Merge rsync://client.linux-nfs.org/pub/linux/nfs-2.6

fs/Kconfig

@@ -1268,6 +1268,7 @@ config NFS_FS
 	depends on INET
 	select LOCKD
 	select SUNRPC
+	select NFS_ACL_SUPPORT if NFS_V3_ACL
 	help
 	  If you are connected to some other (usually local) Unix computer
 	  (using SLIP, PLIP, PPP or Ethernet) and want to mount files residing
@@ -1310,6 +1311,16 @@ config NFS_V3
 
 	  If unsure, say Y.
 
+config NFS_V3_ACL
+	bool "Provide client support for the NFSv3 ACL protocol extension"
+	depends on NFS_V3
+	help
+	  Implement the NFSv3 ACL protocol extension for manipulating POSIX
+	  Access Control Lists.  The server should also be compiled with
+	  the NFSv3 ACL protocol extension; see the CONFIG_NFSD_V3_ACL option.
+
+	  If unsure, say N.
+
 config NFS_V4
 	bool "Provide NFSv4 client support (EXPERIMENTAL)"
 	depends on NFS_FS && EXPERIMENTAL
@@ -1353,6 +1364,7 @@ config NFSD
 	select LOCKD
 	select SUNRPC
 	select EXPORTFS
+	select NFS_ACL_SUPPORT if NFSD_V3_ACL || NFSD_V2_ACL
 	help
 	  If you want your Linux box to act as an NFS *server*, so that other
 	  computers on your local network which support NFS can access certain
@@ -1376,6 +1388,10 @@ config NFSD
 	  To compile the NFS server support as a module, choose M here: the
 	  module will be called nfsd.  If unsure, say N.
 
+config NFSD_V2_ACL
+	bool
+	depends on NFSD
+
 config NFSD_V3
 	bool "Provide NFSv3 server support"
 	depends on NFSD
@@ -1383,6 +1399,16 @@ config NFSD_V3
 	  If you would like to include the NFSv3 server as well as the NFSv2
 	  server, say Y here.  If unsure, say Y.
 
+config NFSD_V3_ACL
+	bool "Provide server support for the NFSv3 ACL protocol extension"
+	depends on NFSD_V3
+	select NFSD_V2_ACL
+	help
+	  Implement the NFSv3 ACL protocol extension for manipulating POSIX
+	  Access Control Lists on exported file systems. NFS clients should
+	  be compiled with the NFSv3 ACL protocol extension; see the
+	  CONFIG_NFS_V3_ACL option.  If unsure, say N.
+
 config NFSD_V4
 	bool "Provide NFSv4 server support (EXPERIMENTAL)"
 	depends on NFSD_V3 && EXPERIMENTAL
@@ -1427,6 +1453,15 @@ config LOCKD_V4
 config EXPORTFS
 	tristate
 
+config NFS_ACL_SUPPORT
+	tristate
+	select FS_POSIX_ACL
+
+config NFS_COMMON
+	bool
+	depends on NFSD || NFS_FS
+	default y
+
 config SUNRPC
 	tristate
 

fs/Makefile

@@ -31,6 +31,7 @@ obj-$(CONFIG_BINFMT_FLAT)	+= binfmt_flat.o
 
 obj-$(CONFIG_FS_MBCACHE)	+= mbcache.o
 obj-$(CONFIG_FS_POSIX_ACL)	+= posix_acl.o xattr_acl.o
+obj-$(CONFIG_NFS_COMMON)	+= nfs_common/
 
 obj-$(CONFIG_QUOTA)		+= dquot.o
 obj-$(CONFIG_QFMT_V1)		+= quota_v1.o

fs/lockd/clntlock.c

@@ -31,7 +31,7 @@ static int			reclaimer(void *ptr);
  * This is the representation of a blocked client lock.
  */
 struct nlm_wait {
-	struct nlm_wait *	b_next;		/* linked list */
+	struct list_head	b_list;		/* linked list */
 	wait_queue_head_t	b_wait;		/* where to wait on */
 	struct nlm_host *	b_host;
 	struct file_lock *	b_lock;		/* local file lock */
@@ -39,27 +39,54 @@ struct nlm_wait {
 	u32			b_status;	/* grant callback status */
 };
 
-static struct nlm_wait *	nlm_blocked;
+static LIST_HEAD(nlm_blocked);
+
+/*
+ * Queue up a lock for blocking so that the GRANTED request can see it
+ */
+int nlmclnt_prepare_block(struct nlm_rqst *req, struct nlm_host *host, struct file_lock *fl)
+{
+	struct nlm_wait *block;
+
+	BUG_ON(req->a_block != NULL);
+	block = kmalloc(sizeof(*block), GFP_KERNEL);
+	if (block == NULL)
+		return -ENOMEM;
+	block->b_host = host;
+	block->b_lock = fl;
+	init_waitqueue_head(&block->b_wait);
+	block->b_status = NLM_LCK_BLOCKED;
+
+	list_add(&block->b_list, &nlm_blocked);
+	req->a_block = block;
+
+	return 0;
+}
+
+void nlmclnt_finish_block(struct nlm_rqst *req)
+{
+	struct nlm_wait *block = req->a_block;
+
+	if (block == NULL)
+		return;
+	req->a_block = NULL;
+	list_del(&block->b_list);
+	kfree(block);
+}
 
 /*
  * Block on a lock
  */
-int
+long nlmclnt_block(struct nlm_rqst *req, long timeout)
-nlmclnt_block(struct nlm_host *host, struct file_lock *fl, u32 *statp)
 {
-	struct nlm_wait	block, **head;
+	struct nlm_wait	*block = req->a_block;
-	int		err;
+	long ret;
-	u32		pstate;
 
-	block.b_host   = host;
+	/* A borken server might ask us to block even if we didn't
-	block.b_lock   = fl;
+	 * request it. Just say no!
-	init_waitqueue_head(&block.b_wait);
+	 */
-	block.b_status = NLM_LCK_BLOCKED;
+	if (!req->a_args.block)
-	block.b_next   = nlm_blocked;
+		return -EAGAIN;
-	nlm_blocked    = █
-
-	/* Remember pseudo nsm state */
-	pstate = host->h_state;
 
 	/* Go to sleep waiting for GRANT callback. Some servers seem
 	 * to lose callbacks, however, so we're going to poll from
@@ -69,28 +96,16 @@ nlmclnt_block(struct nlm_host *host, struct file_lock *fl, u32 *statp)
 	 * a 1 minute timeout would do. See the comment before
 	 * nlmclnt_lock for an explanation.
 	 */
-	sleep_on_timeout(&block.b_wait, 30*HZ);
+	ret = wait_event_interruptible_timeout(block->b_wait,
+			block->b_status != NLM_LCK_BLOCKED,
+			timeout);
 
-	for (head = &nlm_blocked; *head; head = &(*head)->b_next) {
+	if (block->b_status != NLM_LCK_BLOCKED) {
-		if (*head == &block) {
+		req->a_res.status = block->b_status;
-			*head = block.b_next;
+		block->b_status = NLM_LCK_BLOCKED;
-			break;
-		}
 	}
 
-	if (!signalled()) {
+	return ret;
-		*statp = block.b_status;
-		return 0;
-	}
-
-	/* Okay, we were interrupted. Cancel the pending request
-	 * unless the server has rebooted.
-	 */
-	if (pstate == host->h_state && (err = nlmclnt_cancel(host, fl)) < 0)
-		printk(KERN_NOTICE
-			"lockd: CANCEL call failed (errno %d)\n", -err);
-
-	return -ERESTARTSYS;
 }
 
 /*
@@ -100,27 +115,23 @@ u32
 nlmclnt_grant(struct nlm_lock *lock)
 {
 	struct nlm_wait	*block;
+	u32 res = nlm_lck_denied;
 
 	/*
 	 * Look up blocked request based on arguments. 
 	 * Warning: must not use cookie to match it!
 	 */
-	for (block = nlm_blocked; block; block = block->b_next) {
+	list_for_each_entry(block, &nlm_blocked, b_list) {
-		if (nlm_compare_locks(block->b_lock, &lock->fl))
+		if (nlm_compare_locks(block->b_lock, &lock->fl)) {
-			break;
+			/* Alright, we found a lock. Set the return status
-	}
+			 * and wake up the caller
-
-	/* Ooops, no blocked request found. */
-	if (block == NULL)
-		return nlm_lck_denied;
-
-	/* Alright, we found the lock. Set the return status and
-	 * wake up the caller.
 			 */
 			block->b_status = NLM_LCK_GRANTED;
 			wake_up(&block->b_wait);
-
+			res = nlm_granted;
-	return nlm_granted;
+		}
+	}
+	return res;
 }
 
 /*
@@ -230,7 +241,7 @@ reclaimer(void *ptr)
 	host->h_reclaiming = 0;
 
 	/* Now, wake up all processes that sleep on a blocked lock */
-	for (block = nlm_blocked; block; block = block->b_next) {
+	list_for_each_entry(block, &nlm_blocked, b_list) {
 		if (block->b_host == host) {
 			block->b_status = NLM_LCK_DENIED_GRACE_PERIOD;
 			wake_up(&block->b_wait);

fs/lockd/clntproc.c

@@ -21,6 +21,7 @@
 
 #define NLMDBG_FACILITY		NLMDBG_CLIENT
 #define NLMCLNT_GRACE_WAIT	(5*HZ)
+#define NLMCLNT_POLL_TIMEOUT	(30*HZ)
 
 static int	nlmclnt_test(struct nlm_rqst *, struct file_lock *);
 static int	nlmclnt_lock(struct nlm_rqst *, struct file_lock *);
@@ -553,6 +554,7 @@ nlmclnt_lock(struct nlm_rqst *req, struct file_lock *fl)
 {
 	struct nlm_host	*host = req->a_host;
 	struct nlm_res	*resp = &req->a_res;
+	long timeout;
 	int status;
 
 	if (!host->h_monitored && nsm_monitor(host) < 0) {
@@ -562,15 +564,32 @@ nlmclnt_lock(struct nlm_rqst *req, struct file_lock *fl)
 		goto out;
 	}
 
-	do {
+	if (req->a_args.block) {
-		if ((status = nlmclnt_call(req, NLMPROC_LOCK)) >= 0) {
+		status = nlmclnt_prepare_block(req, host, fl);
-			if (resp->status != NLM_LCK_BLOCKED)
-				break;
-			status = nlmclnt_block(host, fl, &resp->status);
-		}
 		if (status < 0)
 			goto out;
-	} while (resp->status == NLM_LCK_BLOCKED && req->a_args.block);
+	}
+	for(;;) {
+		status = nlmclnt_call(req, NLMPROC_LOCK);
+		if (status < 0)
+			goto out_unblock;
+		if (resp->status != NLM_LCK_BLOCKED)
+			break;
+		/* Wait on an NLM blocking lock */
+		timeout = nlmclnt_block(req, NLMCLNT_POLL_TIMEOUT);
+		/* Did a reclaimer thread notify us of a server reboot? */
+		if (resp->status ==  NLM_LCK_DENIED_GRACE_PERIOD)
+			continue;
+		if (resp->status != NLM_LCK_BLOCKED)
+			break;
+		if (timeout >= 0)
+			continue;
+		/* We were interrupted. Send a CANCEL request to the server
+		 * and exit
+		 */
+		status = (int)timeout;
+		goto out_unblock;
+	}
 
 	if (resp->status == NLM_LCK_GRANTED) {
 		fl->fl_u.nfs_fl.state = host->h_state;
@@ -579,6 +598,11 @@ nlmclnt_lock(struct nlm_rqst *req, struct file_lock *fl)
 		do_vfs_lock(fl);
 	}
 	status = nlm_stat_to_errno(resp->status);
+out_unblock:
+	nlmclnt_finish_block(req);
+	/* Cancel the blocked request if it is still pending */
+	if (resp->status == NLM_LCK_BLOCKED)
+		nlmclnt_cancel(host, fl);
 out:
 	nlmclnt_release_lockargs(req);
 	return status;

fs/lockd/host.c

@@ -189,17 +189,15 @@ nlm_bind_host(struct nlm_host *host)
 			goto forgetit;
 
 		xprt_set_timeout(&xprt->timeout, 5, nlmsvc_timeout);
+		xprt->nocong = 1;	/* No congestion control for NLM */
+		xprt->resvport = 1;	/* NLM requires a reserved port */
 
 		/* Existing NLM servers accept AUTH_UNIX only */
 		clnt = rpc_create_client(xprt, host->h_name, &nlm_program,
 					host->h_version, RPC_AUTH_UNIX);
-		if (IS_ERR(clnt)) {
+		if (IS_ERR(clnt))
-			xprt_destroy(xprt);
 			goto forgetit;
-		}
 		clnt->cl_autobind = 1;	/* turn on pmap queries */
-		xprt->nocong = 1;	/* No congestion control for NLM */
-		xprt->resvport = 1;	/* NLM requires a reserved port */
 
 		host->h_rpcclnt = clnt;
 	}

fs/lockd/mon.c

@@ -115,20 +115,19 @@ nsm_create(void)
 	xprt = xprt_create_proto(IPPROTO_UDP, &sin, NULL);
 	if (IS_ERR(xprt))
 		return (struct rpc_clnt *)xprt;
+	xprt->resvport = 1;	/* NSM requires a reserved port */
 
 	clnt = rpc_create_client(xprt, "localhost",
 				&nsm_program, SM_VERSION,
 				RPC_AUTH_NULL);
 	if (IS_ERR(clnt))
-		goto out_destroy;
+		goto out_err;
 	clnt->cl_softrtry = 1;
 	clnt->cl_chatty   = 1;
 	clnt->cl_oneshot  = 1;
-	xprt->resvport = 1;	/* NSM requires a reserved port */
 	return clnt;
 
-out_destroy:
+out_err:
-	xprt_destroy(xprt);
 	return clnt;
 }
 

fs/locks.c

@@ -1548,6 +1548,8 @@ int fcntl_getlk(struct file *filp, struct flock __user *l)
 
 	if (filp->f_op && filp->f_op->lock) {
 		error = filp->f_op->lock(filp, F_GETLK, &file_lock);
+		if (file_lock.fl_ops && file_lock.fl_ops->fl_release_private)
+			file_lock.fl_ops->fl_release_private(&file_lock);
 		if (error < 0)
 			goto out;
 		else
@@ -1690,6 +1692,8 @@ int fcntl_getlk64(struct file *filp, struct flock64 __user *l)
 
 	if (filp->f_op && filp->f_op->lock) {
 		error = filp->f_op->lock(filp, F_GETLK, &file_lock);
+		if (file_lock.fl_ops && file_lock.fl_ops->fl_release_private)
+			file_lock.fl_ops->fl_release_private(&file_lock);
 		if (error < 0)
 			goto out;
 		else
@@ -1873,6 +1877,8 @@ void locks_remove_flock(struct file *filp)
 			.fl_end = OFFSET_MAX,
 		};
 		filp->f_op->flock(filp, F_SETLKW, &fl);
+		if (fl.fl_ops && fl.fl_ops->fl_release_private)
+			fl.fl_ops->fl_release_private(&fl);
 	}
 
 	lock_kernel();

fs/nfs/Makefile

@@ -8,6 +8,7 @@ nfs-y 			:= dir.o file.o inode.o nfs2xdr.o pagelist.o \
 			   proc.o read.o symlink.o unlink.o write.o
 nfs-$(CONFIG_ROOT_NFS)	+= nfsroot.o mount_clnt.o      
 nfs-$(CONFIG_NFS_V3)	+= nfs3proc.o nfs3xdr.o
+nfs-$(CONFIG_NFS_V3_ACL)	+= nfs3acl.o
 nfs-$(CONFIG_NFS_V4)	+= nfs4proc.o nfs4xdr.o nfs4state.o nfs4renewd.o \
 			   delegation.o idmap.o \
 			   callback.o callback_xdr.o callback_proc.o

fs/nfs/callback.c

@@ -14,6 +14,7 @@
 #include <linux/sunrpc/svc.h>
 #include <linux/sunrpc/svcsock.h>
 #include <linux/nfs_fs.h>
+#include "nfs4_fs.h"
 #include "callback.h"
 
 #define NFSDBG_FACILITY NFSDBG_CALLBACK

fs/nfs/callback_proc.c

@@ -8,6 +8,7 @@
 #include <linux/config.h>
 #include <linux/nfs4.h>
 #include <linux/nfs_fs.h>
+#include "nfs4_fs.h"
 #include "callback.h"
 #include "delegation.h"
 

fs/nfs/callback_xdr.c

@@ -10,6 +10,7 @@
 #include <linux/sunrpc/svc.h>
 #include <linux/nfs4.h>
 #include <linux/nfs_fs.h>
+#include "nfs4_fs.h"
 #include "callback.h"
 
 #define CB_OP_TAGLEN_MAXSZ	(512)
@@ -410,7 +411,6 @@ static int nfs4_callback_compound(struct svc_rqst *rqstp, void *argp, void *resp
 	xdr_init_decode(&xdr_in, &rqstp->rq_arg, rqstp->rq_arg.head[0].iov_base);
 
 	p = (uint32_t*)((char *)rqstp->rq_res.head[0].iov_base + rqstp->rq_res.head[0].iov_len);
-	rqstp->rq_res.head[0].iov_len = PAGE_SIZE;
 	xdr_init_encode(&xdr_out, &rqstp->rq_res, p);
 
 	decode_compound_hdr_arg(&xdr_in, &hdr_arg);

fs/nfs/delegation.c

@@ -16,6 +16,7 @@
 #include <linux/nfs_fs.h>
 #include <linux/nfs_xdr.h>
 
+#include "nfs4_fs.h"
 #include "delegation.h"
 
 static struct nfs_delegation *nfs_alloc_delegation(void)

fs/nfs/dir.c

@@ -32,6 +32,7 @@
 #include <linux/smp_lock.h>
 #include <linux/namei.h>
 
+#include "nfs4_fs.h"
 #include "delegation.h"
 
 #define NFS_PARANOIA 1
@@ -50,8 +51,10 @@ static int nfs_mknod(struct inode *, struct dentry *, int, dev_t);
 static int nfs_rename(struct inode *, struct dentry *,
 		      struct inode *, struct dentry *);
 static int nfs_fsync_dir(struct file *, struct dentry *, int);
+static loff_t nfs_llseek_dir(struct file *, loff_t, int);
 
 struct file_operations nfs_dir_operations = {
+	.llseek		= nfs_llseek_dir,
 	.read		= generic_read_dir,
 	.readdir	= nfs_readdir,
 	.open		= nfs_opendir,
@@ -74,6 +77,27 @@ struct inode_operations nfs_dir_inode_operations = {
 	.setattr	= nfs_setattr,
 };
 
+#ifdef CONFIG_NFS_V3
+struct inode_operations nfs3_dir_inode_operations = {
+	.create		= nfs_create,
+	.lookup		= nfs_lookup,
+	.link		= nfs_link,
+	.unlink		= nfs_unlink,
+	.symlink	= nfs_symlink,
+	.mkdir		= nfs_mkdir,
+	.rmdir		= nfs_rmdir,
+	.mknod		= nfs_mknod,
+	.rename		= nfs_rename,
+	.permission	= nfs_permission,
+	.getattr	= nfs_getattr,
+	.setattr	= nfs_setattr,
+	.listxattr	= nfs3_listxattr,
+	.getxattr	= nfs3_getxattr,
+	.setxattr	= nfs3_setxattr,
+	.removexattr	= nfs3_removexattr,
+};
+#endif  /* CONFIG_NFS_V3 */
+
 #ifdef CONFIG_NFS_V4
 
 static struct dentry *nfs_atomic_lookup(struct inode *, struct dentry *, struct nameidata *);
@@ -90,6 +114,9 @@ struct inode_operations nfs4_dir_inode_operations = {
 	.permission	= nfs_permission,
 	.getattr	= nfs_getattr,
 	.setattr	= nfs_setattr,
+	.getxattr       = nfs4_getxattr,
+	.setxattr       = nfs4_setxattr,
+	.listxattr      = nfs4_listxattr,
 };
 
 #endif /* CONFIG_NFS_V4 */
@@ -116,7 +143,8 @@ typedef struct {
 	struct page	*page;
 	unsigned long	page_index;
 	u32		*ptr;
-	u64		target;
+	u64		*dir_cookie;
+	loff_t		current_index;
 	struct nfs_entry *entry;
 	decode_dirent_t	decode;
 	int		plus;
@@ -164,12 +192,10 @@ int nfs_readdir_filler(nfs_readdir_descriptor_t *desc, struct page *page)
 	NFS_FLAGS(inode) |= NFS_INO_INVALID_ATIME;
 	/* Ensure consistent page alignment of the data.
 	 * Note: assumes we have exclusive access to this mapping either
-	 *	 throught inode->i_sem or some other mechanism.
+	 *	 through inode->i_sem or some other mechanism.
 	 */
-	if (page->index == 0) {
+	if (page->index == 0)
-		invalidate_inode_pages(inode->i_mapping);
+		invalidate_inode_pages2_range(inode->i_mapping, PAGE_CACHE_SIZE, -1);
-		NFS_I(inode)->readdir_timestamp = timestamp;
-	}
 	unlock_page(page);
 	return 0;
  error:
@@ -202,22 +228,22 @@ void dir_page_release(nfs_readdir_descriptor_t *desc)
 
 /*
  * Given a pointer to a buffer that has already been filled by a call
- * to readdir, find the next entry.
+ * to readdir, find the next entry with cookie '*desc->dir_cookie'.
  *
  * If the end of the buffer has been reached, return -EAGAIN, if not,
  * return the offset within the buffer of the next entry to be
  * read.
  */
 static inline
-int find_dirent(nfs_readdir_descriptor_t *desc, struct page *page)
+int find_dirent(nfs_readdir_descriptor_t *desc)
 {
 	struct nfs_entry *entry = desc->entry;
 	int		loop_count = 0,
 			status;
 
 	while((status = dir_decode(desc)) == 0) {
-		dfprintk(VFS, "NFS: found cookie %Lu\n", (long long)entry->cookie);
+		dfprintk(VFS, "NFS: found cookie %Lu\n", (unsigned long long)entry->cookie);
-		if (entry->prev_cookie == desc->target)
+		if (entry->prev_cookie == *desc->dir_cookie)
 			break;
 		if (loop_count++ > 200) {
 			loop_count = 0;
@@ -229,8 +255,44 @@ int find_dirent(nfs_readdir_descriptor_t *desc, struct page *page)
 }
 
 /*
- * Find the given page, and call find_dirent() in order to try to
+ * Given a pointer to a buffer that has already been filled by a call
- * return the next entry.
+ * to readdir, find the entry at offset 'desc->file->f_pos'.
+ *
+ * If the end of the buffer has been reached, return -EAGAIN, if not,
+ * return the offset within the buffer of the next entry to be
+ * read.
+ */
+static inline
+int find_dirent_index(nfs_readdir_descriptor_t *desc)
+{
+	struct nfs_entry *entry = desc->entry;
+	int		loop_count = 0,
+			status;
+
+	for(;;) {
+		status = dir_decode(desc);
+		if (status)
+			break;
+
+		dfprintk(VFS, "NFS: found cookie %Lu at index %Ld\n", (unsigned long long)entry->cookie, desc->current_index);
+
+		if (desc->file->f_pos == desc->current_index) {
+			*desc->dir_cookie = entry->cookie;
+			break;
+		}
+		desc->current_index++;
+		if (loop_count++ > 200) {
+			loop_count = 0;
+			schedule();
+		}
+	}
+	dfprintk(VFS, "NFS: find_dirent_index() returns %d\n", status);
+	return status;
+}
+
+/*
+ * Find the given page, and call find_dirent() or find_dirent_index in
+ * order to try to return the next entry.
  */
 static inline
 int find_dirent_page(nfs_readdir_descriptor_t *desc)
@@ -253,7 +315,10 @@ int find_dirent_page(nfs_readdir_descriptor_t *desc)
 	/* NOTE: Someone else may have changed the READDIRPLUS flag */
 	desc->page = page;
 	desc->ptr = kmap(page);		/* matching kunmap in nfs_do_filldir */
-	status = find_dirent(desc, page);
+	if (*desc->dir_cookie != 0)
+		status = find_dirent(desc);
+	else
+		status = find_dirent_index(desc);
 	if (status < 0)
 		dir_page_release(desc);
  out:
@@ -268,7 +333,8 @@ int find_dirent_page(nfs_readdir_descriptor_t *desc)
  * Recurse through the page cache pages, and return a
  * filled nfs_entry structure of the next directory entry if possible.
  *
- * The target for the search is 'desc->target'.
+ * The target for the search is '*desc->dir_cookie' if non-0,
+ * 'desc->file->f_pos' otherwise
  */
 static inline
 int readdir_search_pagecache(nfs_readdir_descriptor_t *desc)
@@ -276,7 +342,16 @@ int readdir_search_pagecache(nfs_readdir_descriptor_t *desc)
 	int		loop_count = 0;
 	int		res;
 
-	dfprintk(VFS, "NFS: readdir_search_pagecache() searching for cookie %Lu\n", (long long)desc->target);
+	/* Always search-by-index from the beginning of the cache */
+	if (*desc->dir_cookie == 0) {
+		dfprintk(VFS, "NFS: readdir_search_pagecache() searching for offset %Ld\n", (long long)desc->file->f_pos);
+		desc->page_index = 0;
+		desc->entry->cookie = desc->entry->prev_cookie = 0;
+		desc->entry->eof = 0;
+		desc->current_index = 0;
+	} else
+		dfprintk(VFS, "NFS: readdir_search_pagecache() searching for cookie %Lu\n", (unsigned long long)*desc->dir_cookie);
+
 	for (;;) {
 		res = find_dirent_page(desc);
 		if (res != -EAGAIN)
@@ -313,7 +388,7 @@ int nfs_do_filldir(nfs_readdir_descriptor_t *desc, void *dirent,
 	int		loop_count = 0,
 			res;
 
-	dfprintk(VFS, "NFS: nfs_do_filldir() filling starting @ cookie %Lu\n", (long long)desc->target);
+	dfprintk(VFS, "NFS: nfs_do_filldir() filling starting @ cookie %Lu\n", (long long)entry->cookie);
 
 	for(;;) {
 		unsigned d_type = DT_UNKNOWN;
@@ -333,10 +408,11 @@ int nfs_do_filldir(nfs_readdir_descriptor_t *desc, void *dirent,
 		}
 
 		res = filldir(dirent, entry->name, entry->len, 
-			      entry->prev_cookie, fileid, d_type);
+			      file->f_pos, fileid, d_type);
 		if (res < 0)
 			break;
-		file->f_pos = desc->target = entry->cookie;
+		file->f_pos++;
+		*desc->dir_cookie = entry->cookie;
 		if (dir_decode(desc) != 0) {
 			desc->page_index ++;
 			break;
@@ -349,7 +425,7 @@ int nfs_do_filldir(nfs_readdir_descriptor_t *desc, void *dirent,
 	dir_page_release(desc);
 	if (dentry != NULL)
 		dput(dentry);
-	dfprintk(VFS, "NFS: nfs_do_filldir() filling ended @ cookie %Lu; returning = %d\n", (long long)desc->target, res);
+	dfprintk(VFS, "NFS: nfs_do_filldir() filling ended @ cookie %Lu; returning = %d\n", (unsigned long long)*desc->dir_cookie, res);
 	return res;
 }
 
@@ -375,14 +451,14 @@ int uncached_readdir(nfs_readdir_descriptor_t *desc, void *dirent,
 	struct page	*page = NULL;
 	int		status;
 
-	dfprintk(VFS, "NFS: uncached_readdir() searching for cookie %Lu\n", (long long)desc->target);
+	dfprintk(VFS, "NFS: uncached_readdir() searching for cookie %Lu\n", (unsigned long long)*desc->dir_cookie);
 
 	page = alloc_page(GFP_HIGHUSER);
 	if (!page) {
 		status = -ENOMEM;
 		goto out;
 	}
-	desc->error = NFS_PROTO(inode)->readdir(file->f_dentry, cred, desc->target,
+	desc->error = NFS_PROTO(inode)->readdir(file->f_dentry, cred, *desc->dir_cookie,
 						page,
 						NFS_SERVER(inode)->dtsize,
 						desc->plus);
@@ -391,7 +467,7 @@ int uncached_readdir(nfs_readdir_descriptor_t *desc, void *dirent,
 	desc->ptr = kmap(page);		/* matching kunmap in nfs_do_filldir */
 	if (desc->error >= 0) {
 		if ((status = dir_decode(desc)) == 0)
-			desc->entry->prev_cookie = desc->target;
+			desc->entry->prev_cookie = *desc->dir_cookie;
 	} else
 		status = -EIO;
 	if (status < 0)
@@ -412,8 +488,9 @@ int uncached_readdir(nfs_readdir_descriptor_t *desc, void *dirent,
 	goto out;
 }
 
-/* The file offset position is now represented as a true offset into the
+/* The file offset position represents the dirent entry number.  A
- * page cache as is the case in most of the other filesystems.
+   last cookie cache takes care of the common case of reading the
+   whole directory.
  */
 static int nfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 {
@@ -435,15 +512,15 @@ static int nfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 	}
 
 	/*
-	 * filp->f_pos points to the file offset in the page cache.
+	 * filp->f_pos points to the dirent entry number.
-	 * but if the cache has meanwhile been zapped, we need to
+	 * *desc->dir_cookie has the cookie for the next entry. We have
-	 * read from the last dirent to revalidate f_pos
+	 * to either find the entry with the appropriate number or
-	 * itself.
+	 * revalidate the cookie.
 	 */
 	memset(desc, 0, sizeof(*desc));
 
 	desc->file = filp;
-	desc->target = filp->f_pos;
+	desc->dir_cookie = &((struct nfs_open_context *)filp->private_data)->dir_cookie;
 	desc->decode = NFS_PROTO(inode)->decode_dirent;
 	desc->plus = NFS_USE_READDIRPLUS(inode);
 
@@ -455,9 +532,10 @@ static int nfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 
 	while(!desc->entry->eof) {
 		res = readdir_search_pagecache(desc);
+
 		if (res == -EBADCOOKIE) {
 			/* This means either end of directory */
-			if (desc->entry->cookie != desc->target) {
+			if (*desc->dir_cookie && desc->entry->cookie != *desc->dir_cookie) {
 				/* Or that the server has 'lost' a cookie */
 				res = uncached_readdir(desc, dirent, filldir);
 				if (res >= 0)
@@ -490,6 +568,28 @@ static int nfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 	return 0;
 }
 
+loff_t nfs_llseek_dir(struct file *filp, loff_t offset, int origin)
+{
+	down(&filp->f_dentry->d_inode->i_sem);
+	switch (origin) {
+		case 1:
+			offset += filp->f_pos;
+		case 0:
+			if (offset >= 0)
+				break;
+		default:
+			offset = -EINVAL;
+			goto out;
+	}
+	if (offset != filp->f_pos) {
+		filp->f_pos = offset;
+		((struct nfs_open_context *)filp->private_data)->dir_cookie = 0;
+	}
+out:
+	up(&filp->f_dentry->d_inode->i_sem);
+	return offset;
+}
+
 /*
  * All directory operations under NFS are synchronous, so fsync()
  * is a dummy operation.

fs/nfs/direct.c

@@ -517,7 +517,7 @@ static ssize_t nfs_direct_write_seg(struct inode *inode,
 	result = tot_bytes;
 
 out:
-	nfs_end_data_update_defer(inode);
+	nfs_end_data_update(inode);
 	nfs_writedata_free(wdata);
 	return result;
 

fs/nfs/file.c

@@ -71,6 +71,18 @@ struct inode_operations nfs_file_inode_operations = {
 	.setattr	= nfs_setattr,
 };
 
+#ifdef CONFIG_NFS_V3
+struct inode_operations nfs3_file_inode_operations = {
+	.permission	= nfs_permission,
+	.getattr	= nfs_getattr,
+	.setattr	= nfs_setattr,
+	.listxattr	= nfs3_listxattr,
+	.getxattr	= nfs3_getxattr,
+	.setxattr	= nfs3_setxattr,
+	.removexattr	= nfs3_removexattr,
+};
+#endif  /* CONFIG_NFS_v3 */
+
 /* Hack for future NFS swap support */
 #ifndef IS_SWAPFILE
 # define IS_SWAPFILE(inode)	(0)
@@ -115,6 +127,21 @@ nfs_file_release(struct inode *inode, struct file *filp)
 	return NFS_PROTO(inode)->file_release(inode, filp);
 }
 
+/**
+ * nfs_revalidate_file - Revalidate the page cache & related metadata
+ * @inode - pointer to inode struct
+ * @file - pointer to file
+ */
+static int nfs_revalidate_file(struct inode *inode, struct file *filp)
+{
+	int retval = 0;
+
+	if ((NFS_FLAGS(inode) & NFS_INO_REVAL_PAGECACHE) || nfs_attribute_timeout(inode))
+		retval = __nfs_revalidate_inode(NFS_SERVER(inode), inode);
+	nfs_revalidate_mapping(inode, filp->f_mapping);
+	return 0;
+}
+
 /**
  * nfs_revalidate_size - Revalidate the file size
  * @inode - pointer to inode struct
@@ -137,7 +164,8 @@ static int nfs_revalidate_file_size(struct inode *inode, struct file *filp)
 		goto force_reval;
 	if (nfsi->npages != 0)
 		return 0;
-	return nfs_revalidate_inode(server, inode);
+	if (!(NFS_FLAGS(inode) & NFS_INO_REVAL_PAGECACHE) && !nfs_attribute_timeout(inode))
+		return 0;
 force_reval:
 	return __nfs_revalidate_inode(server, inode);
 }
@@ -198,7 +226,7 @@ nfs_file_read(struct kiocb *iocb, char __user * buf, size_t count, loff_t pos)
 		dentry->d_parent->d_name.name, dentry->d_name.name,
 		(unsigned long) count, (unsigned long) pos);
 
-	result = nfs_revalidate_inode(NFS_SERVER(inode), inode);
+	result = nfs_revalidate_file(inode, iocb->ki_filp);
 	if (!result)
 		result = generic_file_aio_read(iocb, buf, count, pos);
 	return result;
@@ -216,7 +244,7 @@ nfs_file_sendfile(struct file *filp, loff_t *ppos, size_t count,
 		dentry->d_parent->d_name.name, dentry->d_name.name,
 		(unsigned long) count, (unsigned long long) *ppos);
 
-	res = nfs_revalidate_inode(NFS_SERVER(inode), inode);
+	res = nfs_revalidate_file(inode, filp);
 	if (!res)
 		res = generic_file_sendfile(filp, ppos, count, actor, target);
 	return res;
@@ -232,7 +260,7 @@ nfs_file_mmap(struct file * file, struct vm_area_struct * vma)
 	dfprintk(VFS, "nfs: mmap(%s/%s)\n",
 		dentry->d_parent->d_name.name, dentry->d_name.name);
 
-	status = nfs_revalidate_inode(NFS_SERVER(inode), inode);
+	status = nfs_revalidate_file(inode, file);
 	if (!status)
 		status = generic_file_mmap(file, vma);
 	return status;
@@ -321,9 +349,15 @@ nfs_file_write(struct kiocb *iocb, const char __user *buf, size_t count, loff_t
 	result = -EBUSY;
 	if (IS_SWAPFILE(inode))
 		goto out_swapfile;
-	result = nfs_revalidate_inode(NFS_SERVER(inode), inode);
+	/*
+	 * O_APPEND implies that we must revalidate the file length.
+	 */
+	if (iocb->ki_filp->f_flags & O_APPEND) {
+		result = nfs_revalidate_file_size(inode, iocb->ki_filp);
 		if (result)
 			goto out;
+	}
+	nfs_revalidate_mapping(inode, iocb->ki_filp->f_mapping);
 
 	result = count;
 	if (!count)

fs/nfs/idmap.c

@@ -50,6 +50,7 @@
 #include <linux/nfs_fs.h>
 
 #include <linux/nfs_idmap.h>
+#include "nfs4_fs.h"
 
 #define IDMAP_HASH_SZ          128
 

fs/nfs/inode.c

@@ -39,6 +39,7 @@
 #include <asm/system.h>
 #include <asm/uaccess.h>
 
+#include "nfs4_fs.h"
 #include "delegation.h"
 
 #define NFSDBG_FACILITY		NFSDBG_VFS
@@ -63,6 +64,7 @@ static void nfs_clear_inode(struct inode *);
 static void nfs_umount_begin(struct super_block *);
 static int  nfs_statfs(struct super_block *, struct kstatfs *);
 static int  nfs_show_options(struct seq_file *, struct vfsmount *);
+static void nfs_zap_acl_cache(struct inode *);
 
 static struct rpc_program	nfs_program;
 
@@ -106,6 +108,21 @@ static struct rpc_program	nfs_program = {
 	.pipe_dir_name		= "/nfs",
 };
 
+#ifdef CONFIG_NFS_V3_ACL
+static struct rpc_stat		nfsacl_rpcstat = { &nfsacl_program };
+static struct rpc_version *	nfsacl_version[] = {
+	[3]			= &nfsacl_version3,
+};
+
+struct rpc_program		nfsacl_program = {
+	.name =			"nfsacl",
+	.number =		NFS_ACL_PROGRAM,
+	.nrvers =		sizeof(nfsacl_version) / sizeof(nfsacl_version[0]),
+	.version =		nfsacl_version,
+	.stats =		&nfsacl_rpcstat,
+};
+#endif  /* CONFIG_NFS_V3_ACL */
+
 static inline unsigned long
 nfs_fattr_to_ino_t(struct nfs_fattr *fattr)
 {
@@ -118,7 +135,7 @@ nfs_write_inode(struct inode *inode, int sync)
 	int flags = sync ? FLUSH_WAIT : 0;
 	int ret;
 
-	ret = nfs_commit_inode(inode, 0, 0, flags);
+	ret = nfs_commit_inode(inode, flags);
 	if (ret < 0)
 		return ret;
 	return 0;
@@ -140,10 +157,6 @@ nfs_delete_inode(struct inode * inode)
 	clear_inode(inode);
 }
 
-/*
- * For the moment, the only task for the NFS clear_inode method is to
- * release the mmap credential
- */
 static void
 nfs_clear_inode(struct inode *inode)
 {
@@ -152,6 +165,7 @@ nfs_clear_inode(struct inode *inode)
 
 	nfs_wb_all(inode);
 	BUG_ON (!list_empty(&nfsi->open_files));
+	nfs_zap_acl_cache(inode);
 	cred = nfsi->cache_access.cred;
 	if (cred)
 		put_rpccred(cred);
@@ -161,11 +175,13 @@ nfs_clear_inode(struct inode *inode)
 void
 nfs_umount_begin(struct super_block *sb)
 {
-	struct nfs_server *server = NFS_SB(sb);
+	struct rpc_clnt	*rpc = NFS_SB(sb)->client;
-	struct rpc_clnt	*rpc;
 
 	/* -EIO all pending I/O */
-	if ((rpc = server->client) != NULL)
+	if (!IS_ERR(rpc))
+		rpc_killall_tasks(rpc);
+	rpc = NFS_SB(sb)->client_acl;
+	if (!IS_ERR(rpc))
 		rpc_killall_tasks(rpc);
 }
 
@@ -366,13 +382,15 @@ nfs_create_client(struct nfs_server *server, const struct nfs_mount_data *data)
 	xprt = xprt_create_proto(tcp ? IPPROTO_TCP : IPPROTO_UDP,
 				 &server->addr, &timeparms);
 	if (IS_ERR(xprt)) {
-		printk(KERN_WARNING "NFS: cannot create RPC transport.\n");
+		dprintk("%s: cannot create RPC transport. Error = %ld\n",
+				__FUNCTION__, PTR_ERR(xprt));
 		return (struct rpc_clnt *)xprt;
 	}
 	clnt = rpc_create_client(xprt, server->hostname, &nfs_program,
 				 server->rpc_ops->version, data->pseudoflavor);
 	if (IS_ERR(clnt)) {
-		printk(KERN_WARNING "NFS: cannot create RPC client.\n");
+		dprintk("%s: cannot create RPC client. Error = %ld\n",
+				__FUNCTION__, PTR_ERR(xprt));
 		goto out_fail;
 	}
 
@@ -383,7 +401,6 @@ nfs_create_client(struct nfs_server *server, const struct nfs_mount_data *data)
 	return clnt;
 
 out_fail:
-	xprt_destroy(xprt);
 	return clnt;
 }
 
@@ -427,21 +444,16 @@ nfs_fill_super(struct super_block *sb, struct nfs_mount_data *data, int silent)
 
 	/* Check NFS protocol revision and initialize RPC op vector
 	 * and file handle pool. */
-	if (server->flags & NFS_MOUNT_VER3) {
 #ifdef CONFIG_NFS_V3
+	if (server->flags & NFS_MOUNT_VER3) {
 		server->rpc_ops = &nfs_v3_clientops;
 		server->caps |= NFS_CAP_READDIRPLUS;
-		if (data->version < 4) {
-			printk(KERN_NOTICE "NFS: NFSv3 not supported by mount program.\n");
-			return -EIO;
-		}
-#else
-		printk(KERN_NOTICE "NFS: NFSv3 not supported.\n");
-		return -EIO;
-#endif
 	} else {
 		server->rpc_ops = &nfs_v2_clientops;
 	}
+#else
+	server->rpc_ops = &nfs_v2_clientops;
+#endif
 
 	/* Fill in pseudoflavor for mount version < 5 */
 	if (!(data->flags & NFS_MOUNT_SECFLAVOUR))
@@ -455,17 +467,34 @@ nfs_fill_super(struct super_block *sb, struct nfs_mount_data *data, int silent)
 		return PTR_ERR(server->client);
 	/* RFC 2623, sec 2.3.2 */
 	if (authflavor != RPC_AUTH_UNIX) {
+		struct rpc_auth *auth;
+
 		server->client_sys = rpc_clone_client(server->client);
 		if (IS_ERR(server->client_sys))
 			return PTR_ERR(server->client_sys);
-		if (!rpcauth_create(RPC_AUTH_UNIX, server->client_sys))
+		auth = rpcauth_create(RPC_AUTH_UNIX, server->client_sys);
-			return -ENOMEM;
+		if (IS_ERR(auth))
+			return PTR_ERR(auth);
 	} else {
 		atomic_inc(&server->client->cl_count);
 		server->client_sys = server->client;
 	}
-
 	if (server->flags & NFS_MOUNT_VER3) {
+#ifdef CONFIG_NFS_V3_ACL
+		if (!(server->flags & NFS_MOUNT_NOACL)) {
+			server->client_acl = rpc_bind_new_program(server->client, &nfsacl_program, 3);
+			/* No errors! Assume that Sun nfsacls are supported */
+			if (!IS_ERR(server->client_acl))
+				server->caps |= NFS_CAP_ACLS;
+		}
+#else
+		server->flags &= ~NFS_MOUNT_NOACL;
+#endif /* CONFIG_NFS_V3_ACL */
+		/*
+		 * The VFS shouldn't apply the umask to mode bits. We will
+		 * do so ourselves when necessary.
+		 */
+		sb->s_flags |= MS_POSIXACL;
 		if (server->namelen == 0 || server->namelen > NFS3_MAXNAMLEN)
 			server->namelen = NFS3_MAXNAMLEN;
 		sb->s_time_gran = 1;
@@ -549,6 +578,7 @@ static int nfs_show_options(struct seq_file *m, struct vfsmount *mnt)
 		{ NFS_MOUNT_NOCTO, ",nocto", "" },
 		{ NFS_MOUNT_NOAC, ",noac", "" },
 		{ NFS_MOUNT_NONLM, ",nolock", ",lock" },
+		{ NFS_MOUNT_NOACL, ",noacl", "" },
 		{ 0, NULL, NULL }
 	};
 	struct proc_nfs_info *nfs_infop;
@@ -590,9 +620,19 @@ nfs_zap_caches(struct inode *inode)
 
 	memset(NFS_COOKIEVERF(inode), 0, sizeof(NFS_COOKIEVERF(inode)));
 	if (S_ISREG(mode) || S_ISDIR(mode) || S_ISLNK(mode))
-		nfsi->flags |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA|NFS_INO_INVALID_ACCESS;
+		nfsi->flags |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL|NFS_INO_REVAL_PAGECACHE;
 	else
-		nfsi->flags |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS;
+		nfsi->flags |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL|NFS_INO_REVAL_PAGECACHE;
+}
+
+static void nfs_zap_acl_cache(struct inode *inode)
+{
+	void (*clear_acl_cache)(struct inode *);
+
+	clear_acl_cache = NFS_PROTO(inode)->clear_acl_cache;
+	if (clear_acl_cache != NULL)
+		clear_acl_cache(inode);
+	NFS_I(inode)->flags &= ~NFS_INO_INVALID_ACL;
 }
 
 /*
@@ -689,7 +729,7 @@ nfs_fhget(struct super_block *sb, struct nfs_fh *fh, struct nfs_fattr *fattr)
 		/* Why so? Because we want revalidate for devices/FIFOs, and
 		 * that's precisely what we have in nfs_file_inode_operations.
 		 */
-		inode->i_op = &nfs_file_inode_operations;
+		inode->i_op = NFS_SB(sb)->rpc_ops->file_inode_ops;
 		if (S_ISREG(inode->i_mode)) {
 			inode->i_fop = &nfs_file_operations;
 			inode->i_data.a_ops = &nfs_file_aops;
@@ -792,7 +832,7 @@ nfs_setattr(struct dentry *dentry, struct iattr *attr)
 		}
 	}
 	if ((attr->ia_valid & (ATTR_MODE|ATTR_UID|ATTR_GID)) != 0)
-		NFS_FLAGS(inode) |= NFS_INO_INVALID_ACCESS;
+		NFS_FLAGS(inode) |= NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
 	nfs_end_data_update(inode);
 	unlock_kernel();
 	return error;
@@ -851,7 +891,7 @@ struct nfs_open_context *alloc_nfs_open_context(struct dentry *dentry, struct rp
 		ctx->state = NULL;
 		ctx->lockowner = current->files;
 		ctx->error = 0;
-		init_waitqueue_head(&ctx->waitq);
+		ctx->dir_cookie = 0;
 	}
 	return ctx;
 }
@@ -1015,6 +1055,7 @@ __nfs_revalidate_inode(struct nfs_server *server, struct inode *inode)
 		goto out;
 	}
 	flags = nfsi->flags;
+	nfsi->flags &= ~NFS_INO_REVAL_PAGECACHE;
 	/*
 	 * We may need to keep the attributes marked as invalid if
 	 * we raced with nfs_end_attr_update().
@@ -1022,21 +1063,9 @@ __nfs_revalidate_inode(struct nfs_server *server, struct inode *inode)
 	if (verifier == nfsi->cache_change_attribute)
 		nfsi->flags &= ~(NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ATIME);
 	/* Do the page cache invalidation */
-	if (flags & NFS_INO_INVALID_DATA) {
+	nfs_revalidate_mapping(inode, inode->i_mapping);
-		if (S_ISREG(inode->i_mode)) {
+	if (flags & NFS_INO_INVALID_ACL)
-			if (filemap_fdatawrite(inode->i_mapping) == 0)
+		nfs_zap_acl_cache(inode);
-				filemap_fdatawait(inode->i_mapping);
-			nfs_wb_all(inode);
-		}
-		nfsi->flags &= ~NFS_INO_INVALID_DATA;
-		invalidate_inode_pages2(inode->i_mapping);
-		memset(NFS_COOKIEVERF(inode), 0, sizeof(NFS_COOKIEVERF(inode)));
-		dfprintk(PAGECACHE, "NFS: (%s/%Ld) data cache invalidated\n",
-				inode->i_sb->s_id,
-				(long long)NFS_FILEID(inode));
-		/* This ensures we revalidate dentries */
-		nfsi->cache_change_attribute++;
-	}
 	dfprintk(PAGECACHE, "NFS: (%s/%Ld) revalidation complete\n",
 		inode->i_sb->s_id,
 		(long long)NFS_FILEID(inode));
@@ -1073,6 +1102,34 @@ int nfs_revalidate_inode(struct nfs_server *server, struct inode *inode)
 	return __nfs_revalidate_inode(server, inode);
 }
 
+/**
+ * nfs_revalidate_mapping - Revalidate the pagecache
+ * @inode - pointer to host inode
+ * @mapping - pointer to mapping
+ */
+void nfs_revalidate_mapping(struct inode *inode, struct address_space *mapping)
+{
+	struct nfs_inode *nfsi = NFS_I(inode);
+
+	if (nfsi->flags & NFS_INO_INVALID_DATA) {
+		if (S_ISREG(inode->i_mode)) {
+			if (filemap_fdatawrite(mapping) == 0)
+				filemap_fdatawait(mapping);
+			nfs_wb_all(inode);
+		}
+		invalidate_inode_pages2(mapping);
+		nfsi->flags &= ~NFS_INO_INVALID_DATA;
+		if (S_ISDIR(inode->i_mode)) {
+			memset(nfsi->cookieverf, 0, sizeof(nfsi->cookieverf));
+			/* This ensures we revalidate child dentries */
+			nfsi->cache_change_attribute++;
+		}
+		dfprintk(PAGECACHE, "NFS: (%s/%Ld) data cache invalidated\n",
+				inode->i_sb->s_id,
+				(long long)NFS_FILEID(inode));
+	}
+}
+
 /**
  * nfs_begin_data_update
  * @inode - pointer to inode
@@ -1105,27 +1162,6 @@ void nfs_end_data_update(struct inode *inode)
 	atomic_dec(&nfsi->data_updates);
 }
 
-/**
- * nfs_end_data_update_defer
- * @inode - pointer to inode
- * Declare end of the operations that will update file data
- * This will defer marking the inode as needing revalidation
- * unless there are no other pending updates.
- */
-void nfs_end_data_update_defer(struct inode *inode)
-{
-	struct nfs_inode *nfsi = NFS_I(inode);
-
-	if (atomic_dec_and_test(&nfsi->data_updates)) {
-		/* Mark the attribute cache for revalidation */
-		nfsi->flags |= NFS_INO_INVALID_ATTR;
-		/* Directories and symlinks: invalidate page cache too */
-		if (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode))
-			nfsi->flags |= NFS_INO_INVALID_DATA;
-		nfsi->cache_change_attribute ++;
-	}
-}
-
 /**
  * nfs_refresh_inode - verify consistency of the inode attribute cache
  * @inode - pointer to inode
@@ -1152,8 +1188,11 @@ int nfs_refresh_inode(struct inode *inode, struct nfs_fattr *fattr)
 		if ((fattr->valid & NFS_ATTR_PRE_CHANGE) != 0
 				&& nfsi->change_attr == fattr->pre_change_attr)
 			nfsi->change_attr = fattr->change_attr;
-		if (!data_unstable && nfsi->change_attr != fattr->change_attr)
+		if (nfsi->change_attr != fattr->change_attr) {
 			nfsi->flags |= NFS_INO_INVALID_ATTR;
+			if (!data_unstable)
+				nfsi->flags |= NFS_INO_REVAL_PAGECACHE;
+		}
 	}
 
 	if ((fattr->valid & NFS_ATTR_FATTR) == 0)
@@ -1176,18 +1215,22 @@ int nfs_refresh_inode(struct inode *inode, struct nfs_fattr *fattr)
 	}
 
 	/* Verify a few of the more important attributes */
-	if (!data_unstable) {
+	if (!timespec_equal(&inode->i_mtime, &fattr->mtime)) {
-		if (!timespec_equal(&inode->i_mtime, &fattr->mtime)
-				|| cur_size != new_isize)
 		nfsi->flags |= NFS_INO_INVALID_ATTR;
-	} else if (S_ISREG(inode->i_mode) && new_isize > cur_size)
+		if (!data_unstable)
+			nfsi->flags |= NFS_INO_REVAL_PAGECACHE;
+	}
+	if (cur_size != new_isize) {
 		nfsi->flags |= NFS_INO_INVALID_ATTR;
+		if (nfsi->npages == 0)
+			nfsi->flags |= NFS_INO_REVAL_PAGECACHE;
+	}
 
 	/* Have any file permissions changed? */
 	if ((inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO)
 			|| inode->i_uid != fattr->uid
 			|| inode->i_gid != fattr->gid)
-		nfsi->flags |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS;
+		nfsi->flags |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL;
 
 	/* Has the link count changed? */
 	if (inode->i_nlink != fattr->nlink)
@@ -1215,10 +1258,8 @@ int nfs_refresh_inode(struct inode *inode, struct nfs_fattr *fattr)
 static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr, unsigned long verifier)
 {
 	struct nfs_inode *nfsi = NFS_I(inode);
-	__u64		new_size;
+	loff_t cur_isize, new_isize;
-	loff_t		new_isize;
 	unsigned int	invalid = 0;
-	loff_t		cur_isize;
 	int data_unstable;
 
 	dfprintk(VFS, "NFS: %s(%s/%ld ct=%d info=0x%x)\n",
@@ -1251,61 +1292,56 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr, unsign
 	/* Are we racing with known updates of the metadata on the server? */
 	data_unstable = ! nfs_verify_change_attribute(inode, verifier);
 
-	/* Check if the file size agrees */
+	/* Check if our cached file size is stale */
-	new_size = fattr->size;
  	new_isize = nfs_size_to_loff_t(fattr->size);
 	cur_isize = i_size_read(inode);
-	if (cur_isize != new_size) {
+	if (new_isize != cur_isize) {
-#ifdef NFS_DEBUG_VERBOSE
+		/* Do we perhaps have any outstanding writes? */
-		printk(KERN_DEBUG "NFS: isize change on %s/%ld\n", inode->i_sb->s_id, inode->i_ino);
+		if (nfsi->npages == 0) {
-#endif
+			/* No, but did we race with nfs_end_data_update()? */
-		/*
+			if (verifier  ==  nfsi->cache_change_attribute) {
-		 * If we have pending writebacks, things can get
+				inode->i_size = new_isize;
-		 * messy.
+				invalid |= NFS_INO_INVALID_DATA;
-		 */
+			}
-		if (S_ISREG(inode->i_mode) && data_unstable) {
+			invalid |= NFS_INO_INVALID_ATTR;
-			if (new_isize > cur_isize) {
+		} else if (new_isize > cur_isize) {
-				inode->i_size = new_isize;
-				invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA;
-			}
-		} else {
 			inode->i_size = new_isize;
 			invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA;
 		}
+		dprintk("NFS: isize change on server for file %s/%ld\n",
+				inode->i_sb->s_id, inode->i_ino);
 	}
 
-	/*
+	/* Check if the mtime agrees */
-	 * Note: we don't check inode->i_mtime since pipes etc.
-	 *       can change this value in VFS without requiring a
-	 *	 cache revalidation.
-	 */
 	if (!timespec_equal(&inode->i_mtime, &fattr->mtime)) {
 		memcpy(&inode->i_mtime, &fattr->mtime, sizeof(inode->i_mtime));
-#ifdef NFS_DEBUG_VERBOSE
+		dprintk("NFS: mtime change on server for file %s/%ld\n",
-		printk(KERN_DEBUG "NFS: mtime change on %s/%ld\n", inode->i_sb->s_id, inode->i_ino);
+				inode->i_sb->s_id, inode->i_ino);
-#endif
 		if (!data_unstable)
 			invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA;
 	}
 
 	if ((fattr->valid & NFS_ATTR_FATTR_V4)
 	    && nfsi->change_attr != fattr->change_attr) {
-#ifdef NFS_DEBUG_VERBOSE
+		dprintk("NFS: change_attr change on server for file %s/%ld\n",
-		printk(KERN_DEBUG "NFS: change_attr change on %s/%ld\n",
 		       inode->i_sb->s_id, inode->i_ino);
-#endif
 		nfsi->change_attr = fattr->change_attr;
 		if (!data_unstable)
-			invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA|NFS_INO_INVALID_ACCESS;
+			invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
 	}
 
+	/* If ctime has changed we should definitely clear access+acl caches */
+	if (!timespec_equal(&inode->i_ctime, &fattr->ctime)) {
+		if (!data_unstable)
+			invalid |= NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
 		memcpy(&inode->i_ctime, &fattr->ctime, sizeof(inode->i_ctime));
+	}
 	memcpy(&inode->i_atime, &fattr->atime, sizeof(inode->i_atime));
 
 	if ((inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO) ||
 	    inode->i_uid != fattr->uid ||
 	    inode->i_gid != fattr->gid)
-		invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS;
+		invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
 
 	inode->i_mode = fattr->mode;
 	inode->i_nlink = fattr->nlink;
@@ -1385,75 +1421,96 @@ static struct super_block *nfs_get_sb(struct file_system_type *fs_type,
 	int flags, const char *dev_name, void *raw_data)
 {
 	int error;
-	struct nfs_server *server;
+	struct nfs_server *server = NULL;
 	struct super_block *s;
 	struct nfs_fh *root;
 	struct nfs_mount_data *data = raw_data;
 
-	if (!data) {
+	s = ERR_PTR(-EINVAL);
-		printk("nfs_read_super: missing data argument\n");
+	if (data == NULL) {
-		return ERR_PTR(-EINVAL);
+		dprintk("%s: missing data argument\n", __FUNCTION__);
+		goto out_err;
 	}
+	if (data->version <= 0 || data->version > NFS_MOUNT_VERSION) {
+		dprintk("%s: bad mount version\n", __FUNCTION__);
+		goto out_err;
+	}
+	switch (data->version) {
+		case 1:
+			data->namlen = 0;
+		case 2:
+			data->bsize  = 0;
+		case 3:
+			if (data->flags & NFS_MOUNT_VER3) {
+				dprintk("%s: mount structure version %d does not support NFSv3\n",
+						__FUNCTION__,
+						data->version);
+				goto out_err;
+			}
+			data->root.size = NFS2_FHSIZE;
+			memcpy(data->root.data, data->old_root.data, NFS2_FHSIZE);
+		case 4:
+			if (data->flags & NFS_MOUNT_SECFLAVOUR) {
+				dprintk("%s: mount structure version %d does not support strong security\n",
+						__FUNCTION__,
+						data->version);
+				goto out_err;
+			}
+		case 5:
+			memset(data->context, 0, sizeof(data->context));
+	}
+#ifndef CONFIG_NFS_V3
+	/* If NFSv3 is not compiled in, return -EPROTONOSUPPORT */
+	s = ERR_PTR(-EPROTONOSUPPORT);
+	if (data->flags & NFS_MOUNT_VER3) {
+		dprintk("%s: NFSv3 not compiled into kernel\n", __FUNCTION__);
+		goto out_err;
+	}
+#endif /* CONFIG_NFS_V3 */
 
+	s = ERR_PTR(-ENOMEM);
 	server = kmalloc(sizeof(struct nfs_server), GFP_KERNEL);
 	if (!server)
-		return ERR_PTR(-ENOMEM);
+		goto out_err;
 	memset(server, 0, sizeof(struct nfs_server));
 	/* Zero out the NFS state stuff */
 	init_nfsv4_state(server);
-
+	server->client = server->client_sys = server->client_acl = ERR_PTR(-EINVAL);
-	if (data->version != NFS_MOUNT_VERSION) {
-		printk("nfs warning: mount version %s than kernel\n",
-			data->version < NFS_MOUNT_VERSION ? "older" : "newer");
-		if (data->version < 2)
-			data->namlen = 0;
-		if (data->version < 3)
-			data->bsize  = 0;
-		if (data->version < 4) {
-			data->flags &= ~NFS_MOUNT_VER3;
-			data->root.size = NFS2_FHSIZE;
-			memcpy(data->root.data, data->old_root.data, NFS2_FHSIZE);
-		}
-		if (data->version < 5)
-			data->flags &= ~NFS_MOUNT_SECFLAVOUR;
-	}
 
 	root = &server->fh;
 	if (data->flags & NFS_MOUNT_VER3)
 		root->size = data->root.size;
 	else
 		root->size = NFS2_FHSIZE;
+	s = ERR_PTR(-EINVAL);
 	if (root->size > sizeof(root->data)) {
-		printk("nfs_get_sb: invalid root filehandle\n");
+		dprintk("%s: invalid root filehandle\n", __FUNCTION__);
-		kfree(server);
+		goto out_err;
-		return ERR_PTR(-EINVAL);
 	}
 	memcpy(root->data, data->root.data, root->size);
 
 	/* We now require that the mount process passes the remote address */
 	memcpy(&server->addr, &data->addr, sizeof(server->addr));
 	if (server->addr.sin_addr.s_addr == INADDR_ANY) {
-		printk("NFS: mount program didn't pass remote address!\n");
+		dprintk("%s: mount program didn't pass remote address!\n",
-		kfree(server);
+				__FUNCTION__);
-		return ERR_PTR(-EINVAL);
+		goto out_err;
+	}
+
+	/* Fire up rpciod if not yet running */
+	s = ERR_PTR(rpciod_up());
+	if (IS_ERR(s)) {
+		dprintk("%s: couldn't start rpciod! Error = %ld\n",
+				__FUNCTION__, PTR_ERR(s));
+		goto out_err;
 	}
 
 	s = sget(fs_type, nfs_compare_super, nfs_set_super, server);
-
+	if (IS_ERR(s) || s->s_root)
-	if (IS_ERR(s) || s->s_root) {
+		goto out_rpciod_down;
-		kfree(server);
-		return s;
-	}
 
 	s->s_flags = flags;
 
-	/* Fire up rpciod if not yet running */
-	if (rpciod_up() != 0) {
-		printk(KERN_WARNING "NFS: couldn't start rpciod!\n");
-		kfree(server);
-		return ERR_PTR(-EIO);
-	}
-
 	error = nfs_fill_super(s, data, flags & MS_VERBOSE ? 1 : 0);
 	if (error) {
 		up_write(&s->s_umount);
@@ -1462,6 +1519,11 @@ static struct super_block *nfs_get_sb(struct file_system_type *fs_type,
 	}
 	s->s_flags |= MS_ACTIVE;
 	return s;
+out_rpciod_down:
+	rpciod_down();
+out_err:
+	kfree(server);
+	return s;
 }
 
 static void nfs_kill_super(struct super_block *s)
@@ -1470,10 +1532,12 @@ static void nfs_kill_super(struct super_block *s)
 
 	kill_anon_super(s);
 
-	if (server->client != NULL && !IS_ERR(server->client))
+	if (!IS_ERR(server->client))
 		rpc_shutdown_client(server->client);
-	if (server->client_sys != NULL && !IS_ERR(server->client_sys))
+	if (!IS_ERR(server->client_sys))
 		rpc_shutdown_client(server->client_sys);
+	if (!IS_ERR(server->client_acl))
+		rpc_shutdown_client(server->client_acl);
 
 	if (!(server->flags & NFS_MOUNT_NONLM))
 		lockd_down();	/* release rpc.lockd */
@@ -1594,15 +1658,19 @@ static int nfs4_fill_super(struct super_block *sb, struct nfs4_mount_data *data,
 
 	clp = nfs4_get_client(&server->addr.sin_addr);
 	if (!clp) {
-		printk(KERN_WARNING "NFS: failed to create NFS4 client.\n");
+		dprintk("%s: failed to create NFS4 client.\n", __FUNCTION__);
 		return -EIO;
 	}
 
 	/* Now create transport and client */
 	authflavour = RPC_AUTH_UNIX;
 	if (data->auth_flavourlen != 0) {
-		if (data->auth_flavourlen > 1)
+		if (data->auth_flavourlen != 1) {
-			printk(KERN_INFO "NFS: cannot yet deal with multiple auth flavours.\n");
+			dprintk("%s: Invalid number of RPC auth flavours %d.\n",
+					__FUNCTION__, data->auth_flavourlen);
+			err = -EINVAL;
+			goto out_fail;
+		}
 		if (copy_from_user(&authflavour, data->auth_flavours, sizeof(authflavour))) {
 			err = -EFAULT;
 			goto out_fail;
@@ -1610,21 +1678,22 @@ static int nfs4_fill_super(struct super_block *sb, struct nfs4_mount_data *data,
 	}
 
 	down_write(&clp->cl_sem);
-	if (clp->cl_rpcclient == NULL) {
+	if (IS_ERR(clp->cl_rpcclient)) {
 		xprt = xprt_create_proto(proto, &server->addr, &timeparms);
 		if (IS_ERR(xprt)) {
 			up_write(&clp->cl_sem);
-			printk(KERN_WARNING "NFS: cannot create RPC transport.\n");
 			err = PTR_ERR(xprt);
+			dprintk("%s: cannot create RPC transport. Error = %d\n",
+					__FUNCTION__, err);
 			goto out_fail;
 		}
 		clnt = rpc_create_client(xprt, server->hostname, &nfs_program,
 				server->rpc_ops->version, authflavour);
 		if (IS_ERR(clnt)) {
 			up_write(&clp->cl_sem);
-			printk(KERN_WARNING "NFS: cannot create RPC client.\n");
-			xprt_destroy(xprt);
 			err = PTR_ERR(clnt);
+			dprintk("%s: cannot create RPC client. Error = %d\n",
+					__FUNCTION__, err);
 			goto out_fail;
 		}
 		clnt->cl_intr     = 1;
@@ -1656,21 +1725,26 @@ static int nfs4_fill_super(struct super_block *sb, struct nfs4_mount_data *data,
 	clp = NULL;
 
 	if (IS_ERR(clnt)) {
-		printk(KERN_WARNING "NFS: cannot create RPC client.\n");
+		err = PTR_ERR(clnt);
-		return PTR_ERR(clnt);
+		dprintk("%s: cannot create RPC client. Error = %d\n",
+				__FUNCTION__, err);
+		return err;
 	}
 
 	server->client    = clnt;
 
 	if (server->nfs4_state->cl_idmap == NULL) {
-		printk(KERN_WARNING "NFS: failed to create idmapper.\n");
+		dprintk("%s: failed to create idmapper.\n", __FUNCTION__);
 		return -ENOMEM;
 	}
 
 	if (clnt->cl_auth->au_flavor != authflavour) {
-		if (rpcauth_create(authflavour, clnt) == NULL) {
+		struct rpc_auth *auth;
-			printk(KERN_WARNING "NFS: couldn't create credcache!\n");
+
-			return -ENOMEM;
+		auth = rpcauth_create(authflavour, clnt);
+		if (IS_ERR(auth)) {
+			dprintk("%s: couldn't create credcache!\n", __FUNCTION__);
+			return PTR_ERR(auth);
 		}
 	}
 
@@ -1730,8 +1804,12 @@ static struct super_block *nfs4_get_sb(struct file_system_type *fs_type,
 	struct nfs4_mount_data *data = raw_data;
 	void *p;
 
-	if (!data) {
+	if (data == NULL) {
-		printk("nfs_read_super: missing data argument\n");
+		dprintk("%s: missing data argument\n", __FUNCTION__);
+		return ERR_PTR(-EINVAL);
+	}
+	if (data->version <= 0 || data->version > NFS4_MOUNT_VERSION) {
+		dprintk("%s: bad mount version\n", __FUNCTION__);
 		return ERR_PTR(-EINVAL);
 	}
 
@@ -1741,11 +1819,7 @@ static struct super_block *nfs4_get_sb(struct file_system_type *fs_type,
 	memset(server, 0, sizeof(struct nfs_server));
 	/* Zero out the NFS state stuff */
 	init_nfsv4_state(server);
-
+	server->client = server->client_sys = server->client_acl = ERR_PTR(-EINVAL);
-	if (data->version != NFS4_MOUNT_VERSION) {
-		printk("nfs warning: mount version %s than kernel\n",
-			data->version < NFS4_MOUNT_VERSION ? "older" : "newer");
-	}
 
 	p = nfs_copy_user_string(NULL, &data->hostname, 256);
 	if (IS_ERR(p))
@@ -1773,11 +1847,20 @@ static struct super_block *nfs4_get_sb(struct file_system_type *fs_type,
 	}
 	if (server->addr.sin_family != AF_INET ||
 	    server->addr.sin_addr.s_addr == INADDR_ANY) {
-		printk("NFS: mount program didn't pass remote IP address!\n");
+		dprintk("%s: mount program didn't pass remote IP address!\n",
+				__FUNCTION__);
 		s = ERR_PTR(-EINVAL);
 		goto out_free;
 	}
 
+	/* Fire up rpciod if not yet running */
+	s = ERR_PTR(rpciod_up());
+	if (IS_ERR(s)) {
+		dprintk("%s: couldn't start rpciod! Error = %ld\n",
+				__FUNCTION__, PTR_ERR(s));
+		goto out_free;
+	}
+
 	s = sget(fs_type, nfs4_compare_super, nfs_set_super, server);
 
 	if (IS_ERR(s) || s->s_root)
@@ -1785,13 +1868,6 @@ static struct super_block *nfs4_get_sb(struct file_system_type *fs_type,
 
 	s->s_flags = flags;
 
-	/* Fire up rpciod if not yet running */
-	if (rpciod_up() != 0) {
-		printk(KERN_WARNING "NFS: couldn't start rpciod!\n");
-		s = ERR_PTR(-EIO);
-		goto out_free;
-	}
-
 	error = nfs4_fill_super(s, data, flags & MS_VERBOSE ? 1 : 0);
 	if (error) {
 		up_write(&s->s_umount);
@@ -1875,6 +1951,13 @@ static struct inode *nfs_alloc_inode(struct super_block *sb)
 	if (!nfsi)
 		return NULL;
 	nfsi->flags = 0;
+#ifdef CONFIG_NFS_V3_ACL
+	nfsi->acl_access = ERR_PTR(-EAGAIN);
+	nfsi->acl_default = ERR_PTR(-EAGAIN);
+#endif
+#ifdef CONFIG_NFS_V4
+	nfsi->nfs4_acl = NULL;
+#endif /* CONFIG_NFS_V4 */
 	return &nfsi->vfs_inode;
 }
 

fs/nfs/mount_clnt.c

@@ -80,9 +80,7 @@ mnt_create(char *hostname, struct sockaddr_in *srvaddr, int version,
 	clnt = rpc_create_client(xprt, hostname,
 				&mnt_program, version,
 				RPC_AUTH_UNIX);
-	if (IS_ERR(clnt)) {
+	if (!IS_ERR(clnt)) {
-		xprt_destroy(xprt);
-	} else {
 		clnt->cl_softrtry = 1;
 		clnt->cl_chatty   = 1;
 		clnt->cl_oneshot  = 1;

fs/nfs/nfs3acl.c

@@ -0,0 +1,403 @@
+#include <linux/fs.h>
+#include <linux/nfs.h>
+#include <linux/nfs3.h>
+#include <linux/nfs_fs.h>
+#include <linux/xattr_acl.h>
+#include <linux/nfsacl.h>
+
+#define NFSDBG_FACILITY	NFSDBG_PROC
+
+ssize_t nfs3_listxattr(struct dentry *dentry, char *buffer, size_t size)
+{
+	struct inode *inode = dentry->d_inode;
+	struct posix_acl *acl;
+	int pos=0, len=0;
+
+#	define output(s) do {						\
+			if (pos + sizeof(s) <= size) {			\
+				memcpy(buffer + pos, s, sizeof(s));	\
+				pos += sizeof(s);			\
+			}						\
+			len += sizeof(s);				\
+		} while(0)
+
+	acl = nfs3_proc_getacl(inode, ACL_TYPE_ACCESS);
+	if (IS_ERR(acl))
+		return PTR_ERR(acl);
+	if (acl) {
+		output("system.posix_acl_access");
+		posix_acl_release(acl);
+	}
+
+	if (S_ISDIR(inode->i_mode)) {
+		acl = nfs3_proc_getacl(inode, ACL_TYPE_DEFAULT);
+		if (IS_ERR(acl))
+			return PTR_ERR(acl);
+		if (acl) {
+			output("system.posix_acl_default");
+			posix_acl_release(acl);
+		}
+	}
+
+#	undef output
+
+	if (!buffer || len <= size)
+		return len;
+	return -ERANGE;
+}
+
+ssize_t nfs3_getxattr(struct dentry *dentry, const char *name,
+		void *buffer, size_t size)
+{
+	struct inode *inode = dentry->d_inode;
+	struct posix_acl *acl;
+	int type, error = 0;
+
+	if (strcmp(name, XATTR_NAME_ACL_ACCESS) == 0)
+		type = ACL_TYPE_ACCESS;
+	else if (strcmp(name, XATTR_NAME_ACL_DEFAULT) == 0)
+		type = ACL_TYPE_DEFAULT;
+	else
+		return -EOPNOTSUPP;
+
+	acl = nfs3_proc_getacl(inode, type);
+	if (IS_ERR(acl))
+		return PTR_ERR(acl);
+	else if (acl) {
+		if (type == ACL_TYPE_ACCESS && acl->a_count == 0)
+			error = -ENODATA;
+		else
+			error = posix_acl_to_xattr(acl, buffer, size);
+		posix_acl_release(acl);
+	} else
+		error = -ENODATA;
+
+	return error;
+}
+
+int nfs3_setxattr(struct dentry *dentry, const char *name,
+	     const void *value, size_t size, int flags)
+{
+	struct inode *inode = dentry->d_inode;
+	struct posix_acl *acl;
+	int type, error;
+
+	if (strcmp(name, XATTR_NAME_ACL_ACCESS) == 0)
+		type = ACL_TYPE_ACCESS;
+	else if (strcmp(name, XATTR_NAME_ACL_DEFAULT) == 0)
+		type = ACL_TYPE_DEFAULT;
+	else
+		return -EOPNOTSUPP;
+
+	acl = posix_acl_from_xattr(value, size);
+	if (IS_ERR(acl))
+		return PTR_ERR(acl);
+	error = nfs3_proc_setacl(inode, type, acl);
+	posix_acl_release(acl);
+
+	return error;
+}
+
+int nfs3_removexattr(struct dentry *dentry, const char *name)
+{
+	struct inode *inode = dentry->d_inode;
+	int type;
+
+	if (strcmp(name, XATTR_NAME_ACL_ACCESS) == 0)
+		type = ACL_TYPE_ACCESS;
+	else if (strcmp(name, XATTR_NAME_ACL_DEFAULT) == 0)
+		type = ACL_TYPE_DEFAULT;
+	else
+		return -EOPNOTSUPP;
+
+	return nfs3_proc_setacl(inode, type, NULL);
+}
+
+static void __nfs3_forget_cached_acls(struct nfs_inode *nfsi)
+{
+	if (!IS_ERR(nfsi->acl_access)) {
+		posix_acl_release(nfsi->acl_access);
+		nfsi->acl_access = ERR_PTR(-EAGAIN);
+	}
+	if (!IS_ERR(nfsi->acl_default)) {
+		posix_acl_release(nfsi->acl_default);
+		nfsi->acl_default = ERR_PTR(-EAGAIN);
+	}
+}
+
+void nfs3_forget_cached_acls(struct inode *inode)
+{
+	dprintk("NFS: nfs3_forget_cached_acls(%s/%ld)\n", inode->i_sb->s_id,
+		inode->i_ino);
+	spin_lock(&inode->i_lock);
+	__nfs3_forget_cached_acls(NFS_I(inode));
+	spin_unlock(&inode->i_lock);
+}
+
+static struct posix_acl *nfs3_get_cached_acl(struct inode *inode, int type)
+{
+	struct nfs_inode *nfsi = NFS_I(inode);
+	struct posix_acl *acl = ERR_PTR(-EINVAL);
+
+	spin_lock(&inode->i_lock);
+	switch(type) {
+		case ACL_TYPE_ACCESS:
+			acl = nfsi->acl_access;
+			break;
+
+		case ACL_TYPE_DEFAULT:
+			acl = nfsi->acl_default;
+			break;
+
+		default:
+			goto out;
+	}
+	if (IS_ERR(acl))
+		acl = ERR_PTR(-EAGAIN);
+	else
+		acl = posix_acl_dup(acl);
+out:
+	spin_unlock(&inode->i_lock);
+	dprintk("NFS: nfs3_get_cached_acl(%s/%ld, %d) = %p\n", inode->i_sb->s_id,
+		inode->i_ino, type, acl);
+	return acl;
+}
+
+static void nfs3_cache_acls(struct inode *inode, struct posix_acl *acl,
+		    struct posix_acl *dfacl)
+{
+	struct nfs_inode *nfsi = NFS_I(inode);
+
+	dprintk("nfs3_cache_acls(%s/%ld, %p, %p)\n", inode->i_sb->s_id,
+		inode->i_ino, acl, dfacl);
+	spin_lock(&inode->i_lock);
+	__nfs3_forget_cached_acls(NFS_I(inode));
+	nfsi->acl_access = posix_acl_dup(acl);
+	nfsi->acl_default = posix_acl_dup(dfacl);
+	spin_unlock(&inode->i_lock);
+}
+
+struct posix_acl *nfs3_proc_getacl(struct inode *inode, int type)
+{
+	struct nfs_server *server = NFS_SERVER(inode);
+	struct nfs_fattr fattr;
+	struct page *pages[NFSACL_MAXPAGES] = { };
+	struct nfs3_getaclargs args = {
+		.fh = NFS_FH(inode),
+		/* The xdr layer may allocate pages here. */
+		.pages = pages,
+	};
+	struct nfs3_getaclres res = {
+		.fattr =	&fattr,
+	};
+	struct posix_acl *acl;
+	int status, count;
+
+	if (!nfs_server_capable(inode, NFS_CAP_ACLS))
+		return ERR_PTR(-EOPNOTSUPP);
+
+	status = nfs_revalidate_inode(server, inode);
+	if (status < 0)
+		return ERR_PTR(status);
+	acl = nfs3_get_cached_acl(inode, type);
+	if (acl != ERR_PTR(-EAGAIN))
+		return acl;
+	acl = NULL;
+
+	/*
+	 * Only get the access acl when explicitly requested: We don't
+	 * need it for access decisions, and only some applications use
+	 * it. Applications which request the access acl first are not
+	 * penalized from this optimization.
+	 */
+	if (type == ACL_TYPE_ACCESS)
+		args.mask |= NFS_ACLCNT|NFS_ACL;
+	if (S_ISDIR(inode->i_mode))
+		args.mask |= NFS_DFACLCNT|NFS_DFACL;
+	if (args.mask == 0)
+		return NULL;
+
+	dprintk("NFS call getacl\n");
+	status = rpc_call(server->client_acl, ACLPROC3_GETACL,
+			  &args, &res, 0);
+	dprintk("NFS reply getacl: %d\n", status);
+
+	/* pages may have been allocated at the xdr layer. */
+	for (count = 0; count < NFSACL_MAXPAGES && args.pages[count]; count++)
+		__free_page(args.pages[count]);
+
+	switch (status) {
+		case 0:
+			status = nfs_refresh_inode(inode, &fattr);
+			break;
+		case -EPFNOSUPPORT:
+		case -EPROTONOSUPPORT:
+			dprintk("NFS_V3_ACL extension not supported; disabling\n");
+			server->caps &= ~NFS_CAP_ACLS;
+		case -ENOTSUPP:
+			status = -EOPNOTSUPP;
+		default:
+			goto getout;
+	}
+	if ((args.mask & res.mask) != args.mask) {
+		status = -EIO;
+		goto getout;
+	}
+
+	if (res.acl_access != NULL) {
+		if (posix_acl_equiv_mode(res.acl_access, NULL) == 0) {
+			posix_acl_release(res.acl_access);
+			res.acl_access = NULL;
+		}
+	}
+	nfs3_cache_acls(inode, res.acl_access, res.acl_default);
+
+	switch(type) {
+		case ACL_TYPE_ACCESS:
+			acl = res.acl_access;
+			res.acl_access = NULL;
+			break;
+
+		case ACL_TYPE_DEFAULT:
+			acl = res.acl_default;
+			res.acl_default = NULL;
+	}
+
+getout:
+	posix_acl_release(res.acl_access);
+	posix_acl_release(res.acl_default);
+
+	if (status != 0) {
+		posix_acl_release(acl);
+		acl = ERR_PTR(status);
+	}
+	return acl;
+}
+
+static int nfs3_proc_setacls(struct inode *inode, struct posix_acl *acl,
+		  struct posix_acl *dfacl)
+{
+	struct nfs_server *server = NFS_SERVER(inode);
+	struct nfs_fattr fattr;
+	struct page *pages[NFSACL_MAXPAGES] = { };
+	struct nfs3_setaclargs args = {
+		.inode = inode,
+		.mask = NFS_ACL,
+		.acl_access = acl,
+		.pages = pages,
+	};
+	int status, count;
+
+	status = -EOPNOTSUPP;
+	if (!nfs_server_capable(inode, NFS_CAP_ACLS))
+		goto out;
+
+	/* We are doing this here, because XDR marshalling can only
+	   return -ENOMEM. */
+	status = -ENOSPC;
+	if (acl != NULL && acl->a_count > NFS_ACL_MAX_ENTRIES)
+		goto out;
+	if (dfacl != NULL && dfacl->a_count > NFS_ACL_MAX_ENTRIES)
+		goto out;
+	if (S_ISDIR(inode->i_mode)) {
+		args.mask |= NFS_DFACL;
+		args.acl_default = dfacl;
+	}
+
+	dprintk("NFS call setacl\n");
+	nfs_begin_data_update(inode);
+	status = rpc_call(server->client_acl, ACLPROC3_SETACL,
+			  &args, &fattr, 0);
+	NFS_FLAGS(inode) |= NFS_INO_INVALID_ACCESS;
+	nfs_end_data_update(inode);
+	dprintk("NFS reply setacl: %d\n", status);
+
+	/* pages may have been allocated at the xdr layer. */
+	for (count = 0; count < NFSACL_MAXPAGES && args.pages[count]; count++)
+		__free_page(args.pages[count]);
+
+	switch (status) {
+		case 0:
+			status = nfs_refresh_inode(inode, &fattr);
+			break;
+		case -EPFNOSUPPORT:
+		case -EPROTONOSUPPORT:
+			dprintk("NFS_V3_ACL SETACL RPC not supported"
+					"(will not retry)\n");
+			server->caps &= ~NFS_CAP_ACLS;
+		case -ENOTSUPP:
+			status = -EOPNOTSUPP;
+	}
+out:
+	return status;
+}
+
+int nfs3_proc_setacl(struct inode *inode, int type, struct posix_acl *acl)
+{
+	struct posix_acl *alloc = NULL, *dfacl = NULL;
+	int status;
+
+	if (S_ISDIR(inode->i_mode)) {
+		switch(type) {
+			case ACL_TYPE_ACCESS:
+				alloc = dfacl = nfs3_proc_getacl(inode,
+						ACL_TYPE_DEFAULT);
+				if (IS_ERR(alloc))
+					goto fail;
+				break;
+
+			case ACL_TYPE_DEFAULT:
+				dfacl = acl;
+				alloc = acl = nfs3_proc_getacl(inode,
+						ACL_TYPE_ACCESS);
+				if (IS_ERR(alloc))
+					goto fail;
+				break;
+
+			default:
+				return -EINVAL;
+		}
+	} else if (type != ACL_TYPE_ACCESS)
+			return -EINVAL;
+
+	if (acl == NULL) {
+		alloc = acl = posix_acl_from_mode(inode->i_mode, GFP_KERNEL);
+		if (IS_ERR(alloc))
+			goto fail;
+	}
+	status = nfs3_proc_setacls(inode, acl, dfacl);
+	posix_acl_release(alloc);
+	return status;
+
+fail:
+	return PTR_ERR(alloc);
+}
+
+int nfs3_proc_set_default_acl(struct inode *dir, struct inode *inode,
+		mode_t mode)
+{
+	struct posix_acl *dfacl, *acl;
+	int error = 0;
+
+	dfacl = nfs3_proc_getacl(dir, ACL_TYPE_DEFAULT);
+	if (IS_ERR(dfacl)) {
+		error = PTR_ERR(dfacl);
+		return (error == -EOPNOTSUPP) ? 0 : error;
+	}
+	if (!dfacl)
+		return 0;
+	acl = posix_acl_clone(dfacl, GFP_KERNEL);
+	error = -ENOMEM;
+	if (!acl)
+		goto out_release_dfacl;
+	error = posix_acl_create_masq(acl, &mode);
+	if (error < 0)
+		goto out_release_acl;
+	error = nfs3_proc_setacls(inode, acl, S_ISDIR(inode->i_mode) ?
+						      dfacl : NULL);
+out_release_acl:
+	posix_acl_release(acl);
+out_release_dfacl:
+	posix_acl_release(dfacl);
+	return error;
+}

fs/nfs/nfs3proc.c

@@ -17,6 +17,7 @@
 #include <linux/nfs_page.h>
 #include <linux/lockd/bind.h>
 #include <linux/smp_lock.h>
+#include <linux/nfs_mount.h>
 
 #define NFSDBG_FACILITY		NFSDBG_PROC
 
@@ -45,7 +46,7 @@ static inline int
 nfs3_rpc_call_wrapper(struct rpc_clnt *clnt, u32 proc, void *argp, void *resp, int flags)
 {
 	struct rpc_message msg = {
-		.rpc_proc	= &nfs3_procedures[proc],
+		.rpc_proc	= &clnt->cl_procinfo[proc],
 		.rpc_argp	= argp,
 		.rpc_resp	= resp,
 	};
@@ -313,6 +314,7 @@ nfs3_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
 		.fh		= &fhandle,
 		.fattr		= &fattr
 	};
+	mode_t mode = sattr->ia_mode;
 	int status;
 
 	dprintk("NFS call  create %s\n", dentry->d_name.name);
@@ -323,6 +325,8 @@ nfs3_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
 		arg.verifier[1] = current->pid;
 	}
 
+	sattr->ia_mode &= ~current->fs->umask;
+
 again:
 	dir_attr.valid = 0;
 	fattr.valid = 0;
@@ -369,6 +373,9 @@ nfs3_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
 		nfs_refresh_inode(dentry->d_inode, &fattr);
 		dprintk("NFS reply setattr (post-create): %d\n", status);
 	}
+	if (status != 0)
+		goto out;
+	status = nfs3_proc_set_default_acl(dir, dentry->d_inode, mode);
 out:
 	dprintk("NFS reply create: %d\n", status);
 	return status;
@@ -538,15 +545,24 @@ nfs3_proc_mkdir(struct inode *dir, struct dentry *dentry, struct iattr *sattr)
 		.fh		= &fhandle,
 		.fattr		= &fattr
 	};
+	int mode = sattr->ia_mode;
 	int status;
 
 	dprintk("NFS call  mkdir %s\n", dentry->d_name.name);
 	dir_attr.valid = 0;
 	fattr.valid = 0;
+
+	sattr->ia_mode &= ~current->fs->umask;
+
 	status = rpc_call(NFS_CLIENT(dir), NFS3PROC_MKDIR, &arg, &res, 0);
 	nfs_refresh_inode(dir, &dir_attr);
-	if (status == 0)
+	if (status != 0)
+		goto out;
 	status = nfs_instantiate(dentry, &fhandle, &fattr);
+	if (status != 0)
+		goto out;
+	status = nfs3_proc_set_default_acl(dir, dentry->d_inode, mode);
+out:
 	dprintk("NFS reply mkdir: %d\n", status);
 	return status;
 }
@@ -641,6 +657,7 @@ nfs3_proc_mknod(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
 		.fh		= &fh,
 		.fattr		= &fattr
 	};
+	mode_t mode = sattr->ia_mode;
 	int status;
 
 	switch (sattr->ia_mode & S_IFMT) {
@@ -653,12 +670,20 @@ nfs3_proc_mknod(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
 
 	dprintk("NFS call  mknod %s %u:%u\n", dentry->d_name.name,
 			MAJOR(rdev), MINOR(rdev));
+
+	sattr->ia_mode &= ~current->fs->umask;
+
 	dir_attr.valid = 0;
 	fattr.valid = 0;
 	status = rpc_call(NFS_CLIENT(dir), NFS3PROC_MKNOD, &arg, &res, 0);
 	nfs_refresh_inode(dir, &dir_attr);
-	if (status == 0)
+	if (status != 0)
+		goto out;
 	status = nfs_instantiate(dentry, &fh, &fattr);
+	if (status != 0)
+		goto out;
+	status = nfs3_proc_set_default_acl(dir, dentry->d_inode, mode);
+out:
 	dprintk("NFS reply mknod: %d\n", status);
 	return status;
 }
@@ -825,7 +850,8 @@ nfs3_proc_lock(struct file *filp, int cmd, struct file_lock *fl)
 struct nfs_rpc_ops	nfs_v3_clientops = {
 	.version	= 3,			/* protocol version */
 	.dentry_ops	= &nfs_dentry_operations,
-	.dir_inode_ops	= &nfs_dir_inode_operations,
+	.dir_inode_ops	= &nfs3_dir_inode_operations,
+	.file_inode_ops	= &nfs3_file_inode_operations,
 	.getroot	= nfs3_proc_get_root,
 	.getattr	= nfs3_proc_getattr,
 	.setattr	= nfs3_proc_setattr,
@@ -856,4 +882,5 @@ struct nfs_rpc_ops	nfs_v3_clientops = {
 	.file_open	= nfs_open,
 	.file_release	= nfs_release,
 	.lock		= nfs3_proc_lock,
+	.clear_acl_cache = nfs3_forget_cached_acls,
 };

fs/nfs/nfs3xdr.c

@@ -21,6 +21,7 @@
 #include <linux/nfs.h>
 #include <linux/nfs3.h>
 #include <linux/nfs_fs.h>
+#include <linux/nfsacl.h>
 
 #define NFSDBG_FACILITY		NFSDBG_XDR
 
@@ -79,6 +80,11 @@ extern int			nfs_stat_to_errno(int);
 #define NFS3_pathconfres_sz	(1+NFS3_post_op_attr_sz+6)
 #define NFS3_commitres_sz	(1+NFS3_wcc_data_sz+2)
 
+#define ACL3_getaclargs_sz	(NFS3_fh_sz+1)
+#define ACL3_setaclargs_sz	(NFS3_fh_sz+1+2*(2+5*3))
+#define ACL3_getaclres_sz	(1+NFS3_post_op_attr_sz+1+2*(2+5*3))
+#define ACL3_setaclres_sz	(1+NFS3_post_op_attr_sz)
+
 /*
  * Map file type to S_IFMT bits
  */
@@ -627,6 +633,74 @@ nfs3_xdr_commitargs(struct rpc_rqst *req, u32 *p, struct nfs_writeargs *args)
 	return 0;
 }
 
+#ifdef CONFIG_NFS_V3_ACL
+/*
+ * Encode GETACL arguments
+ */
+static int
+nfs3_xdr_getaclargs(struct rpc_rqst *req, u32 *p,
+		    struct nfs3_getaclargs *args)
+{
+	struct rpc_auth *auth = req->rq_task->tk_auth;
+	unsigned int replen;
+
+	p = xdr_encode_fhandle(p, args->fh);
+	*p++ = htonl(args->mask);
+	req->rq_slen = xdr_adjust_iovec(req->rq_svec, p);
+
+	if (args->mask & (NFS_ACL | NFS_DFACL)) {
+		/* Inline the page array */
+		replen = (RPC_REPHDRSIZE + auth->au_rslack +
+			  ACL3_getaclres_sz) << 2;
+		xdr_inline_pages(&req->rq_rcv_buf, replen, args->pages, 0,
+				 NFSACL_MAXPAGES << PAGE_SHIFT);
+	}
+	return 0;
+}
+
+/*
+ * Encode SETACL arguments
+ */
+static int
+nfs3_xdr_setaclargs(struct rpc_rqst *req, u32 *p,
+                   struct nfs3_setaclargs *args)
+{
+	struct xdr_buf *buf = &req->rq_snd_buf;
+	unsigned int base, len_in_head, len = nfsacl_size(
+		(args->mask & NFS_ACL)   ? args->acl_access  : NULL,
+		(args->mask & NFS_DFACL) ? args->acl_default : NULL);
+	int count, err;
+
+	p = xdr_encode_fhandle(p, NFS_FH(args->inode));
+	*p++ = htonl(args->mask);
+	base = (char *)p - (char *)buf->head->iov_base;
+	/* put as much of the acls into head as possible. */
+	len_in_head = min_t(unsigned int, buf->head->iov_len - base, len);
+	len -= len_in_head;
+	req->rq_slen = xdr_adjust_iovec(req->rq_svec, p + (len_in_head >> 2));
+
+	for (count = 0; (count << PAGE_SHIFT) < len; count++) {
+		args->pages[count] = alloc_page(GFP_KERNEL);
+		if (!args->pages[count]) {
+			while (count)
+				__free_page(args->pages[--count]);
+			return -ENOMEM;
+		}
+	}
+	xdr_encode_pages(buf, args->pages, 0, len);
+
+	err = nfsacl_encode(buf, base, args->inode,
+			    (args->mask & NFS_ACL) ?
+			    args->acl_access : NULL, 1, 0);
+	if (err > 0)
+		err = nfsacl_encode(buf, base + err, args->inode,
+				    (args->mask & NFS_DFACL) ?
+				    args->acl_default : NULL, 1,
+				    NFS_ACL_DEFAULT);
+	return (err > 0) ? 0 : err;
+}
+#endif  /* CONFIG_NFS_V3_ACL */
+
 /*
  * NFS XDR decode functions
  */
@@ -978,6 +1052,54 @@ nfs3_xdr_commitres(struct rpc_rqst *req, u32 *p, struct nfs_writeres *res)
 	return 0;
 }
 
+#ifdef CONFIG_NFS_V3_ACL
+/*
+ * Decode GETACL reply
+ */
+static int
+nfs3_xdr_getaclres(struct rpc_rqst *req, u32 *p,
+		   struct nfs3_getaclres *res)
+{
+	struct xdr_buf *buf = &req->rq_rcv_buf;
+	int status = ntohl(*p++);
+	struct posix_acl **acl;
+	unsigned int *aclcnt;
+	int err, base;
+
+	if (status != 0)
+		return -nfs_stat_to_errno(status);
+	p = xdr_decode_post_op_attr(p, res->fattr);
+	res->mask = ntohl(*p++);
+	if (res->mask & ~(NFS_ACL|NFS_ACLCNT|NFS_DFACL|NFS_DFACLCNT))
+		return -EINVAL;
+	base = (char *)p - (char *)req->rq_rcv_buf.head->iov_base;
+
+	acl = (res->mask & NFS_ACL) ? &res->acl_access : NULL;
+	aclcnt = (res->mask & NFS_ACLCNT) ? &res->acl_access_count : NULL;
+	err = nfsacl_decode(buf, base, aclcnt, acl);
+
+	acl = (res->mask & NFS_DFACL) ? &res->acl_default : NULL;
+	aclcnt = (res->mask & NFS_DFACLCNT) ? &res->acl_default_count : NULL;
+	if (err > 0)
+		err = nfsacl_decode(buf, base + err, aclcnt, acl);
+	return (err > 0) ? 0 : err;
+}
+
+/*
+ * Decode setacl reply.
+ */
+static int
+nfs3_xdr_setaclres(struct rpc_rqst *req, u32 *p, struct nfs_fattr *fattr)
+{
+	int status = ntohl(*p++);
+
+	if (status)
+		return -nfs_stat_to_errno(status);
+	xdr_decode_post_op_attr(p, fattr);
+	return 0;
+}
+#endif  /* CONFIG_NFS_V3_ACL */
+
 #ifndef MAX
 # define MAX(a, b)	(((a) > (b))? (a) : (b))
 #endif
@@ -1021,3 +1143,28 @@ struct rpc_version		nfs_version3 = {
 	.procs			= nfs3_procedures
 };
 
+#ifdef CONFIG_NFS_V3_ACL
+static struct rpc_procinfo	nfs3_acl_procedures[] = {
+	[ACLPROC3_GETACL] = {
+		.p_proc = ACLPROC3_GETACL,
+		.p_encode = (kxdrproc_t) nfs3_xdr_getaclargs,
+		.p_decode = (kxdrproc_t) nfs3_xdr_getaclres,
+		.p_bufsiz = MAX(ACL3_getaclargs_sz, ACL3_getaclres_sz) << 2,
+		.p_timer = 1,
+	},
+	[ACLPROC3_SETACL] = {
+		.p_proc = ACLPROC3_SETACL,
+		.p_encode = (kxdrproc_t) nfs3_xdr_setaclargs,
+		.p_decode = (kxdrproc_t) nfs3_xdr_setaclres,
+		.p_bufsiz = MAX(ACL3_setaclargs_sz, ACL3_setaclres_sz) << 2,
+		.p_timer = 0,
+	},
+};
+
+struct rpc_version		nfsacl_version3 = {
+	.number			= 3,
+	.nrprocs		= sizeof(nfs3_acl_procedures)/
+				  sizeof(nfs3_acl_procedures[0]),
+	.procs			= nfs3_acl_procedures,
+};
+#endif  /* CONFIG_NFS_V3_ACL */

fs/nfs/nfs4_fs.h

@@ -0,0 +1,253 @@
+/*
+ * linux/fs/nfs/nfs4_fs.h
+ *
+ * Copyright (C) 2005 Trond Myklebust
+ *
+ * NFSv4-specific filesystem definitions and declarations
+ */
+
+#ifndef __LINUX_FS_NFS_NFS4_FS_H
+#define __LINUX_FS_NFS_NFS4_FS_H
+
+#ifdef CONFIG_NFS_V4
+
+struct idmap;
+
+/*
+ * In a seqid-mutating op, this macro controls which error return
+ * values trigger incrementation of the seqid.
+ *
+ * from rfc 3010:
+ * The client MUST monotonically increment the sequence number for the
+ * CLOSE, LOCK, LOCKU, OPEN, OPEN_CONFIRM, and OPEN_DOWNGRADE
+ * operations.  This is true even in the event that the previous
+ * operation that used the sequence number received an error.  The only
+ * exception to this rule is if the previous operation received one of
+ * the following errors: NFSERR_STALE_CLIENTID, NFSERR_STALE_STATEID,
+ * NFSERR_BAD_STATEID, NFSERR_BAD_SEQID, NFSERR_BADXDR,
+ * NFSERR_RESOURCE, NFSERR_NOFILEHANDLE.
+ *
+ */
+#define seqid_mutating_err(err)       \
+(((err) != NFSERR_STALE_CLIENTID) &&  \
+ ((err) != NFSERR_STALE_STATEID)  &&  \
+ ((err) != NFSERR_BAD_STATEID)    &&  \
+ ((err) != NFSERR_BAD_SEQID)      &&  \
+ ((err) != NFSERR_BAD_XDR)        &&  \
+ ((err) != NFSERR_RESOURCE)       &&  \
+ ((err) != NFSERR_NOFILEHANDLE))
+
+enum nfs4_client_state {
+	NFS4CLNT_OK  = 0,
+};
+
+/*
+ * The nfs4_client identifies our client state to the server.
+ */
+struct nfs4_client {
+	struct list_head	cl_servers;	/* Global list of servers */
+	struct in_addr		cl_addr;	/* Server identifier */
+	u64			cl_clientid;	/* constant */
+	nfs4_verifier		cl_confirm;
+	unsigned long		cl_state;
+
+	u32			cl_lockowner_id;
+
+	/*
+	 * The following rwsem ensures exclusive access to the server
+	 * while we recover the state following a lease expiration.
+	 */
+	struct rw_semaphore	cl_sem;
+
+	struct list_head	cl_delegations;
+	struct list_head	cl_state_owners;
+	struct list_head	cl_unused;
+	int			cl_nunused;
+	spinlock_t		cl_lock;
+	atomic_t		cl_count;
+
+	struct rpc_clnt *	cl_rpcclient;
+	struct rpc_cred *	cl_cred;
+
+	struct list_head	cl_superblocks;	/* List of nfs_server structs */
+
+	unsigned long		cl_lease_time;
+	unsigned long		cl_last_renewal;
+	struct work_struct	cl_renewd;
+	struct work_struct	cl_recoverd;
+
+	wait_queue_head_t	cl_waitq;
+	struct rpc_wait_queue	cl_rpcwaitq;
+
+	/* used for the setclientid verifier */
+	struct timespec		cl_boot_time;
+
+	/* idmapper */
+	struct idmap *		cl_idmap;
+
+	/* Our own IP address, as a null-terminated string.
+	 * This is used to generate the clientid, and the callback address.
+	 */
+	char			cl_ipaddr[16];
+	unsigned char		cl_id_uniquifier;
+};
+
+/*
+ * NFS4 state_owners and lock_owners are simply labels for ordered
+ * sequences of RPC calls. Their sole purpose is to provide once-only
+ * semantics by allowing the server to identify replayed requests.
+ *
+ * The ->so_sema is held during all state_owner seqid-mutating operations:
+ * OPEN, OPEN_DOWNGRADE, and CLOSE. Its purpose is to properly serialize
+ * so_seqid.
+ */
+struct nfs4_state_owner {
+	struct list_head     so_list;	 /* per-clientid list of state_owners */
+	struct nfs4_client   *so_client;
+	u32                  so_id;      /* 32-bit identifier, unique */
+	struct semaphore     so_sema;
+	u32                  so_seqid;   /* protected by so_sema */
+	atomic_t	     so_count;
+
+	struct rpc_cred	     *so_cred;	 /* Associated cred */
+	struct list_head     so_states;
+	struct list_head     so_delegations;
+};
+
+/*
+ * struct nfs4_state maintains the client-side state for a given
+ * (state_owner,inode) tuple (OPEN) or state_owner (LOCK).
+ *
+ * OPEN:
+ * In order to know when to OPEN_DOWNGRADE or CLOSE the state on the server,
+ * we need to know how many files are open for reading or writing on a
+ * given inode. This information too is stored here.
+ *
+ * LOCK: one nfs4_state (LOCK) to hold the lock stateid nfs4_state(OPEN)
+ */
+
+struct nfs4_lock_state {
+	struct list_head	ls_locks;	/* Other lock stateids */
+	struct nfs4_state *	ls_state;	/* Pointer to open state */
+	fl_owner_t		ls_owner;	/* POSIX lock owner */
+#define NFS_LOCK_INITIALIZED 1
+	int			ls_flags;
+	u32			ls_seqid;
+	u32			ls_id;
+	nfs4_stateid		ls_stateid;
+	atomic_t		ls_count;
+};
+
+/* bits for nfs4_state->flags */
+enum {
+	LK_STATE_IN_USE,
+	NFS_DELEGATED_STATE,
+};
+
+struct nfs4_state {
+	struct list_head open_states;	/* List of states for the same state_owner */
+	struct list_head inode_states;	/* List of states for the same inode */
+	struct list_head lock_states;	/* List of subservient lock stateids */
+
+	struct nfs4_state_owner *owner;	/* Pointer to the open owner */
+	struct inode *inode;		/* Pointer to the inode */
+
+	unsigned long flags;		/* Do we hold any locks? */
+	struct semaphore lock_sema;	/* Serializes file locking operations */
+	spinlock_t state_lock;		/* Protects the lock_states list */
+
+	nfs4_stateid stateid;
+
+	unsigned int nreaders;
+	unsigned int nwriters;
+	int state;			/* State on the server (R,W, or RW) */
+	atomic_t count;
+};
+
+
+struct nfs4_exception {
+	long timeout;
+	int retry;
+};
+
+struct nfs4_state_recovery_ops {
+	int (*recover_open)(struct nfs4_state_owner *, struct nfs4_state *);
+	int (*recover_lock)(struct nfs4_state *, struct file_lock *);
+};
+
+extern struct dentry_operations nfs4_dentry_operations;
+extern struct inode_operations nfs4_dir_inode_operations;
+
+/* inode.c */
+extern ssize_t nfs4_getxattr(struct dentry *, const char *, void *, size_t);
+extern int nfs4_setxattr(struct dentry *, const char *, const void *, size_t, int);
+extern ssize_t nfs4_listxattr(struct dentry *, char *, size_t);
+
+
+/* nfs4proc.c */
+extern int nfs4_map_errors(int err);
+extern int nfs4_proc_setclientid(struct nfs4_client *, u32, unsigned short);
+extern int nfs4_proc_setclientid_confirm(struct nfs4_client *);
+extern int nfs4_proc_async_renew(struct nfs4_client *);
+extern int nfs4_proc_renew(struct nfs4_client *);
+extern int nfs4_do_close(struct inode *inode, struct nfs4_state *state, mode_t mode);
+extern struct inode *nfs4_atomic_open(struct inode *, struct dentry *, struct nameidata *);
+extern int nfs4_open_revalidate(struct inode *, struct dentry *, int);
+
+extern struct nfs4_state_recovery_ops nfs4_reboot_recovery_ops;
+extern struct nfs4_state_recovery_ops nfs4_network_partition_recovery_ops;
+
+extern const u32 nfs4_fattr_bitmap[2];
+extern const u32 nfs4_statfs_bitmap[2];
+extern const u32 nfs4_pathconf_bitmap[2];
+extern const u32 nfs4_fsinfo_bitmap[2];
+
+/* nfs4renewd.c */
+extern void nfs4_schedule_state_renewal(struct nfs4_client *);
+extern void nfs4_renewd_prepare_shutdown(struct nfs_server *);
+extern void nfs4_kill_renewd(struct nfs4_client *);
+extern void nfs4_renew_state(void *);
+
+/* nfs4state.c */
+extern void init_nfsv4_state(struct nfs_server *);
+extern void destroy_nfsv4_state(struct nfs_server *);
+extern struct nfs4_client *nfs4_get_client(struct in_addr *);
+extern void nfs4_put_client(struct nfs4_client *clp);
+extern int nfs4_init_client(struct nfs4_client *clp);
+extern struct nfs4_client *nfs4_find_client(struct in_addr *);
+extern u32 nfs4_alloc_lockowner_id(struct nfs4_client *);
+
+extern struct nfs4_state_owner * nfs4_get_state_owner(struct nfs_server *, struct rpc_cred *);
+extern void nfs4_put_state_owner(struct nfs4_state_owner *);
+extern void nfs4_drop_state_owner(struct nfs4_state_owner *);
+extern struct nfs4_state * nfs4_get_open_state(struct inode *, struct nfs4_state_owner *);
+extern void nfs4_put_open_state(struct nfs4_state *);
+extern void nfs4_close_state(struct nfs4_state *, mode_t);
+extern struct nfs4_state *nfs4_find_state(struct inode *, struct rpc_cred *, mode_t mode);
+extern void nfs4_increment_seqid(int status, struct nfs4_state_owner *sp);
+extern void nfs4_schedule_state_recovery(struct nfs4_client *);
+extern int nfs4_set_lock_state(struct nfs4_state *state, struct file_lock *fl);
+extern void nfs4_increment_lock_seqid(int status, struct nfs4_lock_state *ls);
+extern void nfs4_copy_stateid(nfs4_stateid *, struct nfs4_state *, fl_owner_t);
+
+extern const nfs4_stateid zero_stateid;
+
+/* nfs4xdr.c */
+extern uint32_t *nfs4_decode_dirent(uint32_t *p, struct nfs_entry *entry, int plus);
+extern struct rpc_procinfo nfs4_procedures[];
+
+struct nfs4_mount_data;
+
+/* callback_xdr.c */
+extern struct svc_version nfs4_callback_version1;
+
+#else
+
+#define init_nfsv4_state(server)  do { } while (0)
+#define destroy_nfsv4_state(server)       do { } while (0)
+#define nfs4_put_state_owner(inode, owner) do { } while (0)
+#define nfs4_put_open_state(state) do { } while (0)
+#define nfs4_close_state(a, b) do { } while (0)
+
+#endif /* CONFIG_NFS_V4 */
+#endif /* __LINUX_FS_NFS_NFS4_FS.H */

fs/nfs/nfs4proc.c

@@ -48,6 +48,7 @@
 #include <linux/smp_lock.h>
 #include <linux/namei.h>
 
+#include "nfs4_fs.h"
 #include "delegation.h"
 
 #define NFSDBG_FACILITY		NFSDBG_PROC
@@ -62,8 +63,6 @@ static int nfs4_handle_exception(struct nfs_server *server, int errorcode, struc
 extern u32 *nfs4_decode_dirent(u32 *p, struct nfs_entry *entry, int plus);
 extern struct rpc_procinfo nfs4_procedures[];
 
-extern nfs4_stateid zero_stateid;
-
 /* Prevent leaks of NFSv4 errors into userland */
 int nfs4_map_errors(int err)
 {
@@ -104,7 +103,7 @@ const u32 nfs4_statfs_bitmap[2] = {
 	| FATTR4_WORD1_SPACE_TOTAL
 };
 
-u32 nfs4_pathconf_bitmap[2] = {
+const u32 nfs4_pathconf_bitmap[2] = {
 	FATTR4_WORD0_MAXLINK
 	| FATTR4_WORD0_MAXNAME,
 	0
@@ -124,7 +123,7 @@ static void nfs4_setup_readdir(u64 cookie, u32 *verifier, struct dentry *dentry,
 
 	BUG_ON(readdir->count < 80);
 	if (cookie > 2) {
-		readdir->cookie = (cookie > 2) ? cookie : 0;
+		readdir->cookie = cookie;
 		memcpy(&readdir->verifier, verifier, sizeof(readdir->verifier));
 		return;
 	}
@@ -270,14 +269,9 @@ static int nfs4_open_reclaim(struct nfs4_state_owner *sp, struct nfs4_state *sta
 	int err;
 	do {
 		err = _nfs4_open_reclaim(sp, state);
-		switch (err) {
+		if (err != -NFS4ERR_DELAY)
-			case 0:
+			break;
-			case -NFS4ERR_STALE_CLIENTID:
+		nfs4_handle_exception(server, err, &exception);
-			case -NFS4ERR_STALE_STATEID:
-			case -NFS4ERR_EXPIRED:
-				return err;
-		}
-		err = nfs4_handle_exception(server, err, &exception);
 	} while (exception.retry);
 	return err;
 }
@@ -509,6 +503,20 @@ static int _nfs4_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *st
 	goto out_nodeleg;
 }
 
+static inline int nfs4_do_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state, struct dentry *dentry)
+{
+	struct nfs_server *server = NFS_SERVER(dentry->d_inode);
+	struct nfs4_exception exception = { };
+	int err;
+
+	do {
+		err = _nfs4_open_expired(sp, state, dentry);
+		if (err == -NFS4ERR_DELAY)
+			nfs4_handle_exception(server, err, &exception);
+	} while (exception.retry);
+	return err;
+}
+
 static int nfs4_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
 {
 	struct nfs_inode *nfsi = NFS_I(state->inode);
@@ -521,7 +529,7 @@ static int nfs4_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *sta
 			continue;
 		get_nfs_open_context(ctx);
 		spin_unlock(&state->inode->i_lock);
-		status = _nfs4_open_expired(sp, state, ctx->dentry);
+		status = nfs4_do_open_expired(sp, state, ctx->dentry);
 		put_nfs_open_context(ctx);
 		return status;
 	}
@@ -748,11 +756,10 @@ static int _nfs4_do_setattr(struct nfs_server *server, struct nfs_fattr *fattr,
 
         fattr->valid = 0;
 
-	if (state != NULL)
+	if (state != NULL) {
 		msg.rpc_cred = state->owner->so_cred;
-	if (sattr->ia_valid & ATTR_SIZE)
+		nfs4_copy_stateid(&arg.stateid, state, current->files);
-		nfs4_copy_stateid(&arg.stateid, state, NULL);
+	} else
-	else
 		memcpy(&arg.stateid, &zero_stateid, sizeof(arg.stateid));
 
 	return rpc_call_sync(server->client, &msg, 0);
@@ -1116,47 +1123,31 @@ static int
 nfs4_proc_setattr(struct dentry *dentry, struct nfs_fattr *fattr,
 		  struct iattr *sattr)
 {
+	struct rpc_cred *cred;
 	struct inode *inode = dentry->d_inode;
-	int			size_change = sattr->ia_valid & ATTR_SIZE;
+	struct nfs4_state *state;
-	struct nfs4_state	*state = NULL;
-	int need_iput = 0;
 	int status;
 
 	fattr->valid = 0;
 	
-	if (size_change) {
+	cred = rpcauth_lookupcred(NFS_SERVER(inode)->client->cl_auth, 0);
-		struct rpc_cred *cred = rpcauth_lookupcred(NFS_SERVER(inode)->client->cl_auth, 0);
 	if (IS_ERR(cred))
 		return PTR_ERR(cred);
+	/* Search for an existing WRITE delegation first */
+	state = nfs4_open_delegated(inode, FMODE_WRITE, cred);
+	if (!IS_ERR(state)) {
+		/* NB: nfs4_open_delegated() bumps the inode->i_count */
+		iput(inode);
+	} else {
+		/* Search for an existing open(O_WRITE) stateid */
 		state = nfs4_find_state(inode, cred, FMODE_WRITE);
-		if (state == NULL) {
-			state = nfs4_open_delegated(dentry->d_inode,
-					FMODE_WRITE, cred);
-			if (IS_ERR(state))
-				state = nfs4_do_open(dentry->d_parent->d_inode,
-						dentry, FMODE_WRITE,
-						NULL, cred);
-			need_iput = 1;
 	}
-		put_rpccred(cred);
-		if (IS_ERR(state))
-			return PTR_ERR(state);
 
-		if (state->inode != inode) {
-			printk(KERN_WARNING "nfs: raced in setattr (%p != %p), returning -EIO\n", inode, state->inode);
-			status = -EIO;
-			goto out;
-		}
-	}
 	status = nfs4_do_setattr(NFS_SERVER(inode), fattr,
 			NFS_FH(inode), sattr, state);
-out:
+	if (state != NULL)
-	if (state) {
-		inode = state->inode;
 		nfs4_close_state(state, FMODE_WRITE);
-		if (need_iput)
+	put_rpccred(cred);
-			iput(inode);
-	}
 	return status;
 }
 
@@ -1731,6 +1722,10 @@ static int _nfs4_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
 	};
 	int			status;
 
+	dprintk("%s: dentry = %s/%s, cookie = %Lu\n", __FUNCTION__,
+			dentry->d_parent->d_name.name,
+			dentry->d_name.name,
+			(unsigned long long)cookie);
 	lock_kernel();
 	nfs4_setup_readdir(cookie, NFS_COOKIEVERF(dir), dentry, &args);
 	res.pgbase = args.pgbase;
@@ -1738,6 +1733,7 @@ static int _nfs4_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
 	if (status == 0)
 		memcpy(NFS_COOKIEVERF(dir), res.verifier.data, NFS4_VERIFIER_SIZE);
 	unlock_kernel();
+	dprintk("%s: returns %d\n", __FUNCTION__, status);
 	return status;
 }
 
@@ -2163,6 +2159,193 @@ nfs4_proc_file_release(struct inode *inode, struct file *filp)
 	return 0;
 }
 
+static inline int nfs4_server_supports_acls(struct nfs_server *server)
+{
+	return (server->caps & NFS_CAP_ACLS)
+		&& (server->acl_bitmask & ACL4_SUPPORT_ALLOW_ACL)
+		&& (server->acl_bitmask & ACL4_SUPPORT_DENY_ACL);
+}
+
+/* Assuming that XATTR_SIZE_MAX is a multiple of PAGE_CACHE_SIZE, and that
+ * it's OK to put sizeof(void) * (XATTR_SIZE_MAX/PAGE_CACHE_SIZE) bytes on
+ * the stack.
+ */
+#define NFS4ACL_MAXPAGES (XATTR_SIZE_MAX >> PAGE_CACHE_SHIFT)
+
+static void buf_to_pages(const void *buf, size_t buflen,
+		struct page **pages, unsigned int *pgbase)
+{
+	const void *p = buf;
+
+	*pgbase = offset_in_page(buf);
+	p -= *pgbase;
+	while (p < buf + buflen) {
+		*(pages++) = virt_to_page(p);
+		p += PAGE_CACHE_SIZE;
+	}
+}
+
+struct nfs4_cached_acl {
+	int cached;
+	size_t len;
+	char data[0];
+};
+
+static void nfs4_set_cached_acl(struct inode *inode, struct nfs4_cached_acl *acl)
+{
+	struct nfs_inode *nfsi = NFS_I(inode);
+
+	spin_lock(&inode->i_lock);
+	kfree(nfsi->nfs4_acl);
+	nfsi->nfs4_acl = acl;
+	spin_unlock(&inode->i_lock);
+}
+
+static void nfs4_zap_acl_attr(struct inode *inode)
+{
+	nfs4_set_cached_acl(inode, NULL);
+}
+
+static inline ssize_t nfs4_read_cached_acl(struct inode *inode, char *buf, size_t buflen)
+{
+	struct nfs_inode *nfsi = NFS_I(inode);
+	struct nfs4_cached_acl *acl;
+	int ret = -ENOENT;
+
+	spin_lock(&inode->i_lock);
+	acl = nfsi->nfs4_acl;
+	if (acl == NULL)
+		goto out;
+	if (buf == NULL) /* user is just asking for length */
+		goto out_len;
+	if (acl->cached == 0)
+		goto out;
+	ret = -ERANGE; /* see getxattr(2) man page */
+	if (acl->len > buflen)
+		goto out;
+	memcpy(buf, acl->data, acl->len);
+out_len:
+	ret = acl->len;
+out:
+	spin_unlock(&inode->i_lock);
+	return ret;
+}
+
+static void nfs4_write_cached_acl(struct inode *inode, const char *buf, size_t acl_len)
+{
+	struct nfs4_cached_acl *acl;
+
+	if (buf && acl_len <= PAGE_SIZE) {
+		acl = kmalloc(sizeof(*acl) + acl_len, GFP_KERNEL);
+		if (acl == NULL)
+			goto out;
+		acl->cached = 1;
+		memcpy(acl->data, buf, acl_len);
+	} else {
+		acl = kmalloc(sizeof(*acl), GFP_KERNEL);
+		if (acl == NULL)
+			goto out;
+		acl->cached = 0;
+	}
+	acl->len = acl_len;
+out:
+	nfs4_set_cached_acl(inode, acl);
+}
+
+static inline ssize_t nfs4_get_acl_uncached(struct inode *inode, void *buf, size_t buflen)
+{
+	struct page *pages[NFS4ACL_MAXPAGES];
+	struct nfs_getaclargs args = {
+		.fh = NFS_FH(inode),
+		.acl_pages = pages,
+		.acl_len = buflen,
+	};
+	size_t resp_len = buflen;
+	void *resp_buf;
+	struct rpc_message msg = {
+		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETACL],
+		.rpc_argp = &args,
+		.rpc_resp = &resp_len,
+	};
+	struct page *localpage = NULL;
+	int ret;
+
+	if (buflen < PAGE_SIZE) {
+		/* As long as we're doing a round trip to the server anyway,
+		 * let's be prepared for a page of acl data. */
+		localpage = alloc_page(GFP_KERNEL);
+		resp_buf = page_address(localpage);
+		if (localpage == NULL)
+			return -ENOMEM;
+		args.acl_pages[0] = localpage;
+		args.acl_pgbase = 0;
+		args.acl_len = PAGE_SIZE;
+	} else {
+		resp_buf = buf;
+		buf_to_pages(buf, buflen, args.acl_pages, &args.acl_pgbase);
+	}
+	ret = rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
+	if (ret)
+		goto out_free;
+	if (resp_len > args.acl_len)
+		nfs4_write_cached_acl(inode, NULL, resp_len);
+	else
+		nfs4_write_cached_acl(inode, resp_buf, resp_len);
+	if (buf) {
+		ret = -ERANGE;
+		if (resp_len > buflen)
+			goto out_free;
+		if (localpage)
+			memcpy(buf, resp_buf, resp_len);
+	}
+	ret = resp_len;
+out_free:
+	if (localpage)
+		__free_page(localpage);
+	return ret;
+}
+
+static ssize_t nfs4_proc_get_acl(struct inode *inode, void *buf, size_t buflen)
+{
+	struct nfs_server *server = NFS_SERVER(inode);
+	int ret;
+
+	if (!nfs4_server_supports_acls(server))
+		return -EOPNOTSUPP;
+	ret = nfs_revalidate_inode(server, inode);
+	if (ret < 0)
+		return ret;
+	ret = nfs4_read_cached_acl(inode, buf, buflen);
+	if (ret != -ENOENT)
+		return ret;
+	return nfs4_get_acl_uncached(inode, buf, buflen);
+}
+
+static int nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen)
+{
+	struct nfs_server *server = NFS_SERVER(inode);
+	struct page *pages[NFS4ACL_MAXPAGES];
+	struct nfs_setaclargs arg = {
+		.fh		= NFS_FH(inode),
+		.acl_pages	= pages,
+		.acl_len	= buflen,
+	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_SETACL],
+		.rpc_argp	= &arg,
+		.rpc_resp	= NULL,
+	};
+	int ret;
+
+	if (!nfs4_server_supports_acls(server))
+		return -EOPNOTSUPP;
+	buf_to_pages(buf, buflen, arg.acl_pages, &arg.acl_pgbase);
+	ret = rpc_call_sync(NFS_SERVER(inode)->client, &msg, 0);
+	if (ret == 0)
+		nfs4_write_cached_acl(inode, buf, buflen);
+	return ret;
+}
+
 static int
 nfs4_async_handle_error(struct rpc_task *task, struct nfs_server *server)
 {
@@ -2448,14 +2631,11 @@ static int _nfs4_proc_getlk(struct nfs4_state *state, int cmd, struct file_lock
 	down_read(&clp->cl_sem);
 	nlo.clientid = clp->cl_clientid;
 	down(&state->lock_sema);
-	lsp = nfs4_find_lock_state(state, request->fl_owner);
+	status = nfs4_set_lock_state(state, request);
-	if (lsp)
+	if (status != 0)
+		goto out;
+	lsp = request->fl_u.nfs4_fl.owner;
 	nlo.id = lsp->ls_id; 
-	else {
-		spin_lock(&clp->cl_lock);
-		nlo.id = nfs4_alloc_lockowner_id(clp);
-		spin_unlock(&clp->cl_lock);
-	}
 	arg.u.lockt = &nlo;
 	status = rpc_call_sync(server->client, &msg, 0);
 	if (!status) {
@@ -2476,8 +2656,7 @@ static int _nfs4_proc_getlk(struct nfs4_state *state, int cmd, struct file_lock
 		request->fl_pid = 0;
 		status = 0;
 	}
-	if (lsp)
+out:
-		nfs4_put_lock_state(lsp);
 	up(&state->lock_sema);
 	up_read(&clp->cl_sem);
 	return status;
@@ -2537,28 +2716,26 @@ static int _nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock
 	};
 	struct nfs4_lock_state *lsp;
 	struct nfs_locku_opargs luargs;
-	int status = 0;
+	int status;
 			
 	down_read(&clp->cl_sem);
 	down(&state->lock_sema);
-	lsp = nfs4_find_lock_state(state, request->fl_owner);
+	status = nfs4_set_lock_state(state, request);
-	if (!lsp)
+	if (status != 0)
 		goto out;
+	lsp = request->fl_u.nfs4_fl.owner;
 	/* We might have lost the locks! */
-	if ((lsp->ls_flags & NFS_LOCK_INITIALIZED) != 0) {
+	if ((lsp->ls_flags & NFS_LOCK_INITIALIZED) == 0)
+		goto out;
 	luargs.seqid = lsp->ls_seqid;
 	memcpy(&luargs.stateid, &lsp->ls_stateid, sizeof(luargs.stateid));
 	arg.u.locku = &luargs;
 	status = rpc_call_sync(server->client, &msg, RPC_TASK_NOINTR);
 	nfs4_increment_lock_seqid(status, lsp);
-	}
 
-	if (status == 0) {
+	if (status == 0)
 		memcpy(&lsp->ls_stateid,  &res.u.stateid, 
 				sizeof(lsp->ls_stateid));
-		nfs4_notify_unlck(state, request, lsp);
-	}
-	nfs4_put_lock_state(lsp);
 out:
 	up(&state->lock_sema);
 	if (status == 0)
@@ -2584,7 +2761,7 @@ static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *r
 {
 	struct inode *inode = state->inode;
 	struct nfs_server *server = NFS_SERVER(inode);
-	struct nfs4_lock_state *lsp;
+	struct nfs4_lock_state *lsp = request->fl_u.nfs4_fl.owner;
 	struct nfs_lockargs arg = {
 		.fh = NFS_FH(inode),
 		.type = nfs4_lck_type(cmd, request),
@@ -2606,9 +2783,6 @@ static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *r
 	};
 	int status;
 
-	lsp = nfs4_get_lock_state(state, request->fl_owner);
-	if (lsp == NULL)
-		return -ENOMEM;
 	if (!(lsp->ls_flags & NFS_LOCK_INITIALIZED)) {
 		struct nfs4_state_owner *owner = state->owner;
 		struct nfs_open_to_lock otl = {
@@ -2630,38 +2804,57 @@ static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *r
 		* seqid mutating errors */
 		nfs4_increment_seqid(status, owner);
 		up(&owner->so_sema);
+		if (status == 0) {
+			lsp->ls_flags |= NFS_LOCK_INITIALIZED;
+			lsp->ls_seqid++;
+		}
 	} else {
 		struct nfs_exist_lock el = {
 			.seqid = lsp->ls_seqid,
 		};
 		memcpy(&el.stateid, &lsp->ls_stateid, sizeof(el.stateid));
 		largs.u.exist_lock = &el;
-		largs.new_lock_owner = 0;
 		arg.u.lock = &largs;
 		status = rpc_call_sync(server->client, &msg, RPC_TASK_NOINTR);
-	}
 		/* increment seqid on success, and * seqid mutating errors*/
 		nfs4_increment_lock_seqid(status, lsp);
+	}
 	/* save the returned stateid. */
-	if (status == 0) {
+	if (status == 0)
 		memcpy(&lsp->ls_stateid, &res.u.stateid, sizeof(nfs4_stateid));
-		lsp->ls_flags |= NFS_LOCK_INITIALIZED;
+	else if (status == -NFS4ERR_DENIED)
-		if (!reclaim)
-			nfs4_notify_setlk(state, request, lsp);
-	} else if (status == -NFS4ERR_DENIED)
 		status = -EAGAIN;
-	nfs4_put_lock_state(lsp);
 	return status;
 }
 
 static int nfs4_lock_reclaim(struct nfs4_state *state, struct file_lock *request)
 {
-	return _nfs4_do_setlk(state, F_SETLK, request, 1);
+	struct nfs_server *server = NFS_SERVER(state->inode);
+	struct nfs4_exception exception = { };
+	int err;
+
+	do {
+		err = _nfs4_do_setlk(state, F_SETLK, request, 1);
+		if (err != -NFS4ERR_DELAY)
+			break;
+		nfs4_handle_exception(server, err, &exception);
+	} while (exception.retry);
+	return err;
 }
 
 static int nfs4_lock_expired(struct nfs4_state *state, struct file_lock *request)
 {
-	return _nfs4_do_setlk(state, F_SETLK, request, 0);
+	struct nfs_server *server = NFS_SERVER(state->inode);
+	struct nfs4_exception exception = { };
+	int err;
+
+	do {
+		err = _nfs4_do_setlk(state, F_SETLK, request, 0);
+		if (err != -NFS4ERR_DELAY)
+			break;
+		nfs4_handle_exception(server, err, &exception);
+	} while (exception.retry);
+	return err;
 }
 
 static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
@@ -2671,6 +2864,8 @@ static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock
 
 	down_read(&clp->cl_sem);
 	down(&state->lock_sema);
+	status = nfs4_set_lock_state(state, request);
+	if (status == 0)
 		status = _nfs4_do_setlk(state, cmd, request, 0);
 	up(&state->lock_sema);
 	if (status == 0) {
@@ -2729,10 +2924,53 @@ nfs4_proc_lock(struct file *filp, int cmd, struct file_lock *request)
 		if (signalled())
 			break;
 	} while(status < 0);
-
 	return status;
 }
 
+
+#define XATTR_NAME_NFSV4_ACL "system.nfs4_acl"
+
+int nfs4_setxattr(struct dentry *dentry, const char *key, const void *buf,
+		size_t buflen, int flags)
+{
+	struct inode *inode = dentry->d_inode;
+
+	if (strcmp(key, XATTR_NAME_NFSV4_ACL) != 0)
+		return -EOPNOTSUPP;
+
+	if (!S_ISREG(inode->i_mode) &&
+	    (!S_ISDIR(inode->i_mode) || inode->i_mode & S_ISVTX))
+		return -EPERM;
+
+	return nfs4_proc_set_acl(inode, buf, buflen);
+}
+
+/* The getxattr man page suggests returning -ENODATA for unknown attributes,
+ * and that's what we'll do for e.g. user attributes that haven't been set.
+ * But we'll follow ext2/ext3's lead by returning -EOPNOTSUPP for unsupported
+ * attributes in kernel-managed attribute namespaces. */
+ssize_t nfs4_getxattr(struct dentry *dentry, const char *key, void *buf,
+		size_t buflen)
+{
+	struct inode *inode = dentry->d_inode;
+
+	if (strcmp(key, XATTR_NAME_NFSV4_ACL) != 0)
+		return -EOPNOTSUPP;
+
+	return nfs4_proc_get_acl(inode, buf, buflen);
+}
+
+ssize_t nfs4_listxattr(struct dentry *dentry, char *buf, size_t buflen)
+{
+	size_t len = strlen(XATTR_NAME_NFSV4_ACL) + 1;
+
+	if (buf && buflen < len)
+		return -ERANGE;
+	if (buf)
+		memcpy(buf, XATTR_NAME_NFSV4_ACL, len);
+	return len;
+}
+
 struct nfs4_state_recovery_ops nfs4_reboot_recovery_ops = {
 	.recover_open	= nfs4_open_reclaim,
 	.recover_lock	= nfs4_lock_reclaim,
@@ -2743,10 +2981,20 @@ struct nfs4_state_recovery_ops nfs4_network_partition_recovery_ops = {
 	.recover_lock	= nfs4_lock_expired,
 };
 
+static struct inode_operations nfs4_file_inode_operations = {
+	.permission	= nfs_permission,
+	.getattr	= nfs_getattr,
+	.setattr	= nfs_setattr,
+	.getxattr	= nfs4_getxattr,
+	.setxattr	= nfs4_setxattr,
+	.listxattr	= nfs4_listxattr,
+};
+
 struct nfs_rpc_ops	nfs_v4_clientops = {
 	.version	= 4,			/* protocol version */
 	.dentry_ops	= &nfs4_dentry_operations,
 	.dir_inode_ops	= &nfs4_dir_inode_operations,
+	.file_inode_ops	= &nfs4_file_inode_operations,
 	.getroot	= nfs4_proc_get_root,
 	.getattr	= nfs4_proc_getattr,
 	.setattr	= nfs4_proc_setattr,
@@ -2777,6 +3025,7 @@ struct nfs_rpc_ops	nfs_v4_clientops = {
 	.file_open      = nfs4_proc_file_open,
 	.file_release   = nfs4_proc_file_release,
 	.lock		= nfs4_proc_lock,
+	.clear_acl_cache = nfs4_zap_acl_attr,
 };
 
 /*

fs/nfs/nfs4renewd.c

@@ -53,6 +53,7 @@
 #include <linux/nfs.h>
 #include <linux/nfs4.h>
 #include <linux/nfs_fs.h>
+#include "nfs4_fs.h"
 
 #define NFSDBG_FACILITY	NFSDBG_PROC
 

fs/nfs/nfs4state.c

@@ -46,24 +46,18 @@
 #include <linux/workqueue.h>
 #include <linux/bitops.h>
 
+#include "nfs4_fs.h"
 #include "callback.h"
 #include "delegation.h"
 
 #define OPENOWNER_POOL_SIZE	8
 
+const nfs4_stateid zero_stateid;
+
 static DEFINE_SPINLOCK(state_spinlock);
-
-nfs4_stateid zero_stateid;
-
-#if 0
-nfs4_stateid one_stateid =
-	{ 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff };
-#endif
-
 static LIST_HEAD(nfs4_clientid_list);
 
 static void nfs4_recover_state(void *);
-extern void nfs4_renew_state(void *);
 
 void
 init_nfsv4_state(struct nfs_server *server)
@@ -116,6 +110,7 @@ nfs4_alloc_client(struct in_addr *addr)
 	INIT_LIST_HEAD(&clp->cl_superblocks);
 	init_waitqueue_head(&clp->cl_waitq);
 	rpc_init_wait_queue(&clp->cl_rpcwaitq, "NFS4 client");
+	clp->cl_rpcclient = ERR_PTR(-EINVAL);
 	clp->cl_boot_time = CURRENT_TIME;
 	clp->cl_state = 1 << NFS4CLNT_OK;
 	return clp;
@@ -137,7 +132,7 @@ nfs4_free_client(struct nfs4_client *clp)
 	if (clp->cl_cred)
 		put_rpccred(clp->cl_cred);
 	nfs_idmap_delete(clp);
-	if (clp->cl_rpcclient)
+	if (!IS_ERR(clp->cl_rpcclient))
 		rpc_shutdown_client(clp->cl_rpcclient);
 	kfree(clp);
 	nfs_callback_down();
@@ -365,7 +360,7 @@ nfs4_alloc_open_state(void)
 	atomic_set(&state->count, 1);
 	INIT_LIST_HEAD(&state->lock_states);
 	init_MUTEX(&state->lock_sema);
-	rwlock_init(&state->state_lock);
+	spin_lock_init(&state->state_lock);
 	return state;
 }
 
@@ -547,16 +542,6 @@ __nfs4_find_lock_state(struct nfs4_state *state, fl_owner_t fl_owner)
 	return NULL;
 }
 
-struct nfs4_lock_state *
-nfs4_find_lock_state(struct nfs4_state *state, fl_owner_t fl_owner)
-{
-	struct nfs4_lock_state *lsp;
-	read_lock(&state->state_lock);
-	lsp = __nfs4_find_lock_state(state, fl_owner);
-	read_unlock(&state->state_lock);
-	return lsp;
-}
-
 /*
  * Return a compatible lock_state. If no initialized lock_state structure
  * exists, return an uninitialized one.
@@ -573,14 +558,13 @@ static struct nfs4_lock_state *nfs4_alloc_lock_state(struct nfs4_state *state, f
 		return NULL;
 	lsp->ls_flags = 0;
 	lsp->ls_seqid = 0;	/* arbitrary */
-	lsp->ls_id = -1; 
 	memset(lsp->ls_stateid.data, 0, sizeof(lsp->ls_stateid.data));
 	atomic_set(&lsp->ls_count, 1);
 	lsp->ls_owner = fl_owner;
-	INIT_LIST_HEAD(&lsp->ls_locks);
 	spin_lock(&clp->cl_lock);
 	lsp->ls_id = nfs4_alloc_lockowner_id(clp);
 	spin_unlock(&clp->cl_lock);
+	INIT_LIST_HEAD(&lsp->ls_locks);
 	return lsp;
 }
 
@@ -590,34 +574,103 @@ static struct nfs4_lock_state *nfs4_alloc_lock_state(struct nfs4_state *state, f
  *
  * The caller must be holding state->lock_sema and clp->cl_sem
  */
-struct nfs4_lock_state *nfs4_get_lock_state(struct nfs4_state *state, fl_owner_t owner)
+static struct nfs4_lock_state *nfs4_get_lock_state(struct nfs4_state *state, fl_owner_t owner)
+{
+	struct nfs4_lock_state *lsp, *new = NULL;
+	
+	for(;;) {
+		spin_lock(&state->state_lock);
+		lsp = __nfs4_find_lock_state(state, owner);
+		if (lsp != NULL)
+			break;
+		if (new != NULL) {
+			new->ls_state = state;
+			list_add(&new->ls_locks, &state->lock_states);
+			set_bit(LK_STATE_IN_USE, &state->flags);
+			lsp = new;
+			new = NULL;
+			break;
+		}
+		spin_unlock(&state->state_lock);
+		new = nfs4_alloc_lock_state(state, owner);
+		if (new == NULL)
+			return NULL;
+	}
+	spin_unlock(&state->state_lock);
+	kfree(new);
+	return lsp;
+}
+
+/*
+ * Release reference to lock_state, and free it if we see that
+ * it is no longer in use
+ */
+static void nfs4_put_lock_state(struct nfs4_lock_state *lsp)
+{
+	struct nfs4_state *state;
+
+	if (lsp == NULL)
+		return;
+	state = lsp->ls_state;
+	if (!atomic_dec_and_lock(&lsp->ls_count, &state->state_lock))
+		return;
+	list_del(&lsp->ls_locks);
+	if (list_empty(&state->lock_states))
+		clear_bit(LK_STATE_IN_USE, &state->flags);
+	spin_unlock(&state->state_lock);
+	kfree(lsp);
+}
+
+static void nfs4_fl_copy_lock(struct file_lock *dst, struct file_lock *src)
+{
+	struct nfs4_lock_state *lsp = src->fl_u.nfs4_fl.owner;
+
+	dst->fl_u.nfs4_fl.owner = lsp;
+	atomic_inc(&lsp->ls_count);
+}
+
+static void nfs4_fl_release_lock(struct file_lock *fl)
+{
+	nfs4_put_lock_state(fl->fl_u.nfs4_fl.owner);
+}
+
+static struct file_lock_operations nfs4_fl_lock_ops = {
+	.fl_copy_lock = nfs4_fl_copy_lock,
+	.fl_release_private = nfs4_fl_release_lock,
+};
+
+int nfs4_set_lock_state(struct nfs4_state *state, struct file_lock *fl)
 {
 	struct nfs4_lock_state *lsp;
 
-	lsp = nfs4_find_lock_state(state, owner);
+	if (fl->fl_ops != NULL)
+		return 0;
+	lsp = nfs4_get_lock_state(state, fl->fl_owner);
 	if (lsp == NULL)
-		lsp = nfs4_alloc_lock_state(state, owner);
+		return -ENOMEM;
-	return lsp;
+	fl->fl_u.nfs4_fl.owner = lsp;
+	fl->fl_ops = &nfs4_fl_lock_ops;
+	return 0;
 }
 
 /*
  * Byte-range lock aware utility to initialize the stateid of read/write
  * requests.
  */
-void
+void nfs4_copy_stateid(nfs4_stateid *dst, struct nfs4_state *state, fl_owner_t fl_owner)
-nfs4_copy_stateid(nfs4_stateid *dst, struct nfs4_state *state, fl_owner_t fl_owner)
 {
-	if (test_bit(LK_STATE_IN_USE, &state->flags)) {
 	struct nfs4_lock_state *lsp;
 
-		lsp = nfs4_find_lock_state(state, fl_owner);
-		if (lsp) {
-			memcpy(dst, &lsp->ls_stateid, sizeof(*dst));
-			nfs4_put_lock_state(lsp);
-			return;
-		}
-	}
 	memcpy(dst, &state->stateid, sizeof(*dst));
+	if (test_bit(LK_STATE_IN_USE, &state->flags) == 0)
+		return;
+
+	spin_lock(&state->state_lock);
+	lsp = __nfs4_find_lock_state(state, fl_owner);
+	if (lsp != NULL && (lsp->ls_flags & NFS_LOCK_INITIALIZED) != 0)
+		memcpy(dst, &lsp->ls_stateid, sizeof(*dst));
+	spin_unlock(&state->state_lock);
+	nfs4_put_lock_state(lsp);
 }
 
 /*
@@ -629,84 +682,6 @@ void nfs4_increment_lock_seqid(int status, struct nfs4_lock_state *lsp)
 		lsp->ls_seqid++;
 }
 
-/* 
-* Check to see if the request lock (type FL_UNLK) effects the fl lock.
-*
-* fl and request must have the same posix owner
-*
-* return: 
-* 0 -> fl not effected by request
-* 1 -> fl consumed by request
-*/
-
-static int
-nfs4_check_unlock(struct file_lock *fl, struct file_lock *request)
-{
-	if (fl->fl_start >= request->fl_start && fl->fl_end <= request->fl_end)
-		return 1;
-	return 0;
-}
-
-/*
- * Post an initialized lock_state on the state->lock_states list.
- */
-void nfs4_notify_setlk(struct nfs4_state *state, struct file_lock *request, struct nfs4_lock_state *lsp)
-{
-	if (!list_empty(&lsp->ls_locks))
-		return;
-	atomic_inc(&lsp->ls_count);
-	write_lock(&state->state_lock);
-	list_add(&lsp->ls_locks, &state->lock_states);
-	set_bit(LK_STATE_IN_USE, &state->flags);
-	write_unlock(&state->state_lock);
-}
-
-/* 
- * to decide to 'reap' lock state:
- * 1) search i_flock for file_locks with fl.lock_state = to ls.
- * 2) determine if unlock will consume found lock. 
- * 	if so, reap
- *
- * 	else, don't reap.
- *
- */
-void
-nfs4_notify_unlck(struct nfs4_state *state, struct file_lock *request, struct nfs4_lock_state *lsp)
-{
-	struct inode *inode = state->inode;
-	struct file_lock *fl;
-
-	for (fl = inode->i_flock; fl != NULL; fl = fl->fl_next) {
-		if (!(fl->fl_flags & FL_POSIX))
-			continue;
-		if (fl->fl_owner != lsp->ls_owner)
-			continue;
-		/* Exit if we find at least one lock which is not consumed */
-		if (nfs4_check_unlock(fl,request) == 0)
-			return;
-	}
-
-	write_lock(&state->state_lock);
-	list_del_init(&lsp->ls_locks);
-	if (list_empty(&state->lock_states))
-		clear_bit(LK_STATE_IN_USE, &state->flags);
-	write_unlock(&state->state_lock);
-	nfs4_put_lock_state(lsp);
-}
-
-/*
- * Release reference to lock_state, and free it if we see that
- * it is no longer in use
- */
-void
-nfs4_put_lock_state(struct nfs4_lock_state *lsp)
-{
-	if (!atomic_dec_and_test(&lsp->ls_count))
-		return;
-	BUG_ON (!list_empty(&lsp->ls_locks));
-	kfree(lsp);
-}
-
 /*
 * Called with sp->so_sema and clp->cl_sem held.
 *

fs/nfs/nfs4xdr.c

@@ -51,6 +51,7 @@
 #include <linux/nfs4.h>
 #include <linux/nfs_fs.h>
 #include <linux/nfs_idmap.h>
+#include "nfs4_fs.h"
 
 #define NFSDBG_FACILITY		NFSDBG_XDR
 
@@ -82,12 +83,16 @@ static int nfs_stat_to_errno(int);
 #define encode_getfh_maxsz      (op_encode_hdr_maxsz)
 #define decode_getfh_maxsz      (op_decode_hdr_maxsz + 1 + \
 				((3+NFS4_FHSIZE) >> 2))
-#define encode_getattr_maxsz    (op_encode_hdr_maxsz + 3)
+#define nfs4_fattr_bitmap_maxsz 3
+#define encode_getattr_maxsz    (op_encode_hdr_maxsz + nfs4_fattr_bitmap_maxsz)
 #define nfs4_name_maxsz		(1 + ((3 + NFS4_MAXNAMLEN) >> 2))
 #define nfs4_path_maxsz		(1 + ((3 + NFS4_MAXPATHLEN) >> 2))
-#define nfs4_fattr_bitmap_maxsz (36 + 2 * nfs4_name_maxsz)
+/* This is based on getfattr, which uses the most attributes: */
-#define decode_getattr_maxsz    (op_decode_hdr_maxsz + 3 + \
+#define nfs4_fattr_value_maxsz	(1 + (1 + 2 + 2 + 4 + 2 + 1 + 1 + 2 + 2 + \
-                                nfs4_fattr_bitmap_maxsz)
+				3 + 3 + 3 + 2 * nfs4_name_maxsz))
+#define nfs4_fattr_maxsz	(nfs4_fattr_bitmap_maxsz + \
+				nfs4_fattr_value_maxsz)
+#define decode_getattr_maxsz    (op_decode_hdr_maxsz + nfs4_fattr_maxsz)
 #define encode_savefh_maxsz     (op_encode_hdr_maxsz)
 #define decode_savefh_maxsz     (op_decode_hdr_maxsz)
 #define encode_fsinfo_maxsz	(op_encode_hdr_maxsz + 2)
@@ -122,11 +127,11 @@ static int nfs_stat_to_errno(int);
 #define encode_symlink_maxsz	(op_encode_hdr_maxsz + \
 				1 + nfs4_name_maxsz + \
 				nfs4_path_maxsz + \
-				nfs4_fattr_bitmap_maxsz)
+				nfs4_fattr_maxsz)
 #define decode_symlink_maxsz	(op_decode_hdr_maxsz + 8)
 #define encode_create_maxsz	(op_encode_hdr_maxsz + \
 				2 + nfs4_name_maxsz + \
-				nfs4_fattr_bitmap_maxsz)
+				nfs4_fattr_maxsz)
 #define decode_create_maxsz	(op_decode_hdr_maxsz + 8)
 #define encode_delegreturn_maxsz (op_encode_hdr_maxsz + 4)
 #define decode_delegreturn_maxsz (op_decode_hdr_maxsz)
@@ -205,7 +210,7 @@ static int nfs_stat_to_errno(int);
 #define NFS4_enc_setattr_sz     (compound_encode_hdr_maxsz + \
                                 encode_putfh_maxsz + \
                                 op_encode_hdr_maxsz + 4 + \
-                                nfs4_fattr_bitmap_maxsz + \
+                                nfs4_fattr_maxsz + \
                                 encode_getattr_maxsz)
 #define NFS4_dec_setattr_sz     (compound_decode_hdr_maxsz + \
                                 decode_putfh_maxsz + \
@@ -360,6 +365,20 @@ static int nfs_stat_to_errno(int);
 				encode_delegreturn_maxsz)
 #define NFS4_dec_delegreturn_sz (compound_decode_hdr_maxsz + \
 				decode_delegreturn_maxsz)
+#define NFS4_enc_getacl_sz	(compound_encode_hdr_maxsz + \
+				encode_putfh_maxsz + \
+				encode_getattr_maxsz)
+#define NFS4_dec_getacl_sz	(compound_decode_hdr_maxsz + \
+				decode_putfh_maxsz + \
+				op_decode_hdr_maxsz + \
+				nfs4_fattr_bitmap_maxsz + 1)
+#define NFS4_enc_setacl_sz	(compound_encode_hdr_maxsz + \
+				encode_putfh_maxsz + \
+				op_encode_hdr_maxsz + 4 + \
+				nfs4_fattr_bitmap_maxsz + 1)
+#define NFS4_dec_setacl_sz	(compound_decode_hdr_maxsz + \
+				decode_putfh_maxsz + \
+				op_decode_hdr_maxsz + nfs4_fattr_bitmap_maxsz)
 
 static struct {
 	unsigned int	mode;
@@ -459,7 +478,7 @@ static int encode_attrs(struct xdr_stream *xdr, const struct iattr *iap, const s
 	 * In the worst-case, this would be
 	 *   12(bitmap) + 4(attrlen) + 8(size) + 4(mode) + 4(atime) + 4(mtime)
 	 *          = 36 bytes, plus any contribution from variable-length fields
-	 *            such as owner/group/acl's.
+	 *            such as owner/group.
 	 */
 	len = 16;
 
@@ -660,8 +679,6 @@ static int encode_getattr_two(struct xdr_stream *xdr, uint32_t bm0, uint32_t bm1
 
 static int encode_getfattr(struct xdr_stream *xdr, const u32* bitmask)
 {
-	extern u32 nfs4_fattr_bitmap[];
-
 	return encode_getattr_two(xdr,
 			bitmask[0] & nfs4_fattr_bitmap[0],
 			bitmask[1] & nfs4_fattr_bitmap[1]);
@@ -669,8 +686,6 @@ static int encode_getfattr(struct xdr_stream *xdr, const u32* bitmask)
 
 static int encode_fsinfo(struct xdr_stream *xdr, const u32* bitmask)
 {
-	extern u32 nfs4_fsinfo_bitmap[];
-
 	return encode_getattr_two(xdr, bitmask[0] & nfs4_fsinfo_bitmap[0],
 			bitmask[1] & nfs4_fsinfo_bitmap[1]);
 }
@@ -969,7 +984,6 @@ static int encode_putrootfh(struct xdr_stream *xdr)
 
 static void encode_stateid(struct xdr_stream *xdr, const struct nfs_open_context *ctx)
 {
-	extern nfs4_stateid zero_stateid;
 	nfs4_stateid stateid;
 	uint32_t *p;
 
@@ -1000,6 +1014,10 @@ static int encode_read(struct xdr_stream *xdr, const struct nfs_readargs *args)
 static int encode_readdir(struct xdr_stream *xdr, const struct nfs4_readdir_arg *readdir, struct rpc_rqst *req)
 {
 	struct rpc_auth *auth = req->rq_task->tk_auth;
+	uint32_t attrs[2] = {
+		FATTR4_WORD0_RDATTR_ERROR|FATTR4_WORD0_FILEID,
+		FATTR4_WORD1_MOUNTED_ON_FILEID,
+	};
 	int replen;
 	uint32_t *p;
 
@@ -1010,13 +1028,20 @@ static int encode_readdir(struct xdr_stream *xdr, const struct nfs4_readdir_arg
 	WRITE32(readdir->count >> 1);  /* We're not doing readdirplus */
 	WRITE32(readdir->count);
 	WRITE32(2);
-	if (readdir->bitmask[1] & FATTR4_WORD1_MOUNTED_ON_FILEID) {
+	/* Switch to mounted_on_fileid if the server supports it */
-		WRITE32(0);
+	if (readdir->bitmask[1] & FATTR4_WORD1_MOUNTED_ON_FILEID)
-		WRITE32(FATTR4_WORD1_MOUNTED_ON_FILEID);
+		attrs[0] &= ~FATTR4_WORD0_FILEID;
-	} else {
+	else
-		WRITE32(FATTR4_WORD0_FILEID);
+		attrs[1] &= ~FATTR4_WORD1_MOUNTED_ON_FILEID;
-		WRITE32(0);
+	WRITE32(attrs[0] & readdir->bitmask[0]);
-	}
+	WRITE32(attrs[1] & readdir->bitmask[1]);
+	dprintk("%s: cookie = %Lu, verifier = 0x%x%x, bitmap = 0x%x%x\n",
+			__FUNCTION__,
+			(unsigned long long)readdir->cookie,
+			((u32 *)readdir->verifier.data)[0],
+			((u32 *)readdir->verifier.data)[1],
+			attrs[0] & readdir->bitmask[0],
+			attrs[1] & readdir->bitmask[1]);
 
 	/* set up reply kvec
 	 *    toplevel_status + taglen + rescount + OP_PUTFH + status
@@ -1025,6 +1050,9 @@ static int encode_readdir(struct xdr_stream *xdr, const struct nfs4_readdir_arg
 	replen = (RPC_REPHDRSIZE + auth->au_rslack + 9) << 2;
 	xdr_inline_pages(&req->rq_rcv_buf, replen, readdir->pages,
 			 readdir->pgbase, readdir->count);
+	dprintk("%s: inlined page args = (%u, %p, %u, %u)\n",
+			__FUNCTION__, replen, readdir->pages,
+			readdir->pgbase, readdir->count);
 
 	return 0;
 }
@@ -1088,6 +1116,25 @@ static int encode_renew(struct xdr_stream *xdr, const struct nfs4_client *client
 	return 0;
 }
 
+static int
+encode_setacl(struct xdr_stream *xdr, struct nfs_setaclargs *arg)
+{
+	uint32_t *p;
+
+	RESERVE_SPACE(4+sizeof(zero_stateid.data));
+	WRITE32(OP_SETATTR);
+	WRITEMEM(zero_stateid.data, sizeof(zero_stateid.data));
+	RESERVE_SPACE(2*4);
+	WRITE32(1);
+	WRITE32(FATTR4_WORD0_ACL);
+	if (arg->acl_len % 4)
+		return -EINVAL;
+	RESERVE_SPACE(4);
+	WRITE32(arg->acl_len);
+	xdr_write_pages(xdr, arg->acl_pages, arg->acl_pgbase, arg->acl_len);
+	return 0;
+}
+
 static int
 encode_savefh(struct xdr_stream *xdr)
 {
@@ -1631,6 +1678,34 @@ static int nfs4_xdr_enc_setattr(struct rpc_rqst *req, uint32_t *p, struct nfs_se
         return status;
 }
 
+/*
+ * Encode a GETACL request
+ */
+static int
+nfs4_xdr_enc_getacl(struct rpc_rqst *req, uint32_t *p,
+		struct nfs_getaclargs *args)
+{
+	struct xdr_stream xdr;
+	struct rpc_auth *auth = req->rq_task->tk_auth;
+	struct compound_hdr hdr = {
+		.nops   = 2,
+	};
+	int replen, status;
+
+	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
+	encode_compound_hdr(&xdr, &hdr);
+	status = encode_putfh(&xdr, args->fh);
+	if (status)
+		goto out;
+	status = encode_getattr_two(&xdr, FATTR4_WORD0_ACL, 0);
+	/* set up reply buffer: */
+	replen = (RPC_REPHDRSIZE + auth->au_rslack + NFS4_dec_getacl_sz) << 2;
+	xdr_inline_pages(&req->rq_rcv_buf, replen,
+		args->acl_pages, args->acl_pgbase, args->acl_len);
+out:
+	return status;
+}
+
 /*
  * Encode a WRITE request
  */
@@ -1697,7 +1772,6 @@ static int nfs4_xdr_enc_fsinfo(struct rpc_rqst *req, uint32_t *p, struct nfs4_fs
  */
 static int nfs4_xdr_enc_pathconf(struct rpc_rqst *req, uint32_t *p, const struct nfs4_pathconf_arg *args)
 {
-	extern u32 nfs4_pathconf_bitmap[2];
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
 		.nops = 2,
@@ -1718,7 +1792,6 @@ static int nfs4_xdr_enc_pathconf(struct rpc_rqst *req, uint32_t *p, const struct
  */
 static int nfs4_xdr_enc_statfs(struct rpc_rqst *req, uint32_t *p, const struct nfs4_statfs_arg *args)
 {
-	extern u32 nfs4_statfs_bitmap[];
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
 		.nops = 2,
@@ -3003,6 +3076,11 @@ static int decode_readdir(struct xdr_stream *xdr, struct rpc_rqst *req, struct n
 		return status;
 	READ_BUF(8);
 	COPYMEM(readdir->verifier.data, 8);
+	dprintk("%s: verifier = 0x%x%x\n",
+			__FUNCTION__,
+			((u32 *)readdir->verifier.data)[0],
+			((u32 *)readdir->verifier.data)[1]);
+
 
 	hdrlen = (char *) p - (char *) iov->iov_base;
 	recvd = rcvbuf->len - hdrlen;
@@ -3017,12 +3095,14 @@ static int decode_readdir(struct xdr_stream *xdr, struct rpc_rqst *req, struct n
 	for (nr = 0; *p++; nr++) {
 		if (p + 3 > end)
 			goto short_pkt;
+		dprintk("cookie = %Lu, ", *((unsigned long long *)p));
 		p += 2;			/* cookie */
 		len = ntohl(*p++);	/* filename length */
 		if (len > NFS4_MAXNAMLEN) {
 			printk(KERN_WARNING "NFS: giant filename in readdir (len 0x%x)\n", len);
 			goto err_unmap;
 		}
+		dprintk("filename = %*s\n", len, (char *)p);
 		p += XDR_QUADLEN(len);
 		if (p + 1 > end)
 			goto short_pkt;
@@ -3042,6 +3122,7 @@ static int decode_readdir(struct xdr_stream *xdr, struct rpc_rqst *req, struct n
 	kunmap_atomic(kaddr, KM_USER0);
 	return 0;
 short_pkt:
+	dprintk("%s: short packet at entry %d\n", __FUNCTION__, nr);
 	entry[0] = entry[1] = 0;
 	/* truncate listing ? */
 	if (!nr) {
@@ -3127,6 +3208,47 @@ static int decode_renew(struct xdr_stream *xdr)
 	return decode_op_hdr(xdr, OP_RENEW);
 }
 
+static int decode_getacl(struct xdr_stream *xdr, struct rpc_rqst *req,
+		size_t *acl_len)
+{
+	uint32_t *savep;
+	uint32_t attrlen,
+		 bitmap[2] = {0};
+	struct kvec *iov = req->rq_rcv_buf.head;
+	int status;
+
+	*acl_len = 0;
+	if ((status = decode_op_hdr(xdr, OP_GETATTR)) != 0)
+		goto out;
+	if ((status = decode_attr_bitmap(xdr, bitmap)) != 0)
+		goto out;
+	if ((status = decode_attr_length(xdr, &attrlen, &savep)) != 0)
+		goto out;
+
+	if (unlikely(bitmap[0] & (FATTR4_WORD0_ACL - 1U)))
+		return -EIO;
+	if (likely(bitmap[0] & FATTR4_WORD0_ACL)) {
+		int hdrlen, recvd;
+
+		/* We ignore &savep and don't do consistency checks on
+		 * the attr length.  Let userspace figure it out.... */
+		hdrlen = (u8 *)xdr->p - (u8 *)iov->iov_base;
+		recvd = req->rq_rcv_buf.len - hdrlen;
+		if (attrlen > recvd) {
+			printk(KERN_WARNING "NFS: server cheating in getattr"
+					" acl reply: attrlen %u > recvd %u\n",
+					attrlen, recvd);
+			return -EINVAL;
+		}
+		if (attrlen <= *acl_len)
+			xdr_read_pages(xdr, attrlen);
+		*acl_len = attrlen;
+	}
+
+out:
+	return status;
+}
+
 static int
 decode_savefh(struct xdr_stream *xdr)
 {
@@ -3418,6 +3540,71 @@ static int nfs4_xdr_dec_getattr(struct rpc_rqst *rqstp, uint32_t *p, struct nfs4
 
 }
 
+/*
+ * Encode an SETACL request
+ */
+static int
+nfs4_xdr_enc_setacl(struct rpc_rqst *req, uint32_t *p, struct nfs_setaclargs *args)
+{
+        struct xdr_stream xdr;
+        struct compound_hdr hdr = {
+                .nops   = 2,
+        };
+        int status;
+
+        xdr_init_encode(&xdr, &req->rq_snd_buf, p);
+        encode_compound_hdr(&xdr, &hdr);
+        status = encode_putfh(&xdr, args->fh);
+        if (status)
+                goto out;
+        status = encode_setacl(&xdr, args);
+out:
+        return status;
+}
+/*
+ * Decode SETACL response
+ */
+static int
+nfs4_xdr_dec_setacl(struct rpc_rqst *rqstp, uint32_t *p, void *res)
+{
+	struct xdr_stream xdr;
+	struct compound_hdr hdr;
+	int status;
+
+	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
+	status = decode_compound_hdr(&xdr, &hdr);
+	if (status)
+		goto out;
+	status = decode_putfh(&xdr);
+	if (status)
+		goto out;
+	status = decode_setattr(&xdr, res);
+out:
+	return status;
+}
+
+/*
+ * Decode GETACL response
+ */
+static int
+nfs4_xdr_dec_getacl(struct rpc_rqst *rqstp, uint32_t *p, size_t *acl_len)
+{
+	struct xdr_stream xdr;
+	struct compound_hdr hdr;
+	int status;
+
+	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
+	status = decode_compound_hdr(&xdr, &hdr);
+	if (status)
+		goto out;
+	status = decode_putfh(&xdr);
+	if (status)
+		goto out;
+	status = decode_getacl(&xdr, rqstp, acl_len);
+
+out:
+	return status;
+}
 
 /*
  * Decode CLOSE response
@@ -3895,6 +4082,12 @@ uint32_t *nfs4_decode_dirent(uint32_t *p, struct nfs_entry *entry, int plus)
 	}
 	len = XDR_QUADLEN(ntohl(*p++));	/* attribute buffer length */
 	if (len > 0) {
+		if (bitmap[0] & FATTR4_WORD0_RDATTR_ERROR) {
+			bitmap[0] &= ~FATTR4_WORD0_RDATTR_ERROR;
+			/* Ignore the return value of rdattr_error for now */
+			p++;
+			len--;
+		}
 		if (bitmap[0] == 0 && bitmap[1] == FATTR4_WORD1_MOUNTED_ON_FILEID)
 			xdr_decode_hyper(p, &entry->ino);
 		else if (bitmap[0] == FATTR4_WORD0_FILEID)
@@ -3934,6 +4127,8 @@ static struct {
 	{ NFS4ERR_DQUOT,	EDQUOT		},
 	{ NFS4ERR_STALE,	ESTALE		},
 	{ NFS4ERR_BADHANDLE,	EBADHANDLE	},
+	{ NFS4ERR_BADOWNER,	EINVAL		},
+	{ NFS4ERR_BADNAME,	EINVAL		},
 	{ NFS4ERR_BAD_COOKIE,	EBADCOOKIE	},
 	{ NFS4ERR_NOTSUPP,	ENOTSUPP	},
 	{ NFS4ERR_TOOSMALL,	ETOOSMALL	},
@@ -4019,6 +4214,8 @@ struct rpc_procinfo	nfs4_procedures[] = {
   PROC(READDIR,		enc_readdir,	dec_readdir),
   PROC(SERVER_CAPS,	enc_server_caps, dec_server_caps),
   PROC(DELEGRETURN,	enc_delegreturn, dec_delegreturn),
+  PROC(GETACL,		enc_getacl,	dec_getacl),
+  PROC(SETACL,		enc_setacl,	dec_setacl),
 };
 
 struct rpc_version		nfs_version4 = {

fs/nfs/nfsroot.c

@@ -124,6 +124,7 @@ enum {
 	Opt_soft, Opt_hard, Opt_intr,
 	Opt_nointr, Opt_posix, Opt_noposix, Opt_cto, Opt_nocto, Opt_ac, 
 	Opt_noac, Opt_lock, Opt_nolock, Opt_v2, Opt_v3, Opt_udp, Opt_tcp,
+	Opt_acl, Opt_noacl,
 	/* Error token */
 	Opt_err
 };
@@ -158,6 +159,8 @@ static match_table_t __initdata tokens = {
 	{Opt_udp, "udp"},
 	{Opt_tcp, "proto=tcp"},
 	{Opt_tcp, "tcp"},
+	{Opt_acl, "acl"},
+	{Opt_noacl, "noacl"},
 	{Opt_err, NULL}
 	
 };
@@ -266,6 +269,12 @@ static int __init root_nfs_parse(char *name, char *buf)
 			case Opt_tcp:
 				nfs_data.flags |= NFS_MOUNT_TCP;
 				break;
+			case Opt_acl:
+				nfs_data.flags &= ~NFS_MOUNT_NOACL;
+				break;
+			case Opt_noacl:
+				nfs_data.flags |= NFS_MOUNT_NOACL;
+				break;
 			default : 
 				return 0;
 		}

fs/nfs/pagelist.c

@@ -107,10 +107,37 @@ void nfs_unlock_request(struct nfs_page *req)
 	smp_mb__before_clear_bit();
 	clear_bit(PG_BUSY, &req->wb_flags);
 	smp_mb__after_clear_bit();
-	wake_up_all(&req->wb_context->waitq);
+	wake_up_bit(&req->wb_flags, PG_BUSY);
 	nfs_release_request(req);
 }
 
+/**
+ * nfs_set_page_writeback_locked - Lock a request for writeback
+ * @req:
+ */
+int nfs_set_page_writeback_locked(struct nfs_page *req)
+{
+	struct nfs_inode *nfsi = NFS_I(req->wb_context->dentry->d_inode);
+
+	if (!nfs_lock_request(req))
+		return 0;
+	radix_tree_tag_set(&nfsi->nfs_page_tree, req->wb_index, NFS_PAGE_TAG_WRITEBACK);
+	return 1;
+}
+
+/**
+ * nfs_clear_page_writeback - Unlock request and wake up sleepers
+ */
+void nfs_clear_page_writeback(struct nfs_page *req)
+{
+	struct nfs_inode *nfsi = NFS_I(req->wb_context->dentry->d_inode);
+
+	spin_lock(&nfsi->req_lock);
+	radix_tree_tag_clear(&nfsi->nfs_page_tree, req->wb_index, NFS_PAGE_TAG_WRITEBACK);
+	spin_unlock(&nfsi->req_lock);
+	nfs_unlock_request(req);
+}
+
 /**
  * nfs_clear_request - Free up all resources allocated to the request
  * @req:
@@ -150,34 +177,15 @@ nfs_release_request(struct nfs_page *req)
 	nfs_page_free(req);
 }
 
-/**
+static int nfs_wait_bit_interruptible(void *word)
- * nfs_list_add_request - Insert a request into a sorted list
- * @req: request
- * @head: head of list into which to insert the request.
- *
- * Note that the wb_list is sorted by page index in order to facilitate
- * coalescing of requests.
- * We use an insertion sort that is optimized for the case of appended
- * writes.
- */
-void
-nfs_list_add_request(struct nfs_page *req, struct list_head *head)
 {
-	struct list_head *pos;
+	int ret = 0;
 
-#ifdef NFS_PARANOIA
+	if (signal_pending(current))
-	if (!list_empty(&req->wb_list)) {
+		ret = -ERESTARTSYS;
-		printk(KERN_ERR "NFS: Add to list failed!\n");
+	else
-		BUG();
+		schedule();
-	}
+	return ret;
-#endif
-	list_for_each_prev(pos, head) {
-		struct nfs_page	*p = nfs_list_entry(pos);
-		if (p->wb_index < req->wb_index)
-			break;
-	}
-	list_add(&req->wb_list, pos);
-	req->wb_list_head = head;
 }
 
 /**
@@ -190,12 +198,22 @@ nfs_list_add_request(struct nfs_page *req, struct list_head *head)
 int
 nfs_wait_on_request(struct nfs_page *req)
 {
-	struct inode	*inode = req->wb_context->dentry->d_inode;
+        struct rpc_clnt	*clnt = NFS_CLIENT(req->wb_context->dentry->d_inode);
-        struct rpc_clnt	*clnt = NFS_CLIENT(inode);
+	sigset_t oldmask;
+	int ret = 0;
 
-	if (!NFS_WBACK_BUSY(req))
+	if (!test_bit(PG_BUSY, &req->wb_flags))
-		return 0;
+		goto out;
-	return nfs_wait_event(clnt, req->wb_context->waitq, !NFS_WBACK_BUSY(req));
+	/*
+	 * Note: the call to rpc_clnt_sigmask() suffices to ensure that we
+	 *	 are not interrupted if intr flag is not set
+	 */
+	rpc_clnt_sigmask(clnt, &oldmask);
+	ret = out_of_line_wait_on_bit(&req->wb_flags, PG_BUSY,
+			nfs_wait_bit_interruptible, TASK_INTERRUPTIBLE);
+	rpc_clnt_sigunmask(clnt, &oldmask);
+out:
+	return ret;
 }
 
 /**
@@ -243,6 +261,62 @@ nfs_coalesce_requests(struct list_head *head, struct list_head *dst,
 	return npages;
 }
 
+#define NFS_SCAN_MAXENTRIES 16
+/**
+ * nfs_scan_lock_dirty - Scan the radix tree for dirty requests
+ * @nfsi: NFS inode
+ * @dst: Destination list
+ * @idx_start: lower bound of page->index to scan
+ * @npages: idx_start + npages sets the upper bound to scan.
+ *
+ * Moves elements from one of the inode request lists.
+ * If the number of requests is set to 0, the entire address_space
+ * starting at index idx_start, is scanned.
+ * The requests are *not* checked to ensure that they form a contiguous set.
+ * You must be holding the inode's req_lock when calling this function
+ */
+int
+nfs_scan_lock_dirty(struct nfs_inode *nfsi, struct list_head *dst,
+	      unsigned long idx_start, unsigned int npages)
+{
+	struct nfs_page *pgvec[NFS_SCAN_MAXENTRIES];
+	struct nfs_page *req;
+	unsigned long idx_end;
+	int found, i;
+	int res;
+
+	res = 0;
+	if (npages == 0)
+		idx_end = ~0;
+	else
+		idx_end = idx_start + npages - 1;
+
+	for (;;) {
+		found = radix_tree_gang_lookup_tag(&nfsi->nfs_page_tree,
+				(void **)&pgvec[0], idx_start, NFS_SCAN_MAXENTRIES,
+				NFS_PAGE_TAG_DIRTY);
+		if (found <= 0)
+			break;
+		for (i = 0; i < found; i++) {
+			req = pgvec[i];
+			if (req->wb_index > idx_end)
+				goto out;
+
+			idx_start = req->wb_index + 1;
+
+			if (nfs_set_page_writeback_locked(req)) {
+				radix_tree_tag_clear(&nfsi->nfs_page_tree,
+						req->wb_index, NFS_PAGE_TAG_DIRTY);
+				nfs_list_remove_request(req);
+				nfs_list_add_request(req, dst);
+				res++;
+			}
+		}
+	}
+out:
+	return res;
+}
+
 /**
  * nfs_scan_list - Scan a list for matching requests
  * @head: One of the NFS inode request lists
@@ -280,7 +354,7 @@ nfs_scan_list(struct list_head *head, struct list_head *dst,
 		if (req->wb_index > idx_end)
 			break;
 
-		if (!nfs_lock_request(req))
+		if (!nfs_set_page_writeback_locked(req))
 			continue;
 		nfs_list_remove_request(req);
 		nfs_list_add_request(req, dst);

fs/nfs/proc.c

@@ -622,6 +622,7 @@ struct nfs_rpc_ops	nfs_v2_clientops = {
 	.version	= 2,		       /* protocol version */
 	.dentry_ops	= &nfs_dentry_operations,
 	.dir_inode_ops	= &nfs_dir_inode_operations,
+	.file_inode_ops	= &nfs_file_inode_operations,
 	.getroot	= nfs_proc_get_root,
 	.getattr	= nfs_proc_getattr,
 	.setattr	= nfs_proc_setattr,

fs/nfs/read.c

@@ -173,7 +173,6 @@ static int nfs_readpage_async(struct nfs_open_context *ctx, struct inode *inode,
 	if (len < PAGE_CACHE_SIZE)
 		memclear_highpage_flush(page, len, PAGE_CACHE_SIZE - len);
 
-	nfs_lock_request(new);
 	nfs_list_add_request(new, &one_request);
 	nfs_pagein_one(&one_request, inode);
 	return 0;
@@ -185,7 +184,6 @@ static void nfs_readpage_release(struct nfs_page *req)
 
 	nfs_clear_request(req);
 	nfs_release_request(req);
-	nfs_unlock_request(req);
 
 	dprintk("NFS: read done (%s/%Ld %d@%Ld)\n",
 			req->wb_context->dentry->d_inode->i_sb->s_id,
@@ -553,7 +551,6 @@ readpage_async_filler(void *data, struct page *page)
 	}
 	if (len < PAGE_CACHE_SIZE)
 		memclear_highpage_flush(page, len, PAGE_CACHE_SIZE - len);
-	nfs_lock_request(new);
 	nfs_list_add_request(new, desc->head);
 	return 0;
 }

fs/nfs/write.c

@@ -220,7 +220,7 @@ static int nfs_writepage_sync(struct nfs_open_context *ctx, struct inode *inode,
 		ClearPageError(page);
 
 io_error:
-	nfs_end_data_update_defer(inode);
+	nfs_end_data_update(inode);
 	nfs_writedata_free(wdata);
 	return written ? written : result;
 }
@@ -352,7 +352,7 @@ int nfs_writepages(struct address_space *mapping, struct writeback_control *wbc)
 		if (err < 0)
 			goto out;
 	}
-	err = nfs_commit_inode(inode, 0, 0, wb_priority(wbc));
+	err = nfs_commit_inode(inode, wb_priority(wbc));
 	if (err > 0) {
 		wbc->nr_to_write -= err;
 		err = 0;
@@ -401,7 +401,7 @@ static void nfs_inode_remove_request(struct nfs_page *req)
 	nfsi->npages--;
 	if (!nfsi->npages) {
 		spin_unlock(&nfsi->req_lock);
-		nfs_end_data_update_defer(inode);
+		nfs_end_data_update(inode);
 		iput(inode);
 	} else
 		spin_unlock(&nfsi->req_lock);
@@ -446,6 +446,8 @@ nfs_mark_request_dirty(struct nfs_page *req)
 	struct nfs_inode *nfsi = NFS_I(inode);
 
 	spin_lock(&nfsi->req_lock);
+	radix_tree_tag_set(&nfsi->nfs_page_tree,
+			req->wb_index, NFS_PAGE_TAG_DIRTY);
 	nfs_list_add_request(req, &nfsi->dirty);
 	nfsi->ndirty++;
 	spin_unlock(&nfsi->req_lock);
@@ -503,13 +505,12 @@ nfs_wait_on_requests(struct inode *inode, unsigned long idx_start, unsigned int
 
 	spin_lock(&nfsi->req_lock);
 	next = idx_start;
-	while (radix_tree_gang_lookup(&nfsi->nfs_page_tree, (void **)&req, next, 1)) {
+	while (radix_tree_gang_lookup_tag(&nfsi->nfs_page_tree, (void **)&req, next, 1, NFS_PAGE_TAG_WRITEBACK)) {
 		if (req->wb_index > idx_end)
 			break;
 
 		next = req->wb_index + 1;
-		if (!NFS_WBACK_BUSY(req))
+		BUG_ON(!NFS_WBACK_BUSY(req));
-			continue;
 
 		atomic_inc(&req->wb_count);
 		spin_unlock(&nfsi->req_lock);
@@ -538,12 +539,15 @@ static int
 nfs_scan_dirty(struct inode *inode, struct list_head *dst, unsigned long idx_start, unsigned int npages)
 {
 	struct nfs_inode *nfsi = NFS_I(inode);
-	int	res;
+	int res = 0;
-	res = nfs_scan_list(&nfsi->dirty, dst, idx_start, npages);
+
+	if (nfsi->ndirty != 0) {
+		res = nfs_scan_lock_dirty(nfsi, dst, idx_start, npages);
 		nfsi->ndirty -= res;
 		sub_page_state(nr_dirty,res);
 		if ((nfsi->ndirty == 0) != list_empty(&nfsi->dirty))
 			printk(KERN_ERR "NFS: desynchronized value of nfs_i.ndirty.\n");
+	}
 	return res;
 }
 
@@ -562,11 +566,14 @@ static int
 nfs_scan_commit(struct inode *inode, struct list_head *dst, unsigned long idx_start, unsigned int npages)
 {
 	struct nfs_inode *nfsi = NFS_I(inode);
-	int	res;
+	int res = 0;
+
+	if (nfsi->ncommit != 0) {
 		res = nfs_scan_list(&nfsi->commit, dst, idx_start, npages);
 		nfsi->ncommit -= res;
 		if ((nfsi->ncommit == 0) != list_empty(&nfsi->commit))
 			printk(KERN_ERR "NFS: desynchronized value of nfs_i.ncommit.\n");
+	}
 	return res;
 }
 #endif
@@ -750,7 +757,7 @@ int nfs_updatepage(struct file *file, struct page *page,
 	 * is entirely in cache, it may be more efficient to avoid
 	 * fragmenting write requests.
 	 */
-	if (PageUptodate(page) && inode->i_flock == NULL) {
+	if (PageUptodate(page) && inode->i_flock == NULL && !(file->f_mode & O_SYNC)) {
 		loff_t end_offs = i_size_read(inode) - 1;
 		unsigned long end_index = end_offs >> PAGE_CACHE_SHIFT;
 
@@ -821,7 +828,7 @@ static void nfs_writepage_release(struct nfs_page *req)
 #else
 	nfs_inode_remove_request(req);
 #endif
-	nfs_unlock_request(req);
+	nfs_clear_page_writeback(req);
 }
 
 static inline int flush_task_priority(int how)
@@ -952,7 +959,7 @@ static int nfs_flush_multi(struct list_head *head, struct inode *inode, int how)
 		nfs_writedata_free(data);
 	}
 	nfs_mark_request_dirty(req);
-	nfs_unlock_request(req);
+	nfs_clear_page_writeback(req);
 	return -ENOMEM;
 }
 
@@ -1002,7 +1009,7 @@ static int nfs_flush_one(struct list_head *head, struct inode *inode, int how)
 		struct nfs_page *req = nfs_list_entry(head->next);
 		nfs_list_remove_request(req);
 		nfs_mark_request_dirty(req);
-		nfs_unlock_request(req);
+		nfs_clear_page_writeback(req);
 	}
 	return -ENOMEM;
 }
@@ -1029,7 +1036,7 @@ nfs_flush_list(struct list_head *head, int wpages, int how)
 		req = nfs_list_entry(head->next);
 		nfs_list_remove_request(req);
 		nfs_mark_request_dirty(req);
-		nfs_unlock_request(req);
+		nfs_clear_page_writeback(req);
 	}
 	return error;
 }
@@ -1121,7 +1128,7 @@ static void nfs_writeback_done_full(struct nfs_write_data *data, int status)
 		nfs_inode_remove_request(req);
 #endif
 	next:
-		nfs_unlock_request(req);
+		nfs_clear_page_writeback(req);
 	}
 }
 
@@ -1210,36 +1217,24 @@ static void nfs_commit_rpcsetup(struct list_head *head,
 		struct nfs_write_data *data, int how)
 {
 	struct rpc_task		*task = &data->task;
-	struct nfs_page		*first, *last;
+	struct nfs_page		*first;
 	struct inode		*inode;
-	loff_t			start, end, len;
 
 	/* Set up the RPC argument and reply structs
 	 * NB: take care not to mess about with data->commit et al. */
 
 	list_splice_init(head, &data->pages);
 	first = nfs_list_entry(data->pages.next);
-	last = nfs_list_entry(data->pages.prev);
 	inode = first->wb_context->dentry->d_inode;
 
-	/*
-	 * Determine the offset range of requests in the COMMIT call.
-	 * We rely on the fact that data->pages is an ordered list...
-	 */
-	start = req_offset(first);
-	end = req_offset(last) + last->wb_bytes;
-	len = end - start;
-	/* If 'len' is not a 32-bit quantity, pass '0' in the COMMIT call */
-	if (end >= i_size_read(inode) || len < 0 || len > (~((u32)0) >> 1))
-		len = 0;
-
 	data->inode	  = inode;
 	data->cred	  = first->wb_context->cred;
 
 	data->args.fh     = NFS_FH(data->inode);
-	data->args.offset = start;
+	/* Note: we always request a commit of the entire inode */
-	data->args.count  = len;
+	data->args.offset = 0;
-	data->res.count   = len;
+	data->args.count  = 0;
+	data->res.count   = 0;
 	data->res.fattr   = &data->fattr;
 	data->res.verf    = &data->verf;
 	
@@ -1278,7 +1273,7 @@ nfs_commit_list(struct list_head *head, int how)
 		req = nfs_list_entry(head->next);
 		nfs_list_remove_request(req);
 		nfs_mark_request_commit(req);
-		nfs_unlock_request(req);
+		nfs_clear_page_writeback(req);
 	}
 	return -ENOMEM;
 }
@@ -1324,7 +1319,7 @@ nfs_commit_done(struct rpc_task *task)
 		dprintk(" mismatch\n");
 		nfs_mark_request_dirty(req);
 	next:
-		nfs_unlock_request(req);
+		nfs_clear_page_writeback(req);
 		res++;
 	}
 	sub_page_state(nr_unstable,res);
@@ -1342,16 +1337,23 @@ static int nfs_flush_inode(struct inode *inode, unsigned long idx_start,
 	spin_lock(&nfsi->req_lock);
 	res = nfs_scan_dirty(inode, &head, idx_start, npages);
 	spin_unlock(&nfsi->req_lock);
-	if (res)
+	if (res) {
-		error = nfs_flush_list(&head, NFS_SERVER(inode)->wpages, how);
+		struct nfs_server *server = NFS_SERVER(inode);
+
+		/* For single writes, FLUSH_STABLE is more efficient */
+		if (res == nfsi->npages && nfsi->npages <= server->wpages) {
+			if (res > 1 || nfs_list_entry(head.next)->wb_bytes <= server->wsize)
+				how |= FLUSH_STABLE;
+		}
+		error = nfs_flush_list(&head, server->wpages, how);
+	}
 	if (error < 0)
 		return error;
 	return res;
 }
 
 #if defined(CONFIG_NFS_V3) || defined(CONFIG_NFS_V4)
-int nfs_commit_inode(struct inode *inode, unsigned long idx_start,
+int nfs_commit_inode(struct inode *inode, int how)
-		    unsigned int npages, int how)
 {
 	struct nfs_inode *nfsi = NFS_I(inode);
 	LIST_HEAD(head);
@@ -1359,15 +1361,13 @@ int nfs_commit_inode(struct inode *inode, unsigned long idx_start,
 				error = 0;
 
 	spin_lock(&nfsi->req_lock);
-	res = nfs_scan_commit(inode, &head, idx_start, npages);
+	res = nfs_scan_commit(inode, &head, 0, 0);
+	spin_unlock(&nfsi->req_lock);
 	if (res) {
-		res += nfs_scan_commit(inode, &head, 0, 0);
-		spin_unlock(&nfsi->req_lock);
 		error = nfs_commit_list(&head, how);
-	} else
-		spin_unlock(&nfsi->req_lock);
 		if (error < 0)
 			return error;
+	}
 	return res;
 }
 #endif
@@ -1389,7 +1389,7 @@ int nfs_sync_inode(struct inode *inode, unsigned long idx_start,
 			error = nfs_flush_inode(inode, idx_start, npages, how);
 #if defined(CONFIG_NFS_V3) || defined(CONFIG_NFS_V4)
 		if (error == 0)
-			error = nfs_commit_inode(inode, idx_start, npages, how);
+			error = nfs_commit_inode(inode, how);
 #endif
 	} while (error > 0);
 	return error;

fs/nfs_common/Makefile

@@ -0,0 +1,7 @@
+#
+# Makefile for Linux filesystem routines that are shared by client and server.
+#
+
+obj-$(CONFIG_NFS_ACL_SUPPORT) += nfs_acl.o
+
+nfs_acl-objs := nfsacl.o

fs/nfs_common/nfsacl.c

@@ -0,0 +1,257 @@
+/*
+ * fs/nfs_common/nfsacl.c
+ *
+ *  Copyright (C) 2002-2003 Andreas Gruenbacher <agruen@suse.de>
+ */
+
+/*
+ * The Solaris nfsacl protocol represents some ACLs slightly differently
+ * than POSIX 1003.1e draft 17 does (and we do):
+ *
+ *  - Minimal ACLs always have an ACL_MASK entry, so they have
+ *    four instead of three entries.
+ *  - The ACL_MASK entry in such minimal ACLs always has the same
+ *    permissions as the ACL_GROUP_OBJ entry. (In extended ACLs
+ *    the ACL_MASK and ACL_GROUP_OBJ entries may differ.)
+ *  - The identifier fields of the ACL_USER_OBJ and ACL_GROUP_OBJ
+ *    entries contain the identifiers of the owner and owning group.
+ *    (In POSIX ACLs we always set them to ACL_UNDEFINED_ID).
+ *  - ACL entries in the kernel are kept sorted in ascending order
+ *    of (e_tag, e_id). Solaris ACLs are unsorted.
+ */
+
+#include <linux/module.h>
+#include <linux/fs.h>
+#include <linux/sunrpc/xdr.h>
+#include <linux/nfsacl.h>
+#include <linux/nfs3.h>
+#include <linux/sort.h>
+
+MODULE_LICENSE("GPL");
+
+EXPORT_SYMBOL(nfsacl_encode);
+EXPORT_SYMBOL(nfsacl_decode);
+
+struct nfsacl_encode_desc {
+	struct xdr_array2_desc desc;
+	unsigned int count;
+	struct posix_acl *acl;
+	int typeflag;
+	uid_t uid;
+	gid_t gid;
+};
+
+static int
+xdr_nfsace_encode(struct xdr_array2_desc *desc, void *elem)
+{
+	struct nfsacl_encode_desc *nfsacl_desc =
+		(struct nfsacl_encode_desc *) desc;
+	u32 *p = (u32 *) elem;
+
+	if (nfsacl_desc->count < nfsacl_desc->acl->a_count) {
+		struct posix_acl_entry *entry =
+			&nfsacl_desc->acl->a_entries[nfsacl_desc->count++];
+
+		*p++ = htonl(entry->e_tag | nfsacl_desc->typeflag);
+		switch(entry->e_tag) {
+			case ACL_USER_OBJ:
+				*p++ = htonl(nfsacl_desc->uid);
+				break;
+			case ACL_GROUP_OBJ:
+				*p++ = htonl(nfsacl_desc->gid);
+				break;
+			case ACL_USER:
+			case ACL_GROUP:
+				*p++ = htonl(entry->e_id);
+				break;
+			default:  /* Solaris depends on that! */
+				*p++ = 0;
+				break;
+		}
+		*p++ = htonl(entry->e_perm & S_IRWXO);
+	} else {
+		const struct posix_acl_entry *pa, *pe;
+		int group_obj_perm = ACL_READ|ACL_WRITE|ACL_EXECUTE;
+
+		FOREACH_ACL_ENTRY(pa, nfsacl_desc->acl, pe) {
+			if (pa->e_tag == ACL_GROUP_OBJ) {
+				group_obj_perm = pa->e_perm & S_IRWXO;
+				break;
+			}
+		}
+		/* fake up ACL_MASK entry */
+		*p++ = htonl(ACL_MASK | nfsacl_desc->typeflag);
+		*p++ = htonl(0);
+		*p++ = htonl(group_obj_perm);
+	}
+
+	return 0;
+}
+
+unsigned int
+nfsacl_encode(struct xdr_buf *buf, unsigned int base, struct inode *inode,
+	      struct posix_acl *acl, int encode_entries, int typeflag)
+{
+	int entries = (acl && acl->a_count) ? max_t(int, acl->a_count, 4) : 0;
+	struct nfsacl_encode_desc nfsacl_desc = {
+		.desc = {
+			.elem_size = 12,
+			.array_len = encode_entries ? entries : 0,
+			.xcode = xdr_nfsace_encode,
+		},
+		.acl = acl,
+		.typeflag = typeflag,
+		.uid = inode->i_uid,
+		.gid = inode->i_gid,
+	};
+	int err;
+
+	if (entries > NFS_ACL_MAX_ENTRIES ||
+	    xdr_encode_word(buf, base, entries))
+		return -EINVAL;
+	err = xdr_encode_array2(buf, base + 4, &nfsacl_desc.desc);
+	if (!err)
+		err = 8 + nfsacl_desc.desc.elem_size *
+			  nfsacl_desc.desc.array_len;
+	return err;
+}
+
+struct nfsacl_decode_desc {
+	struct xdr_array2_desc desc;
+	unsigned int count;
+	struct posix_acl *acl;
+};
+
+static int
+xdr_nfsace_decode(struct xdr_array2_desc *desc, void *elem)
+{
+	struct nfsacl_decode_desc *nfsacl_desc =
+		(struct nfsacl_decode_desc *) desc;
+	u32 *p = (u32 *) elem;
+	struct posix_acl_entry *entry;
+
+	if (!nfsacl_desc->acl) {
+		if (desc->array_len > NFS_ACL_MAX_ENTRIES)
+			return -EINVAL;
+		nfsacl_desc->acl = posix_acl_alloc(desc->array_len, GFP_KERNEL);
+		if (!nfsacl_desc->acl)
+			return -ENOMEM;
+		nfsacl_desc->count = 0;
+	}
+
+	entry = &nfsacl_desc->acl->a_entries[nfsacl_desc->count++];
+	entry->e_tag = ntohl(*p++) & ~NFS_ACL_DEFAULT;
+	entry->e_id = ntohl(*p++);
+	entry->e_perm = ntohl(*p++);
+
+	switch(entry->e_tag) {
+		case ACL_USER_OBJ:
+		case ACL_USER:
+		case ACL_GROUP_OBJ:
+		case ACL_GROUP:
+		case ACL_OTHER:
+			if (entry->e_perm & ~S_IRWXO)
+				return -EINVAL;
+			break;
+		case ACL_MASK:
+			/* Solaris sometimes sets additonal bits in the mask */
+			entry->e_perm &= S_IRWXO;
+			break;
+		default:
+			return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int
+cmp_acl_entry(const void *x, const void *y)
+{
+	const struct posix_acl_entry *a = x, *b = y;
+
+	if (a->e_tag != b->e_tag)
+		return a->e_tag - b->e_tag;
+	else if (a->e_id > b->e_id)
+		return 1;
+	else if (a->e_id < b->e_id)
+		return -1;
+	else
+		return 0;
+}
+
+/*
+ * Convert from a Solaris ACL to a POSIX 1003.1e draft 17 ACL.
+ */
+static int
+posix_acl_from_nfsacl(struct posix_acl *acl)
+{
+	struct posix_acl_entry *pa, *pe,
+	       *group_obj = NULL, *mask = NULL;
+
+	if (!acl)
+		return 0;
+
+	sort(acl->a_entries, acl->a_count, sizeof(struct posix_acl_entry),
+	     cmp_acl_entry, NULL);
+
+	/* Clear undefined identifier fields and find the ACL_GROUP_OBJ
+	   and ACL_MASK entries. */
+	FOREACH_ACL_ENTRY(pa, acl, pe) {
+		switch(pa->e_tag) {
+			case ACL_USER_OBJ:
+				pa->e_id = ACL_UNDEFINED_ID;
+				break;
+			case ACL_GROUP_OBJ:
+				pa->e_id = ACL_UNDEFINED_ID;
+				group_obj = pa;
+				break;
+			case ACL_MASK:
+				mask = pa;
+				/* fall through */
+			case ACL_OTHER:
+				pa->e_id = ACL_UNDEFINED_ID;
+				break;
+		}
+	}
+	if (acl->a_count == 4 && group_obj && mask &&
+	    mask->e_perm == group_obj->e_perm) {
+		/* remove bogus ACL_MASK entry */
+		memmove(mask, mask+1, (3 - (mask - acl->a_entries)) *
+				      sizeof(struct posix_acl_entry));
+		acl->a_count = 3;
+	}
+	return 0;
+}
+
+unsigned int
+nfsacl_decode(struct xdr_buf *buf, unsigned int base, unsigned int *aclcnt,
+	      struct posix_acl **pacl)
+{
+	struct nfsacl_decode_desc nfsacl_desc = {
+		.desc = {
+			.elem_size = 12,
+			.xcode = pacl ? xdr_nfsace_decode : NULL,
+		},
+	};
+	u32 entries;
+	int err;
+
+	if (xdr_decode_word(buf, base, &entries) ||
+	    entries > NFS_ACL_MAX_ENTRIES)
+		return -EINVAL;
+	err = xdr_decode_array2(buf, base + 4, &nfsacl_desc.desc);
+	if (err)
+		return err;
+	if (pacl) {
+		if (entries != nfsacl_desc.desc.array_len ||
+		    posix_acl_from_nfsacl(nfsacl_desc.acl) != 0) {
+			posix_acl_release(nfsacl_desc.acl);
+			return -EINVAL;
+		}
+		*pacl = nfsacl_desc.acl;
+	}
+	if (aclcnt)
+		*aclcnt = entries;
+	return 8 + nfsacl_desc.desc.elem_size *
+		   nfsacl_desc.desc.array_len;
+}

fs/nfsd/Makefile

@@ -6,7 +6,9 @@ obj-$(CONFIG_NFSD)	+= nfsd.o
 
 nfsd-y 			:= nfssvc.o nfsctl.o nfsproc.o nfsfh.o vfs.o \
 			   export.o auth.o lockd.o nfscache.o nfsxdr.o stats.o
+nfsd-$(CONFIG_NFSD_V2_ACL) += nfs2acl.o
 nfsd-$(CONFIG_NFSD_V3)	+= nfs3proc.o nfs3xdr.o
+nfsd-$(CONFIG_NFSD_V3_ACL) += nfs3acl.o
 nfsd-$(CONFIG_NFSD_V4)	+= nfs4proc.o nfs4xdr.o nfs4state.o nfs4idmap.o \
 			   nfs4acl.o nfs4callback.o
 nfsd-objs		:= $(nfsd-y)

fs/nfsd/nfs2acl.c

@@ -0,0 +1,336 @@
+/*
+ * linux/fs/nfsd/nfsacl.c
+ *
+ * Process version 2 NFSACL requests.
+ *
+ * Copyright (C) 2002-2003 Andreas Gruenbacher <agruen@suse.de>
+ */
+
+#include <linux/sunrpc/svc.h>
+#include <linux/nfs.h>
+#include <linux/nfsd/nfsd.h>
+#include <linux/nfsd/cache.h>
+#include <linux/nfsd/xdr.h>
+#include <linux/nfsd/xdr3.h>
+#include <linux/posix_acl.h>
+#include <linux/nfsacl.h>
+
+#define NFSDDBG_FACILITY		NFSDDBG_PROC
+#define RETURN_STATUS(st)	{ resp->status = (st); return (st); }
+
+/*
+ * NULL call.
+ */
+static int
+nfsacld_proc_null(struct svc_rqst *rqstp, void *argp, void *resp)
+{
+	return nfs_ok;
+}
+
+/*
+ * Get the Access and/or Default ACL of a file.
+ */
+static int nfsacld_proc_getacl(struct svc_rqst * rqstp,
+		struct nfsd3_getaclargs *argp, struct nfsd3_getaclres *resp)
+{
+	svc_fh *fh;
+	struct posix_acl *acl;
+	int nfserr = 0;
+
+	dprintk("nfsd: GETACL(2acl)   %s\n", SVCFH_fmt(&argp->fh));
+
+	fh = fh_copy(&resp->fh, &argp->fh);
+	if ((nfserr = fh_verify(rqstp, &resp->fh, 0, MAY_NOP)))
+		RETURN_STATUS(nfserr_inval);
+
+	if (argp->mask & ~(NFS_ACL|NFS_ACLCNT|NFS_DFACL|NFS_DFACLCNT))
+		RETURN_STATUS(nfserr_inval);
+	resp->mask = argp->mask;
+
+	if (resp->mask & (NFS_ACL|NFS_ACLCNT)) {
+		acl = nfsd_get_posix_acl(fh, ACL_TYPE_ACCESS);
+		if (IS_ERR(acl)) {
+			int err = PTR_ERR(acl);
+
+			if (err == -ENODATA || err == -EOPNOTSUPP)
+				acl = NULL;
+			else {
+				nfserr = nfserrno(err);
+				goto fail;
+			}
+		}
+		if (acl == NULL) {
+			/* Solaris returns the inode's minimum ACL. */
+
+			struct inode *inode = fh->fh_dentry->d_inode;
+			acl = posix_acl_from_mode(inode->i_mode, GFP_KERNEL);
+		}
+		resp->acl_access = acl;
+	}
+	if (resp->mask & (NFS_DFACL|NFS_DFACLCNT)) {
+		/* Check how Solaris handles requests for the Default ACL
+		   of a non-directory! */
+
+		acl = nfsd_get_posix_acl(fh, ACL_TYPE_DEFAULT);
+		if (IS_ERR(acl)) {
+			int err = PTR_ERR(acl);
+
+			if (err == -ENODATA || err == -EOPNOTSUPP)
+				acl = NULL;
+			else {
+				nfserr = nfserrno(err);
+				goto fail;
+			}
+		}
+		resp->acl_default = acl;
+	}
+
+	/* resp->acl_{access,default} are released in nfssvc_release_getacl. */
+	RETURN_STATUS(0);
+
+fail:
+	posix_acl_release(resp->acl_access);
+	posix_acl_release(resp->acl_default);
+	RETURN_STATUS(nfserr);
+}
+
+/*
+ * Set the Access and/or Default ACL of a file.
+ */
+static int nfsacld_proc_setacl(struct svc_rqst * rqstp,
+		struct nfsd3_setaclargs *argp,
+		struct nfsd_attrstat *resp)
+{
+	svc_fh *fh;
+	int nfserr = 0;
+
+	dprintk("nfsd: SETACL(2acl)   %s\n", SVCFH_fmt(&argp->fh));
+
+	fh = fh_copy(&resp->fh, &argp->fh);
+	nfserr = fh_verify(rqstp, &resp->fh, 0, MAY_NOP);
+
+	if (!nfserr) {
+		nfserr = nfserrno( nfsd_set_posix_acl(
+			fh, ACL_TYPE_ACCESS, argp->acl_access) );
+	}
+	if (!nfserr) {
+		nfserr = nfserrno( nfsd_set_posix_acl(
+			fh, ACL_TYPE_DEFAULT, argp->acl_default) );
+	}
+
+	/* argp->acl_{access,default} may have been allocated in
+	   nfssvc_decode_setaclargs. */
+	posix_acl_release(argp->acl_access);
+	posix_acl_release(argp->acl_default);
+	return nfserr;
+}
+
+/*
+ * Check file attributes
+ */
+static int nfsacld_proc_getattr(struct svc_rqst * rqstp,
+		struct nfsd_fhandle *argp, struct nfsd_attrstat *resp)
+{
+	dprintk("nfsd: GETATTR  %s\n", SVCFH_fmt(&argp->fh));
+
+	fh_copy(&resp->fh, &argp->fh);
+	return fh_verify(rqstp, &resp->fh, 0, MAY_NOP);
+}
+
+/*
+ * Check file access
+ */
+static int nfsacld_proc_access(struct svc_rqst *rqstp, struct nfsd3_accessargs *argp,
+		struct nfsd3_accessres *resp)
+{
+	int nfserr;
+
+	dprintk("nfsd: ACCESS(2acl)   %s 0x%x\n",
+			SVCFH_fmt(&argp->fh),
+			argp->access);
+
+	fh_copy(&resp->fh, &argp->fh);
+	resp->access = argp->access;
+	nfserr = nfsd_access(rqstp, &resp->fh, &resp->access, NULL);
+	return nfserr;
+}
+
+/*
+ * XDR decode functions
+ */
+static int nfsaclsvc_decode_getaclargs(struct svc_rqst *rqstp, u32 *p,
+		struct nfsd3_getaclargs *argp)
+{
+	if (!(p = nfs2svc_decode_fh(p, &argp->fh)))
+		return 0;
+	argp->mask = ntohl(*p); p++;
+
+	return xdr_argsize_check(rqstp, p);
+}
+
+
+static int nfsaclsvc_decode_setaclargs(struct svc_rqst *rqstp, u32 *p,
+		struct nfsd3_setaclargs *argp)
+{
+	struct kvec *head = rqstp->rq_arg.head;
+	unsigned int base;
+	int n;
+
+	if (!(p = nfs2svc_decode_fh(p, &argp->fh)))
+		return 0;
+	argp->mask = ntohl(*p++);
+	if (argp->mask & ~(NFS_ACL|NFS_ACLCNT|NFS_DFACL|NFS_DFACLCNT) ||
+	    !xdr_argsize_check(rqstp, p))
+		return 0;
+
+	base = (char *)p - (char *)head->iov_base;
+	n = nfsacl_decode(&rqstp->rq_arg, base, NULL,
+			  (argp->mask & NFS_ACL) ?
+			  &argp->acl_access : NULL);
+	if (n > 0)
+		n = nfsacl_decode(&rqstp->rq_arg, base + n, NULL,
+				  (argp->mask & NFS_DFACL) ?
+				  &argp->acl_default : NULL);
+	return (n > 0);
+}
+
+static int nfsaclsvc_decode_fhandleargs(struct svc_rqst *rqstp, u32 *p,
+		struct nfsd_fhandle *argp)
+{
+	if (!(p = nfs2svc_decode_fh(p, &argp->fh)))
+		return 0;
+	return xdr_argsize_check(rqstp, p);
+}
+
+static int nfsaclsvc_decode_accessargs(struct svc_rqst *rqstp, u32 *p,
+		struct nfsd3_accessargs *argp)
+{
+	if (!(p = nfs2svc_decode_fh(p, &argp->fh)))
+		return 0;
+	argp->access = ntohl(*p++);
+
+	return xdr_argsize_check(rqstp, p);
+}
+
+/*
+ * XDR encode functions
+ */
+
+/* GETACL */
+static int nfsaclsvc_encode_getaclres(struct svc_rqst *rqstp, u32 *p,
+		struct nfsd3_getaclres *resp)
+{
+	struct dentry *dentry = resp->fh.fh_dentry;
+	struct inode *inode = dentry->d_inode;
+	int w = nfsacl_size(
+		(resp->mask & NFS_ACL)   ? resp->acl_access  : NULL,
+		(resp->mask & NFS_DFACL) ? resp->acl_default : NULL);
+	struct kvec *head = rqstp->rq_res.head;
+	unsigned int base;
+	int n;
+
+	if (dentry == NULL || dentry->d_inode == NULL)
+		return 0;
+	inode = dentry->d_inode;
+
+	p = nfs2svc_encode_fattr(rqstp, p, &resp->fh);
+	*p++ = htonl(resp->mask);
+	if (!xdr_ressize_check(rqstp, p))
+		return 0;
+	base = (char *)p - (char *)head->iov_base;
+
+	rqstp->rq_res.page_len = w;
+	while (w > 0) {
+		if (!svc_take_res_page(rqstp))
+			return 0;
+		w -= PAGE_SIZE;
+	}
+
+	n = nfsacl_encode(&rqstp->rq_res, base, inode,
+			  resp->acl_access,
+			  resp->mask & NFS_ACL, 0);
+	if (n > 0)
+		n = nfsacl_encode(&rqstp->rq_res, base + n, inode,
+				  resp->acl_default,
+				  resp->mask & NFS_DFACL,
+				  NFS_ACL_DEFAULT);
+	if (n <= 0)
+		return 0;
+	return 1;
+}
+
+static int nfsaclsvc_encode_attrstatres(struct svc_rqst *rqstp, u32 *p,
+		struct nfsd_attrstat *resp)
+{
+	p = nfs2svc_encode_fattr(rqstp, p, &resp->fh);
+	return xdr_ressize_check(rqstp, p);
+}
+
+/* ACCESS */
+static int nfsaclsvc_encode_accessres(struct svc_rqst *rqstp, u32 *p,
+		struct nfsd3_accessres *resp)
+{
+	p = nfs2svc_encode_fattr(rqstp, p, &resp->fh);
+	*p++ = htonl(resp->access);
+	return xdr_ressize_check(rqstp, p);
+}
+
+/*
+ * XDR release functions
+ */
+static int nfsaclsvc_release_getacl(struct svc_rqst *rqstp, u32 *p,
+		struct nfsd3_getaclres *resp)
+{
+	fh_put(&resp->fh);
+	posix_acl_release(resp->acl_access);
+	posix_acl_release(resp->acl_default);
+	return 1;
+}
+
+static int nfsaclsvc_release_fhandle(struct svc_rqst *rqstp, u32 *p,
+		struct nfsd_fhandle *resp)
+{
+	fh_put(&resp->fh);
+	return 1;
+}
+
+#define nfsaclsvc_decode_voidargs	NULL
+#define nfsaclsvc_encode_voidres	NULL
+#define nfsaclsvc_release_void		NULL
+#define nfsd3_fhandleargs	nfsd_fhandle
+#define nfsd3_attrstatres	nfsd_attrstat
+#define nfsd3_voidres		nfsd3_voidargs
+struct nfsd3_voidargs { int dummy; };
+
+#define PROC(name, argt, rest, relt, cache, respsize)	\
+ { (svc_procfunc) nfsacld_proc_##name,		\
+   (kxdrproc_t) nfsaclsvc_decode_##argt##args,	\
+   (kxdrproc_t) nfsaclsvc_encode_##rest##res,	\
+   (kxdrproc_t) nfsaclsvc_release_##relt,		\
+   sizeof(struct nfsd3_##argt##args),		\
+   sizeof(struct nfsd3_##rest##res),		\
+   0,						\
+   cache,					\
+   respsize,					\
+ }
+
+#define ST 1		/* status*/
+#define AT 21		/* attributes */
+#define pAT (1+AT)	/* post attributes - conditional */
+#define ACL (1+NFS_ACL_MAX_ENTRIES*3)  /* Access Control List */
+
+static struct svc_procedure		nfsd_acl_procedures2[] = {
+  PROC(null,	void,		void,		void,	  RC_NOCACHE, ST),
+  PROC(getacl,	getacl,		getacl,		getacl,	  RC_NOCACHE, ST+1+2*(1+ACL)),
+  PROC(setacl,	setacl,		attrstat,	fhandle,  RC_NOCACHE, ST+AT),
+  PROC(getattr, fhandle,	attrstat,	fhandle,  RC_NOCACHE, ST+AT),
+  PROC(access,	access,		access,		fhandle,  RC_NOCACHE, ST+AT+1),
+};
+
+struct svc_version	nfsd_acl_version2 = {
+		.vs_vers	= 2,
+		.vs_nproc	= 5,
+		.vs_proc	= nfsd_acl_procedures2,
+		.vs_dispatch	= nfsd_dispatch,
+		.vs_xdrsize	= NFS3_SVC_XDRSIZE,
+};

fs/nfsd/nfs3acl.c

@@ -0,0 +1,267 @@
+/*
+ * linux/fs/nfsd/nfs3acl.c
+ *
+ * Process version 3 NFSACL requests.
+ *
+ * Copyright (C) 2002-2003 Andreas Gruenbacher <agruen@suse.de>
+ */
+
+#include <linux/sunrpc/svc.h>
+#include <linux/nfs3.h>
+#include <linux/nfsd/nfsd.h>
+#include <linux/nfsd/cache.h>
+#include <linux/nfsd/xdr3.h>
+#include <linux/posix_acl.h>
+#include <linux/nfsacl.h>
+
+#define RETURN_STATUS(st)	{ resp->status = (st); return (st); }
+
+/*
+ * NULL call.
+ */
+static int
+nfsd3_proc_null(struct svc_rqst *rqstp, void *argp, void *resp)
+{
+	return nfs_ok;
+}
+
+/*
+ * Get the Access and/or Default ACL of a file.
+ */
+static int nfsd3_proc_getacl(struct svc_rqst * rqstp,
+		struct nfsd3_getaclargs *argp, struct nfsd3_getaclres *resp)
+{
+	svc_fh *fh;
+	struct posix_acl *acl;
+	int nfserr = 0;
+
+	fh = fh_copy(&resp->fh, &argp->fh);
+	if ((nfserr = fh_verify(rqstp, &resp->fh, 0, MAY_NOP)))
+		RETURN_STATUS(nfserr_inval);
+
+	if (argp->mask & ~(NFS_ACL|NFS_ACLCNT|NFS_DFACL|NFS_DFACLCNT))
+		RETURN_STATUS(nfserr_inval);
+	resp->mask = argp->mask;
+
+	if (resp->mask & (NFS_ACL|NFS_ACLCNT)) {
+		acl = nfsd_get_posix_acl(fh, ACL_TYPE_ACCESS);
+		if (IS_ERR(acl)) {
+			int err = PTR_ERR(acl);
+
+			if (err == -ENODATA || err == -EOPNOTSUPP)
+				acl = NULL;
+			else {
+				nfserr = nfserrno(err);
+				goto fail;
+			}
+		}
+		if (acl == NULL) {
+			/* Solaris returns the inode's minimum ACL. */
+
+			struct inode *inode = fh->fh_dentry->d_inode;
+			acl = posix_acl_from_mode(inode->i_mode, GFP_KERNEL);
+		}
+		resp->acl_access = acl;
+	}
+	if (resp->mask & (NFS_DFACL|NFS_DFACLCNT)) {
+		/* Check how Solaris handles requests for the Default ACL
+		   of a non-directory! */
+
+		acl = nfsd_get_posix_acl(fh, ACL_TYPE_DEFAULT);
+		if (IS_ERR(acl)) {
+			int err = PTR_ERR(acl);
+
+			if (err == -ENODATA || err == -EOPNOTSUPP)
+				acl = NULL;
+			else {
+				nfserr = nfserrno(err);
+				goto fail;
+			}
+		}
+		resp->acl_default = acl;
+	}
+
+	/* resp->acl_{access,default} are released in nfs3svc_release_getacl. */
+	RETURN_STATUS(0);
+
+fail:
+	posix_acl_release(resp->acl_access);
+	posix_acl_release(resp->acl_default);
+	RETURN_STATUS(nfserr);
+}
+
+/*
+ * Set the Access and/or Default ACL of a file.
+ */
+static int nfsd3_proc_setacl(struct svc_rqst * rqstp,
+		struct nfsd3_setaclargs *argp,
+		struct nfsd3_attrstat *resp)
+{
+	svc_fh *fh;
+	int nfserr = 0;
+
+	fh = fh_copy(&resp->fh, &argp->fh);
+	nfserr = fh_verify(rqstp, &resp->fh, 0, MAY_NOP);
+
+	if (!nfserr) {
+		nfserr = nfserrno( nfsd_set_posix_acl(
+			fh, ACL_TYPE_ACCESS, argp->acl_access) );
+	}
+	if (!nfserr) {
+		nfserr = nfserrno( nfsd_set_posix_acl(
+			fh, ACL_TYPE_DEFAULT, argp->acl_default) );
+	}
+
+	/* argp->acl_{access,default} may have been allocated in
+	   nfs3svc_decode_setaclargs. */
+	posix_acl_release(argp->acl_access);
+	posix_acl_release(argp->acl_default);
+	RETURN_STATUS(nfserr);
+}
+
+/*
+ * XDR decode functions
+ */
+static int nfs3svc_decode_getaclargs(struct svc_rqst *rqstp, u32 *p,
+		struct nfsd3_getaclargs *args)
+{
+	if (!(p = nfs3svc_decode_fh(p, &args->fh)))
+		return 0;
+	args->mask = ntohl(*p); p++;
+
+	return xdr_argsize_check(rqstp, p);
+}
+
+
+static int nfs3svc_decode_setaclargs(struct svc_rqst *rqstp, u32 *p,
+		struct nfsd3_setaclargs *args)
+{
+	struct kvec *head = rqstp->rq_arg.head;
+	unsigned int base;
+	int n;
+
+	if (!(p = nfs3svc_decode_fh(p, &args->fh)))
+		return 0;
+	args->mask = ntohl(*p++);
+	if (args->mask & ~(NFS_ACL|NFS_ACLCNT|NFS_DFACL|NFS_DFACLCNT) ||
+	    !xdr_argsize_check(rqstp, p))
+		return 0;
+
+	base = (char *)p - (char *)head->iov_base;
+	n = nfsacl_decode(&rqstp->rq_arg, base, NULL,
+			  (args->mask & NFS_ACL) ?
+			  &args->acl_access : NULL);
+	if (n > 0)
+		n = nfsacl_decode(&rqstp->rq_arg, base + n, NULL,
+				  (args->mask & NFS_DFACL) ?
+				  &args->acl_default : NULL);
+	return (n > 0);
+}
+
+/*
+ * XDR encode functions
+ */
+
+/* GETACL */
+static int nfs3svc_encode_getaclres(struct svc_rqst *rqstp, u32 *p,
+		struct nfsd3_getaclres *resp)
+{
+	struct dentry *dentry = resp->fh.fh_dentry;
+
+	p = nfs3svc_encode_post_op_attr(rqstp, p, &resp->fh);
+	if (resp->status == 0 && dentry && dentry->d_inode) {
+		struct inode *inode = dentry->d_inode;
+		int w = nfsacl_size(
+			(resp->mask & NFS_ACL)   ? resp->acl_access  : NULL,
+			(resp->mask & NFS_DFACL) ? resp->acl_default : NULL);
+		struct kvec *head = rqstp->rq_res.head;
+		unsigned int base;
+		int n;
+
+		*p++ = htonl(resp->mask);
+		if (!xdr_ressize_check(rqstp, p))
+			return 0;
+		base = (char *)p - (char *)head->iov_base;
+
+		rqstp->rq_res.page_len = w;
+		while (w > 0) {
+			if (!svc_take_res_page(rqstp))
+				return 0;
+			w -= PAGE_SIZE;
+		}
+
+		n = nfsacl_encode(&rqstp->rq_res, base, inode,
+				  resp->acl_access,
+				  resp->mask & NFS_ACL, 0);
+		if (n > 0)
+			n = nfsacl_encode(&rqstp->rq_res, base + n, inode,
+					  resp->acl_default,
+					  resp->mask & NFS_DFACL,
+					  NFS_ACL_DEFAULT);
+		if (n <= 0)
+			return 0;
+	} else
+		if (!xdr_ressize_check(rqstp, p))
+			return 0;
+
+	return 1;
+}
+
+/* SETACL */
+static int nfs3svc_encode_setaclres(struct svc_rqst *rqstp, u32 *p,
+		struct nfsd3_attrstat *resp)
+{
+	p = nfs3svc_encode_post_op_attr(rqstp, p, &resp->fh);
+
+	return xdr_ressize_check(rqstp, p);
+}
+
+/*
+ * XDR release functions
+ */
+static int nfs3svc_release_getacl(struct svc_rqst *rqstp, u32 *p,
+		struct nfsd3_getaclres *resp)
+{
+	fh_put(&resp->fh);
+	posix_acl_release(resp->acl_access);
+	posix_acl_release(resp->acl_default);
+	return 1;
+}
+
+#define nfs3svc_decode_voidargs		NULL
+#define nfs3svc_release_void		NULL
+#define nfsd3_setaclres			nfsd3_attrstat
+#define nfsd3_voidres			nfsd3_voidargs
+struct nfsd3_voidargs { int dummy; };
+
+#define PROC(name, argt, rest, relt, cache, respsize)	\
+ { (svc_procfunc) nfsd3_proc_##name,		\
+   (kxdrproc_t) nfs3svc_decode_##argt##args,	\
+   (kxdrproc_t) nfs3svc_encode_##rest##res,	\
+   (kxdrproc_t) nfs3svc_release_##relt,		\
+   sizeof(struct nfsd3_##argt##args),		\
+   sizeof(struct nfsd3_##rest##res),		\
+   0,						\
+   cache,					\
+   respsize,					\
+ }
+
+#define ST 1		/* status*/
+#define AT 21		/* attributes */
+#define pAT (1+AT)	/* post attributes - conditional */
+#define ACL (1+NFS_ACL_MAX_ENTRIES*3)  /* Access Control List */
+
+static struct svc_procedure		nfsd_acl_procedures3[] = {
+  PROC(null,	void,		void,		void,	  RC_NOCACHE, ST),
+  PROC(getacl,	getacl,		getacl,		getacl,	  RC_NOCACHE, ST+1+2*(1+ACL)),
+  PROC(setacl,	setacl,		setacl,		fhandle,  RC_NOCACHE, ST+pAT),
+};
+
+struct svc_version	nfsd_acl_version3 = {
+		.vs_vers	= 3,
+		.vs_nproc	= 3,
+		.vs_proc	= nfsd_acl_procedures3,
+		.vs_dispatch	= nfsd_dispatch,
+		.vs_xdrsize	= NFS3_SVC_XDRSIZE,
+};
+

fs/nfsd/nfs3xdr.c

@@ -71,6 +71,12 @@ decode_fh(u32 *p, struct svc_fh *fhp)
 	return p + XDR_QUADLEN(size);
 }
 
+/* Helper function for NFSv3 ACL code */
+u32 *nfs3svc_decode_fh(u32 *p, struct svc_fh *fhp)
+{
+	return decode_fh(p, fhp);
+}
+
 static inline u32 *
 encode_fh(u32 *p, struct svc_fh *fhp)
 {
@@ -233,6 +239,13 @@ encode_post_op_attr(struct svc_rqst *rqstp, u32 *p, struct svc_fh *fhp)
 	return p;
 }
 
+/* Helper for NFSv3 ACLs */
+u32 *
+nfs3svc_encode_post_op_attr(struct svc_rqst *rqstp, u32 *p, struct svc_fh *fhp)
+{
+	return encode_post_op_attr(rqstp, p, fhp);
+}
+
 /*
  * Enocde weak cache consistency data
  */

fs/nfsd/nfs4callback.c

@@ -430,7 +430,7 @@ nfsd4_probe_callback(struct nfs4_client *clp)
 	clnt = rpc_create_client(xprt, hostname, program, 1, RPC_AUTH_UNIX);
 	if (IS_ERR(clnt)) {
 		dprintk("NFSD: couldn't create callback client\n");
-		goto out_xprt;
+		goto out_err;
 	}
 	clnt->cl_intr = 0;
 	clnt->cl_softrtry = 1;
@@ -465,8 +465,6 @@ nfsd4_probe_callback(struct nfs4_client *clp)
 out_clnt:
 	rpc_shutdown_client(clnt);
 	goto out_err;
-out_xprt:
-	xprt_destroy(xprt);
 out_err:
 	dprintk("NFSD: warning: no callback path to client %.*s\n",
 		(int)clp->cl_name.len, clp->cl_name.data);

fs/nfsd/nfsproc.c

@@ -591,6 +591,7 @@ nfserrno (int errno)
 		{ nfserr_dropit, -ENOMEM },
 		{ nfserr_badname, -ESRCH },
 		{ nfserr_io, -ETXTBSY },
+		{ nfserr_notsupp, -EOPNOTSUPP },
 		{ -1, -EIO }
 	};
 	int	i;

fs/nfsd/nfssvc.c

@@ -31,6 +31,7 @@
 #include <linux/nfsd/stats.h>
 #include <linux/nfsd/cache.h>
 #include <linux/lockd/bind.h>
+#include <linux/nfsacl.h>
 
 #define NFSDDBG_FACILITY	NFSDDBG_SVC
 
@@ -362,6 +363,32 @@ nfsd_dispatch(struct svc_rqst *rqstp, u32 *statp)
 	return 1;
 }
 
+#if defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL)
+static struct svc_stat	nfsd_acl_svcstats;
+static struct svc_version *	nfsd_acl_version[] = {
+	[2] = &nfsd_acl_version2,
+	[3] = &nfsd_acl_version3,
+};
+
+#define NFSD_ACL_NRVERS		(sizeof(nfsd_acl_version)/sizeof(nfsd_acl_version[0]))
+static struct svc_program	nfsd_acl_program = {
+	.pg_prog		= NFS_ACL_PROGRAM,
+	.pg_nvers		= NFSD_ACL_NRVERS,
+	.pg_vers		= nfsd_acl_version,
+	.pg_name		= "nfsd",
+	.pg_class		= "nfsd",
+	.pg_stats		= &nfsd_acl_svcstats,
+};
+
+static struct svc_stat	nfsd_acl_svcstats = {
+	.program	= &nfsd_acl_program,
+};
+
+#define nfsd_acl_program_p	&nfsd_acl_program
+#else
+#define nfsd_acl_program_p	NULL
+#endif /* defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL) */
+
 extern struct svc_version nfsd_version2, nfsd_version3, nfsd_version4;
 
 static struct svc_version *	nfsd_version[] = {
@@ -376,6 +403,7 @@ static struct svc_version *	nfsd_version[] = {
 
 #define NFSD_NRVERS		(sizeof(nfsd_version)/sizeof(nfsd_version[0]))
 struct svc_program		nfsd_program = {
+	.pg_next		= nfsd_acl_program_p,
 	.pg_prog		= NFS_PROGRAM,		/* program number */
 	.pg_nvers		= NFSD_NRVERS,		/* nr of entries in nfsd_version */
 	.pg_vers		= nfsd_version,		/* version table */

fs/nfsd/nfsxdr.c

@@ -49,6 +49,12 @@ decode_fh(u32 *p, struct svc_fh *fhp)
 	return p + (NFS_FHSIZE >> 2);
 }
 
+/* Helper function for NFSv2 ACL code */
+u32 *nfs2svc_decode_fh(u32 *p, struct svc_fh *fhp)
+{
+	return decode_fh(p, fhp);
+}
+
 static inline u32 *
 encode_fh(u32 *p, struct svc_fh *fhp)
 {
@@ -190,6 +196,11 @@ encode_fattr(struct svc_rqst *rqstp, u32 *p, struct svc_fh *fhp)
 	return p;
 }
 
+/* Helper function for NFSv2 ACL code */
+u32 *nfs2svc_encode_fattr(struct svc_rqst *rqstp, u32 *p, struct svc_fh *fhp)
+{
+	return encode_fattr(rqstp, p, fhp);
+}
 
 /*
  * XDR decode functions

fs/nfsd/vfs.c

@@ -46,8 +46,9 @@
 #include <linux/nfsd/nfsfh.h>
 #include <linux/quotaops.h>
 #include <linux/dnotify.h>
-#ifdef CONFIG_NFSD_V4
+#include <linux/xattr_acl.h>
 #include <linux/posix_acl.h>
+#ifdef CONFIG_NFSD_V4
 #include <linux/posix_acl_xattr.h>
 #include <linux/xattr_acl.h>
 #include <linux/xattr.h>
@@ -1857,3 +1858,107 @@ nfsd_racache_init(int cache_size)
 	nfsdstats.ra_size = cache_size;
 	return 0;
 }
+
+#if defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL)
+struct posix_acl *
+nfsd_get_posix_acl(struct svc_fh *fhp, int type)
+{
+	struct inode *inode = fhp->fh_dentry->d_inode;
+	char *name;
+	void *value = NULL;
+	ssize_t size;
+	struct posix_acl *acl;
+
+	if (!IS_POSIXACL(inode) || !inode->i_op || !inode->i_op->getxattr)
+		return ERR_PTR(-EOPNOTSUPP);
+	switch(type) {
+		case ACL_TYPE_ACCESS:
+			name = XATTR_NAME_ACL_ACCESS;
+			break;
+		case ACL_TYPE_DEFAULT:
+			name = XATTR_NAME_ACL_DEFAULT;
+			break;
+		default:
+			return ERR_PTR(-EOPNOTSUPP);
+	}
+
+	size = inode->i_op->getxattr(fhp->fh_dentry, name, NULL, 0);
+
+	if (size < 0) {
+		acl = ERR_PTR(size);
+		goto getout;
+	} else if (size > 0) {
+		value = kmalloc(size, GFP_KERNEL);
+		if (!value) {
+			acl = ERR_PTR(-ENOMEM);
+			goto getout;
+		}
+		size = inode->i_op->getxattr(fhp->fh_dentry, name, value, size);
+		if (size < 0) {
+			acl = ERR_PTR(size);
+			goto getout;
+		}
+	}
+	acl = posix_acl_from_xattr(value, size);
+
+getout:
+	kfree(value);
+	return acl;
+}
+
+int
+nfsd_set_posix_acl(struct svc_fh *fhp, int type, struct posix_acl *acl)
+{
+	struct inode *inode = fhp->fh_dentry->d_inode;
+	char *name;
+	void *value = NULL;
+	size_t size;
+	int error;
+
+	if (!IS_POSIXACL(inode) || !inode->i_op ||
+	    !inode->i_op->setxattr || !inode->i_op->removexattr)
+		return -EOPNOTSUPP;
+	switch(type) {
+		case ACL_TYPE_ACCESS:
+			name = XATTR_NAME_ACL_ACCESS;
+			break;
+		case ACL_TYPE_DEFAULT:
+			name = XATTR_NAME_ACL_DEFAULT;
+			break;
+		default:
+			return -EOPNOTSUPP;
+	}
+
+	if (acl && acl->a_count) {
+		size = xattr_acl_size(acl->a_count);
+		value = kmalloc(size, GFP_KERNEL);
+		if (!value)
+			return -ENOMEM;
+		size = posix_acl_to_xattr(acl, value, size);
+		if (size < 0) {
+			error = size;
+			goto getout;
+		}
+	} else
+		size = 0;
+
+	if (!fhp->fh_locked)
+		fh_lock(fhp);  /* unlocking is done automatically */
+	if (size)
+		error = inode->i_op->setxattr(fhp->fh_dentry, name,
+					      value, size, 0);
+	else {
+		if (!S_ISDIR(inode->i_mode) && type == ACL_TYPE_DEFAULT)
+			error = 0;
+		else {
+			error = inode->i_op->removexattr(fhp->fh_dentry, name);
+			if (error == -ENODATA)
+				error = 0;
+		}
+	}
+
+getout:
+	kfree(value);
+	return error;
+}
+#endif  /* defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL) */

include/linux/fs.h

@@ -674,6 +674,7 @@ struct file_lock {
 	struct lock_manager_operations *fl_lmops;	/* Callbacks for lockmanagers */
 	union {
 		struct nfs_lock_info	nfs_fl;
+		struct nfs4_lock_info	nfs4_fl;
 	} fl_u;
 };
 

include/linux/lockd/lockd.h

@@ -72,6 +72,8 @@ struct nlm_lockowner {
 	uint32_t pid;
 };
 
+struct nlm_wait;
+
 /*
  * Memory chunk for NLM client RPC request.
  */
@@ -81,6 +83,7 @@ struct nlm_rqst {
 	struct nlm_host *	a_host;		/* host handle */
 	struct nlm_args		a_args;		/* arguments */
 	struct nlm_res		a_res;		/* result */
+	struct nlm_wait *	a_block;
 	char			a_owner[NLMCLNT_OHSIZE];
 };
 
@@ -142,7 +145,9 @@ extern unsigned long		nlmsvc_timeout;
  * Lockd client functions
  */
 struct nlm_rqst * nlmclnt_alloc_call(void);
-int		  nlmclnt_block(struct nlm_host *, struct file_lock *, u32 *);
+int		  nlmclnt_prepare_block(struct nlm_rqst *req, struct nlm_host *host, struct file_lock *fl);
+void		  nlmclnt_finish_block(struct nlm_rqst *req);
+long		  nlmclnt_block(struct nlm_rqst *req, long timeout);
 int		  nlmclnt_cancel(struct nlm_host *, struct file_lock *);
 u32		  nlmclnt_grant(struct nlm_lock *);
 void		  nlmclnt_recovery(struct nlm_host *, u32);

include/linux/nfs4.h

@@ -382,6 +382,8 @@ enum {
 	NFSPROC4_CLNT_READDIR,
 	NFSPROC4_CLNT_SERVER_CAPS,
 	NFSPROC4_CLNT_DELEGRETURN,
+	NFSPROC4_CLNT_GETACL,
+	NFSPROC4_CLNT_SETACL,
 };
 
 #endif

include/linux/nfs_fs.h

@@ -15,7 +15,6 @@
 #include <linux/pagemap.h>
 #include <linux/rwsem.h>
 #include <linux/wait.h>
-#include <linux/uio.h>
 
 #include <linux/nfs_fs_sb.h>
 
@@ -29,7 +28,6 @@
 #include <linux/nfs4.h>
 #include <linux/nfs_xdr.h>
 #include <linux/rwsem.h>
-#include <linux/workqueue.h>
 #include <linux/mempool.h>
 
 /*
@@ -43,13 +41,6 @@
 #define NFS_MAX_FILE_IO_BUFFER_SIZE	32768
 #define NFS_DEF_FILE_IO_BUFFER_SIZE	4096
 
-/*
- * The upper limit on timeouts for the exponential backoff algorithm.
- */
-#define NFS_WRITEBACK_DELAY		(5*HZ)
-#define NFS_WRITEBACK_LOCKDELAY		(60*HZ)
-#define NFS_COMMIT_DELAY		(5*HZ)
-
 /*
  * superblock magic number for NFS
  */
@@ -60,9 +51,6 @@
  */
 #define NFS_RPC_SWAPFLAGS		(RPC_TASK_SWAPPER|RPC_TASK_ROOTCREDS)
 
-#define NFS_RW_SYNC		0x0001	/* O_SYNC handling */
-#define NFS_RW_SWAP		0x0002	/* This is a swap request */
-
 /*
  * When flushing a cluster of dirty pages, there can be different
  * strategies:
@@ -96,7 +84,8 @@ struct nfs_open_context {
 	int error;
 
 	struct list_head list;
-	wait_queue_head_t waitq;
+
+	__u64 dir_cookie;
 };
 
 /*
@@ -104,6 +93,8 @@ struct nfs_open_context {
  */
 struct nfs_delegation;
 
+struct posix_acl;
+
 /*
  * nfs fs inode data in memory
  */
@@ -140,7 +131,6 @@ struct nfs_inode {
 	 *
 	 *	mtime != read_cache_mtime
 	 */
-	unsigned long		readdir_timestamp;
 	unsigned long		read_cache_jiffies;
 	unsigned long		attrtimeo;
 	unsigned long		attrtimeo_timestamp;
@@ -158,6 +148,10 @@ struct nfs_inode {
 	atomic_t		data_updates;
 
 	struct nfs_access_entry	cache_access;
+#ifdef CONFIG_NFS_V3_ACL
+	struct posix_acl	*acl_access;
+	struct posix_acl	*acl_default;
+#endif
 
 	/*
 	 * This is the cookie verifier used for NFSv3 readdir
@@ -183,13 +177,13 @@ struct nfs_inode {
 	wait_queue_head_t	nfs_i_wait;
 
 #ifdef CONFIG_NFS_V4
+	struct nfs4_cached_acl	*nfs4_acl;
         /* NFSv4 state */
 	struct list_head	open_states;
 	struct nfs_delegation	*delegation;
 	int			 delegation_state;
 	struct rw_semaphore	rwsem;
 #endif /* CONFIG_NFS_V4*/
-
 	struct inode		vfs_inode;
 };
 
@@ -203,6 +197,8 @@ struct nfs_inode {
 #define NFS_INO_INVALID_DATA	0x0010		/* cached data is invalid */
 #define NFS_INO_INVALID_ATIME	0x0020		/* cached atime is invalid */
 #define NFS_INO_INVALID_ACCESS	0x0040		/* cached access cred invalid */
+#define NFS_INO_INVALID_ACL	0x0080		/* cached acls are invalid */
+#define NFS_INO_REVAL_PAGECACHE	0x1000		/* must revalidate pagecache */
 
 static inline struct nfs_inode *NFS_I(struct inode *inode)
 {
@@ -294,12 +290,12 @@ extern int nfs_release(struct inode *, struct file *);
 extern int nfs_attribute_timeout(struct inode *inode);
 extern int nfs_revalidate_inode(struct nfs_server *server, struct inode *inode);
 extern int __nfs_revalidate_inode(struct nfs_server *, struct inode *);
+extern void nfs_revalidate_mapping(struct inode *inode, struct address_space *mapping);
 extern int nfs_setattr(struct dentry *, struct iattr *);
 extern void nfs_begin_attr_update(struct inode *);
 extern void nfs_end_attr_update(struct inode *);
 extern void nfs_begin_data_update(struct inode *);
 extern void nfs_end_data_update(struct inode *);
-extern void nfs_end_data_update_defer(struct inode *);
 extern struct nfs_open_context *alloc_nfs_open_context(struct dentry *dentry, struct rpc_cred *cred);
 extern struct nfs_open_context *get_nfs_open_context(struct nfs_open_context *ctx);
 extern void put_nfs_open_context(struct nfs_open_context *ctx);
@@ -314,6 +310,9 @@ extern u32 root_nfs_parse_addr(char *name); /*__init*/
  * linux/fs/nfs/file.c
  */
 extern struct inode_operations nfs_file_inode_operations;
+#ifdef CONFIG_NFS_V3
+extern struct inode_operations nfs3_file_inode_operations;
+#endif /* CONFIG_NFS_V3 */
 extern struct file_operations nfs_file_operations;
 extern struct address_space_operations nfs_file_aops;
 
@@ -328,6 +327,22 @@ static inline struct rpc_cred *nfs_file_cred(struct file *file)
 	return NULL;
 }
 
+/*
+ * linux/fs/nfs/xattr.c
+ */
+#ifdef CONFIG_NFS_V3_ACL
+extern ssize_t nfs3_listxattr(struct dentry *, char *, size_t);
+extern ssize_t nfs3_getxattr(struct dentry *, const char *, void *, size_t);
+extern int nfs3_setxattr(struct dentry *, const char *,
+			const void *, size_t, int);
+extern int nfs3_removexattr (struct dentry *, const char *name);
+#else
+# define nfs3_listxattr NULL
+# define nfs3_getxattr NULL
+# define nfs3_setxattr NULL
+# define nfs3_removexattr NULL
+#endif
+
 /*
  * linux/fs/nfs/direct.c
  */
@@ -342,6 +357,9 @@ extern ssize_t nfs_file_direct_write(struct kiocb *iocb, const char __user *buf,
  * linux/fs/nfs/dir.c
  */
 extern struct inode_operations nfs_dir_inode_operations;
+#ifdef CONFIG_NFS_V3
+extern struct inode_operations nfs3_dir_inode_operations;
+#endif /* CONFIG_NFS_V3 */
 extern struct file_operations nfs_dir_operations;
 extern struct dentry_operations nfs_dentry_operations;
 
@@ -377,10 +395,10 @@ extern void nfs_commit_done(struct rpc_task *);
  */
 extern int  nfs_sync_inode(struct inode *, unsigned long, unsigned int, int);
 #if defined(CONFIG_NFS_V3) || defined(CONFIG_NFS_V4)
-extern int  nfs_commit_inode(struct inode *, unsigned long, unsigned int, int);
+extern int  nfs_commit_inode(struct inode *, int);
 #else
 static inline int
-nfs_commit_inode(struct inode *inode, unsigned long idx_start, unsigned int npages, int how)
+nfs_commit_inode(struct inode *inode, int how)
 {
 	return 0;
 }
@@ -434,11 +452,6 @@ static inline void nfs_writedata_free(struct nfs_write_data *p)
 	mempool_free(p, nfs_wdata_mempool);
 }
 
-/* Hack for future NFS swap support */
-#ifndef IS_SWAPFILE
-# define IS_SWAPFILE(inode)	(0)
-#endif
-
 /*
  * linux/fs/nfs/read.c
  */
@@ -467,6 +480,29 @@ static inline void nfs_readdata_free(struct nfs_read_data *p)
 
 extern void  nfs_readdata_release(struct rpc_task *task);
 
+/*
+ * linux/fs/nfs3proc.c
+ */
+#ifdef CONFIG_NFS_V3_ACL
+extern struct posix_acl *nfs3_proc_getacl(struct inode *inode, int type);
+extern int nfs3_proc_setacl(struct inode *inode, int type,
+			    struct posix_acl *acl);
+extern int nfs3_proc_set_default_acl(struct inode *dir, struct inode *inode,
+		mode_t mode);
+extern void nfs3_forget_cached_acls(struct inode *inode);
+#else
+static inline int nfs3_proc_set_default_acl(struct inode *dir,
+					    struct inode *inode,
+					    mode_t mode)
+{
+	return 0;
+}
+
+static inline void nfs3_forget_cached_acls(struct inode *inode)
+{
+}
+#endif /* CONFIG_NFS_V3_ACL */
+
 /*
  * linux/fs/mount_clnt.c
  * (Used only by nfsroot module)
@@ -515,230 +551,6 @@ extern void * nfs_root_data(void);
 
 #define NFS_JUKEBOX_RETRY_TIME (5 * HZ)
 
-#ifdef CONFIG_NFS_V4
-
-struct idmap;
-
-/*
- * In a seqid-mutating op, this macro controls which error return
- * values trigger incrementation of the seqid.
- *
- * from rfc 3010:
- * The client MUST monotonically increment the sequence number for the
- * CLOSE, LOCK, LOCKU, OPEN, OPEN_CONFIRM, and OPEN_DOWNGRADE
- * operations.  This is true even in the event that the previous
- * operation that used the sequence number received an error.  The only
- * exception to this rule is if the previous operation received one of
- * the following errors: NFSERR_STALE_CLIENTID, NFSERR_STALE_STATEID,
- * NFSERR_BAD_STATEID, NFSERR_BAD_SEQID, NFSERR_BADXDR,
- * NFSERR_RESOURCE, NFSERR_NOFILEHANDLE.
- *
- */
-#define seqid_mutating_err(err)       \
-(((err) != NFSERR_STALE_CLIENTID) &&  \
- ((err) != NFSERR_STALE_STATEID)  &&  \
- ((err) != NFSERR_BAD_STATEID)    &&  \
- ((err) != NFSERR_BAD_SEQID)      &&  \
- ((err) != NFSERR_BAD_XDR)        &&  \
- ((err) != NFSERR_RESOURCE)       &&  \
- ((err) != NFSERR_NOFILEHANDLE))
-
-enum nfs4_client_state {
-	NFS4CLNT_OK  = 0,
-};
-
-/*
- * The nfs4_client identifies our client state to the server.
- */
-struct nfs4_client {
-	struct list_head	cl_servers;	/* Global list of servers */
-	struct in_addr		cl_addr;	/* Server identifier */
-	u64			cl_clientid;	/* constant */
-	nfs4_verifier		cl_confirm;
-	unsigned long		cl_state;
-
-	u32			cl_lockowner_id;
-
-	/*
-	 * The following rwsem ensures exclusive access to the server
-	 * while we recover the state following a lease expiration.
-	 */
-	struct rw_semaphore	cl_sem;
-
-	struct list_head	cl_delegations;
-	struct list_head	cl_state_owners;
-	struct list_head	cl_unused;
-	int			cl_nunused;
-	spinlock_t		cl_lock;
-	atomic_t		cl_count;
-
-	struct rpc_clnt *	cl_rpcclient;
-	struct rpc_cred *	cl_cred;
-
-	struct list_head	cl_superblocks;	/* List of nfs_server structs */
-
-	unsigned long		cl_lease_time;
-	unsigned long		cl_last_renewal;
-	struct work_struct	cl_renewd;
-	struct work_struct	cl_recoverd;
-
-	wait_queue_head_t	cl_waitq;
-	struct rpc_wait_queue	cl_rpcwaitq;
-
-	/* used for the setclientid verifier */
-	struct timespec		cl_boot_time;
-
-	/* idmapper */
-	struct idmap *		cl_idmap;
-
-	/* Our own IP address, as a null-terminated string.
-	 * This is used to generate the clientid, and the callback address.
-	 */
-	char			cl_ipaddr[16];
-	unsigned char		cl_id_uniquifier;
-};
-
-/*
- * NFS4 state_owners and lock_owners are simply labels for ordered
- * sequences of RPC calls. Their sole purpose is to provide once-only
- * semantics by allowing the server to identify replayed requests.
- *
- * The ->so_sema is held during all state_owner seqid-mutating operations:
- * OPEN, OPEN_DOWNGRADE, and CLOSE. Its purpose is to properly serialize
- * so_seqid.
- */
-struct nfs4_state_owner {
-	struct list_head     so_list;	 /* per-clientid list of state_owners */
-	struct nfs4_client   *so_client;
-	u32                  so_id;      /* 32-bit identifier, unique */
-	struct semaphore     so_sema;
-	u32                  so_seqid;   /* protected by so_sema */
-	atomic_t	     so_count;
-
-	struct rpc_cred	     *so_cred;	 /* Associated cred */
-	struct list_head     so_states;
-	struct list_head     so_delegations;
-};
-
-/*
- * struct nfs4_state maintains the client-side state for a given
- * (state_owner,inode) tuple (OPEN) or state_owner (LOCK).
- *
- * OPEN:
- * In order to know when to OPEN_DOWNGRADE or CLOSE the state on the server,
- * we need to know how many files are open for reading or writing on a
- * given inode. This information too is stored here.
- *
- * LOCK: one nfs4_state (LOCK) to hold the lock stateid nfs4_state(OPEN)
- */
-
-struct nfs4_lock_state {
-	struct list_head	ls_locks;	/* Other lock stateids */
-	fl_owner_t		ls_owner;	/* POSIX lock owner */
-#define NFS_LOCK_INITIALIZED 1
-	int			ls_flags;
-	u32			ls_seqid;
-	u32			ls_id;
-	nfs4_stateid		ls_stateid;
-	atomic_t		ls_count;
-};
-
-/* bits for nfs4_state->flags */
-enum {
-	LK_STATE_IN_USE,
-	NFS_DELEGATED_STATE,
-};
-
-struct nfs4_state {
-	struct list_head open_states;	/* List of states for the same state_owner */
-	struct list_head inode_states;	/* List of states for the same inode */
-	struct list_head lock_states;	/* List of subservient lock stateids */
-
-	struct nfs4_state_owner *owner;	/* Pointer to the open owner */
-	struct inode *inode;		/* Pointer to the inode */
-
-	unsigned long flags;		/* Do we hold any locks? */
-	struct semaphore lock_sema;	/* Serializes file locking operations */
-	rwlock_t state_lock;		/* Protects the lock_states list */
-
-	nfs4_stateid stateid;
-
-	unsigned int nreaders;
-	unsigned int nwriters;
-	int state;			/* State on the server (R,W, or RW) */
-	atomic_t count;
-};
-
-
-struct nfs4_exception {
-	long timeout;
-	int retry;
-};
-
-struct nfs4_state_recovery_ops {
-	int (*recover_open)(struct nfs4_state_owner *, struct nfs4_state *);
-	int (*recover_lock)(struct nfs4_state *, struct file_lock *);
-};
-
-extern struct dentry_operations nfs4_dentry_operations;
-extern struct inode_operations nfs4_dir_inode_operations;
-
-/* nfs4proc.c */
-extern int nfs4_map_errors(int err);
-extern int nfs4_proc_setclientid(struct nfs4_client *, u32, unsigned short);
-extern int nfs4_proc_setclientid_confirm(struct nfs4_client *);
-extern int nfs4_proc_async_renew(struct nfs4_client *);
-extern int nfs4_proc_renew(struct nfs4_client *);
-extern int nfs4_do_close(struct inode *inode, struct nfs4_state *state, mode_t mode);
-extern struct inode *nfs4_atomic_open(struct inode *, struct dentry *, struct nameidata *);
-extern int nfs4_open_revalidate(struct inode *, struct dentry *, int);
-
-extern struct nfs4_state_recovery_ops nfs4_reboot_recovery_ops;
-extern struct nfs4_state_recovery_ops nfs4_network_partition_recovery_ops;
-
-/* nfs4renewd.c */
-extern void nfs4_schedule_state_renewal(struct nfs4_client *);
-extern void nfs4_renewd_prepare_shutdown(struct nfs_server *);
-extern void nfs4_kill_renewd(struct nfs4_client *);
-
-/* nfs4state.c */
-extern void init_nfsv4_state(struct nfs_server *);
-extern void destroy_nfsv4_state(struct nfs_server *);
-extern struct nfs4_client *nfs4_get_client(struct in_addr *);
-extern void nfs4_put_client(struct nfs4_client *clp);
-extern int nfs4_init_client(struct nfs4_client *clp);
-extern struct nfs4_client *nfs4_find_client(struct in_addr *);
-extern u32 nfs4_alloc_lockowner_id(struct nfs4_client *);
-
-extern struct nfs4_state_owner * nfs4_get_state_owner(struct nfs_server *, struct rpc_cred *);
-extern void nfs4_put_state_owner(struct nfs4_state_owner *);
-extern void nfs4_drop_state_owner(struct nfs4_state_owner *);
-extern struct nfs4_state * nfs4_get_open_state(struct inode *, struct nfs4_state_owner *);
-extern void nfs4_put_open_state(struct nfs4_state *);
-extern void nfs4_close_state(struct nfs4_state *, mode_t);
-extern struct nfs4_state *nfs4_find_state(struct inode *, struct rpc_cred *, mode_t mode);
-extern void nfs4_increment_seqid(int status, struct nfs4_state_owner *sp);
-extern void nfs4_schedule_state_recovery(struct nfs4_client *);
-extern struct nfs4_lock_state *nfs4_find_lock_state(struct nfs4_state *state, fl_owner_t);
-extern struct nfs4_lock_state *nfs4_get_lock_state(struct nfs4_state *state, fl_owner_t);
-extern void nfs4_put_lock_state(struct nfs4_lock_state *state);
-extern void nfs4_increment_lock_seqid(int status, struct nfs4_lock_state *ls);
-extern void nfs4_notify_setlk(struct nfs4_state *, struct file_lock *, struct nfs4_lock_state *);
-extern void nfs4_notify_unlck(struct nfs4_state *, struct file_lock *, struct nfs4_lock_state *);
-extern void nfs4_copy_stateid(nfs4_stateid *, struct nfs4_state *, fl_owner_t);
-
-
-
-struct nfs4_mount_data;
-#else
-#define init_nfsv4_state(server)  do { } while (0)
-#define destroy_nfsv4_state(server)       do { } while (0)
-#define nfs4_put_state_owner(inode, owner) do { } while (0)
-#define nfs4_put_open_state(state) do { } while (0)
-#define nfs4_close_state(a, b) do { } while (0)
-#define nfs4_renewd_prepare_shutdown(server) do { } while (0)
-#endif
-
 #endif /* __KERNEL__ */
 
 /*

include/linux/nfs_fs_i.h

@@ -16,6 +16,11 @@ struct nfs_lock_info {
 	struct nlm_lockowner *owner;
 };
 
+struct nfs4_lock_state;
+struct nfs4_lock_info {
+	struct nfs4_lock_state *owner;
+};
+
 /*
  * Lock flag values
  */

include/linux/nfs_fs_sb.h

@@ -10,6 +10,7 @@
 struct nfs_server {
 	struct rpc_clnt *	client;		/* RPC client handle */
 	struct rpc_clnt *	client_sys;	/* 2nd handle for FSINFO */
+	struct rpc_clnt *	client_acl;	/* ACL RPC client handle */
 	struct nfs_rpc_ops *	rpc_ops;	/* NFS protocol vector */
 	struct backing_dev_info	backing_dev_info;
 	int			flags;		/* various flags */

include/linux/nfs_mount.h

@@ -58,6 +58,7 @@ struct nfs_mount_data {
 #define NFS_MOUNT_KERBEROS	0x0100	/* 3 */
 #define NFS_MOUNT_NONLM		0x0200	/* 3 */
 #define NFS_MOUNT_BROKEN_SUID	0x0400	/* 4 */
+#define NFS_MOUNT_NOACL		0x0800	/* 4 */
 #define NFS_MOUNT_STRICTLOCK	0x1000	/* reserved for NFSv4 */
 #define NFS_MOUNT_SECFLAVOUR	0x2000	/* 5 */
 #define NFS_MOUNT_FLAGMASK	0xFFFF

include/linux/nfs_page.h

@@ -19,6 +19,12 @@
 
 #include <asm/atomic.h>
 
+/*
+ * Valid flags for the radix tree
+ */
+#define NFS_PAGE_TAG_DIRTY	0
+#define NFS_PAGE_TAG_WRITEBACK	1
+
 /*
  * Valid flags for a dirty buffer
  */
@@ -26,6 +32,7 @@
 #define PG_NEED_COMMIT		1
 #define PG_NEED_RESCHED		2
 
+struct nfs_inode;
 struct nfs_page {
 	struct list_head	wb_list,	/* Defines state of page: */
 				*wb_list_head;	/*      read/write/commit */
@@ -54,14 +61,17 @@ extern	void nfs_clear_request(struct nfs_page *req);
 extern	void nfs_release_request(struct nfs_page *req);
 
 
-extern	void nfs_list_add_request(struct nfs_page *, struct list_head *);
+extern  int nfs_scan_lock_dirty(struct nfs_inode *nfsi, struct list_head *dst,
-
+				unsigned long idx_start, unsigned int npages);
 extern	int nfs_scan_list(struct list_head *, struct list_head *,
 			  unsigned long, unsigned int);
 extern	int nfs_coalesce_requests(struct list_head *, struct list_head *,
 				  unsigned int);
 extern  int nfs_wait_on_request(struct nfs_page *);
 extern	void nfs_unlock_request(struct nfs_page *req);
+extern  int nfs_set_page_writeback_locked(struct nfs_page *req);
+extern  void nfs_clear_page_writeback(struct nfs_page *req);
+
 
 /*
  * Lock the page of an asynchronous request without incrementing the wb_count
@@ -86,6 +96,18 @@ nfs_lock_request(struct nfs_page *req)
 	return 1;
 }
 
+/**
+ * nfs_list_add_request - Insert a request into a list
+ * @req: request
+ * @head: head of list into which to insert the request.
+ */
+static inline void
+nfs_list_add_request(struct nfs_page *req, struct list_head *head)
+{
+	list_add_tail(&req->wb_list, head);
+	req->wb_list_head = head;
+}
+
 
 /**
  * nfs_list_remove_request - Remove a request from its wb_list
@@ -96,10 +118,6 @@ nfs_list_remove_request(struct nfs_page *req)
 {
 	if (list_empty(&req->wb_list))
 		return;
-	if (!NFS_WBACK_BUSY(req)) {
-		printk(KERN_ERR "NFS: unlocked request attempted removed from list!\n");
-		BUG();
-	}
 	list_del_init(&req->wb_list);
 	req->wb_list_head = NULL;
 }

include/linux/nfs_xdr.h

@@ -2,6 +2,7 @@
 #define _LINUX_NFS_XDR_H
 
 #include <linux/sunrpc/xprt.h>
+#include <linux/nfsacl.h>
 
 struct nfs4_fsid {
 	__u64 major;
@@ -326,6 +327,20 @@ struct nfs_setattrargs {
 	const u32 *			bitmask;
 };
 
+struct nfs_setaclargs {
+	struct nfs_fh *			fh;
+	size_t				acl_len;
+	unsigned int			acl_pgbase;
+	struct page **			acl_pages;
+};
+
+struct nfs_getaclargs {
+	struct nfs_fh *			fh;
+	size_t				acl_len;
+	unsigned int			acl_pgbase;
+	struct page **			acl_pages;
+};
+
 struct nfs_setattrres {
 	struct nfs_fattr *              fattr;
 	const struct nfs_server *	server;
@@ -354,6 +369,20 @@ struct nfs_readdirargs {
 	struct page **		pages;
 };
 
+struct nfs3_getaclargs {
+	struct nfs_fh *		fh;
+	int			mask;
+	struct page **		pages;
+};
+
+struct nfs3_setaclargs {
+	struct inode *		inode;
+	int			mask;
+	struct posix_acl *	acl_access;
+	struct posix_acl *	acl_default;
+	struct page **		pages;
+};
+
 struct nfs_diropok {
 	struct nfs_fh *		fh;
 	struct nfs_fattr *	fattr;
@@ -477,6 +506,15 @@ struct nfs3_readdirres {
 	int			plus;
 };
 
+struct nfs3_getaclres {
+	struct nfs_fattr *	fattr;
+	int			mask;
+	unsigned int		acl_access_count;
+	unsigned int		acl_default_count;
+	struct posix_acl *	acl_access;
+	struct posix_acl *	acl_default;
+};
+
 #ifdef CONFIG_NFS_V4
 
 typedef u64 clientid4;
@@ -667,6 +705,7 @@ struct nfs_rpc_ops {
 	int	version;		/* Protocol version */
 	struct dentry_operations *dentry_ops;
 	struct inode_operations *dir_inode_ops;
+	struct inode_operations *file_inode_ops;
 
 	int	(*getroot) (struct nfs_server *, struct nfs_fh *,
 			    struct nfs_fsinfo *);
@@ -713,6 +752,7 @@ struct nfs_rpc_ops {
 	int	(*file_open)   (struct inode *, struct file *);
 	int	(*file_release) (struct inode *, struct file *);
 	int	(*lock)(struct file *, int, struct file_lock *);
+	void	(*clear_acl_cache)(struct inode *);
 };
 
 /*
@@ -732,4 +772,7 @@ extern struct rpc_version	nfs_version2;
 extern struct rpc_version	nfs_version3;
 extern struct rpc_version	nfs_version4;
 
+extern struct rpc_version	nfsacl_version3;
+extern struct rpc_program	nfsacl_program;
+
 #endif

include/linux/nfsacl.h

@@ -0,0 +1,58 @@
+/*
+ * File: linux/nfsacl.h
+ *
+ * (C) 2003 Andreas Gruenbacher <agruen@suse.de>
+ */
+#ifndef __LINUX_NFSACL_H
+#define __LINUX_NFSACL_H
+
+#define NFS_ACL_PROGRAM	100227
+
+#define ACLPROC2_GETACL		1
+#define ACLPROC2_SETACL		2
+#define ACLPROC2_GETATTR	3
+#define ACLPROC2_ACCESS		4
+
+#define ACLPROC3_GETACL		1
+#define ACLPROC3_SETACL		2
+
+
+/* Flags for the getacl/setacl mode */
+#define NFS_ACL			0x0001
+#define NFS_ACLCNT		0x0002
+#define NFS_DFACL		0x0004
+#define NFS_DFACLCNT		0x0008
+
+/* Flag for Default ACL entries */
+#define NFS_ACL_DEFAULT		0x1000
+
+#ifdef __KERNEL__
+
+#include <linux/posix_acl.h>
+
+/* Maximum number of ACL entries over NFS */
+#define NFS_ACL_MAX_ENTRIES	1024
+
+#define NFSACL_MAXWORDS		(2*(2+3*NFS_ACL_MAX_ENTRIES))
+#define NFSACL_MAXPAGES		((2*(8+12*NFS_ACL_MAX_ENTRIES) + PAGE_SIZE-1) \
+				 >> PAGE_SHIFT)
+
+static inline unsigned int
+nfsacl_size(struct posix_acl *acl_access, struct posix_acl *acl_default)
+{
+	unsigned int w = 16;
+	w += max(acl_access ? (int)acl_access->a_count : 3, 4) * 12;
+	if (acl_default)
+		w += max((int)acl_default->a_count, 4) * 12;
+	return w;
+}
+
+extern unsigned int
+nfsacl_encode(struct xdr_buf *buf, unsigned int base, struct inode *inode,
+	      struct posix_acl *acl, int encode_entries, int typeflag);
+extern unsigned int
+nfsacl_decode(struct xdr_buf *buf, unsigned int base, unsigned int *aclcnt,
+	      struct posix_acl **pacl);
+
+#endif /* __KERNEL__ */
+#endif  /* __LINUX_NFSACL_H */

include/linux/nfsd/nfsd.h

@@ -15,6 +15,7 @@
 #include <linux/unistd.h>
 #include <linux/dirent.h>
 #include <linux/fs.h>
+#include <linux/posix_acl.h>
 #include <linux/mount.h>
 
 #include <linux/nfsd/debug.h>
@@ -124,6 +125,21 @@ int		nfsd_statfs(struct svc_rqst *, struct svc_fh *,
 int		nfsd_notify_change(struct inode *, struct iattr *);
 int		nfsd_permission(struct svc_export *, struct dentry *, int);
 
+#if defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL)
+#ifdef CONFIG_NFSD_V2_ACL
+extern struct svc_version nfsd_acl_version2;
+#else
+#define nfsd_acl_version2 NULL
+#endif
+#ifdef CONFIG_NFSD_V3_ACL
+extern struct svc_version nfsd_acl_version3;
+#else
+#define nfsd_acl_version3 NULL
+#endif
+struct posix_acl *nfsd_get_posix_acl(struct svc_fh *, int);
+int nfsd_set_posix_acl(struct svc_fh *, int, struct posix_acl *);
+#endif
+
 
 /* 
  * NFSv4 State

include/linux/nfsd/xdr.h

@@ -169,4 +169,8 @@ int nfssvc_encode_entry(struct readdir_cd *, const char *name,
 
 int nfssvc_release_fhandle(struct svc_rqst *, u32 *, struct nfsd_fhandle *);
 
+/* Helper functions for NFSv2 ACL code */
+u32 *nfs2svc_encode_fattr(struct svc_rqst *rqstp, u32 *p, struct svc_fh *fhp);
+u32 *nfs2svc_decode_fh(u32 *p, struct svc_fh *fhp);
+
 #endif /* LINUX_NFSD_H */

include/linux/nfsd/xdr3.h

@@ -110,6 +110,19 @@ struct nfsd3_commitargs {
 	__u32			count;
 };
 
+struct nfsd3_getaclargs {
+	struct svc_fh		fh;
+	int			mask;
+};
+
+struct posix_acl;
+struct nfsd3_setaclargs {
+	struct svc_fh		fh;
+	int			mask;
+	struct posix_acl	*acl_access;
+	struct posix_acl	*acl_default;
+};
+
 struct nfsd3_attrstat {
 	__u32			status;
 	struct svc_fh		fh;
@@ -209,6 +222,14 @@ struct nfsd3_commitres {
 	struct svc_fh		fh;
 };
 
+struct nfsd3_getaclres {
+	__u32			status;
+	struct svc_fh		fh;
+	int			mask;
+	struct posix_acl	*acl_access;
+	struct posix_acl	*acl_default;
+};
+
 /* dummy type for release */
 struct nfsd3_fhandle_pair {
 	__u32			dummy;
@@ -241,6 +262,7 @@ union nfsd3_xdrstore {
 	struct nfsd3_fsinfores		fsinfores;
 	struct nfsd3_pathconfres	pathconfres;
 	struct nfsd3_commitres		commitres;
+	struct nfsd3_getaclres		getaclres;
 };
 
 #define NFS3_SVC_XDRSIZE		sizeof(union nfsd3_xdrstore)
@@ -316,6 +338,10 @@ int nfs3svc_encode_entry(struct readdir_cd *, const char *name,
 int nfs3svc_encode_entry_plus(struct readdir_cd *, const char *name,
 				int namlen, loff_t offset, ino_t ino,
 				unsigned int);
+/* Helper functions for NFSv3 ACL code */
+u32 *nfs3svc_encode_post_op_attr(struct svc_rqst *rqstp, u32 *p,
+				struct svc_fh *fhp);
+u32 *nfs3svc_decode_fh(u32 *p, struct svc_fh *fhp);
 
 
 #endif /* _LINUX_NFSD_XDR3_H */

include/linux/sunrpc/clnt.h

@@ -111,6 +111,11 @@ struct rpc_procinfo {
 struct rpc_clnt *rpc_create_client(struct rpc_xprt *xprt, char *servname,
 				struct rpc_program *info,
 				u32 version, rpc_authflavor_t authflavor);
+struct rpc_clnt *rpc_new_client(struct rpc_xprt *xprt, char *servname,
+				struct rpc_program *info,
+				u32 version, rpc_authflavor_t authflavor);
+struct rpc_clnt	*rpc_bind_new_program(struct rpc_clnt *,
+				struct rpc_program *, int);
 struct rpc_clnt *rpc_clone_client(struct rpc_clnt *);
 int		rpc_shutdown_client(struct rpc_clnt *);
 int		rpc_destroy_client(struct rpc_clnt *);
@@ -129,6 +134,7 @@ void		rpc_clnt_sigmask(struct rpc_clnt *clnt, sigset_t *oldset);
 void		rpc_clnt_sigunmask(struct rpc_clnt *clnt, sigset_t *oldset);
 void		rpc_setbufsize(struct rpc_clnt *, unsigned int, unsigned int);
 size_t		rpc_max_payload(struct rpc_clnt *);
+int		rpc_ping(struct rpc_clnt *clnt, int flags);
 
 static __inline__
 int rpc_call(struct rpc_clnt *clnt, u32 proc, void *argp, void *resp, int flags)

include/linux/sunrpc/sched.h

@@ -31,7 +31,6 @@ struct rpc_wait_queue;
 struct rpc_wait {
 	struct list_head	list;		/* wait queue links */
 	struct list_head	links;		/* Links to related tasks */
-	wait_queue_head_t	waitq;		/* sync: sleep on this q */
 	struct rpc_wait_queue *	rpc_waitq;	/* RPC wait queue we're on */
 };
 

include/linux/sunrpc/svc.h

@@ -185,6 +185,17 @@ xdr_ressize_check(struct svc_rqst *rqstp, u32 *p)
 	return vec->iov_len <= PAGE_SIZE;
 }
 
+static inline struct page *
+svc_take_res_page(struct svc_rqst *rqstp)
+{
+	if (rqstp->rq_arghi <= rqstp->rq_argused)
+		return NULL;
+	rqstp->rq_arghi--;
+	rqstp->rq_respages[rqstp->rq_resused] =
+		rqstp->rq_argpages[rqstp->rq_arghi];
+	return rqstp->rq_respages[rqstp->rq_resused++];
+}
+
 static inline int svc_take_page(struct svc_rqst *rqstp)
 {
 	if (rqstp->rq_arghi <= rqstp->rq_argused)
@@ -240,9 +251,10 @@ struct svc_deferred_req {
 };
 
 /*
- * RPC program
+ * List of RPC programs on the same transport endpoint
  */
 struct svc_program {
+	struct svc_program *	pg_next;	/* other programs (same xprt) */
 	u32			pg_prog;	/* program number */
 	unsigned int		pg_lovers;	/* lowest version */
 	unsigned int		pg_hivers;	/* lowest version */

include/linux/sunrpc/xdr.h

@@ -146,7 +146,8 @@ extern void xdr_shift_buf(struct xdr_buf *, size_t);
 extern void xdr_buf_from_iov(struct kvec *, struct xdr_buf *);
 extern int xdr_buf_subsegment(struct xdr_buf *, struct xdr_buf *, int, int);
 extern int xdr_buf_read_netobj(struct xdr_buf *, struct xdr_netobj *, int);
-extern int read_bytes_from_xdr_buf(struct xdr_buf *buf, int base, void *obj, int len);
+extern int read_bytes_from_xdr_buf(struct xdr_buf *, int, void *, int);
+extern int write_bytes_to_xdr_buf(struct xdr_buf *, int, void *, int);
 
 /*
  * Helper structure for copying from an sk_buff.
@@ -160,7 +161,7 @@ typedef struct {
 
 typedef size_t (*skb_read_actor_t)(skb_reader_t *desc, void *to, size_t len);
 
-extern void xdr_partial_copy_from_skb(struct xdr_buf *, unsigned int,
+extern ssize_t xdr_partial_copy_from_skb(struct xdr_buf *, unsigned int,
 		skb_reader_t *, skb_read_actor_t);
 
 struct socket;
@@ -168,6 +169,22 @@ struct sockaddr;
 extern int xdr_sendpages(struct socket *, struct sockaddr *, int,
 		struct xdr_buf *, unsigned int, int);
 
+extern int xdr_encode_word(struct xdr_buf *, int, u32);
+extern int xdr_decode_word(struct xdr_buf *, int, u32 *);
+
+struct xdr_array2_desc;
+typedef int (*xdr_xcode_elem_t)(struct xdr_array2_desc *desc, void *elem);
+struct xdr_array2_desc {
+	unsigned int elem_size;
+	unsigned int array_len;
+	xdr_xcode_elem_t xcode;
+};
+
+extern int xdr_decode_array2(struct xdr_buf *buf, unsigned int base,
+                             struct xdr_array2_desc *desc);
+extern int xdr_encode_array2(struct xdr_buf *buf, unsigned int base,
+			     struct xdr_array2_desc *desc);
+
 /*
  * Provide some simple tools for XDR buffer overflow-checking etc.
  */

net/sunrpc/auth.c

@@ -66,10 +66,10 @@ rpcauth_create(rpc_authflavor_t pseudoflavor, struct rpc_clnt *clnt)
 	u32			flavor = pseudoflavor_to_flavor(pseudoflavor);
 
 	if (flavor >= RPC_AUTH_MAXFLAVOR || !(ops = auth_flavors[flavor]))
-		return NULL;
+		return ERR_PTR(-EINVAL);
 	auth = ops->create(clnt, pseudoflavor);
-	if (!auth)
+	if (IS_ERR(auth))
-		return NULL;
+		return auth;
 	if (clnt->cl_auth)
 		rpcauth_destroy(clnt->cl_auth);
 	clnt->cl_auth = auth;

net/sunrpc/auth_gss/auth_gss.c

@@ -660,14 +660,16 @@ gss_create(struct rpc_clnt *clnt, rpc_authflavor_t flavor)
 {
 	struct gss_auth *gss_auth;
 	struct rpc_auth * auth;
+	int err = -ENOMEM; /* XXX? */
 
 	dprintk("RPC:      creating GSS authenticator for client %p\n",clnt);
 
 	if (!try_module_get(THIS_MODULE))
-		return NULL;
+		return ERR_PTR(err);
 	if (!(gss_auth = kmalloc(sizeof(*gss_auth), GFP_KERNEL)))
 		goto out_dec;
 	gss_auth->client = clnt;
+	err = -EINVAL;
 	gss_auth->mech = gss_mech_get_by_pseudoflavor(flavor);
 	if (!gss_auth->mech) {
 		printk(KERN_WARNING "%s: Pseudoflavor %d not found!",
@@ -675,9 +677,8 @@ gss_create(struct rpc_clnt *clnt, rpc_authflavor_t flavor)
 		goto err_free;
 	}
 	gss_auth->service = gss_pseudoflavor_to_service(gss_auth->mech, flavor);
-	/* FIXME: Will go away once privacy support is merged in */
+	if (gss_auth->service == 0)
-	if (gss_auth->service == RPC_GSS_SVC_PRIVACY)
+		goto err_put_mech;
-		gss_auth->service = RPC_GSS_SVC_INTEGRITY;
 	INIT_LIST_HEAD(&gss_auth->upcalls);
 	spin_lock_init(&gss_auth->lock);
 	auth = &gss_auth->rpc_auth;
@@ -687,15 +688,18 @@ gss_create(struct rpc_clnt *clnt, rpc_authflavor_t flavor)
 	auth->au_flavor = flavor;
 	atomic_set(&auth->au_count, 1);
 
-	if (rpcauth_init_credcache(auth, GSS_CRED_EXPIRE) < 0)
+	err = rpcauth_init_credcache(auth, GSS_CRED_EXPIRE);
+	if (err)
 		goto err_put_mech;
 
 	snprintf(gss_auth->path, sizeof(gss_auth->path), "%s/%s",
 			clnt->cl_pathname,
 			gss_auth->mech->gm_name);
 	gss_auth->dentry = rpc_mkpipe(gss_auth->path, clnt, &gss_upcall_ops, RPC_PIPE_WAIT_FOR_OPEN);
-	if (IS_ERR(gss_auth->dentry))
+	if (IS_ERR(gss_auth->dentry)) {
+		err = PTR_ERR(gss_auth->dentry);
 		goto err_put_mech;
+	}
 
 	return auth;
 err_put_mech:
@@ -704,7 +708,7 @@ gss_create(struct rpc_clnt *clnt, rpc_authflavor_t flavor)
 	kfree(gss_auth);
 out_dec:
 	module_put(THIS_MODULE);
-	return NULL;
+	return ERR_PTR(err);
 }
 
 static void

net/sunrpc/clnt.c

@@ -97,12 +97,13 @@ rpc_setup_pipedir(struct rpc_clnt *clnt, char *dir_name)
  * made to sleep too long.
  */
 struct rpc_clnt *
-rpc_create_client(struct rpc_xprt *xprt, char *servname,
+rpc_new_client(struct rpc_xprt *xprt, char *servname,
 		  struct rpc_program *program, u32 vers,
 		  rpc_authflavor_t flavor)
 {
 	struct rpc_version	*version;
 	struct rpc_clnt		*clnt = NULL;
+	struct rpc_auth		*auth;
 	int err;
 	int len;
 
@@ -157,10 +158,11 @@ rpc_create_client(struct rpc_xprt *xprt, char *servname,
 	if (err < 0)
 		goto out_no_path;
 
-	err = -ENOMEM;
+	auth = rpcauth_create(flavor, clnt);
-	if (!rpcauth_create(flavor, clnt)) {
+	if (IS_ERR(auth)) {
 		printk(KERN_INFO "RPC: Couldn't create auth handle (flavor %u)\n",
 				flavor);
+		err = PTR_ERR(auth);
 		goto out_no_auth;
 	}
 
@@ -178,6 +180,37 @@ rpc_create_client(struct rpc_xprt *xprt, char *servname,
 		kfree(clnt->cl_server);
 	kfree(clnt);
 out_err:
+	xprt_destroy(xprt);
+	return ERR_PTR(err);
+}
+
+/**
+ * Create an RPC client
+ * @xprt - pointer to xprt struct
+ * @servname - name of server
+ * @info - rpc_program
+ * @version - rpc_program version
+ * @authflavor - rpc_auth flavour to use
+ *
+ * Creates an RPC client structure, then pings the server in order to
+ * determine if it is up, and if it supports this program and version.
+ *
+ * This function should never be called by asynchronous tasks such as
+ * the portmapper.
+ */
+struct rpc_clnt *rpc_create_client(struct rpc_xprt *xprt, char *servname,
+		struct rpc_program *info, u32 version, rpc_authflavor_t authflavor)
+{
+	struct rpc_clnt *clnt;
+	int err;
+	
+	clnt = rpc_new_client(xprt, servname, info, version, authflavor);
+	if (IS_ERR(clnt))
+		return clnt;
+	err = rpc_ping(clnt, RPC_TASK_SOFT|RPC_TASK_NOINTR);
+	if (err == 0)
+		return clnt;
+	rpc_shutdown_client(clnt);
 	return ERR_PTR(err);
 }
 
@@ -208,6 +241,8 @@ rpc_clone_client(struct rpc_clnt *clnt)
 	rpc_init_rtt(&new->cl_rtt_default, clnt->cl_xprt->timeout.to_initval);
 	if (new->cl_auth)
 		atomic_inc(&new->cl_auth->au_count);
+	new->cl_pmap		= &new->cl_pmap_default;
+	rpc_init_wait_queue(&new->cl_pmap_default.pm_bindwait, "bindwait");
 	return new;
 out_no_clnt:
 	printk(KERN_INFO "RPC: out of memory in %s\n", __FUNCTION__);
@@ -296,6 +331,44 @@ rpc_release_client(struct rpc_clnt *clnt)
 		rpc_destroy_client(clnt);
 }
 
+/**
+ * rpc_bind_new_program - bind a new RPC program to an existing client
+ * @old - old rpc_client
+ * @program - rpc program to set
+ * @vers - rpc program version
+ *
+ * Clones the rpc client and sets up a new RPC program. This is mainly
+ * of use for enabling different RPC programs to share the same transport.
+ * The Sun NFSv2/v3 ACL protocol can do this.
+ */
+struct rpc_clnt *rpc_bind_new_program(struct rpc_clnt *old,
+				      struct rpc_program *program,
+				      int vers)
+{
+	struct rpc_clnt *clnt;
+	struct rpc_version *version;
+	int err;
+
+	BUG_ON(vers >= program->nrvers || !program->version[vers]);
+	version = program->version[vers];
+	clnt = rpc_clone_client(old);
+	if (IS_ERR(clnt))
+		goto out;
+	clnt->cl_procinfo = version->procs;
+	clnt->cl_maxproc  = version->nrprocs;
+	clnt->cl_protname = program->name;
+	clnt->cl_prog     = program->number;
+	clnt->cl_vers     = version->number;
+	clnt->cl_stats    = program->stats;
+	err = rpc_ping(clnt, RPC_TASK_SOFT|RPC_TASK_NOINTR);
+	if (err != 0) {
+		rpc_shutdown_client(clnt);
+		clnt = ERR_PTR(err);
+	}
+out:	
+	return clnt;
+}
+
 /*
  * Default callback for async RPC calls
  */
@@ -305,38 +378,41 @@ rpc_default_callback(struct rpc_task *task)
 }
 
 /*
- *	Export the signal mask handling for aysnchronous code that
+ *	Export the signal mask handling for synchronous code that
  *	sleeps on RPC calls
  */
+#define RPC_INTR_SIGNALS (sigmask(SIGINT) | sigmask(SIGQUIT) | sigmask(SIGKILL))
+ 
+static void rpc_save_sigmask(sigset_t *oldset, int intr)
+{
+	unsigned long	sigallow = 0;
+	sigset_t sigmask;
+
+	/* Block all signals except those listed in sigallow */
+	if (intr)
+		sigallow |= RPC_INTR_SIGNALS;
+	siginitsetinv(&sigmask, sigallow);
+	sigprocmask(SIG_BLOCK, &sigmask, oldset);
+}
+
+static inline void rpc_task_sigmask(struct rpc_task *task, sigset_t *oldset)
+{
+	rpc_save_sigmask(oldset, !RPC_TASK_UNINTERRUPTIBLE(task));
+}
+
+static inline void rpc_restore_sigmask(sigset_t *oldset)
+{
+	sigprocmask(SIG_SETMASK, oldset, NULL);
+}
 
 void rpc_clnt_sigmask(struct rpc_clnt *clnt, sigset_t *oldset)
 {
-	unsigned long	sigallow = sigmask(SIGKILL);
+	rpc_save_sigmask(oldset, clnt->cl_intr);
-	unsigned long	irqflags;
-	
-	/* Turn off various signals */
-	if (clnt->cl_intr) {
-		struct k_sigaction *action = current->sighand->action;
-		if (action[SIGINT-1].sa.sa_handler == SIG_DFL)
-			sigallow |= sigmask(SIGINT);
-		if (action[SIGQUIT-1].sa.sa_handler == SIG_DFL)
-			sigallow |= sigmask(SIGQUIT);
-	}
-	spin_lock_irqsave(&current->sighand->siglock, irqflags);
-	*oldset = current->blocked;
-	siginitsetinv(&current->blocked, sigallow & ~oldset->sig[0]);
-	recalc_sigpending();
-	spin_unlock_irqrestore(&current->sighand->siglock, irqflags);
 }
 
 void rpc_clnt_sigunmask(struct rpc_clnt *clnt, sigset_t *oldset)
 {
-	unsigned long	irqflags;
+	rpc_restore_sigmask(oldset);
-	
-	spin_lock_irqsave(&current->sighand->siglock, irqflags);
-	current->blocked = *oldset;
-	recalc_sigpending();
-	spin_unlock_irqrestore(&current->sighand->siglock, irqflags);
 }
 
 /*
@@ -354,26 +430,26 @@ int rpc_call_sync(struct rpc_clnt *clnt, struct rpc_message *msg, int flags)
 
 	BUG_ON(flags & RPC_TASK_ASYNC);
 
-	rpc_clnt_sigmask(clnt, &oldset);		
-
 	status = -ENOMEM;
 	task = rpc_new_task(clnt, NULL, flags);
 	if (task == NULL)
 		goto out;
 
+	/* Mask signals on RPC calls _and_ GSS_AUTH upcalls */
+	rpc_task_sigmask(task, &oldset);
+
 	rpc_call_setup(task, msg, 0);
 
 	/* Set up the call info struct and execute the task */
-	if (task->tk_status == 0)
+	if (task->tk_status == 0) {
 		status = rpc_execute(task);
-	else {
+	} else {
 		status = task->tk_status;
 		rpc_release_task(task);
 	}
 
+	rpc_restore_sigmask(&oldset);
 out:
-	rpc_clnt_sigunmask(clnt, &oldset);		
-
 	return status;
 }
 
@@ -394,8 +470,6 @@ rpc_call_async(struct rpc_clnt *clnt, struct rpc_message *msg, int flags,
 
 	flags |= RPC_TASK_ASYNC;
 
-	rpc_clnt_sigmask(clnt, &oldset);		
-
 	/* Create/initialize a new RPC task */
 	if (!callback)
 		callback = rpc_default_callback;
@@ -404,6 +478,9 @@ rpc_call_async(struct rpc_clnt *clnt, struct rpc_message *msg, int flags,
 		goto out;
 	task->tk_calldata = data;
 
+	/* Mask signals on GSS_AUTH upcalls */
+	rpc_task_sigmask(task, &oldset);		
+
 	rpc_call_setup(task, msg, 0);
 
 	/* Set up the call info struct and execute the task */
@@ -413,9 +490,8 @@ rpc_call_async(struct rpc_clnt *clnt, struct rpc_message *msg, int flags,
 	else
 		rpc_release_task(task);
 
+	rpc_restore_sigmask(&oldset);		
 out:
-	rpc_clnt_sigunmask(clnt, &oldset);		
-
 	return status;
 }
 
@@ -593,7 +669,7 @@ call_allocate(struct rpc_task *task)
 		return;
 	printk(KERN_INFO "RPC: buffer allocation failed for task %p\n", task); 
 
-	if (RPC_IS_ASYNC(task) || !(task->tk_client->cl_intr && signalled())) {
+	if (RPC_IS_ASYNC(task) || !signalled()) {
 		xprt_release(task);
 		task->tk_action = call_reserve;
 		rpc_delay(task, HZ>>4);
@@ -957,7 +1033,9 @@ call_header(struct rpc_task *task)
 	*p++ = htonl(clnt->cl_prog);	/* program number */
 	*p++ = htonl(clnt->cl_vers);	/* program version */
 	*p++ = htonl(task->tk_msg.rpc_proc->p_proc);	/* procedure */
-	return rpcauth_marshcred(task, p);
+	p = rpcauth_marshcred(task, p);
+	req->rq_slen = xdr_adjust_iovec(&req->rq_svec[0], p);
+	return p;
 }
 
 /*
@@ -986,10 +1064,11 @@ call_verify(struct rpc_task *task)
 			case RPC_AUTH_ERROR:
 				break;
 			case RPC_MISMATCH:
-				printk(KERN_WARNING "%s: RPC call version mismatch!\n", __FUNCTION__);
+				dprintk("%s: RPC call version mismatch!\n", __FUNCTION__);
-				goto out_eio;
+				error = -EPROTONOSUPPORT;
+				goto out_err;
 			default:
-				printk(KERN_WARNING "%s: RPC call rejected, unknown error: %x\n", __FUNCTION__, n);
+				dprintk("%s: RPC call rejected, unknown error: %x\n", __FUNCTION__, n);
 				goto out_eio;
 		}
 		if (--len < 0)
@@ -1040,23 +1119,26 @@ call_verify(struct rpc_task *task)
 	case RPC_SUCCESS:
 		return p;
 	case RPC_PROG_UNAVAIL:
-		printk(KERN_WARNING "RPC: call_verify: program %u is unsupported by server %s\n",
+		dprintk("RPC: call_verify: program %u is unsupported by server %s\n",
 				(unsigned int)task->tk_client->cl_prog,
 				task->tk_client->cl_server);
-		goto out_eio;
+		error = -EPFNOSUPPORT;
+		goto out_err;
 	case RPC_PROG_MISMATCH:
-		printk(KERN_WARNING "RPC: call_verify: program %u, version %u unsupported by server %s\n",
+		dprintk("RPC: call_verify: program %u, version %u unsupported by server %s\n",
 				(unsigned int)task->tk_client->cl_prog,
 				(unsigned int)task->tk_client->cl_vers,
 				task->tk_client->cl_server);
-		goto out_eio;
+		error = -EPROTONOSUPPORT;
+		goto out_err;
 	case RPC_PROC_UNAVAIL:
-		printk(KERN_WARNING "RPC: call_verify: proc %p unsupported by program %u, version %u on server %s\n",
+		dprintk("RPC: call_verify: proc %p unsupported by program %u, version %u on server %s\n",
 				task->tk_msg.rpc_proc,
 				task->tk_client->cl_prog,
 				task->tk_client->cl_vers,
 				task->tk_client->cl_server);
-		goto out_eio;
+		error = -EOPNOTSUPP;
+		goto out_err;
 	case RPC_GARBAGE_ARGS:
 		dprintk("RPC: %4d %s: server saw garbage\n", task->tk_pid, __FUNCTION__);
 		break;			/* retry */
@@ -1069,7 +1151,7 @@ call_verify(struct rpc_task *task)
 	task->tk_client->cl_stats->rpcgarbage++;
 	if (task->tk_garb_retry) {
 		task->tk_garb_retry--;
-		dprintk(KERN_WARNING "RPC %s: retrying %4d\n", __FUNCTION__, task->tk_pid);
+		dprintk("RPC %s: retrying %4d\n", __FUNCTION__, task->tk_pid);
 		task->tk_action = call_bind;
 		return NULL;
 	}
@@ -1083,3 +1165,30 @@ call_verify(struct rpc_task *task)
 	printk(KERN_WARNING "RPC %s: server reply was truncated.\n", __FUNCTION__);
 	goto out_retry;
 }
+
+static int rpcproc_encode_null(void *rqstp, u32 *data, void *obj)
+{
+	return 0;
+}
+
+static int rpcproc_decode_null(void *rqstp, u32 *data, void *obj)
+{
+	return 0;
+}
+
+static struct rpc_procinfo rpcproc_null = {
+	.p_encode = rpcproc_encode_null,
+	.p_decode = rpcproc_decode_null,
+};
+
+int rpc_ping(struct rpc_clnt *clnt, int flags)
+{
+	struct rpc_message msg = {
+		.rpc_proc = &rpcproc_null,
+	};
+	int err;
+	msg.rpc_cred = authnull_ops.lookup_cred(NULL, NULL, 0);
+	err = rpc_call_sync(clnt, &msg, flags);
+	put_rpccred(msg.rpc_cred);
+	return err;
+}

net/sunrpc/pmap_clnt.c

@@ -53,6 +53,9 @@ rpc_getport(struct rpc_task *task, struct rpc_clnt *clnt)
 			task->tk_pid, clnt->cl_server,
 			map->pm_prog, map->pm_vers, map->pm_prot);
 
+	/* Autobind on cloned rpc clients is discouraged */
+	BUG_ON(clnt->cl_parent != clnt);
+
 	spin_lock(&pmap_lock);
 	if (map->pm_binding) {
 		rpc_sleep_on(&map->pm_bindwait, task, NULL, NULL);
@@ -207,12 +210,10 @@ pmap_create(char *hostname, struct sockaddr_in *srvaddr, int proto)
 	xprt->addr.sin_port = htons(RPC_PMAP_PORT);
 
 	/* printk("pmap: create clnt\n"); */
-	clnt = rpc_create_client(xprt, hostname,
+	clnt = rpc_new_client(xprt, hostname,
 				&pmap_program, RPC_PMAP_VERSION,
 				RPC_AUTH_UNIX);
-	if (IS_ERR(clnt)) {
+	if (!IS_ERR(clnt)) {
-		xprt_destroy(xprt);
-	} else {
 		clnt->cl_softrtry = 1;
 		clnt->cl_chatty   = 1;
 		clnt->cl_oneshot  = 1;

net/sunrpc/sched.c

@@ -290,7 +290,7 @@ static void rpc_make_runnable(struct rpc_task *task)
 			return;
 		}
 	} else
-		wake_up(&task->u.tk_wait.waitq);
+		wake_up_bit(&task->tk_runstate, RPC_TASK_QUEUED);
 }
 
 /*
@@ -554,6 +554,38 @@ __rpc_atrun(struct rpc_task *task)
 	rpc_wake_up_task(task);
 }
 
+/*
+ * Helper that calls task->tk_exit if it exists and then returns
+ * true if we should exit __rpc_execute.
+ */
+static inline int __rpc_do_exit(struct rpc_task *task)
+{
+	if (task->tk_exit != NULL) {
+		lock_kernel();
+		task->tk_exit(task);
+		unlock_kernel();
+		/* If tk_action is non-null, we should restart the call */
+		if (task->tk_action != NULL) {
+			if (!RPC_ASSASSINATED(task)) {
+				/* Release RPC slot and buffer memory */
+				xprt_release(task);
+				rpc_free(task);
+				return 0;
+			}
+			printk(KERN_ERR "RPC: dead task tried to walk away.\n");
+		}
+	}
+	return 1;
+}
+
+static int rpc_wait_bit_interruptible(void *word)
+{
+	if (signal_pending(current))
+		return -ERESTARTSYS;
+	schedule();
+	return 0;
+}
+
 /*
  * This is the RPC `scheduler' (or rather, the finite state machine).
  */
@@ -566,8 +598,7 @@ static int __rpc_execute(struct rpc_task *task)
 
 	BUG_ON(RPC_IS_QUEUED(task));
 
- restarted:
+	for (;;) {
-	while (1) {
 		/*
 		 * Garbage collection of pending timers...
 		 */
@@ -600,11 +631,12 @@ static int __rpc_execute(struct rpc_task *task)
 		 * by someone else.
 		 */
 		if (!RPC_IS_QUEUED(task)) {
-			if (!task->tk_action)
+			if (task->tk_action != NULL) {
-				break;
 				lock_kernel();
 				task->tk_action(task);
 				unlock_kernel();
+			} else if (__rpc_do_exit(task))
+				break;
 		}
 
 		/*
@@ -624,44 +656,26 @@ static int __rpc_execute(struct rpc_task *task)
 
 		/* sync task: sleep here */
 		dprintk("RPC: %4d sync task going to sleep\n", task->tk_pid);
-		if (RPC_TASK_UNINTERRUPTIBLE(task)) {
+		/* Note: Caller should be using rpc_clnt_sigmask() */
-			__wait_event(task->u.tk_wait.waitq, !RPC_IS_QUEUED(task));
+		status = out_of_line_wait_on_bit(&task->tk_runstate,
-		} else {
+				RPC_TASK_QUEUED, rpc_wait_bit_interruptible,
-			__wait_event_interruptible(task->u.tk_wait.waitq, !RPC_IS_QUEUED(task), status);
+				TASK_INTERRUPTIBLE);
+		if (status == -ERESTARTSYS) {
 			/*
 			 * When a sync task receives a signal, it exits with
 			 * -ERESTARTSYS. In order to catch any callbacks that
 			 * clean up after sleeping on some queue, we don't
 			 * break the loop here, but go around once more.
 			 */
-			if (status == -ERESTARTSYS) {
 			dprintk("RPC: %4d got signal\n", task->tk_pid);
 			task->tk_flags |= RPC_TASK_KILLED;
 			rpc_exit(task, -ERESTARTSYS);
 			rpc_wake_up_task(task);
 		}
-		}
 		rpc_set_running(task);
 		dprintk("RPC: %4d sync task resuming\n", task->tk_pid);
 	}
 
-	if (task->tk_exit) {
-		lock_kernel();
-		task->tk_exit(task);
-		unlock_kernel();
-		/* If tk_action is non-null, the user wants us to restart */
-		if (task->tk_action) {
-			if (!RPC_ASSASSINATED(task)) {
-				/* Release RPC slot and buffer memory */
-				if (task->tk_rqstp)
-					xprt_release(task);
-				rpc_free(task);
-				goto restarted;
-			}
-			printk(KERN_ERR "RPC: dead task tries to walk away.\n");
-		}
-	}
-
 	dprintk("RPC: %4d exit() = %d\n", task->tk_pid, task->tk_status);
 	status = task->tk_status;
 
@@ -759,8 +773,6 @@ void rpc_init_task(struct rpc_task *task, struct rpc_clnt *clnt, rpc_action call
 
 	/* Initialize workqueue for async tasks */
 	task->tk_workqueue = rpciod_workqueue;
-	if (!RPC_IS_ASYNC(task))
-		init_waitqueue_head(&task->u.tk_wait.waitq);
 
 	if (clnt) {
 		atomic_inc(&clnt->cl_users);

net/sunrpc/sunrpc_syms.c

@@ -42,6 +42,7 @@ EXPORT_SYMBOL(rpc_release_task);
 /* RPC client functions */
 EXPORT_SYMBOL(rpc_create_client);
 EXPORT_SYMBOL(rpc_clone_client);
+EXPORT_SYMBOL(rpc_bind_new_program);
 EXPORT_SYMBOL(rpc_destroy_client);
 EXPORT_SYMBOL(rpc_shutdown_client);
 EXPORT_SYMBOL(rpc_release_client);
@@ -61,7 +62,6 @@ EXPORT_SYMBOL(rpc_mkpipe);
 
 /* Client transport */
 EXPORT_SYMBOL(xprt_create_proto);
-EXPORT_SYMBOL(xprt_destroy);
 EXPORT_SYMBOL(xprt_set_timeout);
 EXPORT_SYMBOL(xprt_udp_slot_table_entries);
 EXPORT_SYMBOL(xprt_tcp_slot_table_entries);
@@ -129,6 +129,10 @@ EXPORT_SYMBOL(xdr_encode_netobj);
 EXPORT_SYMBOL(xdr_encode_pages);
 EXPORT_SYMBOL(xdr_inline_pages);
 EXPORT_SYMBOL(xdr_shift_buf);
+EXPORT_SYMBOL(xdr_encode_word);
+EXPORT_SYMBOL(xdr_decode_word);
+EXPORT_SYMBOL(xdr_encode_array2);
+EXPORT_SYMBOL(xdr_decode_array2);
 EXPORT_SYMBOL(xdr_buf_from_iov);
 EXPORT_SYMBOL(xdr_buf_subsegment);
 EXPORT_SYMBOL(xdr_buf_read_netobj);

net/sunrpc/svc.c

@@ -35,12 +35,14 @@ svc_create(struct svc_program *prog, unsigned int bufsize)
 	if (!(serv = (struct svc_serv *) kmalloc(sizeof(*serv), GFP_KERNEL)))
 		return NULL;
 	memset(serv, 0, sizeof(*serv));
+	serv->sv_name      = prog->pg_name;
 	serv->sv_program   = prog;
 	serv->sv_nrthreads = 1;
 	serv->sv_stats     = prog->pg_stats;
 	serv->sv_bufsz	   = bufsize? bufsize : 4096;
-	prog->pg_lovers = prog->pg_nvers-1;
 	xdrsize = 0;
+	while (prog) {
+		prog->pg_lovers = prog->pg_nvers-1;
 		for (vers=0; vers<prog->pg_nvers ; vers++)
 			if (prog->pg_vers[vers]) {
 				prog->pg_hivers = vers;
@@ -49,6 +51,8 @@ svc_create(struct svc_program *prog, unsigned int bufsize)
 				if (prog->pg_vers[vers]->vs_xdrsize > xdrsize)
 					xdrsize = prog->pg_vers[vers]->vs_xdrsize;
 			}
+		prog = prog->pg_next;
+	}
 	serv->sv_xdrsize   = xdrsize;
 	INIT_LIST_HEAD(&serv->sv_threads);
 	INIT_LIST_HEAD(&serv->sv_sockets);
@@ -56,8 +60,6 @@ svc_create(struct svc_program *prog, unsigned int bufsize)
 	INIT_LIST_HEAD(&serv->sv_permsocks);
 	spin_lock_init(&serv->sv_lock);
 
-	serv->sv_name      = prog->pg_name;
-
 	/* Remove any stale portmap registrations */
 	svc_register(serv, 0, 0);
 
@@ -281,6 +283,7 @@ svc_process(struct svc_serv *serv, struct svc_rqst *rqstp)
 	rqstp->rq_res.len = 0;
 	rqstp->rq_res.page_base = 0;
 	rqstp->rq_res.page_len = 0;
+	rqstp->rq_res.buflen = PAGE_SIZE;
 	rqstp->rq_res.tail[0].iov_len = 0;
 	/* tcp needs a space for the record length... */
 	if (rqstp->rq_prot == IPPROTO_TCP)
@@ -338,7 +341,10 @@ svc_process(struct svc_serv *serv, struct svc_rqst *rqstp)
 		goto sendit;
 	}
 		
-	if (prog != progp->pg_prog)
+	for (progp = serv->sv_program; progp; progp = progp->pg_next)
+		if (prog == progp->pg_prog)
+			break;
+	if (progp == NULL)
 		goto err_bad_prog;
 
 	if (vers >= progp->pg_nvers ||
@@ -451,11 +457,7 @@ svc_process(struct svc_serv *serv, struct svc_rqst *rqstp)
 	goto sendit;
 
 err_bad_prog:
-#ifdef RPC_PARANOIA
+	dprintk("svc: unknown program %d\n", prog);
-	if (prog != 100227 || progp->pg_prog != 100003)
-		printk("svc: unknown program %d (me %d)\n", prog, progp->pg_prog);
-	/* else it is just a Solaris client seeing if ACLs are supported */
-#endif
 	serv->sv_stats->rpcbadfmt++;
 	svc_putu32(resv, rpc_prog_unavail);
 	goto sendit;

net/sunrpc/xdr.c

@@ -176,21 +176,23 @@ xdr_inline_pages(struct xdr_buf *xdr, unsigned int offset,
 	xdr->buflen += len;
 }
 
-void
+ssize_t
 xdr_partial_copy_from_skb(struct xdr_buf *xdr, unsigned int base,
 			  skb_reader_t *desc,
 			  skb_read_actor_t copy_actor)
 {
 	struct page	**ppage = xdr->pages;
 	unsigned int	len, pglen = xdr->page_len;
+	ssize_t		copied = 0;
 	int		ret;
 
 	len = xdr->head[0].iov_len;
 	if (base < len) {
 		len -= base;
 		ret = copy_actor(desc, (char *)xdr->head[0].iov_base + base, len);
+		copied += ret;
 		if (ret != len || !desc->count)
-			return;
+			goto out;
 		base = 0;
 	} else
 		base -= len;
@@ -210,6 +212,17 @@ xdr_partial_copy_from_skb(struct xdr_buf *xdr, unsigned int base,
 	do {
 		char *kaddr;
 
+		/* ACL likes to be lazy in allocating pages - ACLs
+		 * are small by default but can get huge. */
+		if (unlikely(*ppage == NULL)) {
+			*ppage = alloc_page(GFP_ATOMIC);
+			if (unlikely(*ppage == NULL)) {
+				if (copied == 0)
+					copied = -ENOMEM;
+				goto out;
+			}
+		}
+
 		len = PAGE_CACHE_SIZE;
 		kaddr = kmap_atomic(*ppage, KM_SKB_SUNRPC_DATA);
 		if (base) {
@@ -225,14 +238,17 @@ xdr_partial_copy_from_skb(struct xdr_buf *xdr, unsigned int base,
 		}
 		flush_dcache_page(*ppage);
 		kunmap_atomic(kaddr, KM_SKB_SUNRPC_DATA);
+		copied += ret;
 		if (ret != len || !desc->count)
-			return;
+			goto out;
 		ppage++;
 	} while ((pglen -= len) != 0);
 copy_tail:
 	len = xdr->tail[0].iov_len;
 	if (base < len)
-		copy_actor(desc, (char *)xdr->tail[0].iov_base + base, len - base);
+		copied += copy_actor(desc, (char *)xdr->tail[0].iov_base + base, len - base);
+out:
+	return copied;
 }
 
 
@@ -616,12 +632,24 @@ xdr_shift_buf(struct xdr_buf *buf, size_t len)
 void xdr_init_encode(struct xdr_stream *xdr, struct xdr_buf *buf, uint32_t *p)
 {
 	struct kvec *iov = buf->head;
+	int scratch_len = buf->buflen - buf->page_len - buf->tail[0].iov_len;
 
+	BUG_ON(scratch_len < 0);
 	xdr->buf = buf;
 	xdr->iov = iov;
-	xdr->end = (uint32_t *)((char *)iov->iov_base + iov->iov_len);
+	xdr->p = (uint32_t *)((char *)iov->iov_base + iov->iov_len);
-	buf->len = iov->iov_len = (char *)p - (char *)iov->iov_base;
+	xdr->end = (uint32_t *)((char *)iov->iov_base + scratch_len);
+	BUG_ON(iov->iov_len > scratch_len);
+
+	if (p != xdr->p && p != NULL) {
+		size_t len;
+
+		BUG_ON(p < xdr->p || p > xdr->end);
+		len = (char *)p - (char *)xdr->p;
 		xdr->p = p;
+		buf->len += len;
+		iov->iov_len += len;
+	}
 }
 EXPORT_SYMBOL(xdr_init_encode);
 
@@ -859,8 +887,34 @@ read_bytes_from_xdr_buf(struct xdr_buf *buf, int base, void *obj, int len)
 	return status;
 }
 
-static int
+/* obj is assumed to point to allocated memory of size at least len: */
-read_u32_from_xdr_buf(struct xdr_buf *buf, int base, u32 *obj)
+int
+write_bytes_to_xdr_buf(struct xdr_buf *buf, int base, void *obj, int len)
+{
+	struct xdr_buf subbuf;
+	int this_len;
+	int status;
+
+	status = xdr_buf_subsegment(buf, &subbuf, base, len);
+	if (status)
+		goto out;
+	this_len = min(len, (int)subbuf.head[0].iov_len);
+	memcpy(subbuf.head[0].iov_base, obj, this_len);
+	len -= this_len;
+	obj += this_len;
+	this_len = min(len, (int)subbuf.page_len);
+	if (this_len)
+		_copy_to_pages(subbuf.pages, subbuf.page_base, obj, this_len);
+	len -= this_len;
+	obj += this_len;
+	this_len = min(len, (int)subbuf.tail[0].iov_len);
+	memcpy(subbuf.tail[0].iov_base, obj, this_len);
+out:
+	return status;
+}
+
+int
+xdr_decode_word(struct xdr_buf *buf, int base, u32 *obj)
 {
 	u32	raw;
 	int	status;
@@ -872,6 +926,14 @@ read_u32_from_xdr_buf(struct xdr_buf *buf, int base, u32 *obj)
 	return 0;
 }
 
+int
+xdr_encode_word(struct xdr_buf *buf, int base, u32 obj)
+{
+	u32	raw = htonl(obj);
+
+	return write_bytes_to_xdr_buf(buf, base, &raw, sizeof(obj));
+}
+
 /* If the netobj starting offset bytes from the start of xdr_buf is contained
  * entirely in the head or the tail, set object to point to it; otherwise
  * try to find space for it at the end of the tail, copy it there, and
@@ -882,7 +944,7 @@ xdr_buf_read_netobj(struct xdr_buf *buf, struct xdr_netobj *obj, int offset)
 	u32	tail_offset = buf->head[0].iov_len + buf->page_len;
 	u32	obj_end_offset;
 
-	if (read_u32_from_xdr_buf(buf, offset, &obj->len))
+	if (xdr_decode_word(buf, offset, &obj->len))
 		goto out;
 	obj_end_offset = offset + 4 + obj->len;
 
@@ -915,3 +977,219 @@ xdr_buf_read_netobj(struct xdr_buf *buf, struct xdr_netobj *obj, int offset)
 out:
 	return -1;
 }
+
+/* Returns 0 on success, or else a negative error code. */
+static int
+xdr_xcode_array2(struct xdr_buf *buf, unsigned int base,
+		 struct xdr_array2_desc *desc, int encode)
+{
+	char *elem = NULL, *c;
+	unsigned int copied = 0, todo, avail_here;
+	struct page **ppages = NULL;
+	int err;
+
+	if (encode) {
+		if (xdr_encode_word(buf, base, desc->array_len) != 0)
+			return -EINVAL;
+	} else {
+		if (xdr_decode_word(buf, base, &desc->array_len) != 0 ||
+		    (unsigned long) base + 4 + desc->array_len *
+				    desc->elem_size > buf->len)
+			return -EINVAL;
+	}
+	base += 4;
+
+	if (!desc->xcode)
+		return 0;
+
+	todo = desc->array_len * desc->elem_size;
+
+	/* process head */
+	if (todo && base < buf->head->iov_len) {
+		c = buf->head->iov_base + base;
+		avail_here = min_t(unsigned int, todo,
+				   buf->head->iov_len - base);
+		todo -= avail_here;
+
+		while (avail_here >= desc->elem_size) {
+			err = desc->xcode(desc, c);
+			if (err)
+				goto out;
+			c += desc->elem_size;
+			avail_here -= desc->elem_size;
+		}
+		if (avail_here) {
+			if (!elem) {
+				elem = kmalloc(desc->elem_size, GFP_KERNEL);
+				err = -ENOMEM;
+				if (!elem)
+					goto out;
+			}
+			if (encode) {
+				err = desc->xcode(desc, elem);
+				if (err)
+					goto out;
+				memcpy(c, elem, avail_here);
+			} else
+				memcpy(elem, c, avail_here);
+			copied = avail_here;
+		}
+		base = buf->head->iov_len;  /* align to start of pages */
+	}
+
+	/* process pages array */
+	base -= buf->head->iov_len;
+	if (todo && base < buf->page_len) {
+		unsigned int avail_page;
+
+		avail_here = min(todo, buf->page_len - base);
+		todo -= avail_here;
+
+		base += buf->page_base;
+		ppages = buf->pages + (base >> PAGE_CACHE_SHIFT);
+		base &= ~PAGE_CACHE_MASK;
+		avail_page = min_t(unsigned int, PAGE_CACHE_SIZE - base,
+					avail_here);
+		c = kmap(*ppages) + base;
+
+		while (avail_here) {
+			avail_here -= avail_page;
+			if (copied || avail_page < desc->elem_size) {
+				unsigned int l = min(avail_page,
+					desc->elem_size - copied);
+				if (!elem) {
+					elem = kmalloc(desc->elem_size,
+						       GFP_KERNEL);
+					err = -ENOMEM;
+					if (!elem)
+						goto out;
+				}
+				if (encode) {
+					if (!copied) {
+						err = desc->xcode(desc, elem);
+						if (err)
+							goto out;
+					}
+					memcpy(c, elem + copied, l);
+					copied += l;
+					if (copied == desc->elem_size)
+						copied = 0;
+				} else {
+					memcpy(elem + copied, c, l);
+					copied += l;
+					if (copied == desc->elem_size) {
+						err = desc->xcode(desc, elem);
+						if (err)
+							goto out;
+						copied = 0;
+					}
+				}
+				avail_page -= l;
+				c += l;
+			}
+			while (avail_page >= desc->elem_size) {
+				err = desc->xcode(desc, c);
+				if (err)
+					goto out;
+				c += desc->elem_size;
+				avail_page -= desc->elem_size;
+			}
+			if (avail_page) {
+				unsigned int l = min(avail_page,
+					    desc->elem_size - copied);
+				if (!elem) {
+					elem = kmalloc(desc->elem_size,
+						       GFP_KERNEL);
+					err = -ENOMEM;
+					if (!elem)
+						goto out;
+				}
+				if (encode) {
+					if (!copied) {
+						err = desc->xcode(desc, elem);
+						if (err)
+							goto out;
+					}
+					memcpy(c, elem + copied, l);
+					copied += l;
+					if (copied == desc->elem_size)
+						copied = 0;
+				} else {
+					memcpy(elem + copied, c, l);
+					copied += l;
+					if (copied == desc->elem_size) {
+						err = desc->xcode(desc, elem);
+						if (err)
+							goto out;
+						copied = 0;
+					}
+				}
+			}
+			if (avail_here) {
+				kunmap(*ppages);
+				ppages++;
+				c = kmap(*ppages);
+			}
+
+			avail_page = min(avail_here,
+				 (unsigned int) PAGE_CACHE_SIZE);
+		}
+		base = buf->page_len;  /* align to start of tail */
+	}
+
+	/* process tail */
+	base -= buf->page_len;
+	if (todo) {
+		c = buf->tail->iov_base + base;
+		if (copied) {
+			unsigned int l = desc->elem_size - copied;
+
+			if (encode)
+				memcpy(c, elem + copied, l);
+			else {
+				memcpy(elem + copied, c, l);
+				err = desc->xcode(desc, elem);
+				if (err)
+					goto out;
+			}
+			todo -= l;
+			c += l;
+		}
+		while (todo) {
+			err = desc->xcode(desc, c);
+			if (err)
+				goto out;
+			c += desc->elem_size;
+			todo -= desc->elem_size;
+		}
+	}
+	err = 0;
+
+out:
+	if (elem)
+		kfree(elem);
+	if (ppages)
+		kunmap(*ppages);
+	return err;
+}
+
+int
+xdr_decode_array2(struct xdr_buf *buf, unsigned int base,
+		  struct xdr_array2_desc *desc)
+{
+	if (base >= buf->len)
+		return -EINVAL;
+
+	return xdr_xcode_array2(buf, base, desc, 0);
+}
+
+int
+xdr_encode_array2(struct xdr_buf *buf, unsigned int base,
+		  struct xdr_array2_desc *desc)
+{
+	if ((unsigned long) base + 4 + desc->array_len * desc->elem_size >
+	    buf->head->iov_len + buf->page_len + buf->tail->iov_len)
+		return -EINVAL;
+
+	return xdr_xcode_array2(buf, base, desc, 1);
+}

net/sunrpc/xprt.c

@@ -569,8 +569,11 @@ void xprt_connect(struct rpc_task *task)
 		if (xprt->sock != NULL)
 			schedule_delayed_work(&xprt->sock_connect,
 					RPC_REESTABLISH_TIMEOUT);
-		else
+		else {
 			schedule_work(&xprt->sock_connect);
+			if (!RPC_IS_ASYNC(task))
+				flush_scheduled_work();
+		}
 	}
 	return;
  out_write:
@@ -725,7 +728,8 @@ csum_partial_copy_to_xdr(struct xdr_buf *xdr, struct sk_buff *skb)
 		goto no_checksum;
 
 	desc.csum = csum_partial(skb->data, desc.offset, skb->csum);
-	xdr_partial_copy_from_skb(xdr, 0, &desc, skb_read_and_csum_bits);
+	if (xdr_partial_copy_from_skb(xdr, 0, &desc, skb_read_and_csum_bits) < 0)
+		return -1;
 	if (desc.offset != skb->len) {
 		unsigned int csum2;
 		csum2 = skb_checksum(skb, desc.offset, skb->len - desc.offset, 0);
@@ -737,7 +741,8 @@ csum_partial_copy_to_xdr(struct xdr_buf *xdr, struct sk_buff *skb)
 		return -1;
 	return 0;
 no_checksum:
-	xdr_partial_copy_from_skb(xdr, 0, &desc, skb_read_bits);
+	if (xdr_partial_copy_from_skb(xdr, 0, &desc, skb_read_bits) < 0)
+		return -1;
 	if (desc.count)
 		return -1;
 	return 0;
@@ -821,10 +826,15 @@ tcp_copy_data(skb_reader_t *desc, void *p, size_t len)
 {
 	if (len > desc->count)
 		len = desc->count;
-	if (skb_copy_bits(desc->skb, desc->offset, p, len))
+	if (skb_copy_bits(desc->skb, desc->offset, p, len)) {
+		dprintk("RPC:      failed to copy %zu bytes from skb. %zu bytes remain\n",
+				len, desc->count);
 		return 0;
+	}
 	desc->offset += len;
 	desc->count -= len;
+	dprintk("RPC:      copied %zu bytes from skb. %zu bytes remain\n",
+			len, desc->count);
 	return len;
 }
 
@@ -863,6 +873,8 @@ tcp_read_fraghdr(struct rpc_xprt *xprt, skb_reader_t *desc)
 static void
 tcp_check_recm(struct rpc_xprt *xprt)
 {
+	dprintk("RPC:      xprt = %p, tcp_copied = %lu, tcp_offset = %u, tcp_reclen = %u, tcp_flags = %lx\n",
+			xprt, xprt->tcp_copied, xprt->tcp_offset, xprt->tcp_reclen, xprt->tcp_flags);
 	if (xprt->tcp_offset == xprt->tcp_reclen) {
 		xprt->tcp_flags |= XPRT_COPY_RECM;
 		xprt->tcp_offset = 0;
@@ -907,6 +919,7 @@ tcp_read_request(struct rpc_xprt *xprt, skb_reader_t *desc)
 	struct rpc_rqst *req;
 	struct xdr_buf *rcvbuf;
 	size_t len;
+	ssize_t r;
 
 	/* Find and lock the request corresponding to this xid */
 	spin_lock(&xprt->sock_lock);
@@ -927,15 +940,40 @@ tcp_read_request(struct rpc_xprt *xprt, skb_reader_t *desc)
 		len = xprt->tcp_reclen - xprt->tcp_offset;
 		memcpy(&my_desc, desc, sizeof(my_desc));
 		my_desc.count = len;
-		xdr_partial_copy_from_skb(rcvbuf, xprt->tcp_copied,
+		r = xdr_partial_copy_from_skb(rcvbuf, xprt->tcp_copied,
 					  &my_desc, tcp_copy_data);
-		desc->count -= len;
+		desc->count -= r;
-		desc->offset += len;
+		desc->offset += r;
 	} else
-		xdr_partial_copy_from_skb(rcvbuf, xprt->tcp_copied,
+		r = xdr_partial_copy_from_skb(rcvbuf, xprt->tcp_copied,
 					  desc, tcp_copy_data);
-	xprt->tcp_copied += len;
+
-	xprt->tcp_offset += len;
+	if (r > 0) {
+		xprt->tcp_copied += r;
+		xprt->tcp_offset += r;
+	}
+	if (r != len) {
+		/* Error when copying to the receive buffer,
+		 * usually because we weren't able to allocate
+		 * additional buffer pages. All we can do now
+		 * is turn off XPRT_COPY_DATA, so the request
+		 * will not receive any additional updates,
+		 * and time out.
+		 * Any remaining data from this record will
+		 * be discarded.
+		 */
+		xprt->tcp_flags &= ~XPRT_COPY_DATA;
+		dprintk("RPC:      XID %08x truncated request\n",
+				ntohl(xprt->tcp_xid));
+		dprintk("RPC:      xprt = %p, tcp_copied = %lu, tcp_offset = %u, tcp_reclen = %u\n",
+				xprt, xprt->tcp_copied, xprt->tcp_offset, xprt->tcp_reclen);
+		goto out;
+	}
+
+	dprintk("RPC:      XID %08x read %u bytes\n",
+			ntohl(xprt->tcp_xid), r);
+	dprintk("RPC:      xprt = %p, tcp_copied = %lu, tcp_offset = %u, tcp_reclen = %u\n",
+			xprt, xprt->tcp_copied, xprt->tcp_offset, xprt->tcp_reclen);
 
 	if (xprt->tcp_copied == req->rq_private_buf.buflen)
 		xprt->tcp_flags &= ~XPRT_COPY_DATA;
@@ -944,6 +982,7 @@ tcp_read_request(struct rpc_xprt *xprt, skb_reader_t *desc)
 			xprt->tcp_flags &= ~XPRT_COPY_DATA;
 	}
 
+out:
 	if (!(xprt->tcp_flags & XPRT_COPY_DATA)) {
 		dprintk("RPC: %4d received reply complete\n",
 				req->rq_task->tk_pid);
@@ -967,6 +1006,7 @@ tcp_read_discard(struct rpc_xprt *xprt, skb_reader_t *desc)
 	desc->count -= len;
 	desc->offset += len;
 	xprt->tcp_offset += len;
+	dprintk("RPC:      discarded %u bytes\n", len);
 	tcp_check_recm(xprt);
 }
 
@@ -1064,8 +1104,7 @@ tcp_state_change(struct sock *sk)
 	case TCP_SYN_RECV:
 		break;
 	default:
-		if (xprt_test_and_clear_connected(xprt))
+		xprt_disconnect(xprt);
-			rpc_wake_up_status(&xprt->pending, -ENOTCONN);
 		break;
 	}
  out:
@@ -1203,6 +1242,8 @@ xprt_transmit(struct rpc_task *task)
 			list_add_tail(&req->rq_list, &xprt->recv);
 			spin_unlock_bh(&xprt->sock_lock);
 			xprt_reset_majortimeo(req);
+			/* Turn off autodisconnect */
+			del_singleshot_timer_sync(&xprt->timer);
 		}
 	} else if (!req->rq_bytes_sent)
 		return;
@@ -1333,8 +1374,6 @@ xprt_reserve(struct rpc_task *task)
 		spin_lock(&xprt->xprt_lock);
 		do_xprt_reserve(task);
 		spin_unlock(&xprt->xprt_lock);
-		if (task->tk_rqstp)
-			del_timer_sync(&xprt->timer);
 	}
 }
 
@@ -1649,6 +1688,10 @@ xprt_shutdown(struct rpc_xprt *xprt)
 	rpc_wake_up(&xprt->backlog);
 	wake_up(&xprt->cong_wait);
 	del_timer_sync(&xprt->timer);
+
+	/* synchronously wait for connect worker to finish */
+	cancel_delayed_work(&xprt->sock_connect);
+	flush_scheduled_work();
 }
 
 /*