args.count, rdata->args.offset);
@@ -1243,6 +1258,44 @@ pnfs_try_to_read_data(struct nfs_read_data *rdata,
 	return trypnfs;
 }
 
+static void
+pnfs_do_multiple_reads(struct nfs_pageio_descriptor *desc, struct list_head *head)
+{
+	struct nfs_read_data *data;
+	const struct rpc_call_ops *call_ops = desc->pg_rpc_callops;
+	struct pnfs_layout_segment *lseg = desc->pg_lseg;
+
+	desc->pg_lseg = NULL;
+	while (!list_empty(head)) {
+		enum pnfs_try_status trypnfs;
+
+		data = list_entry(head->next, struct nfs_read_data, list);
+		list_del_init(&data->list);
+
+		trypnfs = pnfs_try_to_read_data(data, call_ops, lseg);
+		if (trypnfs == PNFS_NOT_ATTEMPTED)
+			pnfs_read_through_mds(desc, data);
+	}
+	put_lseg(lseg);
+}
+
+int
+pnfs_generic_pg_readpages(struct nfs_pageio_descriptor *desc)
+{
+	LIST_HEAD(head);
+	int ret;
+
+	ret = nfs_generic_pagein(desc, &head);
+	if (ret != 0) {
+		put_lseg(desc->pg_lseg);
+		desc->pg_lseg = NULL;
+		return ret;
+	}
+	pnfs_do_multiple_reads(desc, &head);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(pnfs_generic_pg_readpages);
+
 /*
  * Currently there is only one (whole file) write lseg.
  */
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index a59736eae6e..c40ffa52c1a 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -157,9 +157,8 @@ 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_generic_pg_init_read(struct nfs_pageio_descriptor *, struct nfs_page *);
+int pnfs_generic_pg_readpages(struct nfs_pageio_descriptor *desc);
 void pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *, struct nfs_page *);
 bool pnfs_generic_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev, struct nfs_page *req);
 int pnfs_layout_process(struct nfs4_layoutget *lgp);
@@ -329,13 +328,6 @@ static inline void put_lseg(struct pnfs_layout_segment *lseg)
 {
 }
 
-static inline enum pnfs_try_status
-pnfs_try_to_read_data(struct nfs_read_data *data,
-		      const struct rpc_call_ops *call_ops)
-{
-	return PNFS_NOT_ATTEMPTED;
-}
-
 static inline enum pnfs_try_status
 pnfs_try_to_write_data(struct nfs_write_data *data,
 		       const struct rpc_call_ops *call_ops, int how)
diff --git a/fs/nfs/read.c b/fs/nfs/read.c
index 75088f58183..d2f53ddd826 100644
--- a/fs/nfs/read.c
+++ b/fs/nfs/read.c
@@ -67,7 +67,7 @@ void nfs_readdata_free(struct nfs_read_data *p)
 	mempool_free(p, nfs_rdata_mempool);
 }
 
-static void nfs_readdata_release(struct nfs_read_data *rdata)
+void nfs_readdata_release(struct nfs_read_data *rdata)
 {
 	put_lseg(rdata->lseg);
 	put_nfs_open_context(rdata->args.context);
@@ -120,6 +120,12 @@ void nfs_pageio_init_read_mds(struct nfs_pageio_descriptor *pgio,
 }
 EXPORT_SYMBOL_GPL(nfs_pageio_init_read_mds);
 
+void nfs_pageio_reset_read_mds(struct nfs_pageio_descriptor *pgio)
+{
+	pgio->pg_ops = &nfs_pageio_read_ops;
+	pgio->pg_bsize = NFS_SERVER(pgio->pg_inode)->rsize;
+}
+
 static void nfs_pageio_init_read(struct nfs_pageio_descriptor *pgio,
 		struct inode *inode)
 {
@@ -235,26 +241,16 @@ static void nfs_read_rpcsetup(struct nfs_page *req, struct nfs_read_data *data,
 }
 
 static int nfs_do_read(struct nfs_read_data *data,
-		const struct rpc_call_ops *call_ops,
-		struct pnfs_layout_segment *lseg)
+		const struct rpc_call_ops *call_ops)
 {
 	struct inode *inode = data->args.context->path.dentry->d_inode;
 
-	if (lseg) {
-		data->lseg = get_lseg(lseg);
-		if (pnfs_try_to_read_data(data, call_ops) == PNFS_ATTEMPTED)
-			return 0;
-		put_lseg(data->lseg);
-		data->lseg = NULL;
-	}
-
 	return nfs_initiate_read(data, NFS_CLIENT(inode), call_ops);
 }
 
 static int
 nfs_do_multiple_reads(struct list_head *head,
-		const struct rpc_call_ops *call_ops,
-		struct pnfs_layout_segment *lseg)
+		const struct rpc_call_ops *call_ops)
 {
 	struct nfs_read_data *data;
 	int ret = 0;
@@ -265,7 +261,7 @@ nfs_do_multiple_reads(struct list_head *head,
 		data = list_entry(head->next, struct nfs_read_data, list);
 		list_del_init(&data->list);
 
-		ret2 = nfs_do_read(data, call_ops, lseg);
+		ret2 = nfs_do_read(data, call_ops);
 		if (ret == 0)
 			ret = ret2;
 	}
@@ -373,25 +369,23 @@ out:
 	return ret;
 }
 
-int nfs_generic_pg_readpages(struct nfs_pageio_descriptor *desc)
+int nfs_generic_pagein(struct nfs_pageio_descriptor *desc, struct list_head *head)
+{
+	if (desc->pg_bsize < PAGE_CACHE_SIZE)
+		return nfs_pagein_multi(desc, head);
+	return nfs_pagein_one(desc, head);
+}
+
+static int nfs_generic_pg_readpages(struct nfs_pageio_descriptor *desc)
 {
 	LIST_HEAD(head);
 	int ret;
 
-	if (desc->pg_bsize < PAGE_CACHE_SIZE)
-		ret = nfs_pagein_multi(desc, &head);
-	else
-		ret = nfs_pagein_one(desc, &head);
-
+	ret = nfs_generic_pagein(desc, &head);
 	if (ret == 0)
-		ret = nfs_do_multiple_reads(&head, desc->pg_rpc_callops,
-				desc->pg_lseg);
-	put_lseg(desc->pg_lseg);
-	desc->pg_lseg = NULL;
+		ret = nfs_do_multiple_reads(&head, desc->pg_rpc_callops);
 	return ret;
 }
-EXPORT_SYMBOL_GPL(nfs_generic_pg_readpages);
-
 
 static const struct nfs_pageio_ops nfs_pageio_read_ops = {
 	.pg_test = nfs_generic_pg_test,
diff --git a/include/linux/nfs_page.h b/include/linux/nfs_page.h
index 7241b2a2a4d..0a48f842f83 100644
--- a/include/linux/nfs_page.h
+++ b/include/linux/nfs_page.h
@@ -108,7 +108,6 @@ extern	void nfs_unlock_request(struct nfs_page *req);
 extern	int nfs_set_page_tag_locked(struct nfs_page *req);
 extern  void nfs_clear_page_tag_locked(struct nfs_page *req);
 
-extern	int nfs_generic_pg_readpages(struct nfs_pageio_descriptor *desc);
 extern	int nfs_generic_pg_writepages(struct nfs_pageio_descriptor *desc);
 
 
-- 
cgit v1.2.3-70-g09d2


From dce81290eed64d24493989bb7a08f9e20495e184 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
Date: Wed, 13 Jul 2011 15:59:19 -0400
Subject: NFS: Move the pnfs write code into pnfs.c

...and ensure that we recoalese to take into account differences in
differences in block sizes when falling back to write through the MDS.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/internal.h            |  4 ++++
 fs/nfs/nfs4filelayout.c      |  2 +-
 fs/nfs/objlayout/objio_osd.c |  2 +-
 fs/nfs/pnfs.c                | 57 ++++++++++++++++++++++++++++++++++++++++++--
 fs/nfs/pnfs.h                | 10 +-------
 fs/nfs/write.c               | 39 ++++++++++++++----------------
 include/linux/nfs_page.h     |  3 ---
 7 files changed, 80 insertions(+), 37 deletions(-)

(limited to 'fs/nfs/nfs4filelayout.c')

diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h
index d74d7dea917..09e20de3190 100644
--- a/fs/nfs/internal.h
+++ b/fs/nfs/internal.h
@@ -305,8 +305,12 @@ extern void nfs_pageio_reset_read_mds(struct nfs_pageio_descriptor *pgio);
 extern void nfs_readdata_release(struct nfs_read_data *rdata);
 
 /* write.c */
+extern int nfs_generic_flush(struct nfs_pageio_descriptor *desc,
+		struct list_head *head);
 extern void nfs_pageio_init_write_mds(struct nfs_pageio_descriptor *pgio,
 				  struct inode *inode, int ioflags);
+extern void nfs_pageio_reset_write_mds(struct nfs_pageio_descriptor *pgio);
+extern void nfs_writedata_release(struct nfs_write_data *wdata);
 extern void nfs_commit_free(struct nfs_write_data *p);
 extern int nfs_initiate_write(struct nfs_write_data *data,
 			      struct rpc_clnt *clnt,
diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c
index fc556d6b90f..fbc5b42d197 100644
--- a/fs/nfs/nfs4filelayout.c
+++ b/fs/nfs/nfs4filelayout.c
@@ -741,7 +741,7 @@ static const struct nfs_pageio_ops filelayout_pg_read_ops = {
 static const struct nfs_pageio_ops filelayout_pg_write_ops = {
 	.pg_init = filelayout_pg_init_write,
 	.pg_test = filelayout_pg_test,
-	.pg_doio = nfs_generic_pg_writepages,
+	.pg_doio = pnfs_generic_pg_writepages,
 };
 
 static bool filelayout_mark_pnfs_commit(struct pnfs_layout_segment *lseg)
diff --git a/fs/nfs/objlayout/objio_osd.c b/fs/nfs/objlayout/objio_osd.c
index add62894f36..7d49bb160b4 100644
--- a/fs/nfs/objlayout/objio_osd.c
+++ b/fs/nfs/objlayout/objio_osd.c
@@ -1013,7 +1013,7 @@ static const struct nfs_pageio_ops objio_pg_read_ops = {
 static const struct nfs_pageio_ops objio_pg_write_ops = {
 	.pg_init = pnfs_generic_pg_init_write,
 	.pg_test = objio_pg_test,
-	.pg_doio = nfs_generic_pg_writepages,
+	.pg_doio = pnfs_generic_pg_writepages,
 };
 
 static struct pnfs_layoutdriver_type objlayout_type = {
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 9eca5a8cdbd..93c73299588 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -1170,15 +1170,30 @@ pnfs_ld_write_done(struct nfs_write_data *data)
 }
 EXPORT_SYMBOL_GPL(pnfs_ld_write_done);
 
-enum pnfs_try_status
+static void
+pnfs_write_through_mds(struct nfs_pageio_descriptor *desc,
+		struct nfs_write_data *data)
+{
+	list_splice_tail_init(&data->pages, &desc->pg_list);
+	if (data->req && list_empty(&data->req->wb_list))
+		nfs_list_add_request(data->req, &desc->pg_list);
+	nfs_pageio_reset_write_mds(desc);
+	desc->pg_recoalesce = 1;
+	nfs_writedata_release(data);
+}
+
+static enum pnfs_try_status
 pnfs_try_to_write_data(struct nfs_write_data *wdata,
-			const struct rpc_call_ops *call_ops, int how)
+			const struct rpc_call_ops *call_ops,
+			struct pnfs_layout_segment *lseg,
+			int how)
 {
 	struct inode *inode = wdata->inode;
 	enum pnfs_try_status trypnfs;
 	struct nfs_server *nfss = NFS_SERVER(inode);
 
 	wdata->mds_ops = call_ops;
+	wdata->lseg = get_lseg(lseg);
 
 	dprintk("%s: Writing ino:%lu %u@%llu (how %d)\n", __func__,
 		inode->i_ino, wdata->args.count, wdata->args.offset, how);
@@ -1194,6 +1209,44 @@ pnfs_try_to_write_data(struct nfs_write_data *wdata,
 	return trypnfs;
 }
 
+static void
+pnfs_do_multiple_writes(struct nfs_pageio_descriptor *desc, struct list_head *head, int how)
+{
+	struct nfs_write_data *data;
+	const struct rpc_call_ops *call_ops = desc->pg_rpc_callops;
+	struct pnfs_layout_segment *lseg = desc->pg_lseg;
+
+	desc->pg_lseg = NULL;
+	while (!list_empty(head)) {
+		enum pnfs_try_status trypnfs;
+
+		data = list_entry(head->next, struct nfs_write_data, list);
+		list_del_init(&data->list);
+
+		trypnfs = pnfs_try_to_write_data(data, call_ops, lseg, how);
+		if (trypnfs == PNFS_NOT_ATTEMPTED)
+			pnfs_write_through_mds(desc, data);
+	}
+	put_lseg(lseg);
+}
+
+int
+pnfs_generic_pg_writepages(struct nfs_pageio_descriptor *desc)
+{
+	LIST_HEAD(head);
+	int ret;
+
+	ret = nfs_generic_flush(desc, &head);
+	if (ret != 0) {
+		put_lseg(desc->pg_lseg);
+		desc->pg_lseg = NULL;
+		return ret;
+	}
+	pnfs_do_multiple_writes(desc, &head, desc->pg_ioflags);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(pnfs_generic_pg_writepages);
+
 /*
  * Called by non rpc-based layout drivers
  */
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index c40ffa52c1a..078670dfbe0 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -155,11 +155,10 @@ bool pnfs_pageio_init_write(struct nfs_pageio_descriptor *, struct inode *, int)
 
 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);
 void pnfs_generic_pg_init_read(struct nfs_pageio_descriptor *, struct nfs_page *);
 int pnfs_generic_pg_readpages(struct nfs_pageio_descriptor *desc);
 void pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *, struct nfs_page *);
+int pnfs_generic_pg_writepages(struct nfs_pageio_descriptor *desc);
 bool pnfs_generic_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev, struct nfs_page *req);
 int pnfs_layout_process(struct nfs4_layoutget *lgp);
 void pnfs_free_lseg_list(struct list_head *tmp_list);
@@ -328,13 +327,6 @@ static inline void put_lseg(struct pnfs_layout_segment *lseg)
 {
 }
 
-static inline enum pnfs_try_status
-pnfs_try_to_write_data(struct nfs_write_data *data,
-		       const struct rpc_call_ops *call_ops, int how)
-{
-	return PNFS_NOT_ATTEMPTED;
-}
-
 static inline int pnfs_return_layout(struct inode *ino)
 {
 	return 0;
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index c88a1ab42c3..cabe5f6611b 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -97,7 +97,7 @@ void nfs_writedata_free(struct nfs_write_data *p)
 	mempool_free(p, nfs_wdata_mempool);
 }
 
-static void nfs_writedata_release(struct nfs_write_data *wdata)
+void nfs_writedata_release(struct nfs_write_data *wdata)
 {
 	put_lseg(wdata->lseg);
 	put_nfs_open_context(wdata->args.context);
@@ -887,25 +887,15 @@ static void nfs_write_rpcsetup(struct nfs_page *req,
 
 static int nfs_do_write(struct nfs_write_data *data,
 		const struct rpc_call_ops *call_ops,
-		struct pnfs_layout_segment *lseg,
 		int how)
 {
 	struct inode *inode = data->args.context->path.dentry->d_inode;
 
-	if (lseg != NULL) {
-		data->lseg = get_lseg(lseg);
-		if (pnfs_try_to_write_data(data, call_ops, how) == PNFS_ATTEMPTED)
-			return 0;
-		put_lseg(data->lseg);
-		data->lseg = NULL;
-	}
-
 	return nfs_initiate_write(data, NFS_CLIENT(inode), call_ops, how);
 }
 
 static int nfs_do_multiple_writes(struct list_head *head,
 		const struct rpc_call_ops *call_ops,
-		struct pnfs_layout_segment *lseg,
 		int how)
 {
 	struct nfs_write_data *data;
@@ -917,7 +907,7 @@ static int nfs_do_multiple_writes(struct list_head *head,
 		data = list_entry(head->next, struct nfs_write_data, list);
 		list_del_init(&data->list);
 		
-		ret2 = nfs_do_write(data, call_ops, lseg, how);
+		ret2 = nfs_do_write(data, call_ops, how);
 		 if (ret == 0)
 			 ret = ret2;
 	}
@@ -1037,23 +1027,24 @@ out:
 	return ret;
 }
 
-int nfs_generic_pg_writepages(struct nfs_pageio_descriptor *desc)
+int nfs_generic_flush(struct nfs_pageio_descriptor *desc, struct list_head *head)
+{
+	if (desc->pg_bsize < PAGE_CACHE_SIZE)
+		return nfs_flush_multi(desc, head);
+	return nfs_flush_one(desc, head);
+}
+
+static int nfs_generic_pg_writepages(struct nfs_pageio_descriptor *desc)
 {
 	LIST_HEAD(head);
 	int ret;
 
-	if (desc->pg_bsize < PAGE_CACHE_SIZE)
-		ret = nfs_flush_multi(desc, &head);
-	else
-		ret = nfs_flush_one(desc, &head);
+	ret = nfs_generic_flush(desc, &head);
 	if (ret == 0)
 		ret = nfs_do_multiple_writes(&head, desc->pg_rpc_callops,
-				desc->pg_lseg, desc->pg_ioflags);
-	put_lseg(desc->pg_lseg);
-	desc->pg_lseg = NULL;
+				desc->pg_ioflags);
 	return ret;
 }
-EXPORT_SYMBOL_GPL(nfs_generic_pg_writepages);
 
 static const struct nfs_pageio_ops nfs_pageio_write_ops = {
 	.pg_test = nfs_generic_pg_test,
@@ -1068,6 +1059,12 @@ void nfs_pageio_init_write_mds(struct nfs_pageio_descriptor *pgio,
 }
 EXPORT_SYMBOL_GPL(nfs_pageio_init_write_mds);
 
+void nfs_pageio_reset_write_mds(struct nfs_pageio_descriptor *pgio)
+{
+	pgio->pg_ops = &nfs_pageio_write_ops;
+	pgio->pg_bsize = NFS_SERVER(pgio->pg_inode)->wsize;
+}
+
 static void nfs_pageio_init_write(struct nfs_pageio_descriptor *pgio,
 				  struct inode *inode, int ioflags)
 {
diff --git a/include/linux/nfs_page.h b/include/linux/nfs_page.h
index 0a48f842f83..e2791a27a90 100644
--- a/include/linux/nfs_page.h
+++ b/include/linux/nfs_page.h
@@ -108,9 +108,6 @@ extern	void nfs_unlock_request(struct nfs_page *req);
 extern	int nfs_set_page_tag_locked(struct nfs_page *req);
 extern  void nfs_clear_page_tag_locked(struct nfs_page *req);
 
-extern	int nfs_generic_pg_writepages(struct nfs_pageio_descriptor *desc);
-
-
 /*
  * Lock the page of an asynchronous request without getting a new reference
  */
-- 
cgit v1.2.3-70-g09d2


From 1f9453578f059d2651aa6c6b16756627fc9f2a74 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
Date: Wed, 13 Jul 2011 15:59:57 -0400
Subject: NFS: Clean up - simplify the switch to read/write-through-MDS

Use nfs_pageio_reset_read_mds and nfs_pageio_reset_write_mds instead of
completely reinitialising the struct nfs_pageio_descriptor.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/internal.h       | 4 ----
 fs/nfs/nfs4filelayout.c | 5 ++---
 fs/nfs/pnfs.c           | 4 ++--
 fs/nfs/read.c           | 4 ++--
 fs/nfs/write.c          | 4 ++--
 5 files changed, 8 insertions(+), 13 deletions(-)

(limited to 'fs/nfs/nfs4filelayout.c')

diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h
index 09e20de3190..ab12913dd47 100644
--- a/fs/nfs/internal.h
+++ b/fs/nfs/internal.h
@@ -299,16 +299,12 @@ extern void nfs_read_prepare(struct rpc_task *task, void *calldata);
 extern int nfs_generic_pagein(struct nfs_pageio_descriptor *desc,
 		struct list_head *head);
 
-extern void nfs_pageio_init_read_mds(struct nfs_pageio_descriptor *pgio,
-		struct inode *inode);
 extern void nfs_pageio_reset_read_mds(struct nfs_pageio_descriptor *pgio);
 extern void nfs_readdata_release(struct nfs_read_data *rdata);
 
 /* write.c */
 extern int nfs_generic_flush(struct nfs_pageio_descriptor *desc,
 		struct list_head *head);
-extern void nfs_pageio_init_write_mds(struct nfs_pageio_descriptor *pgio,
-				  struct inode *inode, int ioflags);
 extern void nfs_pageio_reset_write_mds(struct nfs_pageio_descriptor *pgio);
 extern void nfs_writedata_release(struct nfs_write_data *wdata);
 extern void nfs_commit_free(struct nfs_write_data *p);
diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c
index fbc5b42d197..f0b37e155ce 100644
--- a/fs/nfs/nfs4filelayout.c
+++ b/fs/nfs/nfs4filelayout.c
@@ -711,7 +711,7 @@ filelayout_pg_init_read(struct nfs_pageio_descriptor *pgio,
 					   GFP_KERNEL);
 	/* If no lseg, fall back to read through mds */
 	if (pgio->pg_lseg == NULL)
-		nfs_pageio_init_read_mds(pgio, pgio->pg_inode);
+		nfs_pageio_reset_read_mds(pgio);
 }
 
 void
@@ -728,8 +728,7 @@ filelayout_pg_init_write(struct nfs_pageio_descriptor *pgio,
 					   GFP_NOFS);
 	/* If no lseg, fall back to write through mds */
 	if (pgio->pg_lseg == NULL)
-		nfs_pageio_init_write_mds(pgio, pgio->pg_inode,
-					  pgio->pg_ioflags);
+		nfs_pageio_reset_write_mds(pgio);
 }
 
 static const struct nfs_pageio_ops filelayout_pg_read_ops = {
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 93c73299588..38e5508555c 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -1075,7 +1075,7 @@ pnfs_generic_pg_init_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *r
 					   GFP_KERNEL);
 	/* If no lseg, fall back to read through mds */
 	if (pgio->pg_lseg == NULL)
-		nfs_pageio_init_read_mds(pgio, pgio->pg_inode);
+		nfs_pageio_reset_read_mds(pgio);
 
 }
 EXPORT_SYMBOL_GPL(pnfs_generic_pg_init_read);
@@ -1093,7 +1093,7 @@ pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *pgio, struct nfs_page *
 					   GFP_NOFS);
 	/* If no lseg, fall back to write through mds */
 	if (pgio->pg_lseg == NULL)
-		nfs_pageio_init_write_mds(pgio, pgio->pg_inode, pgio->pg_ioflags);
+		nfs_pageio_reset_write_mds(pgio);
 }
 EXPORT_SYMBOL_GPL(pnfs_generic_pg_init_write);
 
diff --git a/fs/nfs/read.c b/fs/nfs/read.c
index d2f53ddd826..7cba2280e2b 100644
--- a/fs/nfs/read.c
+++ b/fs/nfs/read.c
@@ -112,19 +112,19 @@ static void nfs_readpage_truncate_uninitialised_page(struct nfs_read_data *data)
 	}
 }
 
-void nfs_pageio_init_read_mds(struct nfs_pageio_descriptor *pgio,
+static void nfs_pageio_init_read_mds(struct nfs_pageio_descriptor *pgio,
 		struct inode *inode)
 {
 	nfs_pageio_init(pgio, inode, &nfs_pageio_read_ops,
 			NFS_SERVER(inode)->rsize, 0);
 }
-EXPORT_SYMBOL_GPL(nfs_pageio_init_read_mds);
 
 void nfs_pageio_reset_read_mds(struct nfs_pageio_descriptor *pgio)
 {
 	pgio->pg_ops = &nfs_pageio_read_ops;
 	pgio->pg_bsize = NFS_SERVER(pgio->pg_inode)->rsize;
 }
+EXPORT_SYMBOL_GPL(nfs_pageio_reset_read_mds);
 
 static void nfs_pageio_init_read(struct nfs_pageio_descriptor *pgio,
 		struct inode *inode)
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index cabe5f6611b..9fba5270b1a 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -1051,19 +1051,19 @@ static const struct nfs_pageio_ops nfs_pageio_write_ops = {
 	.pg_doio = nfs_generic_pg_writepages,
 };
 
-void nfs_pageio_init_write_mds(struct nfs_pageio_descriptor *pgio,
+static void nfs_pageio_init_write_mds(struct nfs_pageio_descriptor *pgio,
 				  struct inode *inode, int ioflags)
 {
 	nfs_pageio_init(pgio, inode, &nfs_pageio_write_ops,
 				NFS_SERVER(inode)->wsize, ioflags);
 }
-EXPORT_SYMBOL_GPL(nfs_pageio_init_write_mds);
 
 void nfs_pageio_reset_write_mds(struct nfs_pageio_descriptor *pgio)
 {
 	pgio->pg_ops = &nfs_pageio_write_ops;
 	pgio->pg_bsize = NFS_SERVER(pgio->pg_inode)->wsize;
 }
+EXPORT_SYMBOL_GPL(nfs_pageio_reset_write_mds);
 
 static void nfs_pageio_init_write(struct nfs_pageio_descriptor *pgio,
 				  struct inode *inode, int ioflags)
-- 
cgit v1.2.3-70-g09d2


From f85ef69ce08bc2209858135328335f668ba35bdb Mon Sep 17 00:00:00 2001
From: "J. Bruce Fields" <bfields@redhat.com>
Date: Fri, 15 Jul 2011 19:18:42 -0400
Subject: pnfs: simplify pnfs files module autoloading

Embed the necessary alias into the module rather than waiting for
someone to add it to /etc/modprobe.conf

Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/nfs4filelayout.c      | 2 ++
 fs/nfs/objlayout/objio_osd.c | 2 ++
 2 files changed, 4 insertions(+)

(limited to 'fs/nfs/nfs4filelayout.c')

diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c
index f0b37e155ce..be93a622872 100644
--- a/fs/nfs/nfs4filelayout.c
+++ b/fs/nfs/nfs4filelayout.c
@@ -964,5 +964,7 @@ static void __exit nfs4filelayout_exit(void)
 	pnfs_unregister_layoutdriver(&filelayout_type);
 }
 
+MODULE_ALIAS("nfs-layouttype4-1");
+
 module_init(nfs4filelayout_init);
 module_exit(nfs4filelayout_exit);
diff --git a/fs/nfs/objlayout/objio_osd.c b/fs/nfs/objlayout/objio_osd.c
index 7d49bb160b4..9383ca7245b 100644
--- a/fs/nfs/objlayout/objio_osd.c
+++ b/fs/nfs/objlayout/objio_osd.c
@@ -1065,5 +1065,7 @@ objlayout_exit(void)
 	       __func__);
 }
 
+MODULE_ALIAS("nfs-layouttype4-2");
+
 module_init(objlayout_init);
 module_exit(objlayout_exit);
-- 
cgit v1.2.3-70-g09d2


From acff5880539fe33897d016c0f3dcf062e67c61b6 Mon Sep 17 00:00:00 2001
From: Peng Tao <bergwolf@gmail.com>
Date: Sat, 30 Jul 2011 20:52:31 -0400
Subject: pnfs: save layoutcommit lwb at layout header

No need to save it for every lseg.

[Needed in v3.0]
CC: Stable Tree <stable@kernel.org>
Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/nfs4filelayout.c |  2 +-
 fs/nfs/pnfs.c           | 10 ++++++----
 fs/nfs/pnfs.h           |  2 +-
 3 files changed, 8 insertions(+), 6 deletions(-)

(limited to 'fs/nfs/nfs4filelayout.c')

diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c
index be93a622872..e8915d4840a 100644
--- a/fs/nfs/nfs4filelayout.c
+++ b/fs/nfs/nfs4filelayout.c
@@ -170,7 +170,7 @@ filelayout_set_layoutcommit(struct nfs_write_data *wdata)
 
 	pnfs_set_layoutcommit(wdata);
 	dprintk("%s ionde %lu pls_end_pos %lu\n", __func__, wdata->inode->i_ino,
-		(unsigned long) wdata->lseg->pls_end_pos);
+		(unsigned long) NFS_I(wdata->inode)->layout->plh_lwb);
 }
 
 /*
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 38e5508555c..7a948ab4e73 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -1379,9 +1379,11 @@ pnfs_set_layoutcommit(struct nfs_write_data *wdata)
 		dprintk("%s: Set layoutcommit for inode %lu ",
 			__func__, wdata->inode->i_ino);
 	}
-	if (end_pos > wdata->lseg->pls_end_pos)
-		wdata->lseg->pls_end_pos = end_pos;
+	if (end_pos > nfsi->layout->plh_lwb)
+		nfsi->layout->plh_lwb = end_pos;
 	spin_unlock(&nfsi->vfs_inode.i_lock);
+	dprintk("%s: lseg %p end_pos %llu\n",
+		__func__, wdata->lseg, nfsi->layout->plh_lwb);
 
 	/* if pnfs_layoutcommit_inode() runs between inode locks, the next one
 	 * will be a noop because NFS_INO_LAYOUTCOMMIT will not be set */
@@ -1433,9 +1435,9 @@ pnfs_layoutcommit_inode(struct inode *inode, bool sync)
 	 */
 	lseg = pnfs_list_write_lseg(inode);
 
-	end_pos = lseg->pls_end_pos;
+	end_pos = nfsi->layout->plh_lwb;
 	cred = lseg->pls_lc_cred;
-	lseg->pls_end_pos = 0;
+	nfsi->layout->plh_lwb = 0;
 	lseg->pls_lc_cred = NULL;
 
 	memcpy(&data->args.stateid.data, nfsi->layout->plh_stateid.data,
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index 078670dfbe0..1eb2c043d23 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -45,7 +45,6 @@ struct pnfs_layout_segment {
 	unsigned long pls_flags;
 	struct pnfs_layout_hdr *pls_layout;
 	struct rpc_cred	*pls_lc_cred; /* LAYOUTCOMMIT credential */
-	loff_t pls_end_pos; /* LAYOUTCOMMIT write end */
 };
 
 enum pnfs_try_status {
@@ -125,6 +124,7 @@ struct pnfs_layout_hdr {
 	unsigned long		plh_block_lgets; /* block LAYOUTGET if >0 */
 	u32			plh_barrier; /* ignore lower seqids */
 	unsigned long		plh_flags;
+	loff_t			plh_lwb; /* last write byte for layoutcommit */
 	struct inode		*plh_inode;
 };
 
-- 
cgit v1.2.3-70-g09d2