Merge git://oss.sgi.com:8090/xfs/linux-2.6

* git://oss.sgi.com:8090/xfs/linux-2.6: (45 commits)
  [XFS] Fix use after free in xfs_log_done().
  [XFS] Make xfs_bmap_*_count_leaves void.
  [XFS] Use KM_NOFS for debug trace buffers
  [XFS] use KM_MAYFAIL in xfs_mountfs
  [XFS] refactor xfs_mount_free
  [XFS] don't call xfs_freesb from xfs_unmountfs
  [XFS] xfs_unmountfs should return void
  [XFS] cleanup xfs_mountfs
  [XFS] move root inode IRELE into xfs_unmountfs
  [XFS] stop using file_update_time
  [XFS] optimize xfs_ichgtime
  [XFS] update timestamp in xfs_ialloc manually
  [XFS] remove the sema_t from XFS.
  [XFS] replace dquot flush semaphore with a completion
  [XFS] replace inode flush semaphore with a completion
  [XFS] extend completions to provide XFS object flush requirements
  [XFS] replace the XFS buf iodone semaphore with a completion
  [XFS] clean up stale references to semaphores
  [XFS] use get_unaligned_* helpers
  [XFS] Fix compile failure in xfs_buf_trace()
  ...

fs/xfs/linux-2.6/sema.h

@@ -1,52 +0,0 @@
-/*
- * Copyright (c) 2000-2002,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_SUPPORT_SEMA_H__
-#define __XFS_SUPPORT_SEMA_H__
-
-#include <linux/time.h>
-#include <linux/wait.h>
-#include <linux/semaphore.h>
-#include <asm/atomic.h>
-
-/*
- * sema_t structure just maps to struct semaphore in Linux kernel.
- */
-
-typedef struct semaphore sema_t;
-
-#define initnsema(sp, val, name)	sema_init(sp, val)
-#define psema(sp, b)			down(sp)
-#define vsema(sp)			up(sp)
-#define freesema(sema)			do { } while (0)
-
-static inline int issemalocked(sema_t *sp)
-{
-	return down_trylock(sp) || (up(sp), 0);
-}
-
-/*
- * Map cpsema (try to get the sema) to down_trylock. We need to switch
- * the return values since cpsema returns 1 (acquired) 0 (failed) and
- * down_trylock returns the reverse 0 (acquired) 1 (failed).
- */
-static inline int cpsema(sema_t *sp)
-{
-	return down_trylock(sp) ? 0 : 1;
-}
-
-#endif /* __XFS_SUPPORT_SEMA_H__ */

fs/xfs/linux-2.6/xfs_aops.c

@@ -73,7 +73,6 @@ xfs_page_trace(
 	unsigned long	pgoff)
 {
 	xfs_inode_t	*ip;
-	bhv_vnode_t	*vp = vn_from_inode(inode);
 	loff_t		isize = i_size_read(inode);
 	loff_t		offset = page_offset(page);
 	int		delalloc = -1, unmapped = -1, unwritten = -1;
@@ -81,7 +80,7 @@ xfs_page_trace(
 	if (page_has_buffers(page))
 		xfs_count_page_state(page, &delalloc, &unmapped, &unwritten);
 
-	ip = xfs_vtoi(vp);
+	ip = XFS_I(inode);
 	if (!ip->i_rwtrace)
 		return;
 

fs/xfs/linux-2.6/xfs_buf.c

@@ -58,7 +58,7 @@ xfs_buf_trace(
 		bp, id,
 		(void *)(unsigned long)bp->b_flags,
 		(void *)(unsigned long)bp->b_hold.counter,
-		(void *)(unsigned long)bp->b_sema.count.counter,
+		(void *)(unsigned long)bp->b_sema.count,
 		(void *)current,
 		data, ra,
 		(void *)(unsigned long)((bp->b_file_offset>>32) & 0xffffffff),
@@ -253,7 +253,7 @@ _xfs_buf_initialize(
 
 	memset(bp, 0, sizeof(xfs_buf_t));
 	atomic_set(&bp->b_hold, 1);
-	init_MUTEX_LOCKED(&bp->b_iodonesema);
+	init_completion(&bp->b_iowait);
 	INIT_LIST_HEAD(&bp->b_list);
 	INIT_LIST_HEAD(&bp->b_hash_list);
 	init_MUTEX_LOCKED(&bp->b_sema); /* held, no waiters */
@@ -838,6 +838,7 @@ xfs_buf_rele(
 		return;
 	}
 
+	ASSERT(atomic_read(&bp->b_hold) > 0);
 	if (atomic_dec_and_lock(&bp->b_hold, &hash->bh_lock)) {
 		if (bp->b_relse) {
 			atomic_inc(&bp->b_hold);
@@ -851,11 +852,6 @@ xfs_buf_rele(
 			spin_unlock(&hash->bh_lock);
 			xfs_buf_free(bp);
 		}
-	} else {
-		/*
-		 * Catch reference count leaks
-		 */
-		ASSERT(atomic_read(&bp->b_hold) >= 0);
 	}
 }
 
@@ -1037,7 +1033,7 @@ xfs_buf_ioend(
 			xfs_buf_iodone_work(&bp->b_iodone_work);
 		}
 	} else {
-		up(&bp->b_iodonesema);
+		complete(&bp->b_iowait);
 	}
 }
 
@@ -1275,7 +1271,7 @@ xfs_buf_iowait(
 	XB_TRACE(bp, "iowait", 0);
 	if (atomic_read(&bp->b_io_remaining))
 		blk_run_address_space(bp->b_target->bt_mapping);
-	down(&bp->b_iodonesema);
+	wait_for_completion(&bp->b_iowait);
 	XB_TRACE(bp, "iowaited", (long)bp->b_error);
 	return bp->b_error;
 }
@@ -1799,7 +1795,7 @@ int __init
 xfs_buf_init(void)
 {
 #ifdef XFS_BUF_TRACE
-	xfs_buf_trace_buf = ktrace_alloc(XFS_BUF_TRACE_SIZE, KM_SLEEP);
+	xfs_buf_trace_buf = ktrace_alloc(XFS_BUF_TRACE_SIZE, KM_NOFS);
 #endif
 
 	xfs_buf_zone = kmem_zone_init_flags(sizeof(xfs_buf_t), "xfs_buf",

fs/xfs/linux-2.6/xfs_buf.h

@@ -157,7 +157,7 @@ typedef struct xfs_buf {
 	xfs_buf_iodone_t	b_iodone;	/* I/O completion function */
 	xfs_buf_relse_t		b_relse;	/* releasing function */
 	xfs_buf_bdstrat_t	b_strat;	/* pre-write function */
-	struct semaphore	b_iodonesema;	/* Semaphore for I/O waiters */
+	struct completion	b_iowait;	/* queue for I/O waiters */
 	void			*b_fspriv;
 	void			*b_fspriv2;
 	void			*b_fspriv3;
@@ -352,7 +352,7 @@ extern void xfs_buf_trace(xfs_buf_t *, char *, void *, void *);
 #define XFS_BUF_CPSEMA(bp)	(xfs_buf_cond_lock(bp) == 0)
 #define XFS_BUF_VSEMA(bp)	xfs_buf_unlock(bp)
 #define XFS_BUF_PSEMA(bp,x)	xfs_buf_lock(bp)
-#define XFS_BUF_V_IODONESEMA(bp) up(&bp->b_iodonesema);
+#define XFS_BUF_FINISH_IOWAIT(bp)	complete(&bp->b_iowait);
 
 #define XFS_BUF_SET_TARGET(bp, target)	((bp)->b_target = (target))
 #define XFS_BUF_TARGET(bp)		((bp)->b_target)

fs/xfs/linux-2.6/xfs_export.c

@@ -139,7 +139,7 @@ xfs_nfs_get_inode(
 	}
 
 	xfs_iunlock(ip, XFS_ILOCK_SHARED);
-	return ip->i_vnode;
+	return VFS_I(ip);
 }
 
 STATIC struct dentry *
@@ -167,7 +167,7 @@ xfs_fs_fh_to_dentry(struct super_block *sb, struct fid *fid,
 	if (!inode)
 		return NULL;
 	if (IS_ERR(inode))
-		return ERR_PTR(PTR_ERR(inode));
+		return ERR_CAST(inode);
 	result = d_alloc_anon(inode);
 	if (!result) {
 		iput(inode);
@@ -198,7 +198,7 @@ xfs_fs_fh_to_parent(struct super_block *sb, struct fid *fid,
 	if (!inode)
 		return NULL;
 	if (IS_ERR(inode))
-		return ERR_PTR(PTR_ERR(inode));
+		return ERR_CAST(inode);
 	result = d_alloc_anon(inode);
 	if (!result) {
 		iput(inode);
@@ -219,9 +219,9 @@ xfs_fs_get_parent(
 	if (unlikely(error))
 		return ERR_PTR(-error);
 
-	parent = d_alloc_anon(cip->i_vnode);
+	parent = d_alloc_anon(VFS_I(cip));
 	if (unlikely(!parent)) {
-		iput(cip->i_vnode);
+		iput(VFS_I(cip));
 		return ERR_PTR(-ENOMEM);
 	}
 	return parent;

fs/xfs/linux-2.6/xfs_fs_subr.c

@@ -31,7 +31,7 @@ xfs_tosspages(
 	xfs_off_t	last,
 	int		fiopt)
 {
-	struct address_space *mapping = ip->i_vnode->i_mapping;
+	struct address_space *mapping = VFS_I(ip)->i_mapping;
 
 	if (mapping->nrpages)
 		truncate_inode_pages(mapping, first);
@@ -44,7 +44,7 @@ xfs_flushinval_pages(
 	xfs_off_t	last,
 	int		fiopt)
 {
-	struct address_space *mapping = ip->i_vnode->i_mapping;
+	struct address_space *mapping = VFS_I(ip)->i_mapping;
 	int		ret = 0;
 
 	if (mapping->nrpages) {
@@ -64,7 +64,7 @@ xfs_flush_pages(
 	uint64_t	flags,
 	int		fiopt)
 {
-	struct address_space *mapping = ip->i_vnode->i_mapping;
+	struct address_space *mapping = VFS_I(ip)->i_mapping;
 	int		ret = 0;
 	int		ret2;
 

fs/xfs/linux-2.6/xfs_ioctl.c

@@ -245,7 +245,7 @@ xfs_vget_fsop_handlereq(
 
 	xfs_iunlock(ip, XFS_ILOCK_SHARED);
 
-	*inode = XFS_ITOV(ip);
+	*inode = VFS_I(ip);
 	return 0;
 }
 
@@ -927,7 +927,7 @@ STATIC void
 xfs_diflags_to_linux(
 	struct xfs_inode	*ip)
 {
-	struct inode		*inode = XFS_ITOV(ip);
+	struct inode		*inode = VFS_I(ip);
 	unsigned int		xflags = xfs_ip2xflags(ip);
 
 	if (xflags & XFS_XFLAG_IMMUTABLE)

fs/xfs/linux-2.6/xfs_iops.c

@@ -62,7 +62,7 @@ void
 xfs_synchronize_atime(
 	xfs_inode_t	*ip)
 {
-	struct inode	*inode = ip->i_vnode;
+	struct inode	*inode = VFS_I(ip);
 
 	if (inode) {
 		ip->i_d.di_atime.t_sec = (__int32_t)inode->i_atime.tv_sec;
@@ -79,7 +79,7 @@ void
 xfs_mark_inode_dirty_sync(
 	xfs_inode_t	*ip)
 {
-	struct inode	*inode = ip->i_vnode;
+	struct inode	*inode = VFS_I(ip);
 
 	if (inode)
 		mark_inode_dirty_sync(inode);
@@ -89,36 +89,31 @@ xfs_mark_inode_dirty_sync(
  * Change the requested timestamp in the given inode.
  * We don't lock across timestamp updates, and we don't log them but
  * we do record the fact that there is dirty information in core.
- *
- * NOTE -- callers MUST combine XFS_ICHGTIME_MOD or XFS_ICHGTIME_CHG
- *		with XFS_ICHGTIME_ACC to be sure that access time
- *		update will take.  Calling first with XFS_ICHGTIME_ACC
- *		and then XFS_ICHGTIME_MOD may fail to modify the access
- *		timestamp if the filesystem is mounted noacctm.
  */
 void
 xfs_ichgtime(
 	xfs_inode_t	*ip,
 	int		flags)
 {
-	struct inode	*inode = vn_to_inode(XFS_ITOV(ip));
+	struct inode	*inode = VFS_I(ip);
 	timespec_t	tv;
+	int		sync_it = 0;
+
+	tv = current_fs_time(inode->i_sb);
 
-	nanotime(&tv);
-	if (flags & XFS_ICHGTIME_MOD) {
+	if ((flags & XFS_ICHGTIME_MOD) &&
+	    !timespec_equal(&inode->i_mtime, &tv)) {
 		inode->i_mtime = tv;
 		ip->i_d.di_mtime.t_sec = (__int32_t)tv.tv_sec;
 		ip->i_d.di_mtime.t_nsec = (__int32_t)tv.tv_nsec;
+		sync_it = 1;
 	}
-	if (flags & XFS_ICHGTIME_ACC) {
-		inode->i_atime = tv;
-		ip->i_d.di_atime.t_sec = (__int32_t)tv.tv_sec;
-		ip->i_d.di_atime.t_nsec = (__int32_t)tv.tv_nsec;
-	}
-	if (flags & XFS_ICHGTIME_CHG) {
+	if ((flags & XFS_ICHGTIME_CHG) &&
+	    !timespec_equal(&inode->i_ctime, &tv)) {
 		inode->i_ctime = tv;
 		ip->i_d.di_ctime.t_sec = (__int32_t)tv.tv_sec;
 		ip->i_d.di_ctime.t_nsec = (__int32_t)tv.tv_nsec;
+		sync_it = 1;
 	}
 
 	/*
@@ -130,55 +125,11 @@ xfs_ichgtime(
 	 * ensure that the compiler does not reorder the update
 	 * of i_update_core above the timestamp updates above.
 	 */
-	SYNCHRONIZE();
-	ip->i_update_core = 1;
-	if (!(inode->i_state & I_NEW))
+	if (sync_it) {
+		SYNCHRONIZE();
+		ip->i_update_core = 1;
 		mark_inode_dirty_sync(inode);
-}
-
-/*
- * Variant on the above which avoids querying the system clock
- * in situations where we know the Linux inode timestamps have
- * just been updated (and so we can update our inode cheaply).
- */
-void
-xfs_ichgtime_fast(
-	xfs_inode_t	*ip,
-	struct inode	*inode,
-	int		flags)
-{
-	timespec_t	*tvp;
-
-	/*
-	 * Atime updates for read() & friends are handled lazily now, and
-	 * explicit updates must go through xfs_ichgtime()
-	 */
-	ASSERT((flags & XFS_ICHGTIME_ACC) == 0);
-
-	if (flags & XFS_ICHGTIME_MOD) {
-		tvp = &inode->i_mtime;
-		ip->i_d.di_mtime.t_sec = (__int32_t)tvp->tv_sec;
-		ip->i_d.di_mtime.t_nsec = (__int32_t)tvp->tv_nsec;
 	}
-	if (flags & XFS_ICHGTIME_CHG) {
-		tvp = &inode->i_ctime;
-		ip->i_d.di_ctime.t_sec = (__int32_t)tvp->tv_sec;
-		ip->i_d.di_ctime.t_nsec = (__int32_t)tvp->tv_nsec;
-	}
-
-	/*
-	 * We update the i_update_core field _after_ changing
-	 * the timestamps in order to coordinate properly with
-	 * xfs_iflush() so that we don't lose timestamp updates.
-	 * This keeps us from having to hold the inode lock
-	 * while doing this.  We use the SYNCHRONIZE macro to
-	 * ensure that the compiler does not reorder the update
-	 * of i_update_core above the timestamp updates above.
-	 */
-	SYNCHRONIZE();
-	ip->i_update_core = 1;
-	if (!(inode->i_state & I_NEW))
-		mark_inode_dirty_sync(inode);
 }
 
 /*
@@ -299,7 +250,7 @@ xfs_vn_mknod(
 	if (unlikely(error))
 		goto out_free_acl;
 
-	inode = ip->i_vnode;
+	inode = VFS_I(ip);
 
 	error = xfs_init_security(inode, dir);
 	if (unlikely(error))
@@ -366,7 +317,7 @@ xfs_vn_lookup(
 		return NULL;
 	}
 
-	return d_splice_alias(cip->i_vnode, dentry);
+	return d_splice_alias(VFS_I(cip), dentry);
 }
 
 STATIC struct dentry *
@@ -399,12 +350,12 @@ xfs_vn_ci_lookup(
 
 	/* if exact match, just splice and exit */
 	if (!ci_name.name)
-		return d_splice_alias(ip->i_vnode, dentry);
+		return d_splice_alias(VFS_I(ip), dentry);
 
 	/* else case-insensitive match... */
 	dname.name = ci_name.name;
 	dname.len = ci_name.len;
-	dentry = d_add_ci(ip->i_vnode, dentry, &dname);
+	dentry = d_add_ci(VFS_I(ip), dentry, &dname);
 	kmem_free(ci_name.name);
 	return dentry;
 }
@@ -478,7 +429,7 @@ xfs_vn_symlink(
 	if (unlikely(error))
 		goto out;
 
-	inode = cip->i_vnode;
+	inode = VFS_I(cip);
 
 	error = xfs_init_security(inode, dir);
 	if (unlikely(error))
@@ -710,7 +661,7 @@ out_error:
 	return error;
 }
 
-const struct inode_operations xfs_inode_operations = {
+static const struct inode_operations xfs_inode_operations = {
 	.permission		= xfs_vn_permission,
 	.truncate		= xfs_vn_truncate,
 	.getattr		= xfs_vn_getattr,
@@ -722,7 +673,7 @@ const struct inode_operations xfs_inode_operations = {
 	.fallocate		= xfs_vn_fallocate,
 };
 
-const struct inode_operations xfs_dir_inode_operations = {
+static const struct inode_operations xfs_dir_inode_operations = {
 	.create			= xfs_vn_create,
 	.lookup			= xfs_vn_lookup,
 	.link			= xfs_vn_link,
@@ -747,7 +698,7 @@ const struct inode_operations xfs_dir_inode_operations = {
 	.listxattr		= xfs_vn_listxattr,
 };
 
-const struct inode_operations xfs_dir_ci_inode_operations = {
+static const struct inode_operations xfs_dir_ci_inode_operations = {
 	.create			= xfs_vn_create,
 	.lookup			= xfs_vn_ci_lookup,
 	.link			= xfs_vn_link,
@@ -772,7 +723,7 @@ const struct inode_operations xfs_dir_ci_inode_operations = {
 	.listxattr		= xfs_vn_listxattr,
 };
 
-const struct inode_operations xfs_symlink_inode_operations = {
+static const struct inode_operations xfs_symlink_inode_operations = {
 	.readlink		= generic_readlink,
 	.follow_link		= xfs_vn_follow_link,
 	.put_link		= xfs_vn_put_link,
@@ -784,3 +735,98 @@ const struct inode_operations xfs_symlink_inode_operations = {
 	.removexattr		= generic_removexattr,
 	.listxattr		= xfs_vn_listxattr,
 };
+
+STATIC void
+xfs_diflags_to_iflags(
+	struct inode		*inode,
+	struct xfs_inode	*ip)
+{
+	if (ip->i_d.di_flags & XFS_DIFLAG_IMMUTABLE)
+		inode->i_flags |= S_IMMUTABLE;
+	else
+		inode->i_flags &= ~S_IMMUTABLE;
+	if (ip->i_d.di_flags & XFS_DIFLAG_APPEND)
+		inode->i_flags |= S_APPEND;
+	else
+		inode->i_flags &= ~S_APPEND;
+	if (ip->i_d.di_flags & XFS_DIFLAG_SYNC)
+		inode->i_flags |= S_SYNC;
+	else
+		inode->i_flags &= ~S_SYNC;
+	if (ip->i_d.di_flags & XFS_DIFLAG_NOATIME)
+		inode->i_flags |= S_NOATIME;
+	else
+		inode->i_flags &= ~S_NOATIME;
+}
+
+/*
+ * Initialize the Linux inode, set up the operation vectors and
+ * unlock the inode.
+ *
+ * When reading existing inodes from disk this is called directly
+ * from xfs_iget, when creating a new inode it is called from
+ * xfs_ialloc after setting up the inode.
+ */
+void
+xfs_setup_inode(
+	struct xfs_inode	*ip)
+{
+	struct inode		*inode = ip->i_vnode;
+
+	inode->i_mode	= ip->i_d.di_mode;
+	inode->i_nlink	= ip->i_d.di_nlink;
+	inode->i_uid	= ip->i_d.di_uid;
+	inode->i_gid	= ip->i_d.di_gid;
+
+	switch (inode->i_mode & S_IFMT) {
+	case S_IFBLK:
+	case S_IFCHR:
+		inode->i_rdev =
+			MKDEV(sysv_major(ip->i_df.if_u2.if_rdev) & 0x1ff,
+			      sysv_minor(ip->i_df.if_u2.if_rdev));
+		break;
+	default:
+		inode->i_rdev = 0;
+		break;
+	}
+
+	inode->i_generation = ip->i_d.di_gen;
+	i_size_write(inode, ip->i_d.di_size);
+	inode->i_atime.tv_sec	= ip->i_d.di_atime.t_sec;
+	inode->i_atime.tv_nsec	= ip->i_d.di_atime.t_nsec;
+	inode->i_mtime.tv_sec	= ip->i_d.di_mtime.t_sec;
+	inode->i_mtime.tv_nsec	= ip->i_d.di_mtime.t_nsec;
+	inode->i_ctime.tv_sec	= ip->i_d.di_ctime.t_sec;
+	inode->i_ctime.tv_nsec	= ip->i_d.di_ctime.t_nsec;
+	xfs_diflags_to_iflags(inode, ip);
+	xfs_iflags_clear(ip, XFS_IMODIFIED);
+
+	switch (inode->i_mode & S_IFMT) {
+	case S_IFREG:
+		inode->i_op = &xfs_inode_operations;
+		inode->i_fop = &xfs_file_operations;
+		inode->i_mapping->a_ops = &xfs_address_space_operations;
+		break;
+	case S_IFDIR:
+		if (xfs_sb_version_hasasciici(&XFS_M(inode->i_sb)->m_sb))
+			inode->i_op = &xfs_dir_ci_inode_operations;
+		else
+			inode->i_op = &xfs_dir_inode_operations;
+		inode->i_fop = &xfs_dir_file_operations;
+		break;
+	case S_IFLNK:
+		inode->i_op = &xfs_symlink_inode_operations;
+		if (!(ip->i_df.if_flags & XFS_IFINLINE))
+			inode->i_mapping->a_ops = &xfs_address_space_operations;
+		break;
+	default:
+		inode->i_op = &xfs_inode_operations;
+		init_special_inode(inode, inode->i_mode, inode->i_rdev);
+		break;
+	}
+
+	xfs_iflags_clear(ip, XFS_INEW);
+	barrier();
+
+	unlock_new_inode(inode);
+}

fs/xfs/linux-2.6/xfs_iops.h

@@ -18,10 +18,7 @@
 #ifndef __XFS_IOPS_H__
 #define __XFS_IOPS_H__
 
-extern const struct inode_operations xfs_inode_operations;
-extern const struct inode_operations xfs_dir_inode_operations;
-extern const struct inode_operations xfs_dir_ci_inode_operations;
-extern const struct inode_operations xfs_symlink_inode_operations;
+struct xfs_inode;
 
 extern const struct file_operations xfs_file_operations;
 extern const struct file_operations xfs_dir_file_operations;
@@ -29,14 +26,6 @@ extern const struct file_operations xfs_invis_file_operations;
 
 extern ssize_t xfs_vn_listxattr(struct dentry *, char *data, size_t size);
 
-struct xfs_inode;
-extern void xfs_ichgtime(struct xfs_inode *, int);
-extern void xfs_ichgtime_fast(struct xfs_inode *, struct inode *, int);
-
-#define xfs_vtoi(vp) \
-	((struct xfs_inode *)vn_to_inode(vp)->i_private)
-
-#define XFS_I(inode) \
-	((struct xfs_inode *)(inode)->i_private)
+extern void xfs_setup_inode(struct xfs_inode *);
 
 #endif /* __XFS_IOPS_H__ */

fs/xfs/linux-2.6/xfs_linux.h

@@ -45,13 +45,13 @@
 #include <mrlock.h>
 #include <sv.h>
 #include <mutex.h>
-#include <sema.h>
 #include <time.h>
 
 #include <support/ktrace.h>
 #include <support/debug.h>
 #include <support/uuid.h>
 
+#include <linux/semaphore.h>
 #include <linux/mm.h>
 #include <linux/kernel.h>
 #include <linux/blkdev.h>
@@ -180,7 +180,7 @@
 #define xfs_sort(a,n,s,fn)	sort(a,n,s,fn,NULL)
 #define xfs_stack_trace()	dump_stack()
 #define xfs_itruncate_data(ip, off)	\
-	(-vmtruncate(vn_to_inode(XFS_ITOV(ip)), (off)))
+	(-vmtruncate(VFS_I(ip), (off)))
 
 
 /* Move the kernel do_div definition off to one side */

fs/xfs/linux-2.6/xfs_lrw.c

@@ -137,7 +137,7 @@ xfs_iozero(
 	struct address_space	*mapping;
 	int			status;
 
-	mapping = ip->i_vnode->i_mapping;
+	mapping = VFS_I(ip)->i_mapping;
 	do {
 		unsigned offset, bytes;
 		void *fsdata;
@@ -674,9 +674,7 @@ start:
 	 */
 	if (likely(!(ioflags & IO_INVIS) &&
 		   !mnt_want_write(file->f_path.mnt))) {
-		file_update_time(file);
-		xfs_ichgtime_fast(xip, inode,
-				  XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
+		xfs_ichgtime(xip, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
 		mnt_drop_write(file->f_path.mnt);
 	}
 

fs/xfs/linux-2.6/xfs_super.c

@@ -581,118 +581,6 @@ xfs_max_file_offset(
 	return (((__uint64_t)pagefactor) << bitshift) - 1;
 }
 
-STATIC_INLINE void
-xfs_set_inodeops(
-	struct inode		*inode)
-{
-	switch (inode->i_mode & S_IFMT) {
-	case S_IFREG:
-		inode->i_op = &xfs_inode_operations;
-		inode->i_fop = &xfs_file_operations;
-		inode->i_mapping->a_ops = &xfs_address_space_operations;
-		break;
-	case S_IFDIR:
-		if (xfs_sb_version_hasasciici(&XFS_M(inode->i_sb)->m_sb))
-			inode->i_op = &xfs_dir_ci_inode_operations;
-		else
-			inode->i_op = &xfs_dir_inode_operations;
-		inode->i_fop = &xfs_dir_file_operations;
-		break;
-	case S_IFLNK:
-		inode->i_op = &xfs_symlink_inode_operations;
-		if (!(XFS_I(inode)->i_df.if_flags & XFS_IFINLINE))
-			inode->i_mapping->a_ops = &xfs_address_space_operations;
-		break;
-	default:
-		inode->i_op = &xfs_inode_operations;
-		init_special_inode(inode, inode->i_mode, inode->i_rdev);
-		break;
-	}
-}
-
-STATIC_INLINE void
-xfs_revalidate_inode(
-	xfs_mount_t		*mp,
-	bhv_vnode_t		*vp,
-	xfs_inode_t		*ip)
-{
-	struct inode		*inode = vn_to_inode(vp);
-
-	inode->i_mode	= ip->i_d.di_mode;
-	inode->i_nlink	= ip->i_d.di_nlink;
-	inode->i_uid	= ip->i_d.di_uid;
-	inode->i_gid	= ip->i_d.di_gid;
-
-	switch (inode->i_mode & S_IFMT) {
-	case S_IFBLK:
-	case S_IFCHR:
-		inode->i_rdev =
-			MKDEV(sysv_major(ip->i_df.if_u2.if_rdev) & 0x1ff,
-			      sysv_minor(ip->i_df.if_u2.if_rdev));
-		break;
-	default:
-		inode->i_rdev = 0;
-		break;
-	}
-
-	inode->i_generation = ip->i_d.di_gen;
-	i_size_write(inode, ip->i_d.di_size);
-	inode->i_atime.tv_sec	= ip->i_d.di_atime.t_sec;
-	inode->i_atime.tv_nsec	= ip->i_d.di_atime.t_nsec;
-	inode->i_mtime.tv_sec	= ip->i_d.di_mtime.t_sec;
-	inode->i_mtime.tv_nsec	= ip->i_d.di_mtime.t_nsec;
-	inode->i_ctime.tv_sec	= ip->i_d.di_ctime.t_sec;
-	inode->i_ctime.tv_nsec	= ip->i_d.di_ctime.t_nsec;
-	if (ip->i_d.di_flags & XFS_DIFLAG_IMMUTABLE)
-		inode->i_flags |= S_IMMUTABLE;
-	else
-		inode->i_flags &= ~S_IMMUTABLE;
-	if (ip->i_d.di_flags & XFS_DIFLAG_APPEND)
-		inode->i_flags |= S_APPEND;
-	else
-		inode->i_flags &= ~S_APPEND;
-	if (ip->i_d.di_flags & XFS_DIFLAG_SYNC)
-		inode->i_flags |= S_SYNC;
-	else
-		inode->i_flags &= ~S_SYNC;
-	if (ip->i_d.di_flags & XFS_DIFLAG_NOATIME)
-		inode->i_flags |= S_NOATIME;
-	else
-		inode->i_flags &= ~S_NOATIME;
-	xfs_iflags_clear(ip, XFS_IMODIFIED);
-}
-
-void
-xfs_initialize_vnode(
-	struct xfs_mount	*mp,
-	bhv_vnode_t		*vp,
-	struct xfs_inode	*ip)
-{
-	struct inode		*inode = vn_to_inode(vp);
-
-	if (!ip->i_vnode) {
-		ip->i_vnode = vp;
-		inode->i_private = ip;
-	}
-
-	/*
-	 * We need to set the ops vectors, and unlock the inode, but if
-	 * we have been called during the new inode create process, it is
-	 * too early to fill in the Linux inode.  We will get called a
-	 * second time once the inode is properly set up, and then we can
-	 * finish our work.
-	 */
-	if (ip->i_d.di_mode != 0 && (inode->i_state & I_NEW)) {
-		xfs_revalidate_inode(mp, vp, ip);
-		xfs_set_inodeops(inode);
-
-		xfs_iflags_clear(ip, XFS_INEW);
-		barrier();
-
-		unlock_new_inode(inode);
-	}
-}
-
 int
 xfs_blkdev_get(
 	xfs_mount_t		*mp,
@@ -982,26 +870,21 @@ STATIC struct inode *
 xfs_fs_alloc_inode(
 	struct super_block	*sb)
 {
-	bhv_vnode_t		*vp;
-
-	vp = kmem_zone_alloc(xfs_vnode_zone, KM_SLEEP);
-	if (unlikely(!vp))
-		return NULL;
-	return vn_to_inode(vp);
+	return kmem_zone_alloc(xfs_vnode_zone, KM_SLEEP);
 }
 
 STATIC void
 xfs_fs_destroy_inode(
 	struct inode		*inode)
 {
-	kmem_zone_free(xfs_vnode_zone, vn_from_inode(inode));
+	kmem_zone_free(xfs_vnode_zone, inode);
 }
 
 STATIC void
 xfs_fs_inode_init_once(
 	void			*vnode)
 {
-	inode_init_once(vn_to_inode((bhv_vnode_t *)vnode));
+	inode_init_once((struct inode *)vnode);
 }
 
 /*
@@ -1106,7 +989,7 @@ void
 xfs_flush_inode(
 	xfs_inode_t	*ip)
 {
-	struct inode	*inode = ip->i_vnode;
+	struct inode	*inode = VFS_I(ip);
 
 	igrab(inode);
 	xfs_syncd_queue_work(ip->i_mount, inode, xfs_flush_inode_work);
@@ -1131,7 +1014,7 @@ void
 xfs_flush_device(
 	xfs_inode_t	*ip)
 {
-	struct inode	*inode = vn_to_inode(XFS_ITOV(ip));
+	struct inode	*inode = VFS_I(ip);
 
 	igrab(inode);
 	xfs_syncd_queue_work(ip->i_mount, inode, xfs_flush_device_work);
@@ -1200,6 +1083,15 @@ xfssyncd(
 	return 0;
 }
 
+STATIC void
+xfs_free_fsname(
+	struct xfs_mount	*mp)
+{
+	kfree(mp->m_fsname);
+	kfree(mp->m_rtname);
+	kfree(mp->m_logname);
+}
+
 STATIC void
 xfs_fs_put_super(
 	struct super_block	*sb)
@@ -1239,8 +1131,6 @@ xfs_fs_put_super(
 	error = xfs_unmount_flush(mp, 0);
 	WARN_ON(error);
 
-	IRELE(rip);
-
 	/*
 	 * If we're forcing a shutdown, typically because of a media error,
 	 * we want to make sure we invalidate dirty pages that belong to
@@ -1257,10 +1147,12 @@ xfs_fs_put_super(
 	}
 
 	xfs_unmountfs(mp);
+	xfs_freesb(mp);
 	xfs_icsb_destroy_counters(mp);
 	xfs_close_devices(mp);
 	xfs_qmops_put(mp);
 	xfs_dmops_put(mp);
+	xfs_free_fsname(mp);
 	kfree(mp);
 }
 
@@ -1517,6 +1409,8 @@ xfs_start_flags(
 	struct xfs_mount_args	*ap,
 	struct xfs_mount	*mp)
 {
+	int			error;
+
 	/* Values are in BBs */
 	if ((ap->flags & XFSMNT_NOALIGN) != XFSMNT_NOALIGN) {
 		/*
@@ -1549,17 +1443,27 @@ xfs_start_flags(
 			ap->logbufsize);
 		return XFS_ERROR(EINVAL);
 	}
+
+	error = ENOMEM;
+
 	mp->m_logbsize = ap->logbufsize;
 	mp->m_fsname_len = strlen(ap->fsname) + 1;
-	mp->m_fsname = kmem_alloc(mp->m_fsname_len, KM_SLEEP);
-	strcpy(mp->m_fsname, ap->fsname);
+
+	mp->m_fsname = kstrdup(ap->fsname, GFP_KERNEL);
+	if (!mp->m_fsname)
+		goto out;
+
 	if (ap->rtname[0]) {
-		mp->m_rtname = kmem_alloc(strlen(ap->rtname) + 1, KM_SLEEP);
-		strcpy(mp->m_rtname, ap->rtname);
+		mp->m_rtname = kstrdup(ap->rtname, GFP_KERNEL);
+		if (!mp->m_rtname)
+			goto out_free_fsname;
+
 	}
+
 	if (ap->logname[0]) {
-		mp->m_logname = kmem_alloc(strlen(ap->logname) + 1, KM_SLEEP);
-		strcpy(mp->m_logname, ap->logname);
+		mp->m_logname = kstrdup(ap->logname, GFP_KERNEL);
+		if (!mp->m_logname)
+			goto out_free_rtname;
 	}
 
 	if (ap->flags & XFSMNT_WSYNC)
@@ -1632,6 +1536,14 @@ xfs_start_flags(
 	if (ap->flags & XFSMNT_DMAPI)
 		mp->m_flags |= XFS_MOUNT_DMAPI;
 	return 0;
+
+
+ out_free_rtname:
+	kfree(mp->m_rtname);
+ out_free_fsname:
+	kfree(mp->m_fsname);
+ out:
+	return error;
 }
 
 /*
@@ -1792,10 +1704,10 @@ xfs_fs_fill_super(
 	 */
 	error = xfs_start_flags(args, mp);
 	if (error)
-		goto out_destroy_counters;
+		goto out_free_fsname;
 	error = xfs_readsb(mp, flags);
 	if (error)
-		goto out_destroy_counters;
+		goto out_free_fsname;
 	error = xfs_finish_flags(args, mp);
 	if (error)
 		goto out_free_sb;
@@ -1811,7 +1723,7 @@ xfs_fs_fill_super(
 	if (error)
 		goto out_free_sb;
 
-	error = xfs_mountfs(mp, flags);
+	error = xfs_mountfs(mp);
 	if (error)
 		goto out_filestream_unmount;
 
@@ -1825,7 +1737,7 @@ xfs_fs_fill_super(
 	sb->s_time_gran = 1;
 	set_posix_acl_flag(sb);
 
-	root = igrab(mp->m_rootip->i_vnode);
+	root = igrab(VFS_I(mp->m_rootip));
 	if (!root) {
 		error = ENOENT;
 		goto fail_unmount;
@@ -1857,7 +1769,8 @@ xfs_fs_fill_super(
 	xfs_filestream_unmount(mp);
  out_free_sb:
 	xfs_freesb(mp);
- out_destroy_counters:
+ out_free_fsname:
+	xfs_free_fsname(mp);
 	xfs_icsb_destroy_counters(mp);
 	xfs_close_devices(mp);
  out_put_qmops:
@@ -1890,10 +1803,8 @@ xfs_fs_fill_super(
 	error = xfs_unmount_flush(mp, 0);
 	WARN_ON(error);
 
-	IRELE(mp->m_rootip);
-
 	xfs_unmountfs(mp);
-	goto out_destroy_counters;
+	goto out_free_sb;
 }
 
 STATIC int
@@ -2014,7 +1925,7 @@ xfs_free_trace_bufs(void)
 STATIC int __init
 xfs_init_zones(void)
 {
-	xfs_vnode_zone = kmem_zone_init_flags(sizeof(bhv_vnode_t), "xfs_vnode",
+	xfs_vnode_zone = kmem_zone_init_flags(sizeof(struct inode), "xfs_vnode",
 					KM_ZONE_HWALIGN | KM_ZONE_RECLAIM |
 					KM_ZONE_SPREAD,
 					xfs_fs_inode_init_once);

fs/xfs/linux-2.6/xfs_super.h

@@ -101,9 +101,6 @@ struct block_device;
 
 extern __uint64_t xfs_max_file_offset(unsigned int);
 
-extern void xfs_initialize_vnode(struct xfs_mount *mp, bhv_vnode_t *vp,
-		struct xfs_inode *ip);
-
 extern void xfs_flush_inode(struct xfs_inode *);
 extern void xfs_flush_device(struct xfs_inode *);
 

fs/xfs/linux-2.6/xfs_vnode.c

@@ -33,7 +33,7 @@
 
 
 /*
- * Dedicated vnode inactive/reclaim sync semaphores.
+ * Dedicated vnode inactive/reclaim sync wait queues.
  * Prime number of hash buckets since address is used as the key.
  */
 #define NVSYNC                  37
@@ -82,24 +82,6 @@ vn_ioerror(
 		xfs_do_force_shutdown(ip->i_mount, SHUTDOWN_DEVICE_REQ, f, l);
 }
 
-
-/*
- * Add a reference to a referenced vnode.
- */
-bhv_vnode_t *
-vn_hold(
-	bhv_vnode_t	*vp)
-{
-	struct inode	*inode;
-
-	XFS_STATS_INC(vn_hold);
-
-	inode = igrab(vn_to_inode(vp));
-	ASSERT(inode);
-
-	return vp;
-}
-
 #ifdef	XFS_INODE_TRACE
 
 /*
@@ -108,7 +90,7 @@ vn_hold(
  */
 static inline int xfs_icount(struct xfs_inode *ip)
 {
-	bhv_vnode_t *vp = XFS_ITOV_NULL(ip);
+	struct inode *vp = VFS_I(ip);
 
 	if (vp)
 		return vn_count(vp);

fs/xfs/linux-2.6/xfs_vnode.h

@@ -22,20 +22,6 @@ struct file;
 struct xfs_iomap;
 struct attrlist_cursor_kern;
 
-typedef struct inode	bhv_vnode_t;
-
-/*
- * Vnode to Linux inode mapping.
- */
-static inline bhv_vnode_t *vn_from_inode(struct inode *inode)
-{
-	return inode;
-}
-static inline struct inode *vn_to_inode(bhv_vnode_t *vnode)
-{
-	return vnode;
-}
-
 /*
  * Return values for xfs_inactive.  A return value of
  * VN_INACTIVE_NOCACHE implies that the file system behavior
@@ -76,57 +62,52 @@ extern void	vn_iowait(struct xfs_inode *ip);
 extern void	vn_iowake(struct xfs_inode *ip);
 extern void	vn_ioerror(struct xfs_inode *ip, int error, char *f, int l);
 
-static inline int vn_count(bhv_vnode_t *vp)
+static inline int vn_count(struct inode *vp)
 {
-	return atomic_read(&vn_to_inode(vp)->i_count);
+	return atomic_read(&vp->i_count);
 }
 
-/*
- * Vnode reference counting functions (and macros for compatibility).
- */
-extern bhv_vnode_t	*vn_hold(bhv_vnode_t *);
+#define IHOLD(ip) \
+do { \
+	ASSERT(atomic_read(&VFS_I(ip)->i_count) > 0) ; \
+	atomic_inc(&(VFS_I(ip)->i_count)); \
+	xfs_itrace_hold((ip), __FILE__, __LINE__, (inst_t *)__return_address); \
+} while (0)
 
-#if defined(XFS_INODE_TRACE)
-#define VN_HOLD(vp)		\
-	((void)vn_hold(vp),	\
-	  xfs_itrace_hold(xfs_vtoi(vp), __FILE__, __LINE__, (inst_t *)__return_address))
-#define VN_RELE(vp)		\
-	  (xfs_itrace_rele(xfs_vtoi(vp), __FILE__, __LINE__, (inst_t *)__return_address), \
-	   iput(vn_to_inode(vp)))
-#else
-#define VN_HOLD(vp)		((void)vn_hold(vp))
-#define VN_RELE(vp)		(iput(vn_to_inode(vp)))
-#endif
+#define IRELE(ip) \
+do { \
+	xfs_itrace_rele((ip), __FILE__, __LINE__, (inst_t *)__return_address); \
+	iput(VFS_I(ip)); \
+} while (0)
 
-static inline bhv_vnode_t *vn_grab(bhv_vnode_t *vp)
+static inline struct inode *vn_grab(struct inode *vp)
 {
-	struct inode *inode = igrab(vn_to_inode(vp));
-	return inode ? vn_from_inode(inode) : NULL;
+	return igrab(vp);
 }
 
 /*
  * Dealing with bad inodes
  */
-static inline int VN_BAD(bhv_vnode_t *vp)
+static inline int VN_BAD(struct inode *vp)
 {
-	return is_bad_inode(vn_to_inode(vp));
+	return is_bad_inode(vp);
 }
 
 /*
  * Extracting atime values in various formats
  */
-static inline void vn_atime_to_bstime(bhv_vnode_t *vp, xfs_bstime_t *bs_atime)
+static inline void vn_atime_to_bstime(struct inode *vp, xfs_bstime_t *bs_atime)
 {
 	bs_atime->tv_sec = vp->i_atime.tv_sec;
 	bs_atime->tv_nsec = vp->i_atime.tv_nsec;
 }
 
-static inline void vn_atime_to_timespec(bhv_vnode_t *vp, struct timespec *ts)
+static inline void vn_atime_to_timespec(struct inode *vp, struct timespec *ts)
 {
 	*ts = vp->i_atime;
 }
 
-static inline void vn_atime_to_time_t(bhv_vnode_t *vp, time_t *tt)
+static inline void vn_atime_to_time_t(struct inode *vp, time_t *tt)
 {
 	*tt = vp->i_atime.tv_sec;
 }
@@ -134,9 +115,9 @@ static inline void vn_atime_to_time_t(bhv_vnode_t *vp, time_t *tt)
 /*
  * Some useful predicates.
  */
-#define VN_MAPPED(vp)	mapping_mapped(vn_to_inode(vp)->i_mapping)
-#define VN_CACHED(vp)	(vn_to_inode(vp)->i_mapping->nrpages)
-#define VN_DIRTY(vp)	mapping_tagged(vn_to_inode(vp)->i_mapping, \
+#define VN_MAPPED(vp)	mapping_mapped(vp->i_mapping)
+#define VN_CACHED(vp)	(vp->i_mapping->nrpages)
+#define VN_DIRTY(vp)	mapping_tagged(vp->i_mapping, \
 					PAGECACHE_TAG_DIRTY)
 
 

fs/xfs/quota/xfs_dquot.c

@@ -101,11 +101,18 @@ xfs_qm_dqinit(
 	if (brandnewdquot) {
 		dqp->dq_flnext = dqp->dq_flprev = dqp;
 		mutex_init(&dqp->q_qlock);
-		initnsema(&dqp->q_flock, 1, "fdq");
 		sv_init(&dqp->q_pinwait, SV_DEFAULT, "pdq");
 
+		/*
+		 * Because we want to use a counting completion, complete
+		 * the flush completion once to allow a single access to
+		 * the flush completion without blocking.
+		 */
+		init_completion(&dqp->q_flush);
+		complete(&dqp->q_flush);
+
 #ifdef XFS_DQUOT_TRACE
-		dqp->q_trace = ktrace_alloc(DQUOT_TRACE_SIZE, KM_SLEEP);
+		dqp->q_trace = ktrace_alloc(DQUOT_TRACE_SIZE, KM_NOFS);
 		xfs_dqtrace_entry(dqp, "DQINIT");
 #endif
 	} else {
@@ -150,7 +157,6 @@ xfs_qm_dqdestroy(
 	ASSERT(! XFS_DQ_IS_ON_FREELIST(dqp));
 
 	mutex_destroy(&dqp->q_qlock);
-	freesema(&dqp->q_flock);
 	sv_destroy(&dqp->q_pinwait);
 
 #ifdef XFS_DQUOT_TRACE
@@ -431,7 +437,7 @@ xfs_qm_dqalloc(
 	 * when it unlocks the inode. Since we want to keep the quota
 	 * inode around, we bump the vnode ref count now.
 	 */
-	VN_HOLD(XFS_ITOV(quotip));
+	IHOLD(quotip);
 
 	xfs_trans_ijoin(tp, quotip, XFS_ILOCK_EXCL);
 	nmaps = 1;
@@ -1211,7 +1217,7 @@ xfs_qm_dqflush(
 	int			error;
 
 	ASSERT(XFS_DQ_IS_LOCKED(dqp));
-	ASSERT(XFS_DQ_IS_FLUSH_LOCKED(dqp));
+	ASSERT(!completion_done(&dqp->q_flush));
 	xfs_dqtrace_entry(dqp, "DQFLUSH");
 
 	/*
@@ -1348,34 +1354,18 @@ xfs_qm_dqflush_done(
 	xfs_dqfunlock(dqp);
 }
 
-
-int
-xfs_qm_dqflock_nowait(
-	xfs_dquot_t *dqp)
-{
-	int locked;
-
-	locked = cpsema(&((dqp)->q_flock));
-
-	/* XXX ifdef these out */
-	if (locked)
-		(dqp)->dq_flags |= XFS_DQ_FLOCKED;
-	return (locked);
-}
-
-
 int
 xfs_qm_dqlock_nowait(
 	xfs_dquot_t *dqp)
 {
-	return (mutex_trylock(&((dqp)->q_qlock)));
+	return mutex_trylock(&dqp->q_qlock);
 }
 
 void
 xfs_dqlock(
 	xfs_dquot_t *dqp)
 {
-	mutex_lock(&(dqp->q_qlock));
+	mutex_lock(&dqp->q_qlock);
 }
 
 void
@@ -1468,7 +1458,7 @@ xfs_qm_dqpurge(
 	 * if we're turning off quotas. Basically, we need this flush
 	 * lock, and are willing to block on it.
 	 */
-	if (! xfs_qm_dqflock_nowait(dqp)) {
+	if (!xfs_dqflock_nowait(dqp)) {
 		/*
 		 * Block on the flush lock after nudging dquot buffer,
 		 * if it is incore.

fs/xfs/quota/xfs_dquot.h

@@ -82,7 +82,7 @@ typedef struct xfs_dquot {
 	xfs_qcnt_t	 q_res_icount;	/* total inos allocd+reserved */
 	xfs_qcnt_t	 q_res_rtbcount;/* total realtime blks used+reserved */
 	mutex_t		 q_qlock;	/* quota lock */
-	sema_t		 q_flock;	/* flush lock */
+	struct completion q_flush;	/* flush completion queue */
 	uint		 q_pincount;	/* pin count for this dquot */
 	sv_t		 q_pinwait;	/* sync var for pinning */
 #ifdef XFS_DQUOT_TRACE
@@ -113,17 +113,25 @@ XFS_DQ_IS_LOCKED(xfs_dquot_t *dqp)
 
 
 /*
- * The following three routines simply manage the q_flock
- * semaphore embedded in the dquot.  This semaphore synchronizes
- * processes attempting to flush the in-core dquot back to disk.
+ * Manage the q_flush completion queue embedded in the dquot.  This completion
+ * queue synchronizes processes attempting to flush the in-core dquot back to
+ * disk.
  */
-#define xfs_dqflock(dqp)	 { psema(&((dqp)->q_flock), PINOD | PRECALC);\
-				   (dqp)->dq_flags |= XFS_DQ_FLOCKED; }
-#define xfs_dqfunlock(dqp)	 { ASSERT(issemalocked(&((dqp)->q_flock))); \
-				   vsema(&((dqp)->q_flock)); \
-				   (dqp)->dq_flags &= ~(XFS_DQ_FLOCKED); }
+static inline void xfs_dqflock(xfs_dquot_t *dqp)
+{
+	wait_for_completion(&dqp->q_flush);
+}
+
+static inline int xfs_dqflock_nowait(xfs_dquot_t *dqp)
+{
+	return try_wait_for_completion(&dqp->q_flush);
+}
+
+static inline void xfs_dqfunlock(xfs_dquot_t *dqp)
+{
+	complete(&dqp->q_flush);
+}
 
-#define XFS_DQ_IS_FLUSH_LOCKED(dqp) (issemalocked(&((dqp)->q_flock)))
 #define XFS_DQ_IS_ON_FREELIST(dqp)  ((dqp)->dq_flnext != (dqp))
 #define XFS_DQ_IS_DIRTY(dqp)	((dqp)->dq_flags & XFS_DQ_DIRTY)
 #define XFS_QM_ISUDQ(dqp)	((dqp)->dq_flags & XFS_DQ_USER)
@@ -167,7 +175,6 @@ extern int		xfs_qm_dqflush(xfs_dquot_t *, uint);
 extern int		xfs_qm_dqpurge(xfs_dquot_t *);
 extern void		xfs_qm_dqunpin_wait(xfs_dquot_t *);
 extern int		xfs_qm_dqlock_nowait(xfs_dquot_t *);
-extern int		xfs_qm_dqflock_nowait(xfs_dquot_t *);
 extern void		xfs_qm_dqflock_pushbuf_wait(xfs_dquot_t *dqp);
 extern void		xfs_qm_adjust_dqtimers(xfs_mount_t *,
 					xfs_disk_dquot_t *);

fs/xfs/quota/xfs_dquot_item.c

@@ -151,7 +151,7 @@ xfs_qm_dquot_logitem_push(
 	dqp = logitem->qli_dquot;
 
 	ASSERT(XFS_DQ_IS_LOCKED(dqp));
-	ASSERT(XFS_DQ_IS_FLUSH_LOCKED(dqp));
+	ASSERT(!completion_done(&dqp->q_flush));
 
 	/*
 	 * Since we were able to lock the dquot's flush lock and
@@ -245,7 +245,7 @@ xfs_qm_dquot_logitem_pushbuf(
 	 * inode flush completed and the inode was taken off the AIL.
 	 * So, just get out.
 	 */
-	if (!issemalocked(&(dqp->q_flock))  ||
+	if (completion_done(&dqp->q_flush)  ||
 	    ((qip->qli_item.li_flags & XFS_LI_IN_AIL) == 0)) {
 		qip->qli_pushbuf_flag = 0;
 		xfs_dqunlock(dqp);
@@ -258,7 +258,7 @@ xfs_qm_dquot_logitem_pushbuf(
 	if (bp != NULL) {
 		if (XFS_BUF_ISDELAYWRITE(bp)) {
 			dopush = ((qip->qli_item.li_flags & XFS_LI_IN_AIL) &&
-				  issemalocked(&(dqp->q_flock)));
+				  !completion_done(&dqp->q_flush));
 			qip->qli_pushbuf_flag = 0;
 			xfs_dqunlock(dqp);
 
@@ -317,7 +317,7 @@ xfs_qm_dquot_logitem_trylock(
 		return (XFS_ITEM_LOCKED);
 
 	retval = XFS_ITEM_SUCCESS;
-	if (! xfs_qm_dqflock_nowait(dqp)) {
+	if (!xfs_dqflock_nowait(dqp)) {
 		/*
 		 * The dquot is already being flushed.	It may have been
 		 * flushed delayed write, however, and we don't want to

fs/xfs/quota/xfs_qm.c

@@ -310,8 +310,7 @@ xfs_qm_unmount_quotadestroy(
  */
 void
 xfs_qm_mount_quotas(
-	xfs_mount_t	*mp,
-	int		mfsi_flags)
+	xfs_mount_t	*mp)
 {
 	int		error = 0;
 	uint		sbf;
@@ -346,8 +345,7 @@ xfs_qm_mount_quotas(
 	/*
 	 * If any of the quotas are not consistent, do a quotacheck.
 	 */
-	if (XFS_QM_NEED_QUOTACHECK(mp) &&
-	    !(mfsi_flags & XFS_MFSI_NO_QUOTACHECK)) {
+	if (XFS_QM_NEED_QUOTACHECK(mp)) {
 		error = xfs_qm_quotacheck(mp);
 		if (error) {
 			/* Quotacheck failed and disabled quotas. */
@@ -484,7 +482,7 @@ again:
 		xfs_dqtrace_entry(dqp, "FLUSHALL: DQDIRTY");
 		/* XXX a sentinel would be better */
 		recl = XFS_QI_MPLRECLAIMS(mp);
-		if (! xfs_qm_dqflock_nowait(dqp)) {
+		if (!xfs_dqflock_nowait(dqp)) {
 			/*
 			 * If we can't grab the flush lock then check
 			 * to see if the dquot has been flushed delayed
@@ -1062,7 +1060,7 @@ xfs_qm_sync(
 
 		/* XXX a sentinel would be better */
 		recl = XFS_QI_MPLRECLAIMS(mp);
-		if (! xfs_qm_dqflock_nowait(dqp)) {
+		if (!xfs_dqflock_nowait(dqp)) {
 			if (nowait) {
 				xfs_dqunlock(dqp);
 				continue;
@@ -2079,7 +2077,7 @@ xfs_qm_shake_freelist(
 		 * Try to grab the flush lock. If this dquot is in the process of
 		 * getting flushed to disk, we don't want to reclaim it.
 		 */
-		if (! xfs_qm_dqflock_nowait(dqp)) {
+		if (!xfs_dqflock_nowait(dqp)) {
 			xfs_dqunlock(dqp);
 			dqp = dqp->dq_flnext;
 			continue;
@@ -2257,7 +2255,7 @@ xfs_qm_dqreclaim_one(void)
 		 * Try to grab the flush lock. If this dquot is in the process of
 		 * getting flushed to disk, we don't want to reclaim it.
 		 */
-		if (! xfs_qm_dqflock_nowait(dqp)) {
+		if (!xfs_dqflock_nowait(dqp)) {
 			xfs_dqunlock(dqp);
 			continue;
 		}

fs/xfs/quota/xfs_qm.h

@@ -165,7 +165,7 @@ typedef struct xfs_dquot_acct {
 #define XFS_QM_RELE(xqm)	((xqm)->qm_nrefs--)
 
 extern void		xfs_qm_destroy_quotainfo(xfs_mount_t *);
-extern void		xfs_qm_mount_quotas(xfs_mount_t *, int);
+extern void		xfs_qm_mount_quotas(xfs_mount_t *);
 extern int		xfs_qm_quotacheck(xfs_mount_t *);
 extern void		xfs_qm_unmount_quotadestroy(xfs_mount_t *);
 extern int		xfs_qm_unmount_quotas(xfs_mount_t *);

fs/xfs/quota/xfs_qm_bhv.c

@@ -162,7 +162,7 @@ xfs_qm_newmount(
 			 * mounting, and get on with the boring life
 			 * without disk quotas.
 			 */
-			xfs_qm_mount_quotas(mp, 0);
+			xfs_qm_mount_quotas(mp);
 		} else {
 			/*
 			 * Clear the quota flags, but remember them. This
@@ -184,13 +184,12 @@ STATIC int
 xfs_qm_endmount(
 	xfs_mount_t	*mp,
 	uint		needquotamount,
-	uint		quotaflags,
-	int		mfsi_flags)
+	uint		quotaflags)
 {
 	if (needquotamount) {
 		ASSERT(mp->m_qflags == 0);
 		mp->m_qflags = quotaflags;
-		xfs_qm_mount_quotas(mp, mfsi_flags);
+		xfs_qm_mount_quotas(mp);
 	}
 
 #if defined(DEBUG) && defined(XFS_LOUD_RECOVERY)

fs/xfs/quota/xfs_qm_syscalls.c

@@ -1034,7 +1034,7 @@ xfs_qm_dqrele_all_inodes(
 {
 	xfs_inode_t	*ip, *topino;
 	uint		ireclaims;
-	bhv_vnode_t	*vp;
+	struct inode	*vp;
 	boolean_t	vnode_refd;
 
 	ASSERT(mp->m_quotainfo);
@@ -1059,7 +1059,7 @@ again:
 			ip = ip->i_mnext;
 			continue;
 		}
-		vp = XFS_ITOV_NULL(ip);
+		vp = VFS_I(ip);
 		if (!vp) {
 			ASSERT(ip->i_udquot == NULL);
 			ASSERT(ip->i_gdquot == NULL);

fs/xfs/xfs_acl.c

@@ -37,15 +37,15 @@
 #include <linux/capability.h>
 #include <linux/posix_acl_xattr.h>
 
-STATIC int	xfs_acl_setmode(bhv_vnode_t *, xfs_acl_t *, int *);
+STATIC int	xfs_acl_setmode(struct inode *, xfs_acl_t *, int *);
 STATIC void     xfs_acl_filter_mode(mode_t, xfs_acl_t *);
 STATIC void	xfs_acl_get_endian(xfs_acl_t *);
 STATIC int	xfs_acl_access(uid_t, gid_t, xfs_acl_t *, mode_t, cred_t *);
 STATIC int	xfs_acl_invalid(xfs_acl_t *);
 STATIC void	xfs_acl_sync_mode(mode_t, xfs_acl_t *);
-STATIC void	xfs_acl_get_attr(bhv_vnode_t *, xfs_acl_t *, int, int, int *);
-STATIC void	xfs_acl_set_attr(bhv_vnode_t *, xfs_acl_t *, int, int *);
-STATIC int	xfs_acl_allow_set(bhv_vnode_t *, int);
+STATIC void	xfs_acl_get_attr(struct inode *, xfs_acl_t *, int, int, int *);
+STATIC void	xfs_acl_set_attr(struct inode *, xfs_acl_t *, int, int *);
+STATIC int	xfs_acl_allow_set(struct inode *, int);
 
 kmem_zone_t *xfs_acl_zone;
 
@@ -55,7 +55,7 @@ kmem_zone_t *xfs_acl_zone;
  */
 int
 xfs_acl_vhasacl_access(
-	bhv_vnode_t	*vp)
+	struct inode	*vp)
 {
 	int		error;
 
@@ -68,7 +68,7 @@ xfs_acl_vhasacl_access(
  */
 int
 xfs_acl_vhasacl_default(
-	bhv_vnode_t	*vp)
+	struct inode	*vp)
 {
 	int		error;
 
@@ -207,7 +207,7 @@ posix_acl_xfs_to_xattr(
 
 int
 xfs_acl_vget(
-	bhv_vnode_t	*vp,
+	struct inode	*vp,
 	void		*acl,
 	size_t		size,
 	int		kind)
@@ -217,7 +217,6 @@ xfs_acl_vget(
 	posix_acl_xattr_header	*ext_acl = acl;
 	int			flags = 0;
 
-	VN_HOLD(vp);
 	if(size) {
 		if (!(_ACL_ALLOC(xfs_acl))) {
 			error = ENOMEM;
@@ -239,11 +238,10 @@ xfs_acl_vget(
 			goto out;
 		}
 		if (kind == _ACL_TYPE_ACCESS)
-			xfs_acl_sync_mode(xfs_vtoi(vp)->i_d.di_mode, xfs_acl);
+			xfs_acl_sync_mode(XFS_I(vp)->i_d.di_mode, xfs_acl);
 		error = -posix_acl_xfs_to_xattr(xfs_acl, ext_acl, size);
 	}
 out:
-	VN_RELE(vp);
 	if(xfs_acl)
 		_ACL_FREE(xfs_acl);
 	return -error;
@@ -251,28 +249,26 @@ out:
 
 int
 xfs_acl_vremove(
-	bhv_vnode_t	*vp,
+	struct inode	*vp,
 	int		kind)
 {
 	int		error;
 
-	VN_HOLD(vp);
 	error = xfs_acl_allow_set(vp, kind);
 	if (!error) {
-		error = xfs_attr_remove(xfs_vtoi(vp),
+		error = xfs_attr_remove(XFS_I(vp),
 						kind == _ACL_TYPE_DEFAULT?
 						SGI_ACL_DEFAULT: SGI_ACL_FILE,
 						ATTR_ROOT);
 		if (error == ENOATTR)
 			error = 0;	/* 'scool */
 	}
-	VN_RELE(vp);
 	return -error;
 }
 
 int
 xfs_acl_vset(
-	bhv_vnode_t		*vp,
+	struct inode		*vp,
 	void			*acl,
 	size_t			size,
 	int			kind)
@@ -298,7 +294,6 @@ xfs_acl_vset(
 		return 0;
 	}
 
-	VN_HOLD(vp);
 	error = xfs_acl_allow_set(vp, kind);
 
 	/* Incoming ACL exists, set file mode based on its value */
@@ -321,7 +316,6 @@ xfs_acl_vset(
 	}
 
 out:
-	VN_RELE(vp);
 	_ACL_FREE(xfs_acl);
 	return -error;
 }
@@ -363,7 +357,7 @@ xfs_acl_iaccess(
 
 STATIC int
 xfs_acl_allow_set(
-	bhv_vnode_t	*vp,
+	struct inode	*vp,
 	int		kind)
 {
 	if (vp->i_flags & (S_IMMUTABLE|S_APPEND))
@@ -372,7 +366,7 @@ xfs_acl_allow_set(
 		return ENOTDIR;
 	if (vp->i_sb->s_flags & MS_RDONLY)
 		return EROFS;
-	if (xfs_vtoi(vp)->i_d.di_uid != current->fsuid && !capable(CAP_FOWNER))
+	if (XFS_I(vp)->i_d.di_uid != current->fsuid && !capable(CAP_FOWNER))
 		return EPERM;
 	return 0;
 }
@@ -566,7 +560,7 @@ xfs_acl_get_endian(
  */
 STATIC void
 xfs_acl_get_attr(
-	bhv_vnode_t	*vp,
+	struct inode	*vp,
 	xfs_acl_t	*aclp,
 	int		kind,
 	int		flags,
@@ -576,7 +570,7 @@ xfs_acl_get_attr(
 
 	ASSERT((flags & ATTR_KERNOVAL) ? (aclp == NULL) : 1);
 	flags |= ATTR_ROOT;
-	*error = xfs_attr_get(xfs_vtoi(vp),
+	*error = xfs_attr_get(XFS_I(vp),
 					kind == _ACL_TYPE_ACCESS ?
 					SGI_ACL_FILE : SGI_ACL_DEFAULT,
 					(char *)aclp, &len, flags);
@@ -590,7 +584,7 @@ xfs_acl_get_attr(
  */
 STATIC void
 xfs_acl_set_attr(
-	bhv_vnode_t	*vp,
+	struct inode	*vp,
 	xfs_acl_t	*aclp,
 	int		kind,
 	int		*error)
@@ -615,7 +609,7 @@ xfs_acl_set_attr(
 		INT_SET(newace->ae_perm, ARCH_CONVERT, ace->ae_perm);
 	}
 	INT_SET(newacl->acl_cnt, ARCH_CONVERT, aclp->acl_cnt);
-	*error = xfs_attr_set(xfs_vtoi(vp),
+	*error = xfs_attr_set(XFS_I(vp),
 				kind == _ACL_TYPE_ACCESS ?
 				SGI_ACL_FILE: SGI_ACL_DEFAULT,
 				(char *)newacl, len, ATTR_ROOT);
@@ -624,7 +618,7 @@ xfs_acl_set_attr(
 
 int
 xfs_acl_vtoacl(
-	bhv_vnode_t	*vp,
+	struct inode	*vp,
 	xfs_acl_t	*access_acl,
 	xfs_acl_t	*default_acl)
 {
@@ -639,7 +633,7 @@ xfs_acl_vtoacl(
 		if (error)
 			access_acl->acl_cnt = XFS_ACL_NOT_PRESENT;
 		else /* We have a good ACL and the file mode, synchronize. */
-			xfs_acl_sync_mode(xfs_vtoi(vp)->i_d.di_mode, access_acl);
+			xfs_acl_sync_mode(XFS_I(vp)->i_d.di_mode, access_acl);
 	}
 
 	if (default_acl) {
@@ -656,7 +650,7 @@ xfs_acl_vtoacl(
  */
 int
 xfs_acl_inherit(
-	bhv_vnode_t	*vp,
+	struct inode	*vp,
 	mode_t		mode,
 	xfs_acl_t	*pdaclp)
 {
@@ -715,7 +709,7 @@ out_error:
  */
 STATIC int
 xfs_acl_setmode(
-	bhv_vnode_t	*vp,
+	struct inode	*vp,
 	xfs_acl_t	*acl,
 	int		*basicperms)
 {
@@ -734,7 +728,7 @@ xfs_acl_setmode(
 	 * mode.  The m:: bits take precedence over the g:: bits.
 	 */
 	iattr.ia_valid = ATTR_MODE;
-	iattr.ia_mode = xfs_vtoi(vp)->i_d.di_mode;
+	iattr.ia_mode = XFS_I(vp)->i_d.di_mode;
 	iattr.ia_mode &= ~(S_IRWXU|S_IRWXG|S_IRWXO);
 	ap = acl->acl_entry;
 	for (i = 0; i < acl->acl_cnt; ++i) {
@@ -764,7 +758,7 @@ xfs_acl_setmode(
 	if (gap && nomask)
 		iattr.ia_mode |= gap->ae_perm << 3;
 
-	return xfs_setattr(xfs_vtoi(vp), &iattr, 0, sys_cred);
+	return xfs_setattr(XFS_I(vp), &iattr, 0, sys_cred);
 }
 
 /*

fs/xfs/xfs_acl.h

@@ -59,14 +59,14 @@ extern struct kmem_zone *xfs_acl_zone;
 		(zone) = kmem_zone_init(sizeof(xfs_acl_t), (name))
 #define xfs_acl_zone_destroy(zone)	kmem_zone_destroy(zone)
 
-extern int xfs_acl_inherit(bhv_vnode_t *, mode_t mode, xfs_acl_t *);
+extern int xfs_acl_inherit(struct inode *, mode_t mode, xfs_acl_t *);
 extern int xfs_acl_iaccess(struct xfs_inode *, mode_t, cred_t *);
-extern int xfs_acl_vtoacl(bhv_vnode_t *, xfs_acl_t *, xfs_acl_t *);
-extern int xfs_acl_vhasacl_access(bhv_vnode_t *);
-extern int xfs_acl_vhasacl_default(bhv_vnode_t *);
-extern int xfs_acl_vset(bhv_vnode_t *, void *, size_t, int);
-extern int xfs_acl_vget(bhv_vnode_t *, void *, size_t, int);
-extern int xfs_acl_vremove(bhv_vnode_t *, int);
+extern int xfs_acl_vtoacl(struct inode *, xfs_acl_t *, xfs_acl_t *);
+extern int xfs_acl_vhasacl_access(struct inode *);
+extern int xfs_acl_vhasacl_default(struct inode *);
+extern int xfs_acl_vset(struct inode *, void *, size_t, int);
+extern int xfs_acl_vget(struct inode *, void *, size_t, int);
+extern int xfs_acl_vremove(struct inode *, int);
 
 #define _ACL_PERM_INVALID(perm)	((perm) & ~(ACL_READ|ACL_WRITE|ACL_EXECUTE))
 

fs/xfs/xfs_arch.h

@@ -92,16 +92,6 @@
 	((__u8*)(pointer))[1] = (((value)     ) & 0xff); \
     }
 
-/* define generic INT_ macros */
-
-#define INT_GET(reference,arch) \
-    (((arch) == ARCH_NOCONVERT) \
-	? \
-	    (reference) \
-	: \
-	    INT_SWAP((reference),(reference)) \
-    )
-
 /* does not return a value */
 #define INT_SET(reference,arch,valueref) \
     (__builtin_constant_p(valueref) ? \
@@ -112,64 +102,6 @@
 	) \
     )
 
-/* does not return a value */
-#define INT_MOD_EXPR(reference,arch,code) \
-    (((arch) == ARCH_NOCONVERT) \
-	? \
-	    (void)((reference) code) \
-	: \
-	    (void)( \
-		(reference) = INT_GET((reference),arch) , \
-		((reference) code), \
-		INT_SET(reference, arch, reference) \
-	    ) \
-    )
-
-/* does not return a value */
-#define INT_MOD(reference,arch,delta) \
-    (void)( \
-	INT_MOD_EXPR(reference,arch,+=(delta)) \
-    )
-
-/*
- * INT_COPY - copy a value between two locations with the
- *	      _same architecture_ but _potentially different sizes_
- *
- *	    if the types of the two parameters are equal or they are
- *		in native architecture, a simple copy is done
- *
- *	    otherwise, architecture conversions are done
- *
- */
-
-/* does not return a value */
-#define INT_COPY(dst,src,arch) \
-    ( \
-	((sizeof(dst) == sizeof(src)) || ((arch) == ARCH_NOCONVERT)) \
-	    ? \
-		(void)((dst) = (src)) \
-	    : \
-		INT_SET(dst, arch, INT_GET(src, arch)) \
-    )
-
-/*
- * INT_XLATE - copy a value in either direction between two locations
- *	       with different architectures
- *
- *		    dir < 0	- copy from memory to buffer (native to arch)
- *		    dir > 0	- copy from buffer to memory (arch to native)
- */
-
-/* does not return a value */
-#define INT_XLATE(buf,mem,dir,arch) {\
-    ASSERT(dir); \
-    if (dir>0) { \
-	(mem)=INT_GET(buf, arch); \
-    } else { \
-	INT_SET(buf, arch, mem); \
-    } \
-}
-
 /*
  * In directories inode numbers are stored as unaligned arrays of unsigned
  * 8bit integers on disk.

fs/xfs/xfs_attr.c

@@ -194,6 +194,46 @@ xfs_attr_get(
 	return(error);
 }
 
+/*
+ * Calculate how many blocks we need for the new attribute,
+ */
+int
+xfs_attr_calc_size(
+	struct xfs_inode 	*ip,
+	int			namelen,
+	int			valuelen,
+	int			*local)
+{
+	struct xfs_mount 	*mp = ip->i_mount;
+	int			size;
+	int			nblks;
+
+	/*
+	 * Determine space new attribute will use, and if it would be
+	 * "local" or "remote" (note: local != inline).
+	 */
+	size = xfs_attr_leaf_newentsize(namelen, valuelen,
+					mp->m_sb.sb_blocksize, local);
+
+	nblks = XFS_DAENTER_SPACE_RES(mp, XFS_ATTR_FORK);
+	if (*local) {
+		if (size > (mp->m_sb.sb_blocksize >> 1)) {
+			/* Double split possible */
+			nblks *= 2;
+		}
+	} else {
+		/*
+		 * Out of line attribute, cannot double split, but
+		 * make room for the attribute value itself.
+		 */
+		uint	dblocks = XFS_B_TO_FSB(mp, valuelen);
+		nblks += dblocks;
+		nblks += XFS_NEXTENTADD_SPACE_RES(mp, dblocks, XFS_ATTR_FORK);
+	}
+
+	return nblks;
+}
+
 STATIC int
 xfs_attr_set_int(xfs_inode_t *dp, struct xfs_name *name,
 		char *value, int valuelen, int flags)
@@ -202,10 +242,9 @@ xfs_attr_set_int(xfs_inode_t *dp, struct xfs_name *name,
 	xfs_fsblock_t	firstblock;
 	xfs_bmap_free_t flist;
 	int		error, err2, committed;
-	int		local, size;
-	uint		nblks;
 	xfs_mount_t	*mp = dp->i_mount;
 	int             rsvd = (flags & ATTR_ROOT) != 0;
+	int		local;
 
 	/*
 	 * Attach the dquots to the inode.
@@ -241,30 +280,8 @@ xfs_attr_set_int(xfs_inode_t *dp, struct xfs_name *name,
 	args.whichfork = XFS_ATTR_FORK;
 	args.op_flags = XFS_DA_OP_ADDNAME | XFS_DA_OP_OKNOENT;
 
-	/*
-	 * Determine space new attribute will use, and if it would be
-	 * "local" or "remote" (note: local != inline).
-	 */
-	size = xfs_attr_leaf_newentsize(name->len, valuelen,
-					mp->m_sb.sb_blocksize, &local);
-
-	nblks = XFS_DAENTER_SPACE_RES(mp, XFS_ATTR_FORK);
-	if (local) {
-		if (size > (mp->m_sb.sb_blocksize >> 1)) {
-			/* Double split possible */
-			nblks <<= 1;
-		}
-	} else {
-		uint	dblocks = XFS_B_TO_FSB(mp, valuelen);
-		/* Out of line attribute, cannot double split, but make
-		 * room for the attribute value itself.
-		 */
-		nblks += dblocks;
-		nblks += XFS_NEXTENTADD_SPACE_RES(mp, dblocks, XFS_ATTR_FORK);
-	}
-
 	/* Size is now blocks for attribute data */
-	args.total = nblks;
+	args.total = xfs_attr_calc_size(dp, name->len, valuelen, &local);
 
 	/*
 	 * Start our first transaction of the day.
@@ -286,18 +303,17 @@ xfs_attr_set_int(xfs_inode_t *dp, struct xfs_name *name,
 	if (rsvd)
 		args.trans->t_flags |= XFS_TRANS_RESERVE;
 
-	if ((error = xfs_trans_reserve(args.trans, (uint) nblks,
-				      XFS_ATTRSET_LOG_RES(mp, nblks),
-				      0, XFS_TRANS_PERM_LOG_RES,
-				      XFS_ATTRSET_LOG_COUNT))) {
+	if ((error = xfs_trans_reserve(args.trans, args.total,
+			XFS_ATTRSET_LOG_RES(mp, args.total), 0,
+			XFS_TRANS_PERM_LOG_RES, XFS_ATTRSET_LOG_COUNT))) {
 		xfs_trans_cancel(args.trans, 0);
 		return(error);
 	}
 	xfs_ilock(dp, XFS_ILOCK_EXCL);
 
-	error = XFS_TRANS_RESERVE_QUOTA_NBLKS(mp, args.trans, dp, nblks, 0,
-			 rsvd ? XFS_QMOPT_RES_REGBLKS | XFS_QMOPT_FORCE_RES :
-				XFS_QMOPT_RES_REGBLKS);
+	error = XFS_TRANS_RESERVE_QUOTA_NBLKS(mp, args.trans, dp, args.total, 0,
+				rsvd ? XFS_QMOPT_RES_REGBLKS | XFS_QMOPT_FORCE_RES :
+				       XFS_QMOPT_RES_REGBLKS);
 	if (error) {
 		xfs_iunlock(dp, XFS_ILOCK_EXCL);
 		xfs_trans_cancel(args.trans, XFS_TRANS_RELEASE_LOG_RES);
@@ -384,7 +400,9 @@ xfs_attr_set_int(xfs_inode_t *dp, struct xfs_name *name,
 		 * Commit the leaf transformation.  We'll need another (linked)
 		 * transaction to add the new attribute to the leaf.
 		 */
-		if ((error = xfs_attr_rolltrans(&args.trans, dp)))
+
+		error = xfs_trans_roll(&args.trans, dp);
+		if (error)
 			goto out;
 
 	}
@@ -964,7 +982,8 @@ xfs_attr_leaf_addname(xfs_da_args_t *args)
 		 * Commit the current trans (including the inode) and start
 		 * a new one.
 		 */
-		if ((error = xfs_attr_rolltrans(&args->trans, dp)))
+		error = xfs_trans_roll(&args->trans, dp);
+		if (error)
 			return (error);
 
 		/*
@@ -978,7 +997,8 @@ xfs_attr_leaf_addname(xfs_da_args_t *args)
 	 * Commit the transaction that added the attr name so that
 	 * later routines can manage their own transactions.
 	 */
-	if ((error = xfs_attr_rolltrans(&args->trans, dp)))
+	error = xfs_trans_roll(&args->trans, dp);
+	if (error)
 		return (error);
 
 	/*
@@ -1067,7 +1087,7 @@ xfs_attr_leaf_addname(xfs_da_args_t *args)
 		/*
 		 * Commit the remove and start the next trans in series.
 		 */
-		error = xfs_attr_rolltrans(&args->trans, dp);
+		error = xfs_trans_roll(&args->trans, dp);
 
 	} else if (args->rmtblkno > 0) {
 		/*
@@ -1298,7 +1318,8 @@ restart:
 			 * Commit the node conversion and start the next
 			 * trans in the chain.
 			 */
-			if ((error = xfs_attr_rolltrans(&args->trans, dp)))
+			error = xfs_trans_roll(&args->trans, dp);
+			if (error)
 				goto out;
 
 			goto restart;
@@ -1349,7 +1370,8 @@ restart:
 	 * Commit the leaf addition or btree split and start the next
 	 * trans in the chain.
 	 */
-	if ((error = xfs_attr_rolltrans(&args->trans, dp)))
+	error = xfs_trans_roll(&args->trans, dp);
+	if (error)
 		goto out;
 
 	/*
@@ -1449,7 +1471,8 @@ restart:
 		/*
 		 * Commit and start the next trans in the chain.
 		 */
-		if ((error = xfs_attr_rolltrans(&args->trans, dp)))
+		error = xfs_trans_roll(&args->trans, dp);
+		if (error)
 			goto out;
 
 	} else if (args->rmtblkno > 0) {
@@ -1581,7 +1604,8 @@ xfs_attr_node_removename(xfs_da_args_t *args)
 		/*
 		 * Commit the Btree join operation and start a new trans.
 		 */
-		if ((error = xfs_attr_rolltrans(&args->trans, dp)))
+		error = xfs_trans_roll(&args->trans, dp);
+		if (error)
 			goto out;
 	}
 
@@ -2082,7 +2106,8 @@ xfs_attr_rmtval_set(xfs_da_args_t *args)
 		/*
 		 * Start the next trans in the chain.
 		 */
-		if ((error = xfs_attr_rolltrans(&args->trans, dp)))
+		error = xfs_trans_roll(&args->trans, dp);
+		if (error)
 			return (error);
 	}
 
@@ -2232,7 +2257,8 @@ xfs_attr_rmtval_remove(xfs_da_args_t *args)
 		/*
 		 * Close out trans and start the next one in the chain.
 		 */
-		if ((error = xfs_attr_rolltrans(&args->trans, args->dp)))
+		error = xfs_trans_roll(&args->trans, args->dp);
+		if (error)
 			return (error);
 	}
 	return(0);

fs/xfs/xfs_attr.h

@@ -129,6 +129,7 @@ typedef struct xfs_attr_list_context {
 /*
  * Overall external interface routines.
  */
+int xfs_attr_calc_size(struct xfs_inode *, int, int, int *);
 int xfs_attr_inactive(struct xfs_inode *dp);
 int xfs_attr_fetch(struct xfs_inode *, struct xfs_name *, char *, int *, int);
 int xfs_attr_rmtval_get(struct xfs_da_args *args);

fs/xfs/xfs_attr_leaf.c

@@ -2498,9 +2498,7 @@ xfs_attr_leaf_clearflag(xfs_da_args_t *args)
 	/*
 	 * Commit the flag value change and start the next trans in series.
 	 */
-	error = xfs_attr_rolltrans(&args->trans, args->dp);
-
-	return(error);
+	return xfs_trans_roll(&args->trans, args->dp);
 }
 
 /*
@@ -2547,9 +2545,7 @@ xfs_attr_leaf_setflag(xfs_da_args_t *args)
 	/*
 	 * Commit the flag value change and start the next trans in series.
 	 */
-	error = xfs_attr_rolltrans(&args->trans, args->dp);
-
-	return(error);
+	return xfs_trans_roll(&args->trans, args->dp);
 }
 
 /*
@@ -2665,7 +2661,7 @@ xfs_attr_leaf_flipflags(xfs_da_args_t *args)
 	/*
 	 * Commit the flag value change and start the next trans in series.
 	 */
-	error = xfs_attr_rolltrans(&args->trans, args->dp);
+	error = xfs_trans_roll(&args->trans, args->dp);
 
 	return(error);
 }
@@ -2723,7 +2719,7 @@ xfs_attr_root_inactive(xfs_trans_t **trans, xfs_inode_t *dp)
 	/*
 	 * Commit the invalidate and start the next transaction.
 	 */
-	error = xfs_attr_rolltrans(trans, dp);
+	error = xfs_trans_roll(trans, dp);
 
 	return (error);
 }
@@ -2825,7 +2821,8 @@ xfs_attr_node_inactive(xfs_trans_t **trans, xfs_inode_t *dp, xfs_dabuf_t *bp,
 		/*
 		 * Atomically commit the whole invalidate stuff.
 		 */
-		if ((error = xfs_attr_rolltrans(trans, dp)))
+		error = xfs_trans_roll(trans, dp);
+		if (error)
 			return (error);
 	}
 
@@ -2964,7 +2961,8 @@ xfs_attr_leaf_freextent(xfs_trans_t **trans, xfs_inode_t *dp,
 			/*
 			 * Roll to next transaction.
 			 */
-			if ((error = xfs_attr_rolltrans(trans, dp)))
+			error = xfs_trans_roll(trans, dp);
+			if (error)
 				return (error);
 		}
 
@@ -2974,60 +2972,3 @@ xfs_attr_leaf_freextent(xfs_trans_t **trans, xfs_inode_t *dp,
 
 	return(0);
 }
-
-
-/*
- * Roll from one trans in the sequence of PERMANENT transactions to the next.
- */
-int
-xfs_attr_rolltrans(xfs_trans_t **transp, xfs_inode_t *dp)
-{
-	xfs_trans_t *trans;
-	unsigned int logres, count;
-	int	error;
-
-	/*
-	 * Ensure that the inode is always logged.
-	 */
-	trans = *transp;
-	xfs_trans_log_inode(trans, dp, XFS_ILOG_CORE);
-
-	/*
-	 * Copy the critical parameters from one trans to the next.
-	 */
-	logres = trans->t_log_res;
-	count = trans->t_log_count;
-	*transp = xfs_trans_dup(trans);
-
-	/*
-	 * Commit the current transaction.
-	 * If this commit failed, then it'd just unlock those items that
-	 * are not marked ihold. That also means that a filesystem shutdown
-	 * is in progress. The caller takes the responsibility to cancel
-	 * the duplicate transaction that gets returned.
-	 */
-	if ((error = xfs_trans_commit(trans, 0)))
-		return (error);
-
-	trans = *transp;
-
-	/*
-	 * Reserve space in the log for th next transaction.
-	 * This also pushes items in the "AIL", the list of logged items,
-	 * out to disk if they are taking up space at the tail of the log
-	 * that we want to use.  This requires that either nothing be locked
-	 * across this call, or that anything that is locked be logged in
-	 * the prior and the next transactions.
-	 */
-	error = xfs_trans_reserve(trans, 0, logres, 0,
-				  XFS_TRANS_PERM_LOG_RES, count);
-	/*
-	 *  Ensure that the inode is in the new transaction and locked.
-	 */
-	if (!error) {
-		xfs_trans_ijoin(trans, dp, XFS_ILOCK_EXCL);
-		xfs_trans_ihold(trans, dp);
-	}
-	return (error);
-
-}

fs/xfs/xfs_attr_leaf.h

@@ -274,6 +274,4 @@ int	xfs_attr_leaf_order(struct xfs_dabuf *leaf1_bp,
 				   struct xfs_dabuf *leaf2_bp);
 int	xfs_attr_leaf_newentsize(int namelen, int valuelen, int blocksize,
 					int *local);
-int	xfs_attr_rolltrans(struct xfs_trans **transp, struct xfs_inode *dp);
-
 #endif	/* __XFS_ATTR_LEAF_H__ */

fs/xfs/xfs_bit.c

@@ -25,109 +25,6 @@
  * XFS bit manipulation routines, used in non-realtime code.
  */
 
-#ifndef HAVE_ARCH_HIGHBIT
-/*
- * Index of high bit number in byte, -1 for none set, 0..7 otherwise.
- */
-static const char xfs_highbit[256] = {
-       -1, 0, 1, 1, 2, 2, 2, 2,			/* 00 .. 07 */
-	3, 3, 3, 3, 3, 3, 3, 3,			/* 08 .. 0f */
-	4, 4, 4, 4, 4, 4, 4, 4,			/* 10 .. 17 */
-	4, 4, 4, 4, 4, 4, 4, 4,			/* 18 .. 1f */
-	5, 5, 5, 5, 5, 5, 5, 5,			/* 20 .. 27 */
-	5, 5, 5, 5, 5, 5, 5, 5,			/* 28 .. 2f */
-	5, 5, 5, 5, 5, 5, 5, 5,			/* 30 .. 37 */
-	5, 5, 5, 5, 5, 5, 5, 5,			/* 38 .. 3f */
-	6, 6, 6, 6, 6, 6, 6, 6,			/* 40 .. 47 */
-	6, 6, 6, 6, 6, 6, 6, 6,			/* 48 .. 4f */
-	6, 6, 6, 6, 6, 6, 6, 6,			/* 50 .. 57 */
-	6, 6, 6, 6, 6, 6, 6, 6,			/* 58 .. 5f */
-	6, 6, 6, 6, 6, 6, 6, 6,			/* 60 .. 67 */
-	6, 6, 6, 6, 6, 6, 6, 6,			/* 68 .. 6f */
-	6, 6, 6, 6, 6, 6, 6, 6,			/* 70 .. 77 */
-	6, 6, 6, 6, 6, 6, 6, 6,			/* 78 .. 7f */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* 80 .. 87 */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* 88 .. 8f */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* 90 .. 97 */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* 98 .. 9f */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* a0 .. a7 */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* a8 .. af */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* b0 .. b7 */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* b8 .. bf */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* c0 .. c7 */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* c8 .. cf */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* d0 .. d7 */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* d8 .. df */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* e0 .. e7 */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* e8 .. ef */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* f0 .. f7 */
-	7, 7, 7, 7, 7, 7, 7, 7,			/* f8 .. ff */
-};
-#endif
-
-/*
- * xfs_highbit32: get high bit set out of 32-bit argument, -1 if none set.
- */
-inline int
-xfs_highbit32(
-	__uint32_t	v)
-{
-#ifdef HAVE_ARCH_HIGHBIT
-	return highbit32(v);
-#else
-	int		i;
-
-	if (v & 0xffff0000)
-		if (v & 0xff000000)
-			i = 24;
-		else
-			i = 16;
-	else if (v & 0x0000ffff)
-		if (v & 0x0000ff00)
-			i = 8;
-		else
-			i = 0;
-	else
-		return -1;
-	return i + xfs_highbit[(v >> i) & 0xff];
-#endif
-}
-
-/*
- * xfs_lowbit64: get low bit set out of 64-bit argument, -1 if none set.
- */
-int
-xfs_lowbit64(
-	__uint64_t	v)
-{
-	__uint32_t	w = (__uint32_t)v;
-	int		n = 0;
-
-	if (w) {	/* lower bits */
-		n = ffs(w);
-	} else {	/* upper bits */
-		w = (__uint32_t)(v >> 32);
-		if (w && (n = ffs(w)))
-			n += 32;
-	}
-	return n - 1;
-}
-
-/*
- * xfs_highbit64: get high bit set out of 64-bit argument, -1 if none set.
- */
-int
-xfs_highbit64(
-	__uint64_t	v)
-{
-	__uint32_t	h = (__uint32_t)(v >> 32);
-
-	if (h)
-		return xfs_highbit32(h) + 32;
-	return xfs_highbit32((__uint32_t)v);
-}
-
-
 /*
  * Return whether bitmap is empty.
  * Size is number of words in the bitmap, which is padded to word boundary

fs/xfs/xfs_bit.h

@@ -47,13 +47,39 @@ static inline __uint64_t xfs_mask64lo(int n)
 }
 
 /* Get high bit set out of 32-bit argument, -1 if none set */
-extern int xfs_highbit32(__uint32_t v);
+static inline int xfs_highbit32(__uint32_t v)
+{
+	return fls(v) - 1;
+}
+
+/* Get high bit set out of 64-bit argument, -1 if none set */
+static inline int xfs_highbit64(__uint64_t v)
+{
+	return fls64(v) - 1;
+}
+
+/* Get low bit set out of 32-bit argument, -1 if none set */
+static inline int xfs_lowbit32(__uint32_t v)
+{
+	unsigned long	t = v;
+	return (v) ? find_first_bit(&t, 32) : -1;
+}
 
 /* Get low bit set out of 64-bit argument, -1 if none set */
-extern int xfs_lowbit64(__uint64_t v);
+static inline int xfs_lowbit64(__uint64_t v)
+{
+	__uint32_t	w = (__uint32_t)v;
+	int		n = 0;
 
-/* Get high bit set out of 64-bit argument, -1 if none set */
-extern int xfs_highbit64(__uint64_t);
+	if (w) {	/* lower bits */
+		n = ffs(w);
+	} else {	/* upper bits */
+		w = (__uint32_t)(v >> 32);
+		if (w && (n = ffs(w)))
+		n += 32;
+	}
+	return n - 1;
+}
 
 /* Return whether bitmap is empty (1 == empty) */
 extern int xfs_bitmap_empty(uint *map, uint size);

fs/xfs/xfs_bmap.c

@@ -384,14 +384,14 @@ xfs_bmap_count_tree(
 	int             levelin,
 	int		*count);
 
-STATIC int
+STATIC void
 xfs_bmap_count_leaves(
 	xfs_ifork_t		*ifp,
 	xfs_extnum_t		idx,
 	int			numrecs,
 	int			*count);
 
-STATIC int
+STATIC void
 xfs_bmap_disk_count_leaves(
 	xfs_extnum_t		idx,
 	xfs_bmbt_block_t	*block,
@@ -4000,7 +4000,7 @@ xfs_bmap_add_attrfork(
 		ip->i_d.di_aformat = XFS_DINODE_FMT_EXTENTS;
 	}
 	ASSERT(ip->i_d.di_anextents == 0);
-	VN_HOLD(XFS_ITOV(ip));
+	IHOLD(ip);
 	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 	switch (ip->i_d.di_format) {
@@ -6096,7 +6096,7 @@ xfs_bmap_get_bp(
 		tp = cur->bc_tp;
 		licp = &tp->t_items;
 		while (!bp && licp != NULL) {
-			if (XFS_LIC_ARE_ALL_FREE(licp)) {
+			if (xfs_lic_are_all_free(licp)) {
 				licp = licp->lic_next;
 				continue;
 			}
@@ -6106,11 +6106,11 @@ xfs_bmap_get_bp(
 				xfs_buf_log_item_t	*bip;
 				xfs_buf_t		*lbp;
 
-				if (XFS_LIC_ISFREE(licp, i)) {
+				if (xfs_lic_isfree(licp, i)) {
 					continue;
 				}
 
-				lidp = XFS_LIC_SLOT(licp, i);
+				lidp = xfs_lic_slot(licp, i);
 				lip = lidp->lid_item;
 				if (lip->li_type != XFS_LI_BUF)
 					continue;
@@ -6367,13 +6367,9 @@ xfs_bmap_count_blocks(
 	mp = ip->i_mount;
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	if ( XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS ) {
-		if (unlikely(xfs_bmap_count_leaves(ifp, 0,
+		xfs_bmap_count_leaves(ifp, 0,
 			ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t),
-			count) < 0)) {
-			XFS_ERROR_REPORT("xfs_bmap_count_blocks(1)",
-					 XFS_ERRLEVEL_LOW, mp);
-			return XFS_ERROR(EFSCORRUPTED);
-		}
+			count);
 		return 0;
 	}
 
@@ -6454,13 +6450,7 @@ xfs_bmap_count_tree(
 		for (;;) {
 			nextbno = be64_to_cpu(block->bb_rightsib);
 			numrecs = be16_to_cpu(block->bb_numrecs);
-			if (unlikely(xfs_bmap_disk_count_leaves(0,
-					block, numrecs, count) < 0)) {
-				xfs_trans_brelse(tp, bp);
-				XFS_ERROR_REPORT("xfs_bmap_count_tree(2)",
-						 XFS_ERRLEVEL_LOW, mp);
-				return XFS_ERROR(EFSCORRUPTED);
-			}
+			xfs_bmap_disk_count_leaves(0, block, numrecs, count);
 			xfs_trans_brelse(tp, bp);
 			if (nextbno == NULLFSBLOCK)
 				break;
@@ -6478,7 +6468,7 @@ xfs_bmap_count_tree(
 /*
  * Count leaf blocks given a range of extent records.
  */
-STATIC int
+STATIC void
 xfs_bmap_count_leaves(
 	xfs_ifork_t		*ifp,
 	xfs_extnum_t		idx,
@@ -6491,14 +6481,13 @@ xfs_bmap_count_leaves(
 		xfs_bmbt_rec_host_t *frp = xfs_iext_get_ext(ifp, idx + b);
 		*count += xfs_bmbt_get_blockcount(frp);
 	}
-	return 0;
 }
 
 /*
  * Count leaf blocks given a range of extent records originally
  * in btree format.
  */
-STATIC int
+STATIC void
 xfs_bmap_disk_count_leaves(
 	xfs_extnum_t		idx,
 	xfs_bmbt_block_t	*block,
@@ -6512,5 +6501,4 @@ xfs_bmap_disk_count_leaves(
 		frp = XFS_BTREE_REC_ADDR(xfs_bmbt, block, idx + b);
 		*count += xfs_bmbt_disk_get_blockcount(frp);
 	}
-	return 0;
 }

fs/xfs/xfs_btree.c

@@ -46,37 +46,10 @@ kmem_zone_t	*xfs_btree_cur_zone;
 /*
  * Btree magic numbers.
  */
-const __uint32_t xfs_magics[XFS_BTNUM_MAX] =
-{
+const __uint32_t xfs_magics[XFS_BTNUM_MAX] = {
 	XFS_ABTB_MAGIC, XFS_ABTC_MAGIC, XFS_BMAP_MAGIC, XFS_IBT_MAGIC
 };
 
-/*
- * Prototypes for internal routines.
- */
-
-/*
- * Checking routine: return maxrecs for the block.
- */
-STATIC int				/* number of records fitting in block */
-xfs_btree_maxrecs(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_btree_block_t	*block);/* generic btree block pointer */
-
-/*
- * Internal routines.
- */
-
-/*
- * Retrieve the block pointer from the cursor at the given level.
- * This may be a bmap btree root or from a buffer.
- */
-STATIC xfs_btree_block_t *			/* generic btree block pointer */
-xfs_btree_get_block(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level in btree */
-	struct xfs_buf		**bpp);	/* buffer containing the block */
-
 /*
  * Checking routine: return maxrecs for the block.
  */
@@ -456,35 +429,6 @@ xfs_btree_dup_cursor(
 	return 0;
 }
 
-/*
- * Change the cursor to point to the first record at the given level.
- * Other levels are unaffected.
- */
-int					/* success=1, failure=0 */
-xfs_btree_firstrec(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level)	/* level to change */
-{
-	xfs_btree_block_t	*block;	/* generic btree block pointer */
-	xfs_buf_t		*bp;	/* buffer containing block */
-
-	/*
-	 * Get the block pointer for this level.
-	 */
-	block = xfs_btree_get_block(cur, level, &bp);
-	xfs_btree_check_block(cur, block, level, bp);
-	/*
-	 * It's empty, there is no such record.
-	 */
-	if (!block->bb_h.bb_numrecs)
-		return 0;
-	/*
-	 * Set the ptr value to 1, that's the first record/key.
-	 */
-	cur->bc_ptrs[level] = 1;
-	return 1;
-}
-
 /*
  * Retrieve the block pointer from the cursor at the given level.
  * This may be a bmap btree root or from a buffer.
@@ -626,6 +570,13 @@ xfs_btree_init_cursor(
 		cur->bc_private.a.agbp = agbp;
 		cur->bc_private.a.agno = agno;
 		break;
+	case XFS_BTNUM_INO:
+		/*
+		 * Inode allocation btree fields.
+		 */
+		cur->bc_private.a.agbp = agbp;
+		cur->bc_private.a.agno = agno;
+		break;
 	case XFS_BTNUM_BMAP:
 		/*
 		 * Bmap btree fields.
@@ -638,13 +589,6 @@ xfs_btree_init_cursor(
 		cur->bc_private.b.flags = 0;
 		cur->bc_private.b.whichfork = whichfork;
 		break;
-	case XFS_BTNUM_INO:
-		/*
-		 * Inode allocation btree fields.
-		 */
-		cur->bc_private.i.agbp = agbp;
-		cur->bc_private.i.agno = agno;
-		break;
 	default:
 		ASSERT(0);
 	}
@@ -670,6 +614,35 @@ xfs_btree_islastblock(
 		return be32_to_cpu(block->bb_u.s.bb_rightsib) == NULLAGBLOCK;
 }
 
+/*
+ * Change the cursor to point to the first record at the given level.
+ * Other levels are unaffected.
+ */
+int					/* success=1, failure=0 */
+xfs_btree_firstrec(
+	xfs_btree_cur_t		*cur,	/* btree cursor */
+	int			level)	/* level to change */
+{
+	xfs_btree_block_t	*block;	/* generic btree block pointer */
+	xfs_buf_t		*bp;	/* buffer containing block */
+
+	/*
+	 * Get the block pointer for this level.
+	 */
+	block = xfs_btree_get_block(cur, level, &bp);
+	xfs_btree_check_block(cur, block, level, bp);
+	/*
+	 * It's empty, there is no such record.
+	 */
+	if (!block->bb_h.bb_numrecs)
+		return 0;
+	/*
+	 * Set the ptr value to 1, that's the first record/key.
+	 */
+	cur->bc_ptrs[level] = 1;
+	return 1;
+}
+
 /*
  * Change the cursor to point to the last record in the current block
  * at the given level.  Other levels are unaffected.
@@ -890,12 +863,12 @@ xfs_btree_readahead_core(
 	case XFS_BTNUM_INO:
 		i = XFS_BUF_TO_INOBT_BLOCK(cur->bc_bufs[lev]);
 		if ((lr & XFS_BTCUR_LEFTRA) && be32_to_cpu(i->bb_leftsib) != NULLAGBLOCK) {
-			xfs_btree_reada_bufs(cur->bc_mp, cur->bc_private.i.agno,
+			xfs_btree_reada_bufs(cur->bc_mp, cur->bc_private.a.agno,
 				be32_to_cpu(i->bb_leftsib), 1);
 			rval++;
 		}
 		if ((lr & XFS_BTCUR_RIGHTRA) && be32_to_cpu(i->bb_rightsib) != NULLAGBLOCK) {
-			xfs_btree_reada_bufs(cur->bc_mp, cur->bc_private.i.agno,
+			xfs_btree_reada_bufs(cur->bc_mp, cur->bc_private.a.agno,
 				be32_to_cpu(i->bb_rightsib), 1);
 			rval++;
 		}

fs/xfs/xfs_btree.h

@@ -158,8 +158,8 @@ typedef struct xfs_btree_cur
 	__uint8_t	bc_blocklog;	/* log2(blocksize) of btree blocks */
 	xfs_btnum_t	bc_btnum;	/* identifies which btree type */
 	union {
-		struct {			/* needed for BNO, CNT */
-			struct xfs_buf	*agbp;	/* agf buffer pointer */
+		struct {			/* needed for BNO, CNT, INO */
+			struct xfs_buf	*agbp;	/* agf/agi buffer pointer */
 			xfs_agnumber_t	agno;	/* ag number */
 		} a;
 		struct {			/* needed for BMAP */
@@ -172,10 +172,6 @@ typedef struct xfs_btree_cur
 			char		flags;		/* flags */
 #define	XFS_BTCUR_BPRV_WASDEL	1			/* was delayed */
 		} b;
-		struct {			/* needed for INO */
-			struct xfs_buf	*agbp;	/* agi buffer pointer */
-			xfs_agnumber_t	agno;	/* ag number */
-		} i;
 	}		bc_private;	/* per-btree type data */
 } xfs_btree_cur_t;
 

fs/xfs/xfs_buf_item.c

@@ -737,7 +737,7 @@ xfs_buf_item_init(
 	bip->bli_format.blf_len = (ushort)BTOBB(XFS_BUF_COUNT(bp));
 	bip->bli_format.blf_map_size = map_size;
 #ifdef XFS_BLI_TRACE
-	bip->bli_trace = ktrace_alloc(XFS_BLI_TRACE_SIZE, KM_SLEEP);
+	bip->bli_trace = ktrace_alloc(XFS_BLI_TRACE_SIZE, KM_NOFS);
 #endif
 
 #ifdef XFS_TRANS_DEBUG
@@ -1056,7 +1056,7 @@ xfs_buf_iodone_callbacks(
 			   anyway. */
 			XFS_BUF_SET_BRELSE_FUNC(bp,xfs_buf_error_relse);
 			XFS_BUF_DONE(bp);
-			XFS_BUF_V_IODONESEMA(bp);
+			XFS_BUF_FINISH_IOWAIT(bp);
 		}
 		return;
 	}

fs/xfs/xfs_dfrag.c

@@ -128,10 +128,8 @@ xfs_swap_extents(
 	xfs_swapext_t	*sxp)
 {
 	xfs_mount_t	*mp;
-	xfs_inode_t	*ips[2];
 	xfs_trans_t	*tp;
 	xfs_bstat_t	*sbp = &sxp->sx_stat;
-	bhv_vnode_t	*vp, *tvp;
 	xfs_ifork_t	*tempifp, *ifp, *tifp;
 	int		ilf_fields, tilf_fields;
 	static uint	lock_flags = XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL;
@@ -150,19 +148,8 @@ xfs_swap_extents(
 	}
 
 	sbp = &sxp->sx_stat;
-	vp = XFS_ITOV(ip);
-	tvp = XFS_ITOV(tip);
-
-	/* Lock in i_ino order */
-	if (ip->i_ino < tip->i_ino) {
-		ips[0] = ip;
-		ips[1] = tip;
-	} else {
-		ips[0] = tip;
-		ips[1] = ip;
-	}
 
-	xfs_lock_inodes(ips, 2, lock_flags);
+	xfs_lock_two_inodes(ip, tip, lock_flags);
 	locked = 1;
 
 	/* Verify that both files have the same format */
@@ -184,7 +171,7 @@ xfs_swap_extents(
 		goto error0;
 	}
 
-	if (VN_CACHED(tvp) != 0) {
+	if (VN_CACHED(VFS_I(tip)) != 0) {
 		xfs_inval_cached_trace(tip, 0, -1, 0, -1);
 		error = xfs_flushinval_pages(tip, 0, -1,
 				FI_REMAPF_LOCKED);
@@ -193,7 +180,7 @@ xfs_swap_extents(
 	}
 
 	/* Verify O_DIRECT for ftmp */
-	if (VN_CACHED(tvp) != 0) {
+	if (VN_CACHED(VFS_I(tip)) != 0) {
 		error = XFS_ERROR(EINVAL);
 		goto error0;
 	}
@@ -237,7 +224,7 @@ xfs_swap_extents(
 	 * vop_read (or write in the case of autogrow) they block on the iolock
 	 * until we have switched the extents.
 	 */
-	if (VN_MAPPED(vp)) {
+	if (VN_MAPPED(VFS_I(ip))) {
 		error = XFS_ERROR(EBUSY);
 		goto error0;
 	}
@@ -265,7 +252,7 @@ xfs_swap_extents(
 		locked = 0;
 		goto error0;
 	}
-	xfs_lock_inodes(ips, 2, XFS_ILOCK_EXCL);
+	xfs_lock_two_inodes(ip, tip, XFS_ILOCK_EXCL);
 
 	/*
 	 * Count the number of extended attribute blocks
@@ -350,15 +337,11 @@ xfs_swap_extents(
 		break;
 	}
 
-	/*
-	 * Increment vnode ref counts since xfs_trans_commit &
-	 * xfs_trans_cancel will both unlock the inodes and
-	 * decrement the associated ref counts.
-	 */
-	VN_HOLD(vp);
-	VN_HOLD(tvp);
 
+	IHOLD(ip);
 	xfs_trans_ijoin(tp, ip, lock_flags);
+
+	IHOLD(tip);
 	xfs_trans_ijoin(tp, tip, lock_flags);
 
 	xfs_trans_log_inode(tp, ip,  ilf_fields);

fs/xfs/xfs_error.c

@@ -58,9 +58,6 @@ xfs_error_trap(int e)
 	}
 	return e;
 }
-#endif
-
-#if (defined(DEBUG) || defined(INDUCE_IO_ERROR))
 
 int	xfs_etest[XFS_NUM_INJECT_ERROR];
 int64_t	xfs_etest_fsid[XFS_NUM_INJECT_ERROR];
@@ -154,7 +151,7 @@ xfs_errortag_clearall(xfs_mount_t *mp, int loud)
 
 	return 0;
 }
-#endif /* DEBUG || INDUCE_IO_ERROR */
+#endif /* DEBUG */
 
 static void
 xfs_fs_vcmn_err(int level, xfs_mount_t *mp, char *fmt, va_list ap)

fs/xfs/xfs_error.h

@@ -125,22 +125,14 @@ extern void xfs_corruption_error(char *tag, int level, struct xfs_mount *mp,
 #define XFS_RANDOM_DIOWRITE_IOERR			(XFS_RANDOM_DEFAULT/10)
 #define	XFS_RANDOM_BMAPIFORMAT				XFS_RANDOM_DEFAULT
 
-#if (defined(DEBUG) || defined(INDUCE_IO_ERROR))
+#ifdef DEBUG
 extern int xfs_error_test(int, int *, char *, int, char *, unsigned long);
 
 #define	XFS_NUM_INJECT_ERROR				10
-
-#ifdef __ANSI_CPP__
-#define XFS_TEST_ERROR(expr, mp, tag, rf)		\
-	((expr) || \
-	 xfs_error_test((tag), (mp)->m_fixedfsid, #expr, __LINE__, __FILE__, \
-			 (rf)))
-#else
 #define XFS_TEST_ERROR(expr, mp, tag, rf)		\
 	((expr) || \
 	 xfs_error_test((tag), (mp)->m_fixedfsid, "expr", __LINE__, __FILE__, \
 			(rf)))
-#endif /* __ANSI_CPP__ */
 
 extern int xfs_errortag_add(int error_tag, xfs_mount_t *mp);
 extern int xfs_errortag_clearall(xfs_mount_t *mp, int loud);
@@ -148,7 +140,7 @@ extern int xfs_errortag_clearall(xfs_mount_t *mp, int loud);
 #define XFS_TEST_ERROR(expr, mp, tag, rf)	(expr)
 #define xfs_errortag_add(tag, mp)		(ENOSYS)
 #define xfs_errortag_clearall(mp, loud)		(ENOSYS)
-#endif /* (DEBUG || INDUCE_IO_ERROR) */
+#endif /* DEBUG */
 
 /*
  * XFS panic tags -- allow a call to xfs_cmn_err() be turned into

fs/xfs/xfs_filestream.c

@@ -400,7 +400,7 @@ xfs_filestream_init(void)
 	if (!item_zone)
 		return -ENOMEM;
 #ifdef XFS_FILESTREAMS_TRACE
-	xfs_filestreams_trace_buf = ktrace_alloc(XFS_FSTRM_KTRACE_SIZE, KM_SLEEP);
+	xfs_filestreams_trace_buf = ktrace_alloc(XFS_FSTRM_KTRACE_SIZE, KM_NOFS);
 #endif
 	return 0;
 }

fs/xfs/xfs_ialloc_btree.c

@@ -181,7 +181,7 @@ xfs_inobt_delrec(
 		 * then we can get rid of this level.
 		 */
 		if (numrecs == 1 && level > 0) {
-			agbp = cur->bc_private.i.agbp;
+			agbp = cur->bc_private.a.agbp;
 			agi = XFS_BUF_TO_AGI(agbp);
 			/*
 			 * pp is still set to the first pointer in the block.
@@ -194,7 +194,7 @@ xfs_inobt_delrec(
 			 * Free the block.
 			 */
 			if ((error = xfs_free_extent(cur->bc_tp,
-				XFS_AGB_TO_FSB(mp, cur->bc_private.i.agno, bno), 1)))
+				XFS_AGB_TO_FSB(mp, cur->bc_private.a.agno, bno), 1)))
 				return error;
 			xfs_trans_binval(cur->bc_tp, bp);
 			xfs_ialloc_log_agi(cur->bc_tp, agbp,
@@ -379,7 +379,7 @@ xfs_inobt_delrec(
 		rrecs = be16_to_cpu(right->bb_numrecs);
 		rbp = bp;
 		if ((error = xfs_btree_read_bufs(mp, cur->bc_tp,
-				cur->bc_private.i.agno, lbno, 0, &lbp,
+				cur->bc_private.a.agno, lbno, 0, &lbp,
 				XFS_INO_BTREE_REF)))
 			return error;
 		left = XFS_BUF_TO_INOBT_BLOCK(lbp);
@@ -401,7 +401,7 @@ xfs_inobt_delrec(
 		lrecs = be16_to_cpu(left->bb_numrecs);
 		lbp = bp;
 		if ((error = xfs_btree_read_bufs(mp, cur->bc_tp,
-				cur->bc_private.i.agno, rbno, 0, &rbp,
+				cur->bc_private.a.agno, rbno, 0, &rbp,
 				XFS_INO_BTREE_REF)))
 			return error;
 		right = XFS_BUF_TO_INOBT_BLOCK(rbp);
@@ -484,7 +484,7 @@ xfs_inobt_delrec(
 		xfs_buf_t		*rrbp;
 
 		if ((error = xfs_btree_read_bufs(mp, cur->bc_tp,
-				cur->bc_private.i.agno, be32_to_cpu(left->bb_rightsib), 0,
+				cur->bc_private.a.agno, be32_to_cpu(left->bb_rightsib), 0,
 				&rrbp, XFS_INO_BTREE_REF)))
 			return error;
 		rrblock = XFS_BUF_TO_INOBT_BLOCK(rrbp);
@@ -497,7 +497,7 @@ xfs_inobt_delrec(
 	 * Free the deleting block.
 	 */
 	if ((error = xfs_free_extent(cur->bc_tp, XFS_AGB_TO_FSB(mp,
-				     cur->bc_private.i.agno, rbno), 1)))
+				     cur->bc_private.a.agno, rbno), 1)))
 		return error;
 	xfs_trans_binval(cur->bc_tp, rbp);
 	/*
@@ -854,7 +854,7 @@ xfs_inobt_lookup(
 	{
 		xfs_agi_t	*agi;	/* a.g. inode header */
 
-		agi = XFS_BUF_TO_AGI(cur->bc_private.i.agbp);
+		agi = XFS_BUF_TO_AGI(cur->bc_private.a.agbp);
 		agno = be32_to_cpu(agi->agi_seqno);
 		agbno = be32_to_cpu(agi->agi_root);
 	}
@@ -1089,7 +1089,7 @@ xfs_inobt_lshift(
 	 * Set up the left neighbor as "left".
 	 */
 	if ((error = xfs_btree_read_bufs(cur->bc_mp, cur->bc_tp,
-			cur->bc_private.i.agno, be32_to_cpu(right->bb_leftsib),
+			cur->bc_private.a.agno, be32_to_cpu(right->bb_leftsib),
 			0, &lbp, XFS_INO_BTREE_REF)))
 		return error;
 	left = XFS_BUF_TO_INOBT_BLOCK(lbp);
@@ -1207,10 +1207,10 @@ xfs_inobt_newroot(
 	/*
 	 * Get a block & a buffer.
 	 */
-	agi = XFS_BUF_TO_AGI(cur->bc_private.i.agbp);
+	agi = XFS_BUF_TO_AGI(cur->bc_private.a.agbp);
 	args.tp = cur->bc_tp;
 	args.mp = cur->bc_mp;
-	args.fsbno = XFS_AGB_TO_FSB(args.mp, cur->bc_private.i.agno,
+	args.fsbno = XFS_AGB_TO_FSB(args.mp, cur->bc_private.a.agno,
 		be32_to_cpu(agi->agi_root));
 	args.mod = args.minleft = args.alignment = args.total = args.wasdel =
 		args.isfl = args.userdata = args.minalignslop = 0;
@@ -1233,7 +1233,7 @@ xfs_inobt_newroot(
 	 */
 	agi->agi_root = cpu_to_be32(args.agbno);
 	be32_add_cpu(&agi->agi_level, 1);
-	xfs_ialloc_log_agi(args.tp, cur->bc_private.i.agbp,
+	xfs_ialloc_log_agi(args.tp, cur->bc_private.a.agbp,
 		XFS_AGI_ROOT | XFS_AGI_LEVEL);
 	/*
 	 * At the previous root level there are now two blocks: the old
@@ -1376,7 +1376,7 @@ xfs_inobt_rshift(
 	 * Set up the right neighbor as "right".
 	 */
 	if ((error = xfs_btree_read_bufs(cur->bc_mp, cur->bc_tp,
-			cur->bc_private.i.agno, be32_to_cpu(left->bb_rightsib),
+			cur->bc_private.a.agno, be32_to_cpu(left->bb_rightsib),
 			0, &rbp, XFS_INO_BTREE_REF)))
 		return error;
 	right = XFS_BUF_TO_INOBT_BLOCK(rbp);
@@ -1492,7 +1492,7 @@ xfs_inobt_split(
 	 * Allocate the new block.
 	 * If we can't do it, we're toast.  Give up.
 	 */
-	args.fsbno = XFS_AGB_TO_FSB(args.mp, cur->bc_private.i.agno, lbno);
+	args.fsbno = XFS_AGB_TO_FSB(args.mp, cur->bc_private.a.agno, lbno);
 	args.mod = args.minleft = args.alignment = args.total = args.wasdel =
 		args.isfl = args.userdata = args.minalignslop = 0;
 	args.minlen = args.maxlen = args.prod = 1;
@@ -1725,7 +1725,7 @@ xfs_inobt_decrement(
 
 		agbno = be32_to_cpu(*XFS_INOBT_PTR_ADDR(block, cur->bc_ptrs[lev], cur));
 		if ((error = xfs_btree_read_bufs(cur->bc_mp, cur->bc_tp,
-				cur->bc_private.i.agno, agbno, 0, &bp,
+				cur->bc_private.a.agno, agbno, 0, &bp,
 				XFS_INO_BTREE_REF)))
 			return error;
 		lev--;
@@ -1897,7 +1897,7 @@ xfs_inobt_increment(
 
 		agbno = be32_to_cpu(*XFS_INOBT_PTR_ADDR(block, cur->bc_ptrs[lev], cur));
 		if ((error = xfs_btree_read_bufs(cur->bc_mp, cur->bc_tp,
-				cur->bc_private.i.agno, agbno, 0, &bp,
+				cur->bc_private.a.agno, agbno, 0, &bp,
 				XFS_INO_BTREE_REF)))
 			return error;
 		lev--;

fs/xfs/xfs_iget.c

@@ -216,7 +216,14 @@ finish_inode:
 	mrlock_init(&ip->i_iolock, MRLOCK_BARRIER, "xfsio", ip->i_ino);
 	init_waitqueue_head(&ip->i_ipin_wait);
 	atomic_set(&ip->i_pincount, 0);
-	initnsema(&ip->i_flock, 1, "xfsfino");
+
+	/*
+	 * Because we want to use a counting completion, complete
+	 * the flush completion once to allow a single access to
+	 * the flush completion without blocking.
+	 */
+	init_completion(&ip->i_flush);
+	complete(&ip->i_flush);
 
 	if (lock_flags)
 		xfs_ilock(ip, lock_flags);
@@ -287,11 +294,18 @@ finish_inode:
 	xfs_iflags_set(ip, XFS_IMODIFIED);
 	*ipp = ip;
 
+	/*
+	 * Set up the Linux with the Linux inode.
+	 */
+	ip->i_vnode = inode;
+	inode->i_private = ip;
+
 	/*
 	 * If we have a real type for an on-disk inode, we can set ops(&unlock)
 	 * now.	 If it's a new inode being created, xfs_ialloc will handle it.
 	 */
-	xfs_initialize_vnode(mp, inode, ip);
+	if (ip->i_d.di_mode != 0)
+		xfs_setup_inode(ip);
 	return 0;
 }
 
@@ -411,10 +425,11 @@ xfs_iput(xfs_inode_t	*ip,
  * Special iput for brand-new inodes that are still locked
  */
 void
-xfs_iput_new(xfs_inode_t	*ip,
-	     uint		lock_flags)
+xfs_iput_new(
+	xfs_inode_t	*ip,
+	uint		lock_flags)
 {
-	struct inode	*inode = ip->i_vnode;
+	struct inode	*inode = VFS_I(ip);
 
 	xfs_itrace_entry(ip);
 
@@ -775,26 +790,3 @@ xfs_isilocked(
 }
 #endif
 
-/*
- * The following three routines simply manage the i_flock
- * semaphore embedded in the inode.  This semaphore synchronizes
- * processes attempting to flush the in-core inode back to disk.
- */
-void
-xfs_iflock(xfs_inode_t *ip)
-{
-	psema(&(ip->i_flock), PINOD|PLTWAIT);
-}
-
-int
-xfs_iflock_nowait(xfs_inode_t *ip)
-{
-	return (cpsema(&(ip->i_flock)));
-}
-
-void
-xfs_ifunlock(xfs_inode_t *ip)
-{
-	ASSERT(issemalocked(&(ip->i_flock)));
-	vsema(&(ip->i_flock));
-}

fs/xfs/xfs_inode.c

@@ -580,8 +580,8 @@ xfs_iformat_extents(
 		xfs_validate_extents(ifp, nex, XFS_EXTFMT_INODE(ip));
 		for (i = 0; i < nex; i++, dp++) {
 			xfs_bmbt_rec_host_t *ep = xfs_iext_get_ext(ifp, i);
-			ep->l0 = be64_to_cpu(get_unaligned(&dp->l0));
-			ep->l1 = be64_to_cpu(get_unaligned(&dp->l1));
+			ep->l0 = get_unaligned_be64(&dp->l0);
+			ep->l1 = get_unaligned_be64(&dp->l1);
 		}
 		XFS_BMAP_TRACE_EXLIST(ip, nex, whichfork);
 		if (whichfork != XFS_DATA_FORK ||
@@ -835,22 +835,22 @@ xfs_iread(
 	 * Do this before xfs_iformat in case it adds entries.
 	 */
 #ifdef	XFS_INODE_TRACE
-	ip->i_trace = ktrace_alloc(INODE_TRACE_SIZE, KM_SLEEP);
+	ip->i_trace = ktrace_alloc(INODE_TRACE_SIZE, KM_NOFS);
 #endif
 #ifdef XFS_BMAP_TRACE
-	ip->i_xtrace = ktrace_alloc(XFS_BMAP_KTRACE_SIZE, KM_SLEEP);
+	ip->i_xtrace = ktrace_alloc(XFS_BMAP_KTRACE_SIZE, KM_NOFS);
 #endif
 #ifdef XFS_BMBT_TRACE
-	ip->i_btrace = ktrace_alloc(XFS_BMBT_KTRACE_SIZE, KM_SLEEP);
+	ip->i_btrace = ktrace_alloc(XFS_BMBT_KTRACE_SIZE, KM_NOFS);
 #endif
 #ifdef XFS_RW_TRACE
-	ip->i_rwtrace = ktrace_alloc(XFS_RW_KTRACE_SIZE, KM_SLEEP);
+	ip->i_rwtrace = ktrace_alloc(XFS_RW_KTRACE_SIZE, KM_NOFS);
 #endif
 #ifdef XFS_ILOCK_TRACE
-	ip->i_lock_trace = ktrace_alloc(XFS_ILOCK_KTRACE_SIZE, KM_SLEEP);
+	ip->i_lock_trace = ktrace_alloc(XFS_ILOCK_KTRACE_SIZE, KM_NOFS);
 #endif
 #ifdef XFS_DIR2_TRACE
-	ip->i_dir_trace = ktrace_alloc(XFS_DIR2_KTRACE_SIZE, KM_SLEEP);
+	ip->i_dir_trace = ktrace_alloc(XFS_DIR2_KTRACE_SIZE, KM_NOFS);
 #endif
 
 	/*
@@ -1046,9 +1046,9 @@ xfs_ialloc(
 {
 	xfs_ino_t	ino;
 	xfs_inode_t	*ip;
-	bhv_vnode_t	*vp;
 	uint		flags;
 	int		error;
+	timespec_t	tv;
 
 	/*
 	 * Call the space management code to pick
@@ -1077,7 +1077,6 @@ xfs_ialloc(
 	}
 	ASSERT(ip != NULL);
 
-	vp = XFS_ITOV(ip);
 	ip->i_d.di_mode = (__uint16_t)mode;
 	ip->i_d.di_onlink = 0;
 	ip->i_d.di_nlink = nlink;
@@ -1130,7 +1129,13 @@ xfs_ialloc(
 	ip->i_size = 0;
 	ip->i_d.di_nextents = 0;
 	ASSERT(ip->i_d.di_nblocks == 0);
-	xfs_ichgtime(ip, XFS_ICHGTIME_CHG|XFS_ICHGTIME_ACC|XFS_ICHGTIME_MOD);
+
+	nanotime(&tv);
+	ip->i_d.di_mtime.t_sec = (__int32_t)tv.tv_sec;
+	ip->i_d.di_mtime.t_nsec = (__int32_t)tv.tv_nsec;
+	ip->i_d.di_atime = ip->i_d.di_mtime;
+	ip->i_d.di_ctime = ip->i_d.di_mtime;
+
 	/*
 	 * di_gen will have been taken care of in xfs_iread.
 	 */
@@ -1220,7 +1225,7 @@ xfs_ialloc(
 	xfs_trans_log_inode(tp, ip, flags);
 
 	/* now that we have an i_mode we can setup inode ops and unlock */
-	xfs_initialize_vnode(tp->t_mountp, vp, ip);
+	xfs_setup_inode(ip);
 
 	*ipp = ip;
 	return 0;
@@ -1399,7 +1404,6 @@ xfs_itruncate_start(
 	xfs_fsize_t	last_byte;
 	xfs_off_t	toss_start;
 	xfs_mount_t	*mp;
-	bhv_vnode_t	*vp;
 	int		error = 0;
 
 	ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL));
@@ -1408,7 +1412,6 @@ xfs_itruncate_start(
 	       (flags == XFS_ITRUNC_MAYBE));
 
 	mp = ip->i_mount;
-	vp = XFS_ITOV(ip);
 
 	/* wait for the completion of any pending DIOs */
 	if (new_size < ip->i_size)
@@ -1457,7 +1460,7 @@ xfs_itruncate_start(
 
 #ifdef DEBUG
 	if (new_size == 0) {
-		ASSERT(VN_CACHED(vp) == 0);
+		ASSERT(VN_CACHED(VFS_I(ip)) == 0);
 	}
 #endif
 	return error;
@@ -2630,7 +2633,6 @@ xfs_idestroy(
 		xfs_idestroy_fork(ip, XFS_ATTR_FORK);
 	mrfree(&ip->i_lock);
 	mrfree(&ip->i_iolock);
-	freesema(&ip->i_flock);
 
 #ifdef XFS_INODE_TRACE
 	ktrace_free(ip->i_trace);
@@ -3048,10 +3050,10 @@ cluster_corrupt_out:
 /*
  * xfs_iflush() will write a modified inode's changes out to the
  * inode's on disk home.  The caller must have the inode lock held
- * in at least shared mode and the inode flush semaphore must be
- * held as well.  The inode lock will still be held upon return from
+ * in at least shared mode and the inode flush completion must be
+ * active as well.  The inode lock will still be held upon return from
  * the call and the caller is free to unlock it.
- * The inode flush lock will be unlocked when the inode reaches the disk.
+ * The inode flush will be completed when the inode reaches the disk.
  * The flags indicate how the inode's buffer should be written out.
  */
 int
@@ -3070,7 +3072,7 @@ xfs_iflush(
 	XFS_STATS_INC(xs_iflush_count);
 
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED));
-	ASSERT(issemalocked(&(ip->i_flock)));
+	ASSERT(!completion_done(&ip->i_flush));
 	ASSERT(ip->i_d.di_format != XFS_DINODE_FMT_BTREE ||
 	       ip->i_d.di_nextents > ip->i_df.if_ext_max);
 
@@ -3233,7 +3235,7 @@ xfs_iflush_int(
 #endif
 
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED));
-	ASSERT(issemalocked(&(ip->i_flock)));
+	ASSERT(!completion_done(&ip->i_flush));
 	ASSERT(ip->i_d.di_format != XFS_DINODE_FMT_BTREE ||
 	       ip->i_d.di_nextents > ip->i_df.if_ext_max);
 
@@ -3465,7 +3467,6 @@ xfs_iflush_all(
 	xfs_mount_t	*mp)
 {
 	xfs_inode_t	*ip;
-	bhv_vnode_t	*vp;
 
  again:
 	XFS_MOUNT_ILOCK(mp);
@@ -3480,14 +3481,13 @@ xfs_iflush_all(
 			continue;
 		}
 
-		vp = XFS_ITOV_NULL(ip);
-		if (!vp) {
+		if (!VFS_I(ip)) {
 			XFS_MOUNT_IUNLOCK(mp);
 			xfs_finish_reclaim(ip, 0, XFS_IFLUSH_ASYNC);
 			goto again;
 		}
 
-		ASSERT(vn_count(vp) == 0);
+		ASSERT(vn_count(VFS_I(ip)) == 0);
 
 		ip = ip->i_mnext;
 	} while (ip != mp->m_inodes);
@@ -3707,7 +3707,7 @@ xfs_iext_add_indirect_multi(
 	 * (all extents past */
 	if (nex2) {
 		byte_diff = nex2 * sizeof(xfs_bmbt_rec_t);
-		nex2_ep = (xfs_bmbt_rec_t *) kmem_alloc(byte_diff, KM_SLEEP);
+		nex2_ep = (xfs_bmbt_rec_t *) kmem_alloc(byte_diff, KM_NOFS);
 		memmove(nex2_ep, &erp->er_extbuf[idx], byte_diff);
 		erp->er_extcount -= nex2;
 		xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, -nex2);
@@ -4007,8 +4007,7 @@ xfs_iext_realloc_direct(
 			ifp->if_u1.if_extents =
 				kmem_realloc(ifp->if_u1.if_extents,
 						rnew_size,
-						ifp->if_real_bytes,
-						KM_SLEEP);
+						ifp->if_real_bytes, KM_NOFS);
 		}
 		if (rnew_size > ifp->if_real_bytes) {
 			memset(&ifp->if_u1.if_extents[ifp->if_bytes /
@@ -4067,7 +4066,7 @@ xfs_iext_inline_to_direct(
 	xfs_ifork_t	*ifp,		/* inode fork pointer */
 	int		new_size)	/* number of extents in file */
 {
-	ifp->if_u1.if_extents = kmem_alloc(new_size, KM_SLEEP);
+	ifp->if_u1.if_extents = kmem_alloc(new_size, KM_NOFS);
 	memset(ifp->if_u1.if_extents, 0, new_size);
 	if (ifp->if_bytes) {
 		memcpy(ifp->if_u1.if_extents, ifp->if_u2.if_inline_ext,
@@ -4099,7 +4098,7 @@ xfs_iext_realloc_indirect(
 	} else {
 		ifp->if_u1.if_ext_irec = (xfs_ext_irec_t *)
 			kmem_realloc(ifp->if_u1.if_ext_irec,
-				new_size, size, KM_SLEEP);
+				new_size, size, KM_NOFS);
 	}
 }
 
@@ -4341,11 +4340,10 @@ xfs_iext_irec_init(
 	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
 	ASSERT(nextents <= XFS_LINEAR_EXTS);
 
-	erp = (xfs_ext_irec_t *)
-		kmem_alloc(sizeof(xfs_ext_irec_t), KM_SLEEP);
+	erp = kmem_alloc(sizeof(xfs_ext_irec_t), KM_NOFS);
 
 	if (nextents == 0) {
-		ifp->if_u1.if_extents = kmem_alloc(XFS_IEXT_BUFSZ, KM_SLEEP);
+		ifp->if_u1.if_extents = kmem_alloc(XFS_IEXT_BUFSZ, KM_NOFS);
 	} else if (!ifp->if_real_bytes) {
 		xfs_iext_inline_to_direct(ifp, XFS_IEXT_BUFSZ);
 	} else if (ifp->if_real_bytes < XFS_IEXT_BUFSZ) {
@@ -4393,7 +4391,7 @@ xfs_iext_irec_new(
 
 	/* Initialize new extent record */
 	erp = ifp->if_u1.if_ext_irec;
-	erp[erp_idx].er_extbuf = kmem_alloc(XFS_IEXT_BUFSZ, KM_SLEEP);
+	erp[erp_idx].er_extbuf = kmem_alloc(XFS_IEXT_BUFSZ, KM_NOFS);
 	ifp->if_real_bytes = nlists * XFS_IEXT_BUFSZ;
 	memset(erp[erp_idx].er_extbuf, 0, XFS_IEXT_BUFSZ);
 	erp[erp_idx].er_extcount = 0;

fs/xfs/xfs_inode.h

@@ -87,8 +87,7 @@ typedef struct xfs_ifork {
  * Flags for xfs_ichgtime().
  */
 #define	XFS_ICHGTIME_MOD	0x1	/* data fork modification timestamp */
-#define	XFS_ICHGTIME_ACC	0x2	/* data fork access timestamp */
-#define	XFS_ICHGTIME_CHG	0x4	/* inode field change timestamp */
+#define	XFS_ICHGTIME_CHG	0x2	/* inode field change timestamp */
 
 /*
  * Per-fork incore inode flags.
@@ -204,7 +203,7 @@ typedef struct xfs_inode {
 	struct xfs_inode	*i_mprev;	/* ptr to prev inode */
 	struct xfs_mount	*i_mount;	/* fs mount struct ptr */
 	struct list_head	i_reclaim;	/* reclaim list */
-	bhv_vnode_t		*i_vnode;	/* vnode backpointer */
+	struct inode		*i_vnode;	/* vnode backpointer */
 	struct xfs_dquot	*i_udquot;	/* user dquot */
 	struct xfs_dquot	*i_gdquot;	/* group dquot */
 
@@ -223,7 +222,7 @@ typedef struct xfs_inode {
 	struct xfs_inode_log_item *i_itemp;	/* logging information */
 	mrlock_t		i_lock;		/* inode lock */
 	mrlock_t		i_iolock;	/* inode IO lock */
-	sema_t			i_flock;	/* inode flush lock */
+	struct completion	i_flush;	/* inode flush completion q */
 	atomic_t		i_pincount;	/* inode pin count */
 	wait_queue_head_t	i_ipin_wait;	/* inode pinning wait queue */
 	spinlock_t		i_flags_lock;	/* inode i_flags lock */
@@ -263,6 +262,18 @@ typedef struct xfs_inode {
 #define XFS_ISIZE(ip)	(((ip)->i_d.di_mode & S_IFMT) == S_IFREG) ? \
 				(ip)->i_size : (ip)->i_d.di_size;
 
+/* Convert from vfs inode to xfs inode */
+static inline struct xfs_inode *XFS_I(struct inode *inode)
+{
+	return (struct xfs_inode *)inode->i_private;
+}
+
+/* convert from xfs inode to vfs inode */
+static inline struct inode *VFS_I(struct xfs_inode *ip)
+{
+	return (struct inode *)ip->i_vnode;
+}
+
 /*
  * i_flags helper functions
  */
@@ -439,9 +450,6 @@ xfs_iflags_test_and_clear(xfs_inode_t *ip, unsigned short flags)
 #define	XFS_ITRUNC_DEFINITE	0x1
 #define	XFS_ITRUNC_MAYBE	0x2
 
-#define	XFS_ITOV(ip)		((ip)->i_vnode)
-#define	XFS_ITOV_NULL(ip)	((ip)->i_vnode)
-
 /*
  * For multiple groups support: if S_ISGID bit is set in the parent
  * directory, group of new file is set to that of the parent, and
@@ -473,11 +481,8 @@ int		xfs_ilock_nowait(xfs_inode_t *, uint);
 void		xfs_iunlock(xfs_inode_t *, uint);
 void		xfs_ilock_demote(xfs_inode_t *, uint);
 int		xfs_isilocked(xfs_inode_t *, uint);
-void		xfs_iflock(xfs_inode_t *);
-int		xfs_iflock_nowait(xfs_inode_t *);
 uint		xfs_ilock_map_shared(xfs_inode_t *);
 void		xfs_iunlock_map_shared(xfs_inode_t *, uint);
-void		xfs_ifunlock(xfs_inode_t *);
 void		xfs_ireclaim(xfs_inode_t *);
 int		xfs_finish_reclaim(xfs_inode_t *, int, int);
 int		xfs_finish_reclaim_all(struct xfs_mount *, int);
@@ -522,6 +527,7 @@ void		xfs_iflush_all(struct xfs_mount *);
 void		xfs_ichgtime(xfs_inode_t *, int);
 xfs_fsize_t	xfs_file_last_byte(xfs_inode_t *);
 void		xfs_lock_inodes(xfs_inode_t **, int, uint);
+void		xfs_lock_two_inodes(xfs_inode_t *, xfs_inode_t *, uint);
 
 void		xfs_synchronize_atime(xfs_inode_t *);
 void		xfs_mark_inode_dirty_sync(xfs_inode_t *);
@@ -570,6 +576,26 @@ extern struct kmem_zone	*xfs_ifork_zone;
 extern struct kmem_zone	*xfs_inode_zone;
 extern struct kmem_zone	*xfs_ili_zone;
 
+/*
+ * Manage the i_flush queue embedded in the inode.  This completion
+ * queue synchronizes processes attempting to flush the in-core
+ * inode back to disk.
+ */
+static inline void xfs_iflock(xfs_inode_t *ip)
+{
+	wait_for_completion(&ip->i_flush);
+}
+
+static inline int xfs_iflock_nowait(xfs_inode_t *ip)
+{
+	return try_wait_for_completion(&ip->i_flush);
+}
+
+static inline void xfs_ifunlock(xfs_inode_t *ip)
+{
+	complete(&ip->i_flush);
+}
+
 #endif	/* __KERNEL__ */
 
 #endif	/* __XFS_INODE_H__ */

fs/xfs/xfs_inode_item.c

@@ -779,11 +779,10 @@ xfs_inode_item_pushbuf(
 	ASSERT(iip->ili_push_owner == current_pid());
 
 	/*
-	 * If flushlock isn't locked anymore, chances are that the
-	 * inode flush completed and the inode was taken off the AIL.
-	 * So, just get out.
+	 * If a flush is not in progress anymore, chances are that the
+	 * inode was taken off the AIL. So, just get out.
 	 */
-	if (!issemalocked(&(ip->i_flock)) ||
+	if (completion_done(&ip->i_flush) ||
 	    ((iip->ili_item.li_flags & XFS_LI_IN_AIL) == 0)) {
 		iip->ili_pushbuf_flag = 0;
 		xfs_iunlock(ip, XFS_ILOCK_SHARED);
@@ -805,7 +804,7 @@ xfs_inode_item_pushbuf(
 			 * If not, we can flush it async.
 			 */
 			dopush = ((iip->ili_item.li_flags & XFS_LI_IN_AIL) &&
-				  issemalocked(&(ip->i_flock)));
+				  !completion_done(&ip->i_flush));
 			iip->ili_pushbuf_flag = 0;
 			xfs_iunlock(ip, XFS_ILOCK_SHARED);
 			xfs_buftrace("INODE ITEM PUSH", bp);
@@ -858,7 +857,7 @@ xfs_inode_item_push(
 	ip = iip->ili_inode;
 
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_SHARED));
-	ASSERT(issemalocked(&(ip->i_flock)));
+	ASSERT(!completion_done(&ip->i_flush));
 	/*
 	 * Since we were able to lock the inode's flush lock and
 	 * we found it on the AIL, the inode must be dirty.  This

fs/xfs/xfs_itable.c

@@ -59,7 +59,6 @@ xfs_bulkstat_one_iget(
 {
 	xfs_icdinode_t	*dic;	/* dinode core info pointer */
 	xfs_inode_t	*ip;		/* incore inode pointer */
-	bhv_vnode_t	*vp;
 	int		error;
 
 	error = xfs_iget(mp, NULL, ino,
@@ -72,7 +71,6 @@ xfs_bulkstat_one_iget(
 	ASSERT(ip != NULL);
 	ASSERT(ip->i_blkno != (xfs_daddr_t)0);
 
-	vp = XFS_ITOV(ip);
 	dic = &ip->i_d;
 
 	/* xfs_iget returns the following without needing
@@ -85,7 +83,7 @@ xfs_bulkstat_one_iget(
 	buf->bs_uid = dic->di_uid;
 	buf->bs_gid = dic->di_gid;
 	buf->bs_size = dic->di_size;
-	vn_atime_to_bstime(vp, &buf->bs_atime);
+	vn_atime_to_bstime(VFS_I(ip), &buf->bs_atime);
 	buf->bs_mtime.tv_sec = dic->di_mtime.t_sec;
 	buf->bs_mtime.tv_nsec = dic->di_mtime.t_nsec;
 	buf->bs_ctime.tv_sec = dic->di_ctime.t_sec;

fs/xfs/xfs_log.c

@@ -160,7 +160,7 @@ void
 xlog_trace_iclog(xlog_in_core_t *iclog, uint state)
 {
 	if (!iclog->ic_trace)
-		iclog->ic_trace = ktrace_alloc(256, KM_SLEEP);
+		iclog->ic_trace = ktrace_alloc(256, KM_NOFS);
 	ktrace_enter(iclog->ic_trace,
 		     (void *)((unsigned long)state),
 		     (void *)((unsigned long)current_pid()),
@@ -336,15 +336,12 @@ xfs_log_done(xfs_mount_t	*mp,
 	} else {
 		xlog_trace_loggrant(log, ticket, "xfs_log_done: (permanent)");
 		xlog_regrant_reserve_log_space(log, ticket);
-	}
-
-	/* If this ticket was a permanent reservation and we aren't
-	 * trying to release it, reset the inited flags; so next time
-	 * we write, a start record will be written out.
-	 */
-	if ((ticket->t_flags & XLOG_TIC_PERM_RESERV) &&
-	    (flags & XFS_LOG_REL_PERM_RESERV) == 0)
+		/* If this ticket was a permanent reservation and we aren't
+		 * trying to release it, reset the inited flags; so next time
+		 * we write, a start record will be written out.
+		 */
 		ticket->t_flags |= XLOG_TIC_INITED;
+	}
 
 	return lsn;
 }	/* xfs_log_done */
@@ -357,11 +354,11 @@ xfs_log_done(xfs_mount_t	*mp,
  * Asynchronous forces are implemented by setting the WANT_SYNC
  * bit in the appropriate in-core log and then returning.
  *
- * Synchronous forces are implemented with a semaphore.  All callers
- * to force a given lsn to disk will wait on a semaphore attached to the
+ * Synchronous forces are implemented with a signal variable. All callers
+ * to force a given lsn to disk will wait on a the sv attached to the
  * specific in-core log.  When given in-core log finally completes its
  * write to disk, that thread will wake up all threads waiting on the
- * semaphore.
+ * sv.
  */
 int
 _xfs_log_force(
@@ -588,12 +585,12 @@ error:
  * mp		- ubiquitous xfs mount point structure
  */
 int
-xfs_log_mount_finish(xfs_mount_t *mp, int mfsi_flags)
+xfs_log_mount_finish(xfs_mount_t *mp)
 {
 	int	error;
 
 	if (!(mp->m_flags & XFS_MOUNT_NORECOVERY))
-		error = xlog_recover_finish(mp->m_log, mfsi_flags);
+		error = xlog_recover_finish(mp->m_log);
 	else {
 		error = 0;
 		ASSERT(mp->m_flags & XFS_MOUNT_RDONLY);
@@ -707,7 +704,7 @@ xfs_log_unmount_write(xfs_mount_t *mp)
 		if (!(iclog->ic_state == XLOG_STATE_ACTIVE ||
 		      iclog->ic_state == XLOG_STATE_DIRTY)) {
 			if (!XLOG_FORCED_SHUTDOWN(log)) {
-				sv_wait(&iclog->ic_forcesema, PMEM,
+				sv_wait(&iclog->ic_force_wait, PMEM,
 					&log->l_icloglock, s);
 			} else {
 				spin_unlock(&log->l_icloglock);
@@ -748,7 +745,7 @@ xfs_log_unmount_write(xfs_mount_t *mp)
 			|| iclog->ic_state == XLOG_STATE_DIRTY
 			|| iclog->ic_state == XLOG_STATE_IOERROR) ) {
 
-				sv_wait(&iclog->ic_forcesema, PMEM,
+				sv_wait(&iclog->ic_force_wait, PMEM,
 					&log->l_icloglock, s);
 		} else {
 			spin_unlock(&log->l_icloglock);
@@ -838,7 +835,7 @@ xfs_log_move_tail(xfs_mount_t	*mp,
 				break;
 			tail_lsn = 0;
 			free_bytes -= tic->t_unit_res;
-			sv_signal(&tic->t_sema);
+			sv_signal(&tic->t_wait);
 			tic = tic->t_next;
 		} while (tic != log->l_write_headq);
 	}
@@ -859,7 +856,7 @@ xfs_log_move_tail(xfs_mount_t	*mp,
 				break;
 			tail_lsn = 0;
 			free_bytes -= need_bytes;
-			sv_signal(&tic->t_sema);
+			sv_signal(&tic->t_wait);
 			tic = tic->t_next;
 		} while (tic != log->l_reserve_headq);
 	}
@@ -1285,8 +1282,8 @@ xlog_alloc_log(xfs_mount_t	*mp,
 
 		ASSERT(XFS_BUF_ISBUSY(iclog->ic_bp));
 		ASSERT(XFS_BUF_VALUSEMA(iclog->ic_bp) <= 0);
-		sv_init(&iclog->ic_forcesema, SV_DEFAULT, "iclog-force");
-		sv_init(&iclog->ic_writesema, SV_DEFAULT, "iclog-write");
+		sv_init(&iclog->ic_force_wait, SV_DEFAULT, "iclog-force");
+		sv_init(&iclog->ic_write_wait, SV_DEFAULT, "iclog-write");
 
 		iclogp = &iclog->ic_next;
 	}
@@ -1565,8 +1562,8 @@ xlog_dealloc_log(xlog_t *log)
 
 	iclog = log->l_iclog;
 	for (i=0; i<log->l_iclog_bufs; i++) {
-		sv_destroy(&iclog->ic_forcesema);
-		sv_destroy(&iclog->ic_writesema);
+		sv_destroy(&iclog->ic_force_wait);
+		sv_destroy(&iclog->ic_write_wait);
 		xfs_buf_free(iclog->ic_bp);
 #ifdef XFS_LOG_TRACE
 		if (iclog->ic_trace != NULL) {
@@ -1976,7 +1973,7 @@ xlog_write(xfs_mount_t *	mp,
 /* Clean iclogs starting from the head.  This ordering must be
  * maintained, so an iclog doesn't become ACTIVE beyond one that
  * is SYNCING.  This is also required to maintain the notion that we use
- * a counting semaphore to hold off would be writers to the log when every
+ * a ordered wait queue to hold off would be writers to the log when every
  * iclog is trying to sync to disk.
  *
  * State Change: DIRTY -> ACTIVE
@@ -2240,7 +2237,7 @@ xlog_state_do_callback(
 			xlog_state_clean_log(log);
 
 			/* wake up threads waiting in xfs_log_force() */
-			sv_broadcast(&iclog->ic_forcesema);
+			sv_broadcast(&iclog->ic_force_wait);
 
 			iclog = iclog->ic_next;
 		} while (first_iclog != iclog);
@@ -2302,8 +2299,7 @@ xlog_state_do_callback(
  * the second completion goes through.
  *
  * Callbacks could take time, so they are done outside the scope of the
- * global state machine log lock.  Assume that the calls to cvsema won't
- * take a long time.  At least we know it won't sleep.
+ * global state machine log lock.
  */
 STATIC void
 xlog_state_done_syncing(
@@ -2339,7 +2335,7 @@ xlog_state_done_syncing(
 	 * iclog buffer, we wake them all, one will get to do the
 	 * I/O, the others get to wait for the result.
 	 */
-	sv_broadcast(&iclog->ic_writesema);
+	sv_broadcast(&iclog->ic_write_wait);
 	spin_unlock(&log->l_icloglock);
 	xlog_state_do_callback(log, aborted, iclog);	/* also cleans log */
 }	/* xlog_state_done_syncing */
@@ -2347,11 +2343,9 @@ xlog_state_done_syncing(
 
 /*
  * If the head of the in-core log ring is not (ACTIVE or DIRTY), then we must
- * sleep.  The flush semaphore is set to the number of in-core buffers and
- * decremented around disk syncing.  Therefore, if all buffers are syncing,
- * this semaphore will cause new writes to sleep until a sync completes.
- * Otherwise, this code just does p() followed by v().  This approximates
- * a sleep/wakeup except we can't race.
+ * sleep.  We wait on the flush queue on the head iclog as that should be
+ * the first iclog to complete flushing. Hence if all iclogs are syncing,
+ * we will wait here and all new writes will sleep until a sync completes.
  *
  * The in-core logs are used in a circular fashion. They are not used
  * out-of-order even when an iclog past the head is free.
@@ -2508,7 +2502,7 @@ xlog_grant_log_space(xlog_t	   *log,
 			goto error_return;
 
 		XFS_STATS_INC(xs_sleep_logspace);
-		sv_wait(&tic->t_sema, PINOD|PLTWAIT, &log->l_grant_lock, s);
+		sv_wait(&tic->t_wait, PINOD|PLTWAIT, &log->l_grant_lock, s);
 		/*
 		 * If we got an error, and the filesystem is shutting down,
 		 * we'll catch it down below. So just continue...
@@ -2534,7 +2528,7 @@ redo:
 		xlog_trace_loggrant(log, tic,
 				    "xlog_grant_log_space: sleep 2");
 		XFS_STATS_INC(xs_sleep_logspace);
-		sv_wait(&tic->t_sema, PINOD|PLTWAIT, &log->l_grant_lock, s);
+		sv_wait(&tic->t_wait, PINOD|PLTWAIT, &log->l_grant_lock, s);
 
 		if (XLOG_FORCED_SHUTDOWN(log)) {
 			spin_lock(&log->l_grant_lock);
@@ -2633,7 +2627,7 @@ xlog_regrant_write_log_space(xlog_t	   *log,
 			if (free_bytes < ntic->t_unit_res)
 				break;
 			free_bytes -= ntic->t_unit_res;
-			sv_signal(&ntic->t_sema);
+			sv_signal(&ntic->t_wait);
 			ntic = ntic->t_next;
 		} while (ntic != log->l_write_headq);
 
@@ -2644,7 +2638,7 @@ xlog_regrant_write_log_space(xlog_t	   *log,
 			xlog_trace_loggrant(log, tic,
 				    "xlog_regrant_write_log_space: sleep 1");
 			XFS_STATS_INC(xs_sleep_logspace);
-			sv_wait(&tic->t_sema, PINOD|PLTWAIT,
+			sv_wait(&tic->t_wait, PINOD|PLTWAIT,
 				&log->l_grant_lock, s);
 
 			/* If we're shutting down, this tic is already
@@ -2673,7 +2667,7 @@ redo:
 		if ((tic->t_flags & XLOG_TIC_IN_Q) == 0)
 			xlog_ins_ticketq(&log->l_write_headq, tic);
 		XFS_STATS_INC(xs_sleep_logspace);
-		sv_wait(&tic->t_sema, PINOD|PLTWAIT, &log->l_grant_lock, s);
+		sv_wait(&tic->t_wait, PINOD|PLTWAIT, &log->l_grant_lock, s);
 
 		/* If we're shutting down, this tic is already off the queue */
 		if (XLOG_FORCED_SHUTDOWN(log)) {
@@ -2916,7 +2910,7 @@ xlog_state_switch_iclogs(xlog_t		*log,
  *	2. the current iclog is drity, and the previous iclog is in the
  *		active or dirty state.
  *
- * We may sleep (call psema) if:
+ * We may sleep if:
  *
  *	1. the current iclog is not in the active nor dirty state.
  *	2. the current iclog dirty, and the previous iclog is not in the
@@ -3013,7 +3007,7 @@ maybe_sleep:
 			return XFS_ERROR(EIO);
 		}
 		XFS_STATS_INC(xs_log_force_sleep);
-		sv_wait(&iclog->ic_forcesema, PINOD, &log->l_icloglock, s);
+		sv_wait(&iclog->ic_force_wait, PINOD, &log->l_icloglock, s);
 		/*
 		 * No need to grab the log lock here since we're
 		 * only deciding whether or not to return EIO
@@ -3096,7 +3090,7 @@ try_again:
 						 XLOG_STATE_SYNCING))) {
 			ASSERT(!(iclog->ic_state & XLOG_STATE_IOERROR));
 			XFS_STATS_INC(xs_log_force_sleep);
-			sv_wait(&iclog->ic_prev->ic_writesema, PSWP,
+			sv_wait(&iclog->ic_prev->ic_write_wait, PSWP,
 				&log->l_icloglock, s);
 			*log_flushed = 1;
 			already_slept = 1;
@@ -3116,7 +3110,7 @@ try_again:
 	    !(iclog->ic_state & (XLOG_STATE_ACTIVE | XLOG_STATE_DIRTY))) {
 
 		/*
-		 * Don't wait on the forcesema if we know that we've
+		 * Don't wait on completion if we know that we've
 		 * gotten a log write error.
 		 */
 		if (iclog->ic_state & XLOG_STATE_IOERROR) {
@@ -3124,7 +3118,7 @@ try_again:
 			return XFS_ERROR(EIO);
 		}
 		XFS_STATS_INC(xs_log_force_sleep);
-		sv_wait(&iclog->ic_forcesema, PSWP, &log->l_icloglock, s);
+		sv_wait(&iclog->ic_force_wait, PSWP, &log->l_icloglock, s);
 		/*
 		 * No need to grab the log lock here since we're
 		 * only deciding whether or not to return EIO
@@ -3180,7 +3174,7 @@ STATIC void
 xlog_ticket_put(xlog_t		*log,
 		xlog_ticket_t	*ticket)
 {
-	sv_destroy(&ticket->t_sema);
+	sv_destroy(&ticket->t_wait);
 	kmem_zone_free(xfs_log_ticket_zone, ticket);
 }	/* xlog_ticket_put */
 
@@ -3270,7 +3264,7 @@ xlog_ticket_get(xlog_t		*log,
 	tic->t_trans_type	= 0;
 	if (xflags & XFS_LOG_PERM_RESERV)
 		tic->t_flags |= XLOG_TIC_PERM_RESERV;
-	sv_init(&(tic->t_sema), SV_DEFAULT, "logtick");
+	sv_init(&(tic->t_wait), SV_DEFAULT, "logtick");
 
 	xlog_tic_reset_res(tic);
 
@@ -3557,14 +3551,14 @@ xfs_log_force_umount(
 	 */
 	if ((tic = log->l_reserve_headq)) {
 		do {
-			sv_signal(&tic->t_sema);
+			sv_signal(&tic->t_wait);
 			tic = tic->t_next;
 		} while (tic != log->l_reserve_headq);
 	}
 
 	if ((tic = log->l_write_headq)) {
 		do {
-			sv_signal(&tic->t_sema);
+			sv_signal(&tic->t_wait);
 			tic = tic->t_next;
 		} while (tic != log->l_write_headq);
 	}

fs/xfs/xfs_log.h

@@ -149,7 +149,7 @@ int	  xfs_log_mount(struct xfs_mount	*mp,
 			struct xfs_buftarg	*log_target,
 			xfs_daddr_t		start_block,
 			int		 	num_bblocks);
-int	  xfs_log_mount_finish(struct xfs_mount *mp, int);
+int	  xfs_log_mount_finish(struct xfs_mount *mp);
 void	  xfs_log_move_tail(struct xfs_mount	*mp,
 			    xfs_lsn_t		tail_lsn);
 int	  xfs_log_notify(struct xfs_mount	*mp,

fs/xfs/xfs_log_priv.h

@@ -241,7 +241,7 @@ typedef struct xlog_res {
 } xlog_res_t;
 
 typedef struct xlog_ticket {
-	sv_t		   t_sema;	 /* sleep on this semaphore      : 20 */
+	sv_t		   t_wait;	 /* ticket wait queue            : 20 */
 	struct xlog_ticket *t_next;	 /*			         :4|8 */
 	struct xlog_ticket *t_prev;	 /*				 :4|8 */
 	xlog_tid_t	   t_tid;	 /* transaction identifier	 : 4  */
@@ -314,7 +314,7 @@ typedef struct xlog_rec_ext_header {
  *	xlog_rec_header_t into the reserved space.
  * - ic_data follows, so a write to disk can start at the beginning of
  *	the iclog.
- * - ic_forcesema is used to implement synchronous forcing of the iclog to disk.
+ * - ic_forcewait is used to implement synchronous forcing of the iclog to disk.
  * - ic_next is the pointer to the next iclog in the ring.
  * - ic_bp is a pointer to the buffer used to write this incore log to disk.
  * - ic_log is a pointer back to the global log structure.
@@ -339,8 +339,8 @@ typedef struct xlog_rec_ext_header {
  * and move everything else out to subsequent cachelines.
  */
 typedef struct xlog_iclog_fields {
-	sv_t			ic_forcesema;
-	sv_t			ic_writesema;
+	sv_t			ic_force_wait;
+	sv_t			ic_write_wait;
 	struct xlog_in_core	*ic_next;
 	struct xlog_in_core	*ic_prev;
 	struct xfs_buf		*ic_bp;
@@ -377,8 +377,8 @@ typedef struct xlog_in_core {
 /*
  * Defines to save our code from this glop.
  */
-#define	ic_forcesema	hic_fields.ic_forcesema
-#define ic_writesema	hic_fields.ic_writesema
+#define	ic_force_wait	hic_fields.ic_force_wait
+#define ic_write_wait	hic_fields.ic_write_wait
 #define	ic_next		hic_fields.ic_next
 #define	ic_prev		hic_fields.ic_prev
 #define	ic_bp		hic_fields.ic_bp
@@ -468,7 +468,7 @@ extern int	 xlog_find_tail(xlog_t	*log,
 				xfs_daddr_t *head_blk,
 				xfs_daddr_t *tail_blk);
 extern int	 xlog_recover(xlog_t *log);
-extern int	 xlog_recover_finish(xlog_t *log, int mfsi_flags);
+extern int	 xlog_recover_finish(xlog_t *log);
 extern void	 xlog_pack_data(xlog_t *log, xlog_in_core_t *iclog, int);
 extern void	 xlog_recover_process_iunlinks(xlog_t *log);
 

fs/xfs/xfs_log_recover.c

@@ -3940,8 +3940,7 @@ xlog_recover(
  */
 int
 xlog_recover_finish(
-	xlog_t		*log,
-	int		mfsi_flags)
+	xlog_t		*log)
 {
 	/*
 	 * Now we're ready to do the transactions needed for the
@@ -3969,9 +3968,7 @@ xlog_recover_finish(
 		xfs_log_force(log->l_mp, (xfs_lsn_t)0,
 			      (XFS_LOG_FORCE | XFS_LOG_SYNC));
 
-		if ( (mfsi_flags & XFS_MFSI_NOUNLINK) == 0 ) {
-			xlog_recover_process_iunlinks(log);
-		}
+		xlog_recover_process_iunlinks(log);
 
 		xlog_recover_check_summary(log);
 

fs/xfs/xfs_mount.c

@@ -128,7 +128,7 @@ static const struct {
  * initialized.
  */
 STATIC void
-xfs_mount_free(
+xfs_free_perag(
 	xfs_mount_t	*mp)
 {
 	if (mp->m_perag) {
@@ -139,20 +139,6 @@ xfs_mount_free(
 				kmem_free(mp->m_perag[agno].pagb_list);
 		kmem_free(mp->m_perag);
 	}
-
-	spinlock_destroy(&mp->m_ail_lock);
-	spinlock_destroy(&mp->m_sb_lock);
-	mutex_destroy(&mp->m_ilock);
-	mutex_destroy(&mp->m_growlock);
-	if (mp->m_quotainfo)
-		XFS_QM_DONE(mp);
-
-	if (mp->m_fsname != NULL)
-		kmem_free(mp->m_fsname);
-	if (mp->m_rtname != NULL)
-		kmem_free(mp->m_rtname);
-	if (mp->m_logname != NULL)
-		kmem_free(mp->m_logname);
 }
 
 /*
@@ -704,11 +690,11 @@ xfs_initialize_perag_data(xfs_mount_t *mp, xfs_agnumber_t agcount)
  * Update alignment values based on mount options and sb values
  */
 STATIC int
-xfs_update_alignment(xfs_mount_t *mp, int mfsi_flags, __uint64_t *update_flags)
+xfs_update_alignment(xfs_mount_t *mp, __uint64_t *update_flags)
 {
 	xfs_sb_t	*sbp = &(mp->m_sb);
 
-	if (mp->m_dalign && !(mfsi_flags & XFS_MFSI_SECOND)) {
+	if (mp->m_dalign) {
 		/*
 		 * If stripe unit and stripe width are not multiples
 		 * of the fs blocksize turn off alignment.
@@ -864,7 +850,7 @@ xfs_set_inoalignment(xfs_mount_t *mp)
  * Check that the data (and log if separate) are an ok size.
  */
 STATIC int
-xfs_check_sizes(xfs_mount_t *mp, int mfsi_flags)
+xfs_check_sizes(xfs_mount_t *mp)
 {
 	xfs_buf_t	*bp;
 	xfs_daddr_t	d;
@@ -887,8 +873,7 @@ xfs_check_sizes(xfs_mount_t *mp, int mfsi_flags)
 		return error;
 	}
 
-	if (((mfsi_flags & XFS_MFSI_CLIENT) == 0) &&
-	    mp->m_logdev_targp != mp->m_ddev_targp) {
+	if (mp->m_logdev_targp != mp->m_ddev_targp) {
 		d = (xfs_daddr_t)XFS_FSB_TO_BB(mp, mp->m_sb.sb_logblocks);
 		if (XFS_BB_TO_FSB(mp, d) != mp->m_sb.sb_logblocks) {
 			cmn_err(CE_WARN, "XFS: size check 3 failed");
@@ -923,15 +908,13 @@ xfs_check_sizes(xfs_mount_t *mp, int mfsi_flags)
  */
 int
 xfs_mountfs(
-	xfs_mount_t	*mp,
-	int		mfsi_flags)
+	xfs_mount_t	*mp)
 {
 	xfs_sb_t	*sbp = &(mp->m_sb);
 	xfs_inode_t	*rip;
 	__uint64_t	resblks;
 	__int64_t	update_flags = 0LL;
 	uint		quotamount, quotaflags;
-	int		agno;
 	int		uuid_mounted = 0;
 	int		error = 0;
 
@@ -985,7 +968,7 @@ xfs_mountfs(
 	 * allocator alignment is within an ag, therefore ag has
 	 * to be aligned at stripe boundary.
 	 */
-	error = xfs_update_alignment(mp, mfsi_flags, &update_flags);
+	error = xfs_update_alignment(mp, &update_flags);
 	if (error)
 		goto error1;
 
@@ -1004,8 +987,7 @@ xfs_mountfs(
 	 * since a single partition filesystem is identical to a single
 	 * partition volume/filesystem.
 	 */
-	if ((mfsi_flags & XFS_MFSI_SECOND) == 0 &&
-	    (mp->m_flags & XFS_MOUNT_NOUUID) == 0) {
+	if ((mp->m_flags & XFS_MOUNT_NOUUID) == 0) {
 		if (xfs_uuid_mount(mp)) {
 			error = XFS_ERROR(EINVAL);
 			goto error1;
@@ -1033,7 +1015,7 @@ xfs_mountfs(
 	/*
 	 * Check that the data (and log if separate) are an ok size.
 	 */
-	error = xfs_check_sizes(mp, mfsi_flags);
+	error = xfs_check_sizes(mp);
 	if (error)
 		goto error1;
 
@@ -1046,13 +1028,6 @@ xfs_mountfs(
 		goto error1;
 	}
 
-	/*
-	 * For client case we are done now
-	 */
-	if (mfsi_flags & XFS_MFSI_CLIENT) {
-		return 0;
-	}
-
 	/*
 	 *  Copies the low order bits of the timestamp and the randomly
 	 *  set "sequence" number out of a UUID.
@@ -1077,8 +1052,10 @@ xfs_mountfs(
 	 * Allocate and initialize the per-ag data.
 	 */
 	init_rwsem(&mp->m_peraglock);
-	mp->m_perag =
-		kmem_zalloc(sbp->sb_agcount * sizeof(xfs_perag_t), KM_SLEEP);
+	mp->m_perag = kmem_zalloc(sbp->sb_agcount * sizeof(xfs_perag_t),
+				  KM_MAYFAIL);
+	if (!mp->m_perag)
+		goto error1;
 
 	mp->m_maxagi = xfs_initialize_perag(mp, sbp->sb_agcount);
 
@@ -1190,7 +1167,7 @@ xfs_mountfs(
 	 * delayed until after the root and real-time bitmap inodes
 	 * were consistently read in.
 	 */
-	error = xfs_log_mount_finish(mp, mfsi_flags);
+	error = xfs_log_mount_finish(mp);
 	if (error) {
 		cmn_err(CE_WARN, "XFS: log mount finish failed");
 		goto error4;
@@ -1199,7 +1176,7 @@ xfs_mountfs(
 	/*
 	 * Complete the quota initialisation, post-log-replay component.
 	 */
-	error = XFS_QM_MOUNT(mp, quotamount, quotaflags, mfsi_flags);
+	error = XFS_QM_MOUNT(mp, quotamount, quotaflags);
 	if (error)
 		goto error4;
 
@@ -1233,12 +1210,7 @@ xfs_mountfs(
  error3:
 	xfs_log_unmount_dealloc(mp);
  error2:
-	for (agno = 0; agno < sbp->sb_agcount; agno++)
-		if (mp->m_perag[agno].pagb_list)
-			kmem_free(mp->m_perag[agno].pagb_list);
-	kmem_free(mp->m_perag);
-	mp->m_perag = NULL;
-	/* FALLTHROUGH */
+	xfs_free_perag(mp);
  error1:
 	if (uuid_mounted)
 		uuid_table_remove(&mp->m_sb.sb_uuid);
@@ -1246,16 +1218,17 @@ xfs_mountfs(
 }
 
 /*
- * xfs_unmountfs
- *
  * This flushes out the inodes,dquots and the superblock, unmounts the
  * log and makes sure that incore structures are freed.
  */
-int
-xfs_unmountfs(xfs_mount_t *mp)
+void
+xfs_unmountfs(
+	struct xfs_mount	*mp)
 {
-	__uint64_t	resblks;
-	int		error = 0;
+	__uint64_t		resblks;
+	int			error;
+
+	IRELE(mp->m_rootip);
 
 	/*
 	 * We can potentially deadlock here if we have an inode cluster
@@ -1312,8 +1285,6 @@ xfs_unmountfs(xfs_mount_t *mp)
 	xfs_unmountfs_wait(mp); 		/* wait for async bufs */
 	xfs_log_unmount(mp);			/* Done! No more fs ops. */
 
-	xfs_freesb(mp);
-
 	/*
 	 * All inodes from this mount point should be freed.
 	 */
@@ -1322,11 +1293,12 @@ xfs_unmountfs(xfs_mount_t *mp)
 	if ((mp->m_flags & XFS_MOUNT_NOUUID) == 0)
 		uuid_table_remove(&mp->m_sb.sb_uuid);
 
-#if defined(DEBUG) || defined(INDUCE_IO_ERROR)
+#if defined(DEBUG)
 	xfs_errortag_clearall(mp, 0);
 #endif
-	xfs_mount_free(mp);
-	return 0;
+	xfs_free_perag(mp);
+	if (mp->m_quotainfo)
+		XFS_QM_DONE(mp);
 }
 
 STATIC void

fs/xfs/xfs_mount.h

@@ -114,7 +114,7 @@ struct xfs_dqtrxops;
 struct xfs_quotainfo;
 
 typedef int	(*xfs_qminit_t)(struct xfs_mount *, uint *, uint *);
-typedef int	(*xfs_qmmount_t)(struct xfs_mount *, uint, uint, int);
+typedef int	(*xfs_qmmount_t)(struct xfs_mount *, uint, uint);
 typedef int	(*xfs_qmunmount_t)(struct xfs_mount *);
 typedef void	(*xfs_qmdone_t)(struct xfs_mount *);
 typedef void	(*xfs_dqrele_t)(struct xfs_dquot *);
@@ -158,8 +158,8 @@ typedef struct xfs_qmops {
 
 #define XFS_QM_INIT(mp, mnt, fl) \
 	(*(mp)->m_qm_ops->xfs_qminit)(mp, mnt, fl)
-#define XFS_QM_MOUNT(mp, mnt, fl, mfsi_flags) \
-	(*(mp)->m_qm_ops->xfs_qmmount)(mp, mnt, fl, mfsi_flags)
+#define XFS_QM_MOUNT(mp, mnt, fl) \
+	(*(mp)->m_qm_ops->xfs_qmmount)(mp, mnt, fl)
 #define XFS_QM_UNMOUNT(mp) \
 	(*(mp)->m_qm_ops->xfs_qmunmount)(mp)
 #define XFS_QM_DONE(mp) \
@@ -442,13 +442,6 @@ void xfs_do_force_shutdown(struct xfs_mount *mp, int flags, char *fname,
 /*
  * Flags for xfs_mountfs
  */
-#define XFS_MFSI_SECOND		0x01	/* Secondary mount -- skip stuff */
-#define XFS_MFSI_CLIENT		0x02	/* Is a client -- skip lots of stuff */
-/*	XFS_MFSI_RRINODES	*/
-#define XFS_MFSI_NOUNLINK	0x08	/* Skip unlinked inode processing in */
-					/* log recovery */
-#define XFS_MFSI_NO_QUOTACHECK	0x10	/* Skip quotacheck processing */
-/*	XFS_MFSI_CONVERT_SUNIT	*/
 #define XFS_MFSI_QUIET		0x40	/* Be silent if mount errors found */
 
 #define XFS_DADDR_TO_AGNO(mp,d)         xfs_daddr_to_agno(mp,d)
@@ -517,10 +510,10 @@ typedef struct xfs_mod_sb {
 
 extern void	xfs_mod_sb(xfs_trans_t *, __int64_t);
 extern int	xfs_log_sbcount(xfs_mount_t *, uint);
-extern int	xfs_mountfs(xfs_mount_t *mp, int);
+extern int	xfs_mountfs(xfs_mount_t *mp);
 extern void	xfs_mountfs_check_barriers(xfs_mount_t *mp);
 
-extern int	xfs_unmountfs(xfs_mount_t *);
+extern void	xfs_unmountfs(xfs_mount_t *);
 extern int	xfs_unmountfs_writesb(xfs_mount_t *);
 extern int	xfs_unmount_flush(xfs_mount_t *, int);
 extern int	xfs_mod_incore_sb(xfs_mount_t *, xfs_sb_field_t, int64_t, int);

fs/xfs/xfs_rtalloc.c

@@ -73,18 +73,6 @@ STATIC int xfs_rtmodify_summary(xfs_mount_t *, xfs_trans_t *, int,
  * Internal functions.
  */
 
-/*
- * xfs_lowbit32: get low bit set out of 32-bit argument, -1 if none set.
- */
-STATIC int
-xfs_lowbit32(
-	__uint32_t	v)
-{
-	if (v)
-		return ffs(v) - 1;
-	return -1;
-}
-
 /*
  * Allocate space to the bitmap or summary file, and zero it, for growfs.
  */
@@ -450,6 +438,7 @@ xfs_rtallocate_extent_near(
 	}
 	bbno = XFS_BITTOBLOCK(mp, bno);
 	i = 0;
+	ASSERT(minlen != 0);
 	log2len = xfs_highbit32(minlen);
 	/*
 	 * Loop over all bitmap blocks (bbno + i is current block).
@@ -618,6 +607,8 @@ xfs_rtallocate_extent_size(
 	xfs_suminfo_t	sum;		/* summary information for extents */
 
 	ASSERT(minlen % prod == 0 && maxlen % prod == 0);
+	ASSERT(maxlen != 0);
+
 	/*
 	 * Loop over all the levels starting with maxlen.
 	 * At each level, look at all the bitmap blocks, to see if there
@@ -675,6 +666,9 @@ xfs_rtallocate_extent_size(
 		*rtblock = NULLRTBLOCK;
 		return 0;
 	}
+	ASSERT(minlen != 0);
+	ASSERT(maxlen != 0);
+
 	/*
 	 * Loop over sizes, from maxlen down to minlen.
 	 * This time, when we do the allocations, allow smaller ones
@@ -1961,6 +1955,7 @@ xfs_growfs_rt(
 				  nsbp->sb_blocksize * nsbp->sb_rextsize);
 		nsbp->sb_rextents = nsbp->sb_rblocks;
 		do_div(nsbp->sb_rextents, nsbp->sb_rextsize);
+		ASSERT(nsbp->sb_rextents != 0);
 		nsbp->sb_rextslog = xfs_highbit32(nsbp->sb_rextents);
 		nrsumlevels = nmp->m_rsumlevels = nsbp->sb_rextslog + 1;
 		nrsumsize =

fs/xfs/xfs_rw.c

@@ -314,7 +314,7 @@ xfs_bioerror_relse(
 		 * ASYNC buffers.
 		 */
 		XFS_BUF_ERROR(bp, EIO);
-		XFS_BUF_V_IODONESEMA(bp);
+		XFS_BUF_FINISH_IOWAIT(bp);
 	} else {
 		xfs_buf_relse(bp);
 	}

fs/xfs/xfs_trans.c

@@ -43,6 +43,7 @@
 #include "xfs_quota.h"
 #include "xfs_trans_priv.h"
 #include "xfs_trans_space.h"
+#include "xfs_inode_item.h"
 
 
 STATIC void	xfs_trans_apply_sb_deltas(xfs_trans_t *);
@@ -253,7 +254,7 @@ _xfs_trans_alloc(
 	tp->t_mountp = mp;
 	tp->t_items_free = XFS_LIC_NUM_SLOTS;
 	tp->t_busy_free = XFS_LBC_NUM_SLOTS;
-	XFS_LIC_INIT(&(tp->t_items));
+	xfs_lic_init(&(tp->t_items));
 	XFS_LBC_INIT(&(tp->t_busy));
 	return tp;
 }
@@ -282,7 +283,7 @@ xfs_trans_dup(
 	ntp->t_mountp = tp->t_mountp;
 	ntp->t_items_free = XFS_LIC_NUM_SLOTS;
 	ntp->t_busy_free = XFS_LBC_NUM_SLOTS;
-	XFS_LIC_INIT(&(ntp->t_items));
+	xfs_lic_init(&(ntp->t_items));
 	XFS_LBC_INIT(&(ntp->t_busy));
 
 	ASSERT(tp->t_flags & XFS_TRANS_PERM_LOG_RES);
@@ -1169,7 +1170,7 @@ xfs_trans_cancel(
 		while (licp != NULL) {
 			lidp = licp->lic_descs;
 			for (i = 0; i < licp->lic_unused; i++, lidp++) {
-				if (XFS_LIC_ISFREE(licp, i)) {
+				if (xfs_lic_isfree(licp, i)) {
 					continue;
 				}
 
@@ -1216,6 +1217,68 @@ xfs_trans_free(
 	kmem_zone_free(xfs_trans_zone, tp);
 }
 
+/*
+ * Roll from one trans in the sequence of PERMANENT transactions to
+ * the next: permanent transactions are only flushed out when
+ * committed with XFS_TRANS_RELEASE_LOG_RES, but we still want as soon
+ * as possible to let chunks of it go to the log. So we commit the
+ * chunk we've been working on and get a new transaction to continue.
+ */
+int
+xfs_trans_roll(
+	struct xfs_trans	**tpp,
+	struct xfs_inode	*dp)
+{
+	struct xfs_trans	*trans;
+	unsigned int		logres, count;
+	int			error;
+
+	/*
+	 * Ensure that the inode is always logged.
+	 */
+	trans = *tpp;
+	xfs_trans_log_inode(trans, dp, XFS_ILOG_CORE);
+
+	/*
+	 * Copy the critical parameters from one trans to the next.
+	 */
+	logres = trans->t_log_res;
+	count = trans->t_log_count;
+	*tpp = xfs_trans_dup(trans);
+
+	/*
+	 * Commit the current transaction.
+	 * If this commit failed, then it'd just unlock those items that
+	 * are not marked ihold. That also means that a filesystem shutdown
+	 * is in progress. The caller takes the responsibility to cancel
+	 * the duplicate transaction that gets returned.
+	 */
+	error = xfs_trans_commit(trans, 0);
+	if (error)
+		return (error);
+
+	trans = *tpp;
+
+	/*
+	 * Reserve space in the log for th next transaction.
+	 * This also pushes items in the "AIL", the list of logged items,
+	 * out to disk if they are taking up space at the tail of the log
+	 * that we want to use.  This requires that either nothing be locked
+	 * across this call, or that anything that is locked be logged in
+	 * the prior and the next transactions.
+	 */
+	error = xfs_trans_reserve(trans, 0, logres, 0,
+				  XFS_TRANS_PERM_LOG_RES, count);
+	/*
+	 *  Ensure that the inode is in the new transaction and locked.
+	 */
+	if (error)
+		return error;
+
+	xfs_trans_ijoin(trans, dp, XFS_ILOCK_EXCL);
+	xfs_trans_ihold(trans, dp);
+	return 0;
+}
 
 /*
  * THIS SHOULD BE REWRITTEN TO USE xfs_trans_next_item().
@@ -1253,7 +1316,7 @@ xfs_trans_committed(
 	 * Special case the chunk embedded in the transaction.
 	 */
 	licp = &(tp->t_items);
-	if (!(XFS_LIC_ARE_ALL_FREE(licp))) {
+	if (!(xfs_lic_are_all_free(licp))) {
 		xfs_trans_chunk_committed(licp, tp->t_lsn, abortflag);
 	}
 
@@ -1262,7 +1325,7 @@ xfs_trans_committed(
 	 */
 	licp = licp->lic_next;
 	while (licp != NULL) {
-		ASSERT(!XFS_LIC_ARE_ALL_FREE(licp));
+		ASSERT(!xfs_lic_are_all_free(licp));
 		xfs_trans_chunk_committed(licp, tp->t_lsn, abortflag);
 		next_licp = licp->lic_next;
 		kmem_free(licp);
@@ -1325,7 +1388,7 @@ xfs_trans_chunk_committed(
 
 	lidp = licp->lic_descs;
 	for (i = 0; i < licp->lic_unused; i++, lidp++) {
-		if (XFS_LIC_ISFREE(licp, i)) {
+		if (xfs_lic_isfree(licp, i)) {
 			continue;
 		}
 

fs/xfs/xfs_trans.h

@@ -210,62 +210,52 @@ typedef struct xfs_log_item_chunk {
  * lic_unused to the right value (0 matches all free).  The
  * lic_descs.lid_index values are set up as each desc is allocated.
  */
-#define	XFS_LIC_INIT(cp)	xfs_lic_init(cp)
 static inline void xfs_lic_init(xfs_log_item_chunk_t *cp)
 {
 	cp->lic_free = XFS_LIC_FREEMASK;
 }
 
-#define	XFS_LIC_INIT_SLOT(cp,slot)	xfs_lic_init_slot(cp, slot)
 static inline void xfs_lic_init_slot(xfs_log_item_chunk_t *cp, int slot)
 {
 	cp->lic_descs[slot].lid_index = (unsigned char)(slot);
 }
 
-#define	XFS_LIC_VACANCY(cp)		xfs_lic_vacancy(cp)
 static inline int xfs_lic_vacancy(xfs_log_item_chunk_t *cp)
 {
 	return cp->lic_free & XFS_LIC_FREEMASK;
 }
 
-#define	XFS_LIC_ALL_FREE(cp)		xfs_lic_all_free(cp)
 static inline void xfs_lic_all_free(xfs_log_item_chunk_t *cp)
 {
 	cp->lic_free = XFS_LIC_FREEMASK;
 }
 
-#define	XFS_LIC_ARE_ALL_FREE(cp)	xfs_lic_are_all_free(cp)
 static inline int xfs_lic_are_all_free(xfs_log_item_chunk_t *cp)
 {
 	return ((cp->lic_free & XFS_LIC_FREEMASK) == XFS_LIC_FREEMASK);
 }
 
-#define	XFS_LIC_ISFREE(cp,slot)	xfs_lic_isfree(cp,slot)
 static inline int xfs_lic_isfree(xfs_log_item_chunk_t *cp, int slot)
 {
 	return (cp->lic_free & (1 << slot));
 }
 
-#define	XFS_LIC_CLAIM(cp,slot)		xfs_lic_claim(cp,slot)
 static inline void xfs_lic_claim(xfs_log_item_chunk_t *cp, int slot)
 {
 	cp->lic_free &= ~(1 << slot);
 }
 
-#define	XFS_LIC_RELSE(cp,slot)		xfs_lic_relse(cp,slot)
 static inline void xfs_lic_relse(xfs_log_item_chunk_t *cp, int slot)
 {
 	cp->lic_free |= 1 << slot;
 }
 
-#define	XFS_LIC_SLOT(cp,slot)		xfs_lic_slot(cp,slot)
 static inline xfs_log_item_desc_t *
 xfs_lic_slot(xfs_log_item_chunk_t *cp, int slot)
 {
 	return &(cp->lic_descs[slot]);
 }
 
-#define	XFS_LIC_DESC_TO_SLOT(dp)	xfs_lic_desc_to_slot(dp)
 static inline int xfs_lic_desc_to_slot(xfs_log_item_desc_t *dp)
 {
 	return (uint)dp->lid_index;
@@ -278,7 +268,6 @@ static inline int xfs_lic_desc_to_slot(xfs_log_item_desc_t *dp)
  * All of this yields the address of the chunk, which is
  * cast to a chunk pointer.
  */
-#define	XFS_LIC_DESC_TO_CHUNK(dp)	xfs_lic_desc_to_chunk(dp)
 static inline xfs_log_item_chunk_t *
 xfs_lic_desc_to_chunk(xfs_log_item_desc_t *dp)
 {
@@ -986,6 +975,7 @@ int		_xfs_trans_commit(xfs_trans_t *,
 				  int *);
 #define xfs_trans_commit(tp, flags)	_xfs_trans_commit(tp, flags, NULL)
 void		xfs_trans_cancel(xfs_trans_t *, int);
+int		xfs_trans_roll(struct xfs_trans **, struct xfs_inode *);
 int		xfs_trans_ail_init(struct xfs_mount *);
 void		xfs_trans_ail_destroy(struct xfs_mount *);
 void		xfs_trans_push_ail(struct xfs_mount *, xfs_lsn_t);

fs/xfs/xfs_trans_buf.c

@@ -1021,16 +1021,16 @@ xfs_trans_buf_item_match(
 	bp = NULL;
 	len = BBTOB(len);
 	licp = &tp->t_items;
-	if (!XFS_LIC_ARE_ALL_FREE(licp)) {
+	if (!xfs_lic_are_all_free(licp)) {
 		for (i = 0; i < licp->lic_unused; i++) {
 			/*
 			 * Skip unoccupied slots.
 			 */
-			if (XFS_LIC_ISFREE(licp, i)) {
+			if (xfs_lic_isfree(licp, i)) {
 				continue;
 			}
 
-			lidp = XFS_LIC_SLOT(licp, i);
+			lidp = xfs_lic_slot(licp, i);
 			blip = (xfs_buf_log_item_t *)lidp->lid_item;
 			if (blip->bli_item.li_type != XFS_LI_BUF) {
 				continue;
@@ -1074,7 +1074,7 @@ xfs_trans_buf_item_match_all(
 	bp = NULL;
 	len = BBTOB(len);
 	for (licp = &tp->t_items; licp != NULL; licp = licp->lic_next) {
-		if (XFS_LIC_ARE_ALL_FREE(licp)) {
+		if (xfs_lic_are_all_free(licp)) {
 			ASSERT(licp == &tp->t_items);
 			ASSERT(licp->lic_next == NULL);
 			return NULL;
@@ -1083,11 +1083,11 @@ xfs_trans_buf_item_match_all(
 			/*
 			 * Skip unoccupied slots.
 			 */
-			if (XFS_LIC_ISFREE(licp, i)) {
+			if (xfs_lic_isfree(licp, i)) {
 				continue;
 			}
 
-			lidp = XFS_LIC_SLOT(licp, i);
+			lidp = xfs_lic_slot(licp, i);
 			blip = (xfs_buf_log_item_t *)lidp->lid_item;
 			if (blip->bli_item.li_type != XFS_LI_BUF) {
 				continue;

fs/xfs/xfs_trans_item.c

@@ -53,11 +53,11 @@ xfs_trans_add_item(xfs_trans_t *tp, xfs_log_item_t *lip)
 		 * Initialize the chunk, and then
 		 * claim the first slot in the newly allocated chunk.
 		 */
-		XFS_LIC_INIT(licp);
-		XFS_LIC_CLAIM(licp, 0);
+		xfs_lic_init(licp);
+		xfs_lic_claim(licp, 0);
 		licp->lic_unused = 1;
-		XFS_LIC_INIT_SLOT(licp, 0);
-		lidp = XFS_LIC_SLOT(licp, 0);
+		xfs_lic_init_slot(licp, 0);
+		lidp = xfs_lic_slot(licp, 0);
 
 		/*
 		 * Link in the new chunk and update the free count.
@@ -88,14 +88,14 @@ xfs_trans_add_item(xfs_trans_t *tp, xfs_log_item_t *lip)
 	 */
 	licp = &tp->t_items;
 	while (licp != NULL) {
-		if (XFS_LIC_VACANCY(licp)) {
+		if (xfs_lic_vacancy(licp)) {
 			if (licp->lic_unused <= XFS_LIC_MAX_SLOT) {
 				i = licp->lic_unused;
-				ASSERT(XFS_LIC_ISFREE(licp, i));
+				ASSERT(xfs_lic_isfree(licp, i));
 				break;
 			}
 			for (i = 0; i <= XFS_LIC_MAX_SLOT; i++) {
-				if (XFS_LIC_ISFREE(licp, i))
+				if (xfs_lic_isfree(licp, i))
 					break;
 			}
 			ASSERT(i <= XFS_LIC_MAX_SLOT);
@@ -108,12 +108,12 @@ xfs_trans_add_item(xfs_trans_t *tp, xfs_log_item_t *lip)
 	 * If we find a free descriptor, claim it,
 	 * initialize it, and return it.
 	 */
-	XFS_LIC_CLAIM(licp, i);
+	xfs_lic_claim(licp, i);
 	if (licp->lic_unused <= i) {
 		licp->lic_unused = i + 1;
-		XFS_LIC_INIT_SLOT(licp, i);
+		xfs_lic_init_slot(licp, i);
 	}
-	lidp = XFS_LIC_SLOT(licp, i);
+	lidp = xfs_lic_slot(licp, i);
 	tp->t_items_free--;
 	lidp->lid_item = lip;
 	lidp->lid_flags = 0;
@@ -136,9 +136,9 @@ xfs_trans_free_item(xfs_trans_t	*tp, xfs_log_item_desc_t *lidp)
 	xfs_log_item_chunk_t	*licp;
 	xfs_log_item_chunk_t	**licpp;
 
-	slot = XFS_LIC_DESC_TO_SLOT(lidp);
-	licp = XFS_LIC_DESC_TO_CHUNK(lidp);
-	XFS_LIC_RELSE(licp, slot);
+	slot = xfs_lic_desc_to_slot(lidp);
+	licp = xfs_lic_desc_to_chunk(lidp);
+	xfs_lic_relse(licp, slot);
 	lidp->lid_item->li_desc = NULL;
 	tp->t_items_free++;
 
@@ -154,7 +154,7 @@ xfs_trans_free_item(xfs_trans_t	*tp, xfs_log_item_desc_t *lidp)
 	 * Also decrement the transaction structure's count of free items
 	 * by the number in a chunk since we are freeing an empty chunk.
 	 */
-	if (XFS_LIC_ARE_ALL_FREE(licp) && (licp != &(tp->t_items))) {
+	if (xfs_lic_are_all_free(licp) && (licp != &(tp->t_items))) {
 		licpp = &(tp->t_items.lic_next);
 		while (*licpp != licp) {
 			ASSERT(*licpp != NULL);
@@ -207,20 +207,20 @@ xfs_trans_first_item(xfs_trans_t *tp)
 	/*
 	 * If it's not in the first chunk, skip to the second.
 	 */
-	if (XFS_LIC_ARE_ALL_FREE(licp)) {
+	if (xfs_lic_are_all_free(licp)) {
 		licp = licp->lic_next;
 	}
 
 	/*
 	 * Return the first non-free descriptor in the chunk.
 	 */
-	ASSERT(!XFS_LIC_ARE_ALL_FREE(licp));
+	ASSERT(!xfs_lic_are_all_free(licp));
 	for (i = 0; i < licp->lic_unused; i++) {
-		if (XFS_LIC_ISFREE(licp, i)) {
+		if (xfs_lic_isfree(licp, i)) {
 			continue;
 		}
 
-		return XFS_LIC_SLOT(licp, i);
+		return xfs_lic_slot(licp, i);
 	}
 	cmn_err(CE_WARN, "xfs_trans_first_item() -- no first item");
 	return NULL;
@@ -242,18 +242,18 @@ xfs_trans_next_item(xfs_trans_t *tp, xfs_log_item_desc_t *lidp)
 	xfs_log_item_chunk_t	*licp;
 	int			i;
 
-	licp = XFS_LIC_DESC_TO_CHUNK(lidp);
+	licp = xfs_lic_desc_to_chunk(lidp);
 
 	/*
 	 * First search the rest of the chunk. The for loop keeps us
 	 * from referencing things beyond the end of the chunk.
 	 */
-	for (i = (int)XFS_LIC_DESC_TO_SLOT(lidp) + 1; i < licp->lic_unused; i++) {
-		if (XFS_LIC_ISFREE(licp, i)) {
+	for (i = (int)xfs_lic_desc_to_slot(lidp) + 1; i < licp->lic_unused; i++) {
+		if (xfs_lic_isfree(licp, i)) {
 			continue;
 		}
 
-		return XFS_LIC_SLOT(licp, i);
+		return xfs_lic_slot(licp, i);
 	}
 
 	/*
@@ -266,13 +266,13 @@ xfs_trans_next_item(xfs_trans_t *tp, xfs_log_item_desc_t *lidp)
 	}
 
 	licp = licp->lic_next;
-	ASSERT(!XFS_LIC_ARE_ALL_FREE(licp));
+	ASSERT(!xfs_lic_are_all_free(licp));
 	for (i = 0; i < licp->lic_unused; i++) {
-		if (XFS_LIC_ISFREE(licp, i)) {
+		if (xfs_lic_isfree(licp, i)) {
 			continue;
 		}
 
-		return XFS_LIC_SLOT(licp, i);
+		return xfs_lic_slot(licp, i);
 	}
 	ASSERT(0);
 	/* NOTREACHED */
@@ -300,9 +300,9 @@ xfs_trans_free_items(
 	/*
 	 * Special case the embedded chunk so we don't free it below.
 	 */
-	if (!XFS_LIC_ARE_ALL_FREE(licp)) {
+	if (!xfs_lic_are_all_free(licp)) {
 		(void) xfs_trans_unlock_chunk(licp, 1, abort, NULLCOMMITLSN);
-		XFS_LIC_ALL_FREE(licp);
+		xfs_lic_all_free(licp);
 		licp->lic_unused = 0;
 	}
 	licp = licp->lic_next;
@@ -311,7 +311,7 @@ xfs_trans_free_items(
 	 * Unlock each item in each chunk and free the chunks.
 	 */
 	while (licp != NULL) {
-		ASSERT(!XFS_LIC_ARE_ALL_FREE(licp));
+		ASSERT(!xfs_lic_are_all_free(licp));
 		(void) xfs_trans_unlock_chunk(licp, 1, abort, NULLCOMMITLSN);
 		next_licp = licp->lic_next;
 		kmem_free(licp);
@@ -347,7 +347,7 @@ xfs_trans_unlock_items(xfs_trans_t *tp, xfs_lsn_t commit_lsn)
 	/*
 	 * Special case the embedded chunk so we don't free.
 	 */
-	if (!XFS_LIC_ARE_ALL_FREE(licp)) {
+	if (!xfs_lic_are_all_free(licp)) {
 		freed = xfs_trans_unlock_chunk(licp, 0, 0, commit_lsn);
 	}
 	licpp = &(tp->t_items.lic_next);
@@ -358,10 +358,10 @@ xfs_trans_unlock_items(xfs_trans_t *tp, xfs_lsn_t commit_lsn)
 	 * and free empty chunks.
 	 */
 	while (licp != NULL) {
-		ASSERT(!XFS_LIC_ARE_ALL_FREE(licp));
+		ASSERT(!xfs_lic_are_all_free(licp));
 		freed += xfs_trans_unlock_chunk(licp, 0, 0, commit_lsn);
 		next_licp = licp->lic_next;
-		if (XFS_LIC_ARE_ALL_FREE(licp)) {
+		if (xfs_lic_are_all_free(licp)) {
 			*licpp = next_licp;
 			kmem_free(licp);
 			freed -= XFS_LIC_NUM_SLOTS;
@@ -402,7 +402,7 @@ xfs_trans_unlock_chunk(
 	freed = 0;
 	lidp = licp->lic_descs;
 	for (i = 0; i < licp->lic_unused; i++, lidp++) {
-		if (XFS_LIC_ISFREE(licp, i)) {
+		if (xfs_lic_isfree(licp, i)) {
 			continue;
 		}
 		lip = lidp->lid_item;
@@ -421,7 +421,7 @@ xfs_trans_unlock_chunk(
 		 */
 		if (!(freeing_chunk) &&
 		    (!(lidp->lid_flags & XFS_LID_DIRTY) || abort)) {
-			XFS_LIC_RELSE(licp, i);
+			xfs_lic_relse(licp, i);
 			freed++;
 		}
 	}

fs/xfs/xfs_utils.c

@@ -237,7 +237,7 @@ xfs_droplink(
 
 	ASSERT (ip->i_d.di_nlink > 0);
 	ip->i_d.di_nlink--;
-	drop_nlink(ip->i_vnode);
+	drop_nlink(VFS_I(ip));
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 
 	error = 0;
@@ -301,7 +301,7 @@ xfs_bumplink(
 
 	ASSERT(ip->i_d.di_nlink > 0);
 	ip->i_d.di_nlink++;
-	inc_nlink(ip->i_vnode);
+	inc_nlink(VFS_I(ip));
 	if ((ip->i_d.di_version == XFS_DINODE_VERSION_1) &&
 	    (ip->i_d.di_nlink > XFS_MAXLINK_1)) {
 		/*

fs/xfs/xfs_utils.h

@@ -18,9 +18,6 @@
 #ifndef __XFS_UTILS_H__
 #define __XFS_UTILS_H__
 
-#define IRELE(ip)	VN_RELE(XFS_ITOV(ip))
-#define IHOLD(ip)	VN_HOLD(XFS_ITOV(ip))
-
 extern int xfs_truncate_file(xfs_mount_t *, xfs_inode_t *);
 extern int xfs_dir_ialloc(xfs_trans_t **, xfs_inode_t *, mode_t, xfs_nlink_t,
 				xfs_dev_t, cred_t *, prid_t, int,

fs/xfs/xfs_vfsops.c

@@ -128,7 +128,6 @@ xfs_unmount_flush(
 	xfs_inode_t	*rip = mp->m_rootip;
 	xfs_inode_t	*rbmip;
 	xfs_inode_t	*rsumip = NULL;
-	bhv_vnode_t	*rvp = XFS_ITOV(rip);
 	int		error;
 
 	xfs_ilock(rip, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT);
@@ -146,7 +145,7 @@ xfs_unmount_flush(
 		if (error == EFSCORRUPTED)
 			goto fscorrupt_out;
 
-		ASSERT(vn_count(XFS_ITOV(rbmip)) == 1);
+		ASSERT(vn_count(VFS_I(rbmip)) == 1);
 
 		rsumip = mp->m_rsumip;
 		xfs_ilock(rsumip, XFS_ILOCK_EXCL);
@@ -157,7 +156,7 @@ xfs_unmount_flush(
 		if (error == EFSCORRUPTED)
 			goto fscorrupt_out;
 
-		ASSERT(vn_count(XFS_ITOV(rsumip)) == 1);
+		ASSERT(vn_count(VFS_I(rsumip)) == 1);
 	}
 
 	/*
@@ -167,7 +166,7 @@ xfs_unmount_flush(
 	if (error == EFSCORRUPTED)
 		goto fscorrupt_out2;
 
-	if (vn_count(rvp) != 1 && !relocation) {
+	if (vn_count(VFS_I(rip)) != 1 && !relocation) {
 		xfs_iunlock(rip, XFS_ILOCK_EXCL);
 		return XFS_ERROR(EBUSY);
 	}
@@ -284,7 +283,7 @@ xfs_sync_inodes(
 	int             *bypassed)
 {
 	xfs_inode_t	*ip = NULL;
-	bhv_vnode_t	*vp = NULL;
+	struct inode	*vp = NULL;
 	int		error;
 	int		last_error;
 	uint64_t	fflag;
@@ -404,7 +403,7 @@ xfs_sync_inodes(
 			continue;
 		}
 
-		vp = XFS_ITOV_NULL(ip);
+		vp = VFS_I(ip);
 
 		/*
 		 * If the vnode is gone then this is being torn down,
@@ -479,7 +478,7 @@ xfs_sync_inodes(
 			IPOINTER_INSERT(ip, mp);
 			xfs_ilock(ip, lock_flags);
 
-			ASSERT(vp == XFS_ITOV(ip));
+			ASSERT(vp == VFS_I(ip));
 			ASSERT(ip->i_mount == mp);
 
 			vnode_refed = B_TRUE;

fs/xfs/xfs_vnodeops.c

@@ -83,7 +83,7 @@ xfs_setattr(
 	cred_t			*credp)
 {
 	xfs_mount_t		*mp = ip->i_mount;
-	struct inode		*inode = XFS_ITOV(ip);
+	struct inode		*inode = VFS_I(ip);
 	int			mask = iattr->ia_valid;
 	xfs_trans_t		*tp;
 	int			code;
@@ -513,7 +513,6 @@ xfs_setattr(
 			ip->i_d.di_atime.t_sec = iattr->ia_atime.tv_sec;
 			ip->i_d.di_atime.t_nsec = iattr->ia_atime.tv_nsec;
 			ip->i_update_core = 1;
-			timeflags &= ~XFS_ICHGTIME_ACC;
 		}
 		if (mask & ATTR_MTIME) {
 			inode->i_mtime = iattr->ia_mtime;
@@ -714,7 +713,7 @@ xfs_fsync(
 		return XFS_ERROR(EIO);
 
 	/* capture size updates in I/O completion before writing the inode. */
-	error = filemap_fdatawait(vn_to_inode(XFS_ITOV(ip))->i_mapping);
+	error = filemap_fdatawait(VFS_I(ip)->i_mapping);
 	if (error)
 		return XFS_ERROR(error);
 
@@ -1160,7 +1159,6 @@ int
 xfs_release(
 	xfs_inode_t	*ip)
 {
-	bhv_vnode_t	*vp = XFS_ITOV(ip);
 	xfs_mount_t	*mp = ip->i_mount;
 	int		error;
 
@@ -1195,13 +1193,13 @@ xfs_release(
 		 * be exposed to that problem.
 		 */
 		truncated = xfs_iflags_test_and_clear(ip, XFS_ITRUNCATED);
-		if (truncated && VN_DIRTY(vp) && ip->i_delayed_blks > 0)
+		if (truncated && VN_DIRTY(VFS_I(ip)) && ip->i_delayed_blks > 0)
 			xfs_flush_pages(ip, 0, -1, XFS_B_ASYNC, FI_NONE);
 	}
 
 	if (ip->i_d.di_nlink != 0) {
 		if ((((ip->i_d.di_mode & S_IFMT) == S_IFREG) &&
-		     ((ip->i_size > 0) || (VN_CACHED(vp) > 0 ||
+		     ((ip->i_size > 0) || (VN_CACHED(VFS_I(ip)) > 0 ||
 		       ip->i_delayed_blks > 0)) &&
 		     (ip->i_df.if_flags & XFS_IFEXTENTS))  &&
 		    (!(ip->i_d.di_flags &
@@ -1227,7 +1225,6 @@ int
 xfs_inactive(
 	xfs_inode_t	*ip)
 {
-	bhv_vnode_t	*vp = XFS_ITOV(ip);
 	xfs_bmap_free_t	free_list;
 	xfs_fsblock_t	first_block;
 	int		committed;
@@ -1242,7 +1239,7 @@ xfs_inactive(
 	 * If the inode is already free, then there can be nothing
 	 * to clean up here.
 	 */
-	if (ip->i_d.di_mode == 0 || VN_BAD(vp)) {
+	if (ip->i_d.di_mode == 0 || VN_BAD(VFS_I(ip))) {
 		ASSERT(ip->i_df.if_real_bytes == 0);
 		ASSERT(ip->i_df.if_broot_bytes == 0);
 		return VN_INACTIVE_CACHE;
@@ -1272,7 +1269,7 @@ xfs_inactive(
 
 	if (ip->i_d.di_nlink != 0) {
 		if ((((ip->i_d.di_mode & S_IFMT) == S_IFREG) &&
-                     ((ip->i_size > 0) || (VN_CACHED(vp) > 0 ||
+                     ((ip->i_size > 0) || (VN_CACHED(VFS_I(ip)) > 0 ||
                        ip->i_delayed_blks > 0)) &&
 		      (ip->i_df.if_flags & XFS_IFEXTENTS) &&
 		     (!(ip->i_d.di_flags &
@@ -1707,111 +1704,6 @@ std_return:
 	goto std_return;
 }
 
-#ifdef DEBUG
-/*
- * Some counters to see if (and how often) we are hitting some deadlock
- * prevention code paths.
- */
-
-int xfs_rm_locks;
-int xfs_rm_lock_delays;
-int xfs_rm_attempts;
-#endif
-
-/*
- * The following routine will lock the inodes associated with the
- * directory and the named entry in the directory. The locks are
- * acquired in increasing inode number.
- *
- * If the entry is "..", then only the directory is locked. The
- * vnode ref count will still include that from the .. entry in
- * this case.
- *
- * There is a deadlock we need to worry about. If the locked directory is
- * in the AIL, it might be blocking up the log. The next inode we lock
- * could be already locked by another thread waiting for log space (e.g
- * a permanent log reservation with a long running transaction (see
- * xfs_itruncate_finish)). To solve this, we must check if the directory
- * is in the ail and use lock_nowait. If we can't lock, we need to
- * drop the inode lock on the directory and try again. xfs_iunlock will
- * potentially push the tail if we were holding up the log.
- */
-STATIC int
-xfs_lock_dir_and_entry(
-	xfs_inode_t	*dp,
-	xfs_inode_t	*ip)	/* inode of entry 'name' */
-{
-	int		attempts;
-	xfs_ino_t	e_inum;
-	xfs_inode_t	*ips[2];
-	xfs_log_item_t	*lp;
-
-#ifdef DEBUG
-	xfs_rm_locks++;
-#endif
-	attempts = 0;
-
-again:
-	xfs_ilock(dp, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT);
-
-	e_inum = ip->i_ino;
-
-	xfs_itrace_ref(ip);
-
-	/*
-	 * We want to lock in increasing inum. Since we've already
-	 * acquired the lock on the directory, we may need to release
-	 * if if the inum of the entry turns out to be less.
-	 */
-	if (e_inum > dp->i_ino) {
-		/*
-		 * We are already in the right order, so just
-		 * lock on the inode of the entry.
-		 * We need to use nowait if dp is in the AIL.
-		 */
-
-		lp = (xfs_log_item_t *)dp->i_itemp;
-		if (lp && (lp->li_flags & XFS_LI_IN_AIL)) {
-			if (!xfs_ilock_nowait(ip, XFS_ILOCK_EXCL)) {
-				attempts++;
-#ifdef DEBUG
-				xfs_rm_attempts++;
-#endif
-
-				/*
-				 * Unlock dp and try again.
-				 * xfs_iunlock will try to push the tail
-				 * if the inode is in the AIL.
-				 */
-
-				xfs_iunlock(dp, XFS_ILOCK_EXCL);
-
-				if ((attempts % 5) == 0) {
-					delay(1); /* Don't just spin the CPU */
-#ifdef DEBUG
-					xfs_rm_lock_delays++;
-#endif
-				}
-				goto again;
-			}
-		} else {
-			xfs_ilock(ip, XFS_ILOCK_EXCL);
-		}
-	} else if (e_inum < dp->i_ino) {
-		xfs_iunlock(dp, XFS_ILOCK_EXCL);
-
-		ips[0] = ip;
-		ips[1] = dp;
-		xfs_lock_inodes(ips, 2, XFS_ILOCK_EXCL);
-	}
-	/* else	 e_inum == dp->i_ino */
-	/*     This can happen if we're asked to lock /x/..
-	 *     the entry is "..", which is also the parent directory.
-	 */
-
-	return 0;
-}
-
 #ifdef DEBUG
 int xfs_locked_n;
 int xfs_small_retries;
@@ -1946,6 +1838,45 @@ again:
 #endif
 }
 
+void
+xfs_lock_two_inodes(
+	xfs_inode_t		*ip0,
+	xfs_inode_t		*ip1,
+	uint			lock_mode)
+{
+	xfs_inode_t		*temp;
+	int			attempts = 0;
+	xfs_log_item_t		*lp;
+
+	ASSERT(ip0->i_ino != ip1->i_ino);
+
+	if (ip0->i_ino > ip1->i_ino) {
+		temp = ip0;
+		ip0 = ip1;
+		ip1 = temp;
+	}
+
+ again:
+	xfs_ilock(ip0, xfs_lock_inumorder(lock_mode, 0));
+
+	/*
+	 * If the first lock we have locked is in the AIL, we must TRY to get
+	 * the second lock. If we can't get it, we must release the first one
+	 * and try again.
+	 */
+	lp = (xfs_log_item_t *)ip0->i_itemp;
+	if (lp && (lp->li_flags & XFS_LI_IN_AIL)) {
+		if (!xfs_ilock_nowait(ip1, xfs_lock_inumorder(lock_mode, 1))) {
+			xfs_iunlock(ip0, lock_mode);
+			if ((++attempts % 5) == 0)
+				delay(1); /* Don't just spin the CPU */
+			goto again;
+		}
+	} else {
+		xfs_ilock(ip1, xfs_lock_inumorder(lock_mode, 1));
+	}
+}
+
 int
 xfs_remove(
 	xfs_inode_t             *dp,
@@ -2018,9 +1949,7 @@ xfs_remove(
 		goto out_trans_cancel;
 	}
 
-	error = xfs_lock_dir_and_entry(dp, ip);
-	if (error)
-		goto out_trans_cancel;
+	xfs_lock_two_inodes(dp, ip, XFS_ILOCK_EXCL);
 
 	/*
 	 * At this point, we've gotten both the directory and the entry
@@ -2047,9 +1976,6 @@ xfs_remove(
 		}
 	}
 
-	/*
-	 * Entry must exist since we did a lookup in xfs_lock_dir_and_entry.
-	 */
 	XFS_BMAP_INIT(&free_list, &first_block);
 	error = xfs_dir_removename(tp, dp, name, ip->i_ino,
 					&first_block, &free_list, resblks);
@@ -2155,7 +2081,6 @@ xfs_link(
 {
 	xfs_mount_t		*mp = tdp->i_mount;
 	xfs_trans_t		*tp;
-	xfs_inode_t		*ips[2];
 	int			error;
 	xfs_bmap_free_t         free_list;
 	xfs_fsblock_t           first_block;
@@ -2203,15 +2128,7 @@ xfs_link(
 		goto error_return;
 	}
 
-	if (sip->i_ino < tdp->i_ino) {
-		ips[0] = sip;
-		ips[1] = tdp;
-	} else {
-		ips[0] = tdp;
-		ips[1] = sip;
-	}
-
-	xfs_lock_inodes(ips, 2, XFS_ILOCK_EXCL);
+	xfs_lock_two_inodes(sip, tdp, XFS_ILOCK_EXCL);
 
 	/*
 	 * Increment vnode ref counts since xfs_trans_commit &
@@ -2873,14 +2790,13 @@ int
 xfs_reclaim(
 	xfs_inode_t	*ip)
 {
-	bhv_vnode_t	*vp = XFS_ITOV(ip);
 
 	xfs_itrace_entry(ip);
 
-	ASSERT(!VN_MAPPED(vp));
+	ASSERT(!VN_MAPPED(VFS_I(ip)));
 
 	/* bad inode, get out here ASAP */
-	if (VN_BAD(vp)) {
+	if (VN_BAD(VFS_I(ip))) {
 		xfs_ireclaim(ip);
 		return 0;
 	}
@@ -2917,7 +2833,7 @@ xfs_reclaim(
 		XFS_MOUNT_ILOCK(mp);
 		spin_lock(&ip->i_flags_lock);
 		__xfs_iflags_set(ip, XFS_IRECLAIMABLE);
-		vn_to_inode(vp)->i_private = NULL;
+		VFS_I(ip)->i_private = NULL;
 		ip->i_vnode = NULL;
 		spin_unlock(&ip->i_flags_lock);
 		list_add_tail(&ip->i_reclaim, &mp->m_del_inodes);
@@ -2933,7 +2849,7 @@ xfs_finish_reclaim(
 	int		sync_mode)
 {
 	xfs_perag_t	*pag = xfs_get_perag(ip->i_mount, ip->i_ino);
-	bhv_vnode_t	*vp = XFS_ITOV_NULL(ip);
+	struct inode	*vp = VFS_I(ip);
 
 	if (vp && VN_BAD(vp))
 		goto reclaim;
@@ -3321,7 +3237,6 @@ xfs_free_file_space(
 	xfs_off_t		len,
 	int			attr_flags)
 {
-	bhv_vnode_t		*vp;
 	int			committed;
 	int			done;
 	xfs_off_t		end_dmi_offset;
@@ -3341,7 +3256,6 @@ xfs_free_file_space(
 	xfs_trans_t		*tp;
 	int			need_iolock = 1;
 
-	vp = XFS_ITOV(ip);
 	mp = ip->i_mount;
 
 	xfs_itrace_entry(ip);
@@ -3378,7 +3292,7 @@ xfs_free_file_space(
 	rounding = max_t(uint, 1 << mp->m_sb.sb_blocklog, PAGE_CACHE_SIZE);
 	ioffset = offset & ~(rounding - 1);
 
-	if (VN_CACHED(vp) != 0) {
+	if (VN_CACHED(VFS_I(ip)) != 0) {
 		xfs_inval_cached_trace(ip, ioffset, -1, ioffset, -1);
 		error = xfs_flushinval_pages(ip, ioffset, -1, FI_REMAPF_LOCKED);
 		if (error)

include/linux/completion.h

@@ -55,4 +55,49 @@ extern void complete_all(struct completion *);
 
 #define INIT_COMPLETION(x)	((x).done = 0)
 
+
+/**
+ *	try_wait_for_completion - try to decrement a completion without blocking
+ *	@x:	completion structure
+ *
+ *	Returns: 0 if a decrement cannot be done without blocking
+ *		 1 if a decrement succeeded.
+ *
+ *	If a completion is being used as a counting completion,
+ *	attempt to decrement the counter without blocking. This
+ *	enables us to avoid waiting if the resource the completion
+ *	is protecting is not available.
+ */
+static inline bool try_wait_for_completion(struct completion *x)
+{
+	int ret = 1;
+
+	spin_lock_irq(&x->wait.lock);
+	if (!x->done)
+		ret = 0;
+	else
+		x->done--;
+	spin_unlock_irq(&x->wait.lock);
+	return ret;
+}
+
+/**
+ *	completion_done - Test to see if a completion has any waiters
+ *	@x:	completion structure
+ *
+ *	Returns: 0 if there are waiters (wait_for_completion() in progress)
+ *		 1 if there are no waiters.
+ *
+ */
+static inline bool completion_done(struct completion *x)
+{
+	int ret = 1;
+
+	spin_lock_irq(&x->wait.lock);
+	if (!x->done)
+		ret = 0;
+	spin_unlock_irq(&x->wait.lock);
+	return ret;
+}
+
 #endif