Merge branch 'for-linus' of git://oss.sgi.com:8090/xfs/xfs-2.6

* 'for-linus' of git://oss.sgi.com:8090/xfs/xfs-2.6: (59 commits)
  [XFS] eagerly remove vmap mappings to avoid upsetting Xen
  [XFS] simplify validata_fields
  [XFS] no longer using io_vnode, as was remaining from 23 cherrypick
  [XFS] Remove STATIC which was missing from prior manual merge
  [XFS] Put back the QUEUE_ORDERED_NONE test in the barrier check.
  [XFS] Turn off XBF_ASYNC flag before re-reading superblock.
  [XFS] avoid race in sync_inodes() that can fail to write out all dirty data
  [XFS] This fix prevents bulkstat from spinning in an infinite loop.
  [XFS] simplify xfs_create/mknod/symlink prototype
  [XFS] avoid xfs_getattr in XFS_IOC_FSGETXATTR ioctl
  [XFS] get_bulkall() could return incorrect inode state
  [XFS] Kill unused IOMAP_EOF flag
  [XFS] fix when DMAPI mount option processing happens
  [XFS] ensure file size is logged on synchronous writes
  [XFS] growlock should be a mutex
  [XFS] replace some large xfs_log_priv.h macros by proper functions
  [XFS] kill struct bhv_vfs
  [XFS] move syncing related members from struct bhv_vfs to struct xfs_mount
  [XFS] kill the vfs_flags member in struct bhv_vfs
  [XFS] kill the vfs_fsid and vfs_altfsid members in struct bhv_vfs
  ...

fs/xfs/Makefile-linux-2.6

@@ -49,7 +49,6 @@ xfs-y				+= xfs_alloc.o \
 				   xfs_alloc_btree.o \
 				   xfs_attr.o \
 				   xfs_attr_leaf.o \
-				   xfs_behavior.o \
 				   xfs_bit.o \
 				   xfs_bmap.o \
 				   xfs_bmap_btree.o \
@@ -108,13 +107,11 @@ xfs-y				+= $(addprefix $(XFS_LINUX)/, \
 				   xfs_iops.o \
 				   xfs_lrw.o \
 				   xfs_super.o \
-				   xfs_vfs.o \
 				   xfs_vnode.o)
 
 # Objects in support/
 xfs-y				+= $(addprefix support/, \
 				   debug.o \
-				   move.o \
 				   uuid.o)
 
 xfs-$(CONFIG_XFS_TRACE)		+= support/ktrace.o

fs/xfs/linux-2.6/xfs_aops.c

@@ -37,6 +37,7 @@
 #include "xfs_error.h"
 #include "xfs_rw.h"
 #include "xfs_iomap.h"
+#include "xfs_vnodeops.h"
 #include <linux/mpage.h>
 #include <linux/pagevec.h>
 #include <linux/writeback.h>
@@ -139,9 +140,11 @@ xfs_destroy_ioend(
 		next = bh->b_private;
 		bh->b_end_io(bh, !ioend->io_error);
 	}
-	if (unlikely(ioend->io_error))
-		vn_ioerror(ioend->io_vnode, ioend->io_error, __FILE__,__LINE__);
-	vn_iowake(ioend->io_vnode);
+	if (unlikely(ioend->io_error)) {
+		vn_ioerror(XFS_I(ioend->io_inode), ioend->io_error,
+				__FILE__,__LINE__);
+	}
+	vn_iowake(XFS_I(ioend->io_inode));
 	mempool_free(ioend, xfs_ioend_pool);
 }
 
@@ -156,14 +159,10 @@ STATIC void
 xfs_setfilesize(
 	xfs_ioend_t		*ioend)
 {
-	xfs_inode_t		*ip;
+	xfs_inode_t		*ip = XFS_I(ioend->io_inode);
 	xfs_fsize_t		isize;
 	xfs_fsize_t		bsize;
 
-	ip = xfs_vtoi(ioend->io_vnode);
-	if (!ip)
-		return;
-
 	ASSERT((ip->i_d.di_mode & S_IFMT) == S_IFREG);
 	ASSERT(ioend->io_type != IOMAP_READ);
 
@@ -181,7 +180,7 @@ xfs_setfilesize(
 		ip->i_d.di_size = isize;
 		ip->i_update_core = 1;
 		ip->i_update_size = 1;
-		mark_inode_dirty_sync(vn_to_inode(ioend->io_vnode));
+		mark_inode_dirty_sync(ioend->io_inode);
 	}
 
 	xfs_iunlock(ip, XFS_ILOCK_EXCL);
@@ -227,12 +226,12 @@ xfs_end_bio_unwritten(
 {
 	xfs_ioend_t		*ioend =
 		container_of(work, xfs_ioend_t, io_work);
-	bhv_vnode_t		*vp = ioend->io_vnode;
 	xfs_off_t		offset = ioend->io_offset;
 	size_t			size = ioend->io_size;
 
 	if (likely(!ioend->io_error)) {
-		bhv_vop_bmap(vp, offset, size, BMAPI_UNWRITTEN, NULL, NULL);
+		xfs_bmap(XFS_I(ioend->io_inode), offset, size,
+				BMAPI_UNWRITTEN, NULL, NULL);
 		xfs_setfilesize(ioend);
 	}
 	xfs_destroy_ioend(ioend);
@@ -275,10 +274,10 @@ xfs_alloc_ioend(
 	ioend->io_error = 0;
 	ioend->io_list = NULL;
 	ioend->io_type = type;
-	ioend->io_vnode = vn_from_inode(inode);
+	ioend->io_inode = inode;
 	ioend->io_buffer_head = NULL;
 	ioend->io_buffer_tail = NULL;
-	atomic_inc(&ioend->io_vnode->v_iocount);
+	atomic_inc(&XFS_I(ioend->io_inode)->i_iocount);
 	ioend->io_offset = 0;
 	ioend->io_size = 0;
 
@@ -302,12 +301,13 @@ xfs_map_blocks(
 	xfs_iomap_t		*mapp,
 	int			flags)
 {
-	bhv_vnode_t		*vp = vn_from_inode(inode);
+	xfs_inode_t		*ip = XFS_I(inode);
 	int			error, nmaps = 1;
 
-	error = bhv_vop_bmap(vp, offset, count, flags, mapp, &nmaps);
+	error = xfs_bmap(ip, offset, count,
+				flags, mapp, &nmaps);
 	if (!error && (flags & (BMAPI_WRITE|BMAPI_ALLOCATE)))
-		VMODIFY(vp);
+		xfs_iflags_set(ip, XFS_IMODIFIED);
 	return -error;
 }
 
@@ -497,7 +497,7 @@ xfs_cancel_ioend(
 			unlock_buffer(bh);
 		} while ((bh = next_bh) != NULL);
 
-		vn_iowake(ioend->io_vnode);
+		vn_iowake(XFS_I(ioend->io_inode));
 		mempool_free(ioend, xfs_ioend_pool);
 	} while ((ioend = next) != NULL);
 }
@@ -1237,10 +1237,7 @@ xfs_vm_writepages(
 	struct address_space	*mapping,
 	struct writeback_control *wbc)
 {
-	struct bhv_vnode	*vp = vn_from_inode(mapping->host);
-
-	if (VN_TRUNC(vp))
-		VUNTRUNCATE(vp);
+	xfs_iflags_clear(XFS_I(mapping->host), XFS_ITRUNCATED);
 	return generic_writepages(mapping, wbc);
 }
 
@@ -1317,7 +1314,6 @@ __xfs_get_blocks(
 	int			direct,
 	bmapi_flags_t		flags)
 {
-	bhv_vnode_t		*vp = vn_from_inode(inode);
 	xfs_iomap_t		iomap;
 	xfs_off_t		offset;
 	ssize_t			size;
@@ -1327,7 +1323,7 @@ __xfs_get_blocks(
 	offset = (xfs_off_t)iblock << inode->i_blkbits;
 	ASSERT(bh_result->b_size >= (1 << inode->i_blkbits));
 	size = bh_result->b_size;
-	error = bhv_vop_bmap(vp, offset, size,
+	error = xfs_bmap(XFS_I(inode), offset, size,
 			     create ? flags : BMAPI_READ, &iomap, &niomap);
 	if (error)
 		return -error;
@@ -1475,13 +1471,13 @@ xfs_vm_direct_IO(
 {
 	struct file	*file = iocb->ki_filp;
 	struct inode	*inode = file->f_mapping->host;
-	bhv_vnode_t	*vp = vn_from_inode(inode);
 	xfs_iomap_t	iomap;
 	int		maps = 1;
 	int		error;
 	ssize_t		ret;
 
-	error = bhv_vop_bmap(vp, offset, 0, BMAPI_DEVICE, &iomap, &maps);
+	error = xfs_bmap(XFS_I(inode), offset, 0,
+				BMAPI_DEVICE, &iomap, &maps);
 	if (error)
 		return -error;
 
@@ -1527,12 +1523,13 @@ xfs_vm_bmap(
 	sector_t		block)
 {
 	struct inode		*inode = (struct inode *)mapping->host;
-	bhv_vnode_t		*vp = vn_from_inode(inode);
+	struct xfs_inode	*ip = XFS_I(inode);
 
-	vn_trace_entry(vp, __FUNCTION__, (inst_t *)__return_address);
-	bhv_vop_rwlock(vp, VRWLOCK_READ);
-	bhv_vop_flush_pages(vp, (xfs_off_t)0, -1, 0, FI_REMAPF);
-	bhv_vop_rwunlock(vp, VRWLOCK_READ);
+	vn_trace_entry(XFS_I(inode), __FUNCTION__,
+			(inst_t *)__return_address);
+	xfs_rwlock(ip, VRWLOCK_READ);
+	xfs_flush_pages(ip, (xfs_off_t)0, -1, 0, FI_REMAPF);
+	xfs_rwunlock(ip, VRWLOCK_READ);
 	return generic_block_bmap(mapping, block, xfs_get_blocks);
 }
 

fs/xfs/linux-2.6/xfs_aops.h

@@ -32,7 +32,7 @@ typedef struct xfs_ioend {
 	unsigned int		io_type;	/* delalloc / unwritten */
 	int			io_error;	/* I/O error code */
 	atomic_t		io_remaining;	/* hold count */
-	struct bhv_vnode	*io_vnode;	/* file being written to */
+	struct inode		*io_inode;	/* file being written to */
 	struct buffer_head	*io_buffer_head;/* buffer linked list head */
 	struct buffer_head	*io_buffer_tail;/* buffer linked list tail */
 	size_t			io_size;	/* size of the extent */

fs/xfs/linux-2.6/xfs_buf.c

@@ -187,6 +187,19 @@ free_address(
 {
 	a_list_t	*aentry;
 
+#ifdef CONFIG_XEN
+	/*
+	 * Xen needs to be able to make sure it can get an exclusive
+	 * RO mapping of pages it wants to turn into a pagetable.  If
+	 * a newly allocated page is also still being vmap()ed by xfs,
+	 * it will cause pagetable construction to fail.  This is a
+	 * quick workaround to always eagerly unmap pages so that Xen
+	 * is happy.
+	 */
+	vunmap(addr);
+	return;
+#endif
+
 	aentry = kmalloc(sizeof(a_list_t), GFP_NOWAIT);
 	if (likely(aentry)) {
 		spin_lock(&as_lock);
@@ -997,7 +1010,18 @@ xfs_buf_iodone_work(
 	xfs_buf_t		*bp =
 		container_of(work, xfs_buf_t, b_iodone_work);
 
-	if (bp->b_iodone)
+	/*
+	 * We can get an EOPNOTSUPP to ordered writes.  Here we clear the
+	 * ordered flag and reissue them.  Because we can't tell the higher
+	 * layers directly that they should not issue ordered I/O anymore, they
+	 * need to check if the ordered flag was cleared during I/O completion.
+	 */
+	if ((bp->b_error == EOPNOTSUPP) &&
+	    (bp->b_flags & (XBF_ORDERED|XBF_ASYNC)) == (XBF_ORDERED|XBF_ASYNC)) {
+		XB_TRACE(bp, "ordered_retry", bp->b_iodone);
+		bp->b_flags &= ~XBF_ORDERED;
+		xfs_buf_iorequest(bp);
+	} else if (bp->b_iodone)
 		(*(bp->b_iodone))(bp);
 	else if (bp->b_flags & XBF_ASYNC)
 		xfs_buf_relse(bp);

fs/xfs/linux-2.6/xfs_export.c

@@ -17,12 +17,18 @@
  */
 #include "xfs.h"
 #include "xfs_types.h"
-#include "xfs_dmapi.h"
+#include "xfs_inum.h"
 #include "xfs_log.h"
 #include "xfs_trans.h"
 #include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_dmapi.h"
 #include "xfs_mount.h"
 #include "xfs_export.h"
+#include "xfs_vnodeops.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_inode.h"
+#include "xfs_vfsops.h"
 
 static struct dentry dotdot = { .d_name.name = "..", .d_name.len = 2, };
 
@@ -96,9 +102,7 @@ xfs_fs_encode_fh(
 	int			len;
 	int			is64 = 0;
 #if XFS_BIG_INUMS
-	bhv_vfs_t		*vfs = vfs_from_sb(inode->i_sb);
-
-	if (!(vfs->vfs_flag & VFS_32BITINODES)) {
+	if (!(XFS_M(inode->i_sb)->m_flags & XFS_MOUNT_SMALL_INUMS)) {
 		/* filesystem may contain 64bit inode numbers */
 		is64 = XFS_FILEID_TYPE_64FLAG;
 	}
@@ -138,10 +142,9 @@ xfs_fs_get_dentry(
 	bhv_vnode_t		*vp;
 	struct inode		*inode;
 	struct dentry		*result;
-	bhv_vfs_t		*vfsp = vfs_from_sb(sb);
 	int			error;
 
-	error = bhv_vfs_vget(vfsp, &vp, (fid_t *)data);
+	error = xfs_vget(XFS_M(sb), &vp, (fid_t *)data);
 	if (error || vp == NULL)
 		return ERR_PTR(-ESTALE) ;
 
@@ -159,12 +162,11 @@ xfs_fs_get_parent(
 	struct dentry		*child)
 {
 	int			error;
-	bhv_vnode_t		*vp, *cvp;
+	bhv_vnode_t		*cvp;
 	struct dentry		*parent;
 
 	cvp = NULL;
-	vp = vn_from_inode(child->d_inode);
-	error = bhv_vop_lookup(vp, &dotdot, &cvp, 0, NULL, NULL);
+	error = xfs_lookup(XFS_I(child->d_inode), &dotdot, &cvp);
 	if (unlikely(error))
 		return ERR_PTR(-error);
 

fs/xfs/linux-2.6/xfs_file.c

@@ -37,6 +37,7 @@
 #include "xfs_error.h"
 #include "xfs_rw.h"
 #include "xfs_ioctl32.h"
+#include "xfs_vnodeops.h"
 
 #include <linux/dcache.h>
 #include <linux/smp_lock.h>
@@ -55,13 +56,12 @@ __xfs_file_read(
 	loff_t			pos)
 {
 	struct file		*file = iocb->ki_filp;
-	bhv_vnode_t		*vp = vn_from_inode(file->f_path.dentry->d_inode);
 
 	BUG_ON(iocb->ki_pos != pos);
 	if (unlikely(file->f_flags & O_DIRECT))
 		ioflags |= IO_ISDIRECT;
-	return bhv_vop_read(vp, iocb, iov, nr_segs, &iocb->ki_pos,
-				ioflags, NULL);
+	return xfs_read(XFS_I(file->f_path.dentry->d_inode), iocb, iov,
+				nr_segs, &iocb->ki_pos, ioflags);
 }
 
 STATIC ssize_t
@@ -93,14 +93,12 @@ __xfs_file_write(
 	loff_t			pos)
 {
 	struct file	*file = iocb->ki_filp;
-	struct inode	*inode = file->f_mapping->host;
-	bhv_vnode_t	*vp = vn_from_inode(inode);
 
 	BUG_ON(iocb->ki_pos != pos);
 	if (unlikely(file->f_flags & O_DIRECT))
 		ioflags |= IO_ISDIRECT;
-	return bhv_vop_write(vp, iocb, iov, nr_segs, &iocb->ki_pos,
-				ioflags, NULL);
+	return xfs_write(XFS_I(file->f_mapping->host), iocb, iov, nr_segs,
+				&iocb->ki_pos, ioflags);
 }
 
 STATIC ssize_t
@@ -131,8 +129,8 @@ xfs_file_splice_read(
 	size_t			len,
 	unsigned int		flags)
 {
-	return bhv_vop_splice_read(vn_from_inode(infilp->f_path.dentry->d_inode),
-				   infilp, ppos, pipe, len, flags, 0, NULL);
+	return xfs_splice_read(XFS_I(infilp->f_path.dentry->d_inode),
+				   infilp, ppos, pipe, len, flags, 0);
 }
 
 STATIC ssize_t
@@ -143,9 +141,8 @@ xfs_file_splice_read_invis(
 	size_t			len,
 	unsigned int		flags)
 {
-	return bhv_vop_splice_read(vn_from_inode(infilp->f_path.dentry->d_inode),
-				   infilp, ppos, pipe, len, flags, IO_INVIS,
-				   NULL);
+	return xfs_splice_read(XFS_I(infilp->f_path.dentry->d_inode),
+				   infilp, ppos, pipe, len, flags, IO_INVIS);
 }
 
 STATIC ssize_t
@@ -156,8 +153,8 @@ xfs_file_splice_write(
 	size_t			len,
 	unsigned int		flags)
 {
-	return bhv_vop_splice_write(vn_from_inode(outfilp->f_path.dentry->d_inode),
-				    pipe, outfilp, ppos, len, flags, 0, NULL);
+	return xfs_splice_write(XFS_I(outfilp->f_path.dentry->d_inode),
+				    pipe, outfilp, ppos, len, flags, 0);
 }
 
 STATIC ssize_t
@@ -168,9 +165,8 @@ xfs_file_splice_write_invis(
 	size_t			len,
 	unsigned int		flags)
 {
-	return bhv_vop_splice_write(vn_from_inode(outfilp->f_path.dentry->d_inode),
-				    pipe, outfilp, ppos, len, flags, IO_INVIS,
-				    NULL);
+	return xfs_splice_write(XFS_I(outfilp->f_path.dentry->d_inode),
+				    pipe, outfilp, ppos, len, flags, IO_INVIS);
 }
 
 STATIC int
@@ -180,7 +176,7 @@ xfs_file_open(
 {
 	if (!(filp->f_flags & O_LARGEFILE) && i_size_read(inode) > MAX_NON_LFS)
 		return -EFBIG;
-	return -bhv_vop_open(vn_from_inode(inode), NULL);
+	return -xfs_open(XFS_I(inode));
 }
 
 STATIC int
@@ -188,11 +184,7 @@ xfs_file_release(
 	struct inode	*inode,
 	struct file	*filp)
 {
-	bhv_vnode_t	*vp = vn_from_inode(inode);
-
-	if (vp)
-		return -bhv_vop_release(vp);
-	return 0;
+	return -xfs_release(XFS_I(inode));
 }
 
 STATIC int
@@ -201,14 +193,13 @@ xfs_file_fsync(
 	struct dentry	*dentry,
 	int		datasync)
 {
-	bhv_vnode_t	*vp = vn_from_inode(dentry->d_inode);
 	int		flags = FSYNC_WAIT;
 
 	if (datasync)
 		flags |= FSYNC_DATA;
-	if (VN_TRUNC(vp))
-		VUNTRUNCATE(vp);
-	return -bhv_vop_fsync(vp, flags, NULL, (xfs_off_t)0, (xfs_off_t)-1);
+	xfs_iflags_clear(XFS_I(dentry->d_inode), XFS_ITRUNCATED);
+	return -xfs_fsync(XFS_I(dentry->d_inode), flags,
+			(xfs_off_t)0, (xfs_off_t)-1);
 }
 
 #ifdef CONFIG_XFS_DMAPI
@@ -233,74 +224,30 @@ xfs_file_readdir(
 	void		*dirent,
 	filldir_t	filldir)
 {
-	int		error = 0;
-	bhv_vnode_t	*vp = vn_from_inode(filp->f_path.dentry->d_inode);
-	uio_t		uio;
-	iovec_t		iov;
-	int		eof = 0;
-	caddr_t		read_buf;
-	int		namelen, size = 0;
-	size_t		rlen = PAGE_CACHE_SIZE;
-	xfs_off_t	start_offset, curr_offset;
-	xfs_dirent_t	*dbp = NULL;
-
-	/* Try fairly hard to get memory */
-	do {
-		if ((read_buf = kmalloc(rlen, GFP_KERNEL)))
-			break;
-		rlen >>= 1;
-	} while (rlen >= 1024);
-
-	if (read_buf == NULL)
-		return -ENOMEM;
-
-	uio.uio_iov = &iov;
-	uio.uio_segflg = UIO_SYSSPACE;
-	curr_offset = filp->f_pos;
-	if (filp->f_pos != 0x7fffffff)
-		uio.uio_offset = filp->f_pos;
-	else
-		uio.uio_offset = 0xffffffff;
-
-	while (!eof) {
-		uio.uio_resid = iov.iov_len = rlen;
-		iov.iov_base = read_buf;
-		uio.uio_iovcnt = 1;
-
-		start_offset = uio.uio_offset;
-
-		error = bhv_vop_readdir(vp, &uio, NULL, &eof);
-		if ((uio.uio_offset == start_offset) || error) {
-			size = 0;
-			break;
-		}
-
-		size = rlen - uio.uio_resid;
-		dbp = (xfs_dirent_t *)read_buf;
-		while (size > 0) {
-			namelen = strlen(dbp->d_name);
-
-			if (filldir(dirent, dbp->d_name, namelen,
-					(loff_t) curr_offset & 0x7fffffff,
-					(ino_t) dbp->d_ino,
-					DT_UNKNOWN)) {
-				goto done;
-			}
-			size -= dbp->d_reclen;
-			curr_offset = (loff_t)dbp->d_off /* & 0x7fffffff */;
-			dbp = (xfs_dirent_t *)((char *)dbp + dbp->d_reclen);
-		}
-	}
-done:
-	if (!error) {
-		if (size == 0)
-			filp->f_pos = uio.uio_offset & 0x7fffffff;
-		else if (dbp)
-			filp->f_pos = curr_offset;
-	}
+	struct inode	*inode = filp->f_path.dentry->d_inode;
+	xfs_inode_t	*ip = XFS_I(inode);
+	int		error;
+	size_t		bufsize;
+
+	/*
+	 * The Linux API doesn't pass down the total size of the buffer
+	 * we read into down to the filesystem.  With the filldir concept
+	 * it's not needed for correct information, but the XFS dir2 leaf
+	 * code wants an estimate of the buffer size to calculate it's
+	 * readahead window and size the buffers used for mapping to
+	 * physical blocks.
+	 *
+	 * Try to give it an estimate that's good enough, maybe at some
+	 * point we can change the ->readdir prototype to include the
+	 * buffer size.
+	 */
+	bufsize = (size_t)min_t(loff_t, PAGE_SIZE, inode->i_size);
 
-	kfree(read_buf);
-	return -error;
+	error = xfs_readdir(ip, dirent, bufsize,
+				(xfs_off_t *)&filp->f_pos, filldir);
+	if (error)
+		return -error;
+	return 0;
 }
 
 STATIC int
@@ -312,7 +259,7 @@ xfs_file_mmap(
 	vma->vm_flags |= VM_CAN_NONLINEAR;
 
 #ifdef CONFIG_XFS_DMAPI
-	if (vn_from_inode(filp->f_path.dentry->d_inode)->v_vfsp->vfs_flag & VFS_DMI)
+	if (XFS_M(filp->f_path.dentry->d_inode->i_sb)->m_flags & XFS_MOUNT_DMAPI)
 		vma->vm_ops = &xfs_dmapi_file_vm_ops;
 #endif /* CONFIG_XFS_DMAPI */
 
@@ -328,10 +275,9 @@ xfs_file_ioctl(
 {
 	int		error;
 	struct inode	*inode = filp->f_path.dentry->d_inode;
-	bhv_vnode_t	*vp = vn_from_inode(inode);
 
-	error = bhv_vop_ioctl(vp, inode, filp, 0, cmd, (void __user *)p);
-	VMODIFY(vp);
+	error = xfs_ioctl(XFS_I(inode), filp, 0, cmd, (void __user *)p);
+	xfs_iflags_set(XFS_I(inode), XFS_IMODIFIED);
 
 	/* NOTE:  some of the ioctl's return positive #'s as a
 	 *	  byte count indicating success, such as
@@ -350,10 +296,9 @@ xfs_file_ioctl_invis(
 {
 	int		error;
 	struct inode	*inode = filp->f_path.dentry->d_inode;
-	bhv_vnode_t	*vp = vn_from_inode(inode);
 
-	error = bhv_vop_ioctl(vp, inode, filp, IO_INVIS, cmd, (void __user *)p);
-	VMODIFY(vp);
+	error = xfs_ioctl(XFS_I(inode), filp, IO_INVIS, cmd, (void __user *)p);
+	xfs_iflags_set(XFS_I(inode), XFS_IMODIFIED);
 
 	/* NOTE:  some of the ioctl's return positive #'s as a
 	 *	  byte count indicating success, such as
@@ -371,16 +316,14 @@ xfs_vm_mprotect(
 	struct vm_area_struct *vma,
 	unsigned int	newflags)
 {
-	bhv_vnode_t	*vp = vn_from_inode(vma->vm_file->f_path.dentry->d_inode);
+	struct inode	*inode = vma->vm_file->f_path.dentry->d_inode;
+	struct xfs_mount *mp = XFS_M(inode->i_sb);
 	int		error = 0;
 
-	if (vp->v_vfsp->vfs_flag & VFS_DMI) {
+	if (mp->m_flags & XFS_MOUNT_DMAPI) {
 		if ((vma->vm_flags & VM_MAYSHARE) &&
-		    (newflags & VM_WRITE) && !(vma->vm_flags & VM_WRITE)) {
-			xfs_mount_t	*mp = XFS_VFSTOM(vp->v_vfsp);
-
+		    (newflags & VM_WRITE) && !(vma->vm_flags & VM_WRITE))
 			error = XFS_SEND_MMAP(mp, vma, VM_WRITE);
-		    }
 	}
 	return error;
 }
@@ -397,18 +340,17 @@ STATIC int
 xfs_file_open_exec(
 	struct inode	*inode)
 {
-	bhv_vnode_t	*vp = vn_from_inode(inode);
+	struct xfs_mount *mp = XFS_M(inode->i_sb);
 
-	if (unlikely(vp->v_vfsp->vfs_flag & VFS_DMI)) {
-		xfs_mount_t	*mp = XFS_VFSTOM(vp->v_vfsp);
-		xfs_inode_t	*ip = xfs_vtoi(vp);
+	if (unlikely(mp->m_flags & XFS_MOUNT_DMAPI)) {
+		if (DM_EVENT_ENABLED(XFS_I(inode), DM_EVENT_READ)) {
+			bhv_vnode_t *vp = vn_from_inode(inode);
 
-		if (!ip)
-			return -EINVAL;
-		if (DM_EVENT_ENABLED(vp->v_vfsp, ip, DM_EVENT_READ))
-			return -XFS_SEND_DATA(mp, DM_EVENT_READ, vp,
-					       0, 0, 0, NULL);
+			return -XFS_SEND_DATA(mp, DM_EVENT_READ,
+						vp, 0, 0, 0, NULL);
+		}
 	}
+
 	return 0;
 }
 #endif /* HAVE_FOP_OPEN_EXEC */

fs/xfs/linux-2.6/xfs_fs_subr.c

@@ -16,66 +16,78 @@
  * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
  */
 #include "xfs.h"
+#include "xfs_vnodeops.h"
+
+/*
+ * The following six includes are needed so that we can include
+ * xfs_inode.h.  What a mess..
+ */
+#include "xfs_bmap_btree.h"
+#include "xfs_inum.h"
+#include "xfs_dir2.h"
+#include "xfs_dir2_sf.h"
+#include "xfs_attr_sf.h"
+#include "xfs_dinode.h"
+
+#include "xfs_inode.h"
 
 int  fs_noerr(void) { return 0; }
 int  fs_nosys(void) { return ENOSYS; }
 void fs_noval(void) { return; }
 
 void
-fs_tosspages(
-	bhv_desc_t	*bdp,
+xfs_tosspages(
+	xfs_inode_t	*ip,
 	xfs_off_t	first,
 	xfs_off_t	last,
 	int		fiopt)
 {
-	bhv_vnode_t	*vp = BHV_TO_VNODE(bdp);
-	struct inode	*ip = vn_to_inode(vp);
+	bhv_vnode_t	*vp = XFS_ITOV(ip);
+	struct inode	*inode = vn_to_inode(vp);
 
 	if (VN_CACHED(vp))
-		truncate_inode_pages(ip->i_mapping, first);
+		truncate_inode_pages(inode->i_mapping, first);
 }
 
 int
-fs_flushinval_pages(
-	bhv_desc_t	*bdp,
+xfs_flushinval_pages(
+	xfs_inode_t	*ip,
 	xfs_off_t	first,
 	xfs_off_t	last,
 	int		fiopt)
 {
-	bhv_vnode_t	*vp = BHV_TO_VNODE(bdp);
-	struct inode	*ip = vn_to_inode(vp);
+	bhv_vnode_t	*vp = XFS_ITOV(ip);
+	struct inode	*inode = vn_to_inode(vp);
 	int		ret = 0;
 
 	if (VN_CACHED(vp)) {
-		if (VN_TRUNC(vp))
-			VUNTRUNCATE(vp);
-		ret = filemap_write_and_wait(ip->i_mapping);
+		xfs_iflags_clear(ip, XFS_ITRUNCATED);
+		ret = filemap_write_and_wait(inode->i_mapping);
 		if (!ret)
-			truncate_inode_pages(ip->i_mapping, first);
+			truncate_inode_pages(inode->i_mapping, first);
 	}
 	return ret;
 }
 
 int
-fs_flush_pages(
-	bhv_desc_t	*bdp,
+xfs_flush_pages(
+	xfs_inode_t	*ip,
 	xfs_off_t	first,
 	xfs_off_t	last,
 	uint64_t	flags,
 	int		fiopt)
 {
-	bhv_vnode_t	*vp = BHV_TO_VNODE(bdp);
-	struct inode	*ip = vn_to_inode(vp);
+	bhv_vnode_t	*vp = XFS_ITOV(ip);
+	struct inode	*inode = vn_to_inode(vp);
 	int		ret = 0;
 	int		ret2;
 
 	if (VN_DIRTY(vp)) {
-		if (VN_TRUNC(vp))
-			VUNTRUNCATE(vp);
-		ret = filemap_fdatawrite(ip->i_mapping);
+		xfs_iflags_clear(ip, XFS_ITRUNCATED);
+		ret = filemap_fdatawrite(inode->i_mapping);
 		if (flags & XFS_B_ASYNC)
 			return ret;
-		ret2 = filemap_fdatawait(ip->i_mapping);
+		ret2 = filemap_fdatawait(inode->i_mapping);
 		if (!ret)
 			ret = ret2;
 	}

fs/xfs/linux-2.6/xfs_fs_subr.h

@@ -18,12 +18,8 @@
 #ifndef	__XFS_FS_SUBR_H__
 #define __XFS_FS_SUBR_H__
 
-struct cred;
 extern int  fs_noerr(void);
 extern int  fs_nosys(void);
 extern void fs_noval(void);
-extern void fs_tosspages(bhv_desc_t *, xfs_off_t, xfs_off_t, int);
-extern int  fs_flushinval_pages(bhv_desc_t *, xfs_off_t, xfs_off_t, int);
-extern int  fs_flush_pages(bhv_desc_t *, xfs_off_t, xfs_off_t, uint64_t, int);
 
 #endif	/* __XFS_FS_SUBR_H__ */

fs/xfs/linux-2.6/xfs_globals.c

@@ -19,11 +19,6 @@
 #include "xfs_cred.h"
 #include "xfs_sysctl.h"
 
-/*
- * System memory size - used to scale certain data structures in XFS.
- */
-unsigned long xfs_physmem;
-
 /*
  * Tunable XFS parameters.  xfs_params is required even when CONFIG_SYSCTL=n,
  * other XFS code uses these values.  Times are measured in centisecs (i.e.

fs/xfs/linux-2.6/xfs_globals.h

@@ -19,7 +19,6 @@
 #define __XFS_GLOBALS_H__
 
 extern uint64_t	xfs_panic_mask;		/* set to cause more panics */
-extern unsigned long xfs_physmem;
 extern struct cred *sys_cred;
 
 #endif	/* __XFS_GLOBALS_H__ */

fs/xfs/linux-2.6/xfs_ioctl.c

@@ -47,6 +47,7 @@
 #include "xfs_utils.h"
 #include "xfs_dfrag.h"
 #include "xfs_fsops.h"
+#include "xfs_vnodeops.h"
 
 #include <linux/capability.h>
 #include <linux/dcache.h>
@@ -137,7 +138,8 @@ xfs_find_handle(
 	vp = vn_from_inode(inode);
 
 	/* now we can grab the fsid */
-	memcpy(&handle.ha_fsid, vp->v_vfsp->vfs_altfsid, sizeof(xfs_fsid_t));
+	memcpy(&handle.ha_fsid, XFS_I(inode)->i_mount->m_fixedfsid,
+			sizeof(xfs_fsid_t));
 	hsize = sizeof(xfs_fsid_t);
 
 	if (cmd != XFS_IOC_PATH_TO_FSHANDLE) {
@@ -349,19 +351,44 @@ xfs_open_by_handle(
 	return new_fd;
 }
 
+/*
+ * This is a copy from fs/namei.c:vfs_readlink(), except for removing it's
+ * unused first argument.
+ */
+STATIC int
+do_readlink(
+	char __user		*buffer,
+	int			buflen,
+	const char		*link)
+{
+        int len;
+
+	len = PTR_ERR(link);
+	if (IS_ERR(link))
+		goto out;
+
+	len = strlen(link);
+	if (len > (unsigned) buflen)
+		len = buflen;
+	if (copy_to_user(buffer, link, len))
+		len = -EFAULT;
+ out:
+	return len;
+}
+
+
 STATIC int
 xfs_readlink_by_handle(
 	xfs_mount_t		*mp,
 	void			__user *arg,
 	struct inode		*parinode)
 {
-	int			error;
-	struct iovec		aiov;
-	struct uio		auio;
 	struct inode		*inode;
 	xfs_fsop_handlereq_t	hreq;
 	bhv_vnode_t		*vp;
 	__u32			olen;
+	void			*link;
+	int			error;
 
 	if (!capable(CAP_SYS_ADMIN))
 		return -XFS_ERROR(EPERM);
@@ -374,29 +401,31 @@ xfs_readlink_by_handle(
 
 	/* Restrict this handle operation to symlinks only. */
 	if (!S_ISLNK(inode->i_mode)) {
-		VN_RELE(vp);
-		return -XFS_ERROR(EINVAL);
+		error = -XFS_ERROR(EINVAL);
+		goto out_iput;
 	}
 
 	if (copy_from_user(&olen, hreq.ohandlen, sizeof(__u32))) {
-		VN_RELE(vp);
-		return -XFS_ERROR(EFAULT);
+		error = -XFS_ERROR(EFAULT);
+		goto out_iput;
 	}
-	aiov.iov_len	= olen;
-	aiov.iov_base	= hreq.ohandle;
 
-	auio.uio_iov	= (struct kvec *)&aiov;
-	auio.uio_iovcnt	= 1;
-	auio.uio_offset	= 0;
-	auio.uio_segflg	= UIO_USERSPACE;
-	auio.uio_resid	= olen;
+	link = kmalloc(MAXPATHLEN+1, GFP_KERNEL);
+	if (!link)
+		goto out_iput;
 
-	error = bhv_vop_readlink(vp, &auio, IO_INVIS, NULL);
-	VN_RELE(vp);
+	error = -xfs_readlink(XFS_I(inode), link);
 	if (error)
-		return -error;
+		goto out_kfree;
+	error = do_readlink(hreq.ohandle, olen, link);
+	if (error)
+		goto out_kfree;
 
-	return (olen - auio.uio_resid);
+ out_kfree:
+	kfree(link);
+ out_iput:
+	iput(inode);
+	return error;
 }
 
 STATIC int
@@ -409,7 +438,6 @@ xfs_fssetdm_by_handle(
 	struct fsdmidata	fsd;
 	xfs_fsop_setdm_handlereq_t dmhreq;
 	struct inode		*inode;
-	bhv_desc_t		*bdp;
 	bhv_vnode_t		*vp;
 
 	if (!capable(CAP_MKNOD))
@@ -431,8 +459,8 @@ xfs_fssetdm_by_handle(
 		return -XFS_ERROR(EFAULT);
 	}
 
-	bdp = bhv_base_unlocked(VN_BHV_HEAD(vp));
-	error = xfs_set_dmattrs(bdp, fsd.fsd_dmevmask, fsd.fsd_dmstate, NULL);
+	error = xfs_set_dmattrs(xfs_vtoi(vp),
+			fsd.fsd_dmevmask, fsd.fsd_dmstate);
 
 	VN_RELE(vp);
 	if (error)
@@ -470,8 +498,8 @@ xfs_attrlist_by_handle(
 		goto out_vn_rele;
 
 	cursor = (attrlist_cursor_kern_t *)&al_hreq.pos;
-	error = bhv_vop_attr_list(vp, kbuf, al_hreq.buflen, al_hreq.flags,
-					cursor, NULL);
+	error = xfs_attr_list(XFS_I(inode), kbuf, al_hreq.buflen,
+					al_hreq.flags, cursor);
 	if (error)
 		goto out_kfree;
 
@@ -488,7 +516,7 @@ xfs_attrlist_by_handle(
 
 STATIC int
 xfs_attrmulti_attr_get(
-	bhv_vnode_t		*vp,
+	struct inode		*inode,
 	char			*name,
 	char			__user *ubuf,
 	__uint32_t		*len,
@@ -503,7 +531,7 @@ xfs_attrmulti_attr_get(
 	if (!kbuf)
 		return ENOMEM;
 
-	error = bhv_vop_attr_get(vp, name, kbuf, len, flags, NULL);
+	error = xfs_attr_get(XFS_I(inode), name, kbuf, len, flags, NULL);
 	if (error)
 		goto out_kfree;
 
@@ -517,7 +545,7 @@ xfs_attrmulti_attr_get(
 
 STATIC int
 xfs_attrmulti_attr_set(
-	bhv_vnode_t		*vp,
+	struct inode		*inode,
 	char			*name,
 	const char		__user *ubuf,
 	__uint32_t		len,
@@ -526,9 +554,9 @@ xfs_attrmulti_attr_set(
 	char			*kbuf;
 	int			error = EFAULT;
 
-	if (IS_RDONLY(&vp->v_inode))
+	if (IS_RDONLY(inode))
 		return -EROFS;
-	if (IS_IMMUTABLE(&vp->v_inode) || IS_APPEND(&vp->v_inode))
+	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
 		return EPERM;
 	if (len > XATTR_SIZE_MAX)
 		return EINVAL;
@@ -540,7 +568,7 @@ xfs_attrmulti_attr_set(
 	if (copy_from_user(kbuf, ubuf, len))
 		goto out_kfree;
 			
-	error = bhv_vop_attr_set(vp, name, kbuf, len, flags, NULL);
+	error = xfs_attr_set(XFS_I(inode), name, kbuf, len, flags);
 
  out_kfree:
 	kfree(kbuf);
@@ -549,15 +577,15 @@ xfs_attrmulti_attr_set(
 
 STATIC int
 xfs_attrmulti_attr_remove(
-	bhv_vnode_t		*vp,
+	struct inode		*inode,
 	char			*name,
 	__uint32_t		flags)
 {
-	if (IS_RDONLY(&vp->v_inode))
+	if (IS_RDONLY(inode))
 		return -EROFS;
-	if (IS_IMMUTABLE(&vp->v_inode) || IS_APPEND(&vp->v_inode))
+	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
 		return EPERM;
-	return bhv_vop_attr_remove(vp, name, flags, NULL);
+	return xfs_attr_remove(XFS_I(inode), name, flags);
 }
 
 STATIC int
@@ -613,17 +641,17 @@ xfs_attrmulti_by_handle(
 
 		switch (ops[i].am_opcode) {
 		case ATTR_OP_GET:
-			ops[i].am_error = xfs_attrmulti_attr_get(vp,
+			ops[i].am_error = xfs_attrmulti_attr_get(inode,
 					attr_name, ops[i].am_attrvalue,
 					&ops[i].am_length, ops[i].am_flags);
 			break;
 		case ATTR_OP_SET:
-			ops[i].am_error = xfs_attrmulti_attr_set(vp,
+			ops[i].am_error = xfs_attrmulti_attr_set(inode,
 					attr_name, ops[i].am_attrvalue,
 					ops[i].am_length, ops[i].am_flags);
 			break;
 		case ATTR_OP_REMOVE:
-			ops[i].am_error = xfs_attrmulti_attr_remove(vp,
+			ops[i].am_error = xfs_attrmulti_attr_remove(inode,
 					attr_name, ops[i].am_flags);
 			break;
 		default:
@@ -649,7 +677,7 @@ xfs_attrmulti_by_handle(
 
 STATIC int
 xfs_ioc_space(
-	bhv_desc_t		*bdp,
+	struct xfs_inode	*ip,
 	struct inode		*inode,
 	struct file		*filp,
 	int			flags,
@@ -680,38 +708,38 @@ xfs_ioc_xattr(
 	unsigned int		cmd,
 	void			__user *arg);
 
+STATIC int
+xfs_ioc_fsgetxattr(
+	xfs_inode_t		*ip,
+	int			attr,
+	void			__user *arg);
+
 STATIC int
 xfs_ioc_getbmap(
-	bhv_desc_t		*bdp,
+	struct xfs_inode	*ip,
 	int			flags,
 	unsigned int		cmd,
 	void			__user *arg);
 
 STATIC int
 xfs_ioc_getbmapx(
-	bhv_desc_t		*bdp,
+	struct xfs_inode	*ip,
 	void			__user *arg);
 
 int
 xfs_ioctl(
-	bhv_desc_t		*bdp,
-	struct inode		*inode,
+	xfs_inode_t		*ip,
 	struct file		*filp,
 	int			ioflags,
 	unsigned int		cmd,
 	void			__user *arg)
 {
+	struct inode		*inode = filp->f_path.dentry->d_inode;
+	bhv_vnode_t		*vp = vn_from_inode(inode);
+	xfs_mount_t		*mp = ip->i_mount;
 	int			error;
-	bhv_vnode_t		*vp;
-	xfs_inode_t		*ip;
-	xfs_mount_t		*mp;
 
-	vp = vn_from_inode(inode);
-
-	vn_trace_entry(vp, "xfs_ioctl", (inst_t *)__return_address);
-
-	ip = XFS_BHVTOI(bdp);
-	mp = ip->i_mount;
+	vn_trace_entry(XFS_I(inode), "xfs_ioctl", (inst_t *)__return_address);
 
 	switch (cmd) {
 
@@ -731,7 +759,7 @@ xfs_ioctl(
 		    !capable(CAP_SYS_ADMIN))
 			return -EPERM;
 
-		return xfs_ioc_space(bdp, inode, filp, ioflags, cmd, arg);
+		return xfs_ioc_space(ip, inode, filp, ioflags, cmd, arg);
 
 	case XFS_IOC_DIOINFO: {
 		struct dioattr	da;
@@ -761,11 +789,13 @@ xfs_ioctl(
 	case XFS_IOC_GETVERSION:
 		return put_user(inode->i_generation, (int __user *)arg);
 
+	case XFS_IOC_FSGETXATTR:
+		return xfs_ioc_fsgetxattr(ip, 0, arg);
+	case XFS_IOC_FSGETXATTRA:
+		return xfs_ioc_fsgetxattr(ip, 1, arg);
 	case XFS_IOC_GETXFLAGS:
 	case XFS_IOC_SETXFLAGS:
-	case XFS_IOC_FSGETXATTR:
 	case XFS_IOC_FSSETXATTR:
-	case XFS_IOC_FSGETXATTRA:
 		return xfs_ioc_xattr(vp, ip, filp, cmd, arg);
 
 	case XFS_IOC_FSSETDM: {
@@ -774,17 +804,17 @@ xfs_ioctl(
 		if (copy_from_user(&dmi, arg, sizeof(dmi)))
 			return -XFS_ERROR(EFAULT);
 
-		error = xfs_set_dmattrs(bdp, dmi.fsd_dmevmask, dmi.fsd_dmstate,
-							NULL);
+		error = xfs_set_dmattrs(ip, dmi.fsd_dmevmask,
+				dmi.fsd_dmstate);
 		return -error;
 	}
 
 	case XFS_IOC_GETBMAP:
 	case XFS_IOC_GETBMAPA:
-		return xfs_ioc_getbmap(bdp, ioflags, cmd, arg);
+		return xfs_ioc_getbmap(ip, ioflags, cmd, arg);
 
 	case XFS_IOC_GETBMAPX:
-		return xfs_ioc_getbmapx(bdp, arg);
+		return xfs_ioc_getbmapx(ip, arg);
 
 	case XFS_IOC_FD_TO_HANDLE:
 	case XFS_IOC_PATH_TO_HANDLE:
@@ -944,7 +974,7 @@ xfs_ioctl(
 		if (!capable(CAP_SYS_ADMIN))
 			return -EPERM;
 
-		error = xfs_errortag_clearall(mp);
+		error = xfs_errortag_clearall(mp, 1);
 		return -error;
 
 	default:
@@ -954,7 +984,7 @@ xfs_ioctl(
 
 STATIC int
 xfs_ioc_space(
-	bhv_desc_t		*bdp,
+	struct xfs_inode	*ip,
 	struct inode		*inode,
 	struct file		*filp,
 	int			ioflags,
@@ -982,7 +1012,7 @@ xfs_ioc_space(
 	if (ioflags & IO_INVIS)
 		attr_flags |= ATTR_DMI;
 
-	error = xfs_change_file_space(bdp, cmd, &bf, filp->f_pos,
+	error = xfs_change_file_space(ip, cmd, &bf, filp->f_pos,
 					      NULL, attr_flags);
 	return -error;
 }
@@ -1139,6 +1169,42 @@ xfs_di2lxflags(
 	return flags;
 }
 
+STATIC int
+xfs_ioc_fsgetxattr(
+	xfs_inode_t		*ip,
+	int			attr,
+	void			__user *arg)
+{
+	struct fsxattr		fa;
+
+	xfs_ilock(ip, XFS_ILOCK_SHARED);
+	fa.fsx_xflags = xfs_ip2xflags(ip);
+	fa.fsx_extsize = ip->i_d.di_extsize << ip->i_mount->m_sb.sb_blocklog;
+	fa.fsx_projid = ip->i_d.di_projid;
+
+	if (attr) {
+		if (ip->i_afp) {
+			if (ip->i_afp->if_flags & XFS_IFEXTENTS)
+				fa.fsx_nextents = ip->i_afp->if_bytes /
+							sizeof(xfs_bmbt_rec_t);
+			else
+				fa.fsx_nextents = ip->i_d.di_anextents;
+		} else
+			fa.fsx_nextents = 0;
+	} else {
+		if (ip->i_df.if_flags & XFS_IFEXTENTS)
+			fa.fsx_nextents = ip->i_df.if_bytes /
+						sizeof(xfs_bmbt_rec_t);
+		else
+			fa.fsx_nextents = ip->i_d.di_nextents;
+	}
+	xfs_iunlock(ip, XFS_ILOCK_SHARED);
+
+	if (copy_to_user(arg, &fa, sizeof(fa)))
+		return -EFAULT;
+	return 0;
+}
+
 STATIC int
 xfs_ioc_xattr(
 	bhv_vnode_t		*vp,
@@ -1158,27 +1224,6 @@ xfs_ioc_xattr(
 		return -ENOMEM;
 
 	switch (cmd) {
-	case XFS_IOC_FSGETXATTR: {
-		vattr->va_mask = XFS_AT_XFLAGS | XFS_AT_EXTSIZE | \
-				 XFS_AT_NEXTENTS | XFS_AT_PROJID;
-		error = bhv_vop_getattr(vp, vattr, 0, NULL);
-		if (unlikely(error)) {
-			error = -error;
-			break;
-		}
-
-		fa.fsx_xflags	= vattr->va_xflags;
-		fa.fsx_extsize	= vattr->va_extsize;
-		fa.fsx_nextents = vattr->va_nextents;
-		fa.fsx_projid	= vattr->va_projid;
-
-		if (copy_to_user(arg, &fa, sizeof(fa))) {
-			error = -EFAULT;
-			break;
-		}
-		break;
-	}
-
 	case XFS_IOC_FSSETXATTR: {
 		if (copy_from_user(&fa, arg, sizeof(fa))) {
 			error = -EFAULT;
@@ -1194,34 +1239,13 @@ xfs_ioc_xattr(
 		vattr->va_extsize = fa.fsx_extsize;
 		vattr->va_projid  = fa.fsx_projid;
 
-		error = bhv_vop_setattr(vp, vattr, attr_flags, NULL);
+		error = xfs_setattr(ip, vattr, attr_flags, NULL);
 		if (likely(!error))
 			__vn_revalidate(vp, vattr);	/* update flags */
 		error = -error;
 		break;
 	}
 
-	case XFS_IOC_FSGETXATTRA: {
-		vattr->va_mask = XFS_AT_XFLAGS | XFS_AT_EXTSIZE | \
-				 XFS_AT_ANEXTENTS | XFS_AT_PROJID;
-		error = bhv_vop_getattr(vp, vattr, 0, NULL);
-		if (unlikely(error)) {
-			error = -error;
-			break;
-		}
-
-		fa.fsx_xflags	= vattr->va_xflags;
-		fa.fsx_extsize	= vattr->va_extsize;
-		fa.fsx_nextents = vattr->va_anextents;
-		fa.fsx_projid	= vattr->va_projid;
-
-		if (copy_to_user(arg, &fa, sizeof(fa))) {
-			error = -EFAULT;
-			break;
-		}
-		break;
-	}
-
 	case XFS_IOC_GETXFLAGS: {
 		flags = xfs_di2lxflags(ip->i_d.di_flags);
 		if (copy_to_user(arg, &flags, sizeof(flags)))
@@ -1250,7 +1274,7 @@ xfs_ioc_xattr(
 		vattr->va_xflags = xfs_merge_ioc_xflags(flags,
 							xfs_ip2xflags(ip));
 
-		error = bhv_vop_setattr(vp, vattr, attr_flags, NULL);
+		error = xfs_setattr(ip, vattr, attr_flags, NULL);
 		if (likely(!error))
 			__vn_revalidate(vp, vattr);	/* update flags */
 		error = -error;
@@ -1268,7 +1292,7 @@ xfs_ioc_xattr(
 
 STATIC int
 xfs_ioc_getbmap(
-	bhv_desc_t		*bdp,
+	struct xfs_inode	*ip,
 	int			ioflags,
 	unsigned int		cmd,
 	void			__user *arg)
@@ -1287,7 +1311,7 @@ xfs_ioc_getbmap(
 	if (ioflags & IO_INVIS)
 		iflags |= BMV_IF_NO_DMAPI_READ;
 
-	error = xfs_getbmap(bdp, &bm, (struct getbmap __user *)arg+1, iflags);
+	error = xfs_getbmap(ip, &bm, (struct getbmap __user *)arg+1, iflags);
 	if (error)
 		return -error;
 
@@ -1298,7 +1322,7 @@ xfs_ioc_getbmap(
 
 STATIC int
 xfs_ioc_getbmapx(
-	bhv_desc_t		*bdp,
+	struct xfs_inode	*ip,
 	void			__user *arg)
 {
 	struct getbmapx		bmx;
@@ -1325,7 +1349,7 @@ xfs_ioc_getbmapx(
 
 	iflags |= BMV_IF_EXTENDED;
 
-	error = xfs_getbmap(bdp, &bm, (struct getbmapx __user *)arg+1, iflags);
+	error = xfs_getbmap(ip, &bm, (struct getbmapx __user *)arg+1, iflags);
 	if (error)
 		return -error;
 

fs/xfs/linux-2.6/xfs_ioctl32.c

@@ -43,6 +43,7 @@
 #include "xfs_itable.h"
 #include "xfs_error.h"
 #include "xfs_dfrag.h"
+#include "xfs_vnodeops.h"
 
 #define  _NATIVE_IOC(cmd, type) \
 	  _IOC(_IOC_DIR(cmd), _IOC_TYPE(cmd), _IOC_NR(cmd), sizeof(type))
@@ -370,7 +371,6 @@ xfs_compat_ioctl(
 	unsigned long	arg)
 {
 	struct inode	*inode = file->f_path.dentry->d_inode;
-	bhv_vnode_t	*vp = vn_from_inode(inode);
 	int		error;
 
 	switch (cmd) {
@@ -443,7 +443,7 @@ xfs_compat_ioctl(
 	case XFS_IOC_FSBULKSTAT_SINGLE_32:
 	case XFS_IOC_FSINUMBERS_32:
 		cmd = _NATIVE_IOC(cmd, struct xfs_fsop_bulkreq);
-		return xfs_ioc_bulkstat_compat(XFS_BHVTOI(VNHEAD(vp))->i_mount,
+		return xfs_ioc_bulkstat_compat(XFS_I(inode)->i_mount,
 				cmd, (void __user*)arg);
 	case XFS_IOC_FD_TO_HANDLE_32:
 	case XFS_IOC_PATH_TO_HANDLE_32:
@@ -457,8 +457,8 @@ xfs_compat_ioctl(
 		return -ENOIOCTLCMD;
 	}
 
-	error = bhv_vop_ioctl(vp, inode, file, mode, cmd, (void __user *)arg);
-	VMODIFY(vp);
+	error = xfs_ioctl(XFS_I(inode), file, mode, cmd, (void __user *)arg);
+	xfs_iflags_set(XFS_I(inode), XFS_IMODIFIED);
 
 	return error;
 }

fs/xfs/linux-2.6/xfs_iops.c

@@ -46,28 +46,13 @@
 #include "xfs_attr.h"
 #include "xfs_buf_item.h"
 #include "xfs_utils.h"
+#include "xfs_vnodeops.h"
 
 #include <linux/capability.h>
 #include <linux/xattr.h>
 #include <linux/namei.h>
 #include <linux/security.h>
 
-/*
- * Get a XFS inode from a given vnode.
- */
-xfs_inode_t *
-xfs_vtoi(
-	bhv_vnode_t	*vp)
-{
-	bhv_desc_t      *bdp;
-
-	bdp = bhv_lookup_range(VN_BHV_HEAD(vp),
-			VNODE_POSITION_XFS, VNODE_POSITION_XFS);
-	if (unlikely(bdp == NULL))
-		return NULL;
-	return XFS_BHVTOI(bdp);
-}
-
 /*
  * Bring the atime in the XFS inode uptodate.
  * Used before logging the inode to disk or when the Linux inode goes away.
@@ -80,9 +65,8 @@ xfs_synchronize_atime(
 
 	vp = XFS_ITOV_NULL(ip);
 	if (vp) {
-		struct inode *inode = &vp->v_inode;
-		ip->i_d.di_atime.t_sec = (__int32_t)inode->i_atime.tv_sec;
-		ip->i_d.di_atime.t_nsec = (__int32_t)inode->i_atime.tv_nsec;
+		ip->i_d.di_atime.t_sec = (__int32_t)vp->i_atime.tv_sec;
+		ip->i_d.di_atime.t_nsec = (__int32_t)vp->i_atime.tv_nsec;
 	}
 }
 
@@ -195,18 +179,19 @@ xfs_ichgtime_fast(
  */
 STATIC void
 xfs_validate_fields(
-	struct inode	*ip,
-	bhv_vattr_t	*vattr)
+	struct inode		*inode)
 {
-	vattr->va_mask = XFS_AT_NLINK|XFS_AT_SIZE|XFS_AT_NBLOCKS;
-	if (!bhv_vop_getattr(vn_from_inode(ip), vattr, ATTR_LAZY, NULL)) {
-		ip->i_nlink = vattr->va_nlink;
-		ip->i_blocks = vattr->va_nblocks;
-
-		/* we're under i_sem so i_size can't change under us */
-		if (i_size_read(ip) != vattr->va_size)
-			i_size_write(ip, vattr->va_size);
-	}
+	struct xfs_inode	*ip = XFS_I(inode);
+	loff_t size;
+
+	inode->i_nlink = ip->i_d.di_nlink;
+	inode->i_blocks =
+		XFS_FSB_TO_BB(ip->i_mount, ip->i_d.di_nblocks +
+					   ip->i_delayed_blks);
+	/* we're under i_sem so i_size can't change under us */
+	size = XFS_ISIZE(ip);
+	if (i_size_read(inode) != size)
+		i_size_write(inode, size);
 }
 
 /*
@@ -233,9 +218,10 @@ xfs_init_security(
 		return -error;
 	}
 
-	error = bhv_vop_attr_set(vp, name, value, length, ATTR_SECURE, NULL);
+	error = xfs_attr_set(XFS_I(ip), name, value,
+			length, ATTR_SECURE);
 	if (!error)
-		VMODIFY(vp);
+		xfs_iflags_set(XFS_I(ip), XFS_IMODIFIED);
 
 	kfree(name);
 	kfree(value);
@@ -256,7 +242,7 @@ xfs_has_fs_struct(struct task_struct *task)
 
 STATIC void
 xfs_cleanup_inode(
-	bhv_vnode_t	*dvp,
+	struct inode	*dir,
 	bhv_vnode_t	*vp,
 	struct dentry	*dentry,
 	int		mode)
@@ -272,9 +258,9 @@ xfs_cleanup_inode(
 	teardown.d_name = dentry->d_name;
 
 	if (S_ISDIR(mode))
-	  	bhv_vop_rmdir(dvp, &teardown, NULL);
+		xfs_rmdir(XFS_I(dir), &teardown);
 	else
-		bhv_vop_remove(dvp, &teardown, NULL);
+		xfs_remove(XFS_I(dir), &teardown);
 	VN_RELE(vp);
 }
 
@@ -286,7 +272,6 @@ xfs_vn_mknod(
 	dev_t		rdev)
 {
 	struct inode	*ip;
-	bhv_vattr_t	vattr = { 0 };
 	bhv_vnode_t	*vp = NULL, *dvp = vn_from_inode(dir);
 	xfs_acl_t	*default_acl = NULL;
 	attrexists_t	test_default_acl = _ACL_DEFAULT_EXISTS;
@@ -312,19 +297,14 @@ xfs_vn_mknod(
 	if (IS_POSIXACL(dir) && !default_acl && xfs_has_fs_struct(current))
 		mode &= ~current->fs->umask;
 
-	vattr.va_mask = XFS_AT_TYPE|XFS_AT_MODE;
-	vattr.va_mode = mode;
-
 	switch (mode & S_IFMT) {
 	case S_IFCHR: case S_IFBLK: case S_IFIFO: case S_IFSOCK:
-		vattr.va_rdev = sysv_encode_dev(rdev);
-		vattr.va_mask |= XFS_AT_RDEV;
-		/*FALLTHROUGH*/
+		rdev = sysv_encode_dev(rdev);
 	case S_IFREG:
-		error = bhv_vop_create(dvp, dentry, &vattr, &vp, NULL);
+		error = xfs_create(XFS_I(dir), dentry, mode, rdev, &vp, NULL);
 		break;
 	case S_IFDIR:
-		error = bhv_vop_mkdir(dvp, dentry, &vattr, &vp, NULL);
+		error = xfs_mkdir(XFS_I(dir), dentry, mode, &vp, NULL);
 		break;
 	default:
 		error = EINVAL;
@@ -334,16 +314,16 @@ xfs_vn_mknod(
 	if (unlikely(!error)) {
 		error = xfs_init_security(vp, dir);
 		if (error)
-			xfs_cleanup_inode(dvp, vp, dentry, mode);
+			xfs_cleanup_inode(dir, vp, dentry, mode);
 	}
 
 	if (unlikely(default_acl)) {
 		if (!error) {
-			error = _ACL_INHERIT(vp, &vattr, default_acl);
+			error = _ACL_INHERIT(vp, mode, default_acl);
 			if (!error)
-				VMODIFY(vp);
+				xfs_iflags_set(XFS_I(vp), XFS_IMODIFIED);
 			else
-				xfs_cleanup_inode(dvp, vp, dentry, mode);
+				xfs_cleanup_inode(dir, vp, dentry, mode);
 		}
 		_ACL_FREE(default_acl);
 	}
@@ -355,9 +335,9 @@ xfs_vn_mknod(
 		if (S_ISCHR(mode) || S_ISBLK(mode))
 			ip->i_rdev = rdev;
 		else if (S_ISDIR(mode))
-			xfs_validate_fields(ip, &vattr);
+			xfs_validate_fields(ip);
 		d_instantiate(dentry, ip);
-		xfs_validate_fields(dir, &vattr);
+		xfs_validate_fields(dir);
 	}
 	return -error;
 }
@@ -387,13 +367,13 @@ xfs_vn_lookup(
 	struct dentry	*dentry,
 	struct nameidata *nd)
 {
-	bhv_vnode_t	*vp = vn_from_inode(dir), *cvp;
+	bhv_vnode_t	*cvp;
 	int		error;
 
 	if (dentry->d_name.len >= MAXNAMELEN)
 		return ERR_PTR(-ENAMETOOLONG);
 
-	error = bhv_vop_lookup(vp, dentry, &cvp, 0, NULL, NULL);
+	error = xfs_lookup(XFS_I(dir), dentry, &cvp);
 	if (unlikely(error)) {
 		if (unlikely(error != ENOENT))
 			return ERR_PTR(-error);
@@ -411,22 +391,19 @@ xfs_vn_link(
 	struct dentry	*dentry)
 {
 	struct inode	*ip;	/* inode of guy being linked to */
-	bhv_vnode_t	*tdvp;	/* target directory for new name/link */
 	bhv_vnode_t	*vp;	/* vp of name being linked */
-	bhv_vattr_t	vattr;
 	int		error;
 
 	ip = old_dentry->d_inode;	/* inode being linked to */
-	tdvp = vn_from_inode(dir);
 	vp = vn_from_inode(ip);
 
 	VN_HOLD(vp);
-	error = bhv_vop_link(tdvp, vp, dentry, NULL);
+	error = xfs_link(XFS_I(dir), vp, dentry);
 	if (unlikely(error)) {
 		VN_RELE(vp);
 	} else {
-		VMODIFY(tdvp);
-		xfs_validate_fields(ip, &vattr);
+		xfs_iflags_set(XFS_I(dir), XFS_IMODIFIED);
+		xfs_validate_fields(ip);
 		d_instantiate(dentry, ip);
 	}
 	return -error;
@@ -438,17 +415,14 @@ xfs_vn_unlink(
 	struct dentry	*dentry)
 {
 	struct inode	*inode;
-	bhv_vnode_t	*dvp;	/* directory containing name to remove */
-	bhv_vattr_t	vattr;
 	int		error;
 
 	inode = dentry->d_inode;
-	dvp = vn_from_inode(dir);
 
-	error = bhv_vop_remove(dvp, dentry, NULL);
+	error = xfs_remove(XFS_I(dir), dentry);
 	if (likely(!error)) {
-		xfs_validate_fields(dir, &vattr);	/* size needs update */
-		xfs_validate_fields(inode, &vattr);
+		xfs_validate_fields(dir);	/* size needs update */
+		xfs_validate_fields(inode);
 	}
 	return -error;
 }
@@ -460,28 +434,26 @@ xfs_vn_symlink(
 	const char	*symname)
 {
 	struct inode	*ip;
-	bhv_vattr_t	va = { 0 };
-	bhv_vnode_t	*dvp;	/* directory containing name of symlink */
 	bhv_vnode_t	*cvp;	/* used to lookup symlink to put in dentry */
 	int		error;
+	mode_t		mode;
 
-	dvp = vn_from_inode(dir);
 	cvp = NULL;
 
-	va.va_mode = S_IFLNK |
+	mode = S_IFLNK |
 		(irix_symlink_mode ? 0777 & ~current->fs->umask : S_IRWXUGO);
-	va.va_mask = XFS_AT_TYPE|XFS_AT_MODE;
 
-	error = bhv_vop_symlink(dvp, dentry, &va, (char *)symname, &cvp, NULL);
+	error = xfs_symlink(XFS_I(dir), dentry, (char *)symname, mode,
+			    &cvp, NULL);
 	if (likely(!error && cvp)) {
 		error = xfs_init_security(cvp, dir);
 		if (likely(!error)) {
 			ip = vn_to_inode(cvp);
 			d_instantiate(dentry, ip);
-			xfs_validate_fields(dir, &va);
-			xfs_validate_fields(ip, &va);
+			xfs_validate_fields(dir);
+			xfs_validate_fields(ip);
 		} else {
-			xfs_cleanup_inode(dvp, cvp, dentry, 0);
+			xfs_cleanup_inode(dir, cvp, dentry, 0);
 		}
 	}
 	return -error;
@@ -493,14 +465,12 @@ xfs_vn_rmdir(
 	struct dentry	*dentry)
 {
 	struct inode	*inode = dentry->d_inode;
-	bhv_vnode_t	*dvp = vn_from_inode(dir);
-	bhv_vattr_t	vattr;
 	int		error;
 
-	error = bhv_vop_rmdir(dvp, dentry, NULL);
+	error = xfs_rmdir(XFS_I(dir), dentry);
 	if (likely(!error)) {
-		xfs_validate_fields(inode, &vattr);
-		xfs_validate_fields(dir, &vattr);
+		xfs_validate_fields(inode);
+		xfs_validate_fields(dir);
 	}
 	return -error;
 }
@@ -513,21 +483,18 @@ xfs_vn_rename(
 	struct dentry	*ndentry)
 {
 	struct inode	*new_inode = ndentry->d_inode;
-	bhv_vnode_t	*fvp;	/* from directory */
 	bhv_vnode_t	*tvp;	/* target directory */
-	bhv_vattr_t	vattr;
 	int		error;
 
-	fvp = vn_from_inode(odir);
 	tvp = vn_from_inode(ndir);
 
-	error = bhv_vop_rename(fvp, odentry, tvp, ndentry, NULL);
+	error = xfs_rename(XFS_I(odir), odentry, tvp, ndentry);
 	if (likely(!error)) {
 		if (new_inode)
-			xfs_validate_fields(new_inode, &vattr);
-		xfs_validate_fields(odir, &vattr);
+			xfs_validate_fields(new_inode);
+		xfs_validate_fields(odir);
 		if (ndir != odir)
-			xfs_validate_fields(ndir, &vattr);
+			xfs_validate_fields(ndir);
 	}
 	return -error;
 }
@@ -542,50 +509,25 @@ xfs_vn_follow_link(
 	struct dentry		*dentry,
 	struct nameidata	*nd)
 {
-	bhv_vnode_t		*vp;
-	uio_t			*uio;
-	iovec_t			iov;
-	int			error;
 	char			*link;
-
-	ASSERT(dentry);
-	ASSERT(nd);
+	int			error = -ENOMEM;
 
 	link = kmalloc(MAXPATHLEN+1, GFP_KERNEL);
-	if (!link) {
-		nd_set_link(nd, ERR_PTR(-ENOMEM));
-		return NULL;
-	}
-
-	uio = kmalloc(sizeof(uio_t), GFP_KERNEL);
-	if (!uio) {
-		kfree(link);
-		nd_set_link(nd, ERR_PTR(-ENOMEM));
-		return NULL;
-	}
-
-	vp = vn_from_inode(dentry->d_inode);
-
-	iov.iov_base = link;
-	iov.iov_len = MAXPATHLEN;
+	if (!link)
+		goto out_err;
 
-	uio->uio_iov = &iov;
-	uio->uio_offset = 0;
-	uio->uio_segflg = UIO_SYSSPACE;
-	uio->uio_resid = MAXPATHLEN;
-	uio->uio_iovcnt = 1;
-
-	error = bhv_vop_readlink(vp, uio, 0, NULL);
-	if (unlikely(error)) {
-		kfree(link);
-		link = ERR_PTR(-error);
-	} else {
-		link[MAXPATHLEN - uio->uio_resid] = '\0';
-	}
-	kfree(uio);
+	error = -xfs_readlink(XFS_I(dentry->d_inode), link);
+	if (unlikely(error))
+		goto out_kfree;
 
 	nd_set_link(nd, link);
 	return NULL;
+
+ out_kfree:
+	kfree(link);
+ out_err:
+	nd_set_link(nd, ERR_PTR(error));
+	return NULL;
 }
 
 STATIC void
@@ -607,7 +549,7 @@ xfs_vn_permission(
 	int		mode,
 	struct nameidata *nd)
 {
-	return -bhv_vop_access(vn_from_inode(inode), mode << 6, NULL);
+	return -xfs_access(XFS_I(inode), mode << 6, NULL);
 }
 #else
 #define xfs_vn_permission NULL
@@ -620,11 +562,10 @@ xfs_vn_getattr(
 	struct kstat	*stat)
 {
 	struct inode	*inode = dentry->d_inode;
-	bhv_vnode_t	*vp = vn_from_inode(inode);
 	bhv_vattr_t	vattr = { .va_mask = XFS_AT_STAT };
 	int		error;
 
-	error = bhv_vop_getattr(vp, &vattr, ATTR_LAZY, NULL);
+	error = xfs_getattr(XFS_I(inode), &vattr, ATTR_LAZY);
 	if (likely(!error)) {
 		stat->size = i_size_read(inode);
 		stat->dev = inode->i_sb->s_dev;
@@ -652,7 +593,6 @@ xfs_vn_setattr(
 {
 	struct inode	*inode = dentry->d_inode;
 	unsigned int	ia_valid = attr->ia_valid;
-	bhv_vnode_t	*vp = vn_from_inode(inode);
 	bhv_vattr_t	vattr = { 0 };
 	int		flags = 0;
 	int		error;
@@ -696,9 +636,9 @@ xfs_vn_setattr(
 		flags |= ATTR_NONBLOCK;
 #endif
 
-	error = bhv_vop_setattr(vp, &vattr, flags, NULL);
+	error = xfs_setattr(XFS_I(inode), &vattr, flags, NULL);
 	if (likely(!error))
-		__vn_revalidate(vp, &vattr);
+		__vn_revalidate(vn_from_inode(inode), &vattr);
 	return -error;
 }
 

fs/xfs/linux-2.6/xfs_iops.h

@@ -26,11 +26,15 @@ extern const struct file_operations xfs_file_operations;
 extern const struct file_operations xfs_dir_file_operations;
 extern const struct file_operations xfs_invis_file_operations;
 
-extern int xfs_ioctl(struct bhv_desc *, struct inode *, struct file *,
-                        int, unsigned int, void __user *);
 
 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)
+
 #endif /* __XFS_IOPS_H__ */

fs/xfs/linux-2.6/xfs_linux.h

@@ -51,7 +51,6 @@
 
 #include <support/ktrace.h>
 #include <support/debug.h>
-#include <support/move.h>
 #include <support/uuid.h>
 
 #include <linux/mm.h>
@@ -75,6 +74,7 @@
 #include <linux/cpu.h>
 #include <linux/notifier.h>
 #include <linux/delay.h>
+#include <linux/log2.h>
 
 #include <asm/page.h>
 #include <asm/div64.h>
@@ -83,7 +83,6 @@
 #include <asm/byteorder.h>
 #include <asm/unaligned.h>
 
-#include <xfs_behavior.h>
 #include <xfs_vfs.h>
 #include <xfs_cred.h>
 #include <xfs_vnode.h>

fs/xfs/linux-2.6/xfs_lrw.c

@@ -48,6 +48,7 @@
 #include "xfs_buf_item.h"
 #include "xfs_utils.h"
 #include "xfs_iomap.h"
+#include "xfs_vnodeops.h"
 
 #include <linux/capability.h>
 #include <linux/writeback.h>
@@ -169,27 +170,22 @@ xfs_iozero(
 
 ssize_t			/* bytes read, or (-)  error */
 xfs_read(
-	bhv_desc_t		*bdp,
+	xfs_inode_t		*ip,
 	struct kiocb		*iocb,
 	const struct iovec	*iovp,
 	unsigned int		segs,
 	loff_t			*offset,
-	int			ioflags,
-	cred_t			*credp)
+	int			ioflags)
 {
 	struct file		*file = iocb->ki_filp;
 	struct inode		*inode = file->f_mapping->host;
+	bhv_vnode_t		*vp = XFS_ITOV(ip);
+	xfs_mount_t		*mp = ip->i_mount;
 	size_t			size = 0;
 	ssize_t			ret = 0;
 	xfs_fsize_t		n;
-	xfs_inode_t		*ip;
-	xfs_mount_t		*mp;
-	bhv_vnode_t		*vp;
 	unsigned long		seg;
 
-	ip = XFS_BHVTOI(bdp);
-	vp = BHV_TO_VNODE(bdp);
-	mp = ip->i_mount;
 
 	XFS_STATS_INC(xs_read_calls);
 
@@ -234,13 +230,11 @@ xfs_read(
 		mutex_lock(&inode->i_mutex);
 	xfs_ilock(ip, XFS_IOLOCK_SHARED);
 
-	if (DM_EVENT_ENABLED(vp->v_vfsp, ip, DM_EVENT_READ) &&
-	    !(ioflags & IO_INVIS)) {
+	if (DM_EVENT_ENABLED(ip, DM_EVENT_READ) && !(ioflags & IO_INVIS)) {
 		bhv_vrwlock_t locktype = VRWLOCK_READ;
 		int dmflags = FILP_DELAY_FLAG(file) | DM_SEM_FLAG_RD(ioflags);
 
-		ret = -XFS_SEND_DATA(mp, DM_EVENT_READ,
-					BHV_TO_VNODE(bdp), *offset, size,
+		ret = -XFS_SEND_DATA(mp, DM_EVENT_READ, vp, *offset, size,
 					dmflags, &locktype);
 		if (ret) {
 			xfs_iunlock(ip, XFS_IOLOCK_SHARED);
@@ -252,8 +246,9 @@ xfs_read(
 
 	if (unlikely(ioflags & IO_ISDIRECT)) {
 		if (VN_CACHED(vp))
-			ret = bhv_vop_flushinval_pages(vp, ctooff(offtoct(*offset)),
-						 -1, FI_REMAPF_LOCKED);
+			ret = xfs_flushinval_pages(ip,
+					ctooff(offtoct(*offset)),
+					-1, FI_REMAPF_LOCKED);
 		mutex_unlock(&inode->i_mutex);
 		if (ret) {
 			xfs_iunlock(ip, XFS_IOLOCK_SHARED);
@@ -277,16 +272,15 @@ xfs_read(
 
 ssize_t
 xfs_splice_read(
-	bhv_desc_t		*bdp,
+	xfs_inode_t		*ip,
 	struct file		*infilp,
 	loff_t			*ppos,
 	struct pipe_inode_info	*pipe,
 	size_t			count,
 	int			flags,
-	int			ioflags,
-	cred_t			*credp)
+	int			ioflags)
 {
-	xfs_inode_t		*ip = XFS_BHVTOI(bdp);
+	bhv_vnode_t		*vp = XFS_ITOV(ip);
 	xfs_mount_t		*mp = ip->i_mount;
 	ssize_t			ret;
 
@@ -296,13 +290,11 @@ xfs_splice_read(
 
 	xfs_ilock(ip, XFS_IOLOCK_SHARED);
 
-	if (DM_EVENT_ENABLED(BHV_TO_VNODE(bdp)->v_vfsp, ip, DM_EVENT_READ) &&
-	    (!(ioflags & IO_INVIS))) {
+	if (DM_EVENT_ENABLED(ip, DM_EVENT_READ) && !(ioflags & IO_INVIS)) {
 		bhv_vrwlock_t locktype = VRWLOCK_READ;
 		int error;
 
-		error = XFS_SEND_DATA(mp, DM_EVENT_READ, BHV_TO_VNODE(bdp),
-					*ppos, count,
+		error = XFS_SEND_DATA(mp, DM_EVENT_READ, vp, *ppos, count,
 					FILP_DELAY_FLAG(infilp), &locktype);
 		if (error) {
 			xfs_iunlock(ip, XFS_IOLOCK_SHARED);
@@ -321,16 +313,15 @@ xfs_splice_read(
 
 ssize_t
 xfs_splice_write(
-	bhv_desc_t		*bdp,
+	xfs_inode_t		*ip,
 	struct pipe_inode_info	*pipe,
 	struct file		*outfilp,
 	loff_t			*ppos,
 	size_t			count,
 	int			flags,
-	int			ioflags,
-	cred_t			*credp)
+	int			ioflags)
 {
-	xfs_inode_t		*ip = XFS_BHVTOI(bdp);
+	bhv_vnode_t		*vp = XFS_ITOV(ip);
 	xfs_mount_t		*mp = ip->i_mount;
 	xfs_iocore_t		*io = &ip->i_iocore;
 	ssize_t			ret;
@@ -343,13 +334,11 @@ xfs_splice_write(
 
 	xfs_ilock(ip, XFS_IOLOCK_EXCL);
 
-	if (DM_EVENT_ENABLED(BHV_TO_VNODE(bdp)->v_vfsp, ip, DM_EVENT_WRITE) &&
-	    (!(ioflags & IO_INVIS))) {
+	if (DM_EVENT_ENABLED(ip, DM_EVENT_WRITE) && !(ioflags & IO_INVIS)) {
 		bhv_vrwlock_t locktype = VRWLOCK_WRITE;
 		int error;
 
-		error = XFS_SEND_DATA(mp, DM_EVENT_WRITE, BHV_TO_VNODE(bdp),
-					*ppos, count,
+		error = XFS_SEND_DATA(mp, DM_EVENT_WRITE, vp, *ppos, count,
 					FILP_DELAY_FLAG(outfilp), &locktype);
 		if (error) {
 			xfs_iunlock(ip, XFS_IOLOCK_EXCL);
@@ -583,24 +572,22 @@ out_lock:
 
 ssize_t				/* bytes written, or (-) error */
 xfs_write(
-	bhv_desc_t		*bdp,
+	struct xfs_inode	*xip,
 	struct kiocb		*iocb,
 	const struct iovec	*iovp,
 	unsigned int		nsegs,
 	loff_t			*offset,
-	int			ioflags,
-	cred_t			*credp)
+	int			ioflags)
 {
 	struct file		*file = iocb->ki_filp;
 	struct address_space	*mapping = file->f_mapping;
 	struct inode		*inode = mapping->host;
+	bhv_vnode_t		*vp = XFS_ITOV(xip);
 	unsigned long		segs = nsegs;
-	xfs_inode_t		*xip;
 	xfs_mount_t		*mp;
 	ssize_t			ret = 0, error = 0;
 	xfs_fsize_t		isize, new_size;
 	xfs_iocore_t		*io;
-	bhv_vnode_t		*vp;
 	int			iolock;
 	int			eventsent = 0;
 	bhv_vrwlock_t		locktype;
@@ -610,9 +597,6 @@ xfs_write(
 
 	XFS_STATS_INC(xs_write_calls);
 
-	vp = BHV_TO_VNODE(bdp);
-	xip = XFS_BHVTOI(bdp);
-
 	error = generic_segment_checks(iovp, &segs, &ocount, VERIFY_READ);
 	if (error)
 		return error;
@@ -626,7 +610,7 @@ xfs_write(
 	io = &xip->i_iocore;
 	mp = io->io_mount;
 
-	vfs_wait_for_freeze(vp->v_vfsp, SB_FREEZE_WRITE);
+	xfs_wait_for_freeze(mp, SB_FREEZE_WRITE);
 
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return -EIO;
@@ -653,7 +637,7 @@ start:
 		goto out_unlock_mutex;
 	}
 
-	if ((DM_EVENT_ENABLED(vp->v_vfsp, xip, DM_EVENT_WRITE) &&
+	if ((DM_EVENT_ENABLED(xip, DM_EVENT_WRITE) &&
 	    !(ioflags & IO_INVIS) && !eventsent)) {
 		int		dmflags = FILP_DELAY_FLAG(file);
 
@@ -722,7 +706,7 @@ start:
 	 */
 
 	if (pos > xip->i_size) {
-		error = xfs_zero_eof(BHV_TO_VNODE(bdp), io, pos, xip->i_size);
+		error = xfs_zero_eof(vp, io, pos, xip->i_size);
 		if (error) {
 			xfs_iunlock(xip, XFS_ILOCK_EXCL);
 			goto out_unlock_internal;
@@ -758,7 +742,8 @@ retry:
 			WARN_ON(need_i_mutex == 0);
 			xfs_inval_cached_trace(io, pos, -1,
 					ctooff(offtoct(pos)), -1);
-			error = bhv_vop_flushinval_pages(vp, ctooff(offtoct(pos)),
+			error = xfs_flushinval_pages(xip,
+					ctooff(offtoct(pos)),
 					-1, FI_REMAPF_LOCKED);
 			if (error)
 				goto out_unlock_internal;
@@ -805,11 +790,9 @@ retry:
 	if (ret == -EIOCBQUEUED && !(ioflags & IO_ISAIO))
 		ret = wait_on_sync_kiocb(iocb);
 
-	if ((ret == -ENOSPC) &&
-	    DM_EVENT_ENABLED(vp->v_vfsp, xip, DM_EVENT_NOSPACE) &&
-	    !(ioflags & IO_INVIS)) {
-
-		xfs_rwunlock(bdp, locktype);
+	if (ret == -ENOSPC &&
+	    DM_EVENT_ENABLED(xip, DM_EVENT_NOSPACE) && !(ioflags & IO_INVIS)) {
+		xfs_rwunlock(xip, locktype);
 		if (need_i_mutex)
 			mutex_unlock(&inode->i_mutex);
 		error = XFS_SEND_NAMESP(xip->i_mount, DM_EVENT_NOSPACE, vp,
@@ -817,7 +800,7 @@ retry:
 				0, 0, 0); /* Delay flag intentionally  unused */
 		if (need_i_mutex)
 			mutex_lock(&inode->i_mutex);
-		xfs_rwlock(bdp, locktype);
+		xfs_rwlock(xip, locktype);
 		if (error)
 			goto out_unlock_internal;
 		pos = xip->i_size;
@@ -844,20 +827,19 @@ retry:
 
 	/* Handle various SYNC-type writes */
 	if ((file->f_flags & O_SYNC) || IS_SYNC(inode)) {
-		error = xfs_write_sync_logforce(mp, xip);
-		if (error)
-			goto out_unlock_internal;
-
-		xfs_rwunlock(bdp, locktype);
+		int error2;
+		xfs_rwunlock(xip, locktype);
 		if (need_i_mutex)
 			mutex_unlock(&inode->i_mutex);
-
-		error = sync_page_range(inode, mapping, pos, ret);
+		error2 = sync_page_range(inode, mapping, pos, ret);
 		if (!error)
-			error = -ret;
+			error = error2;
 		if (need_i_mutex)
 			mutex_lock(&inode->i_mutex);
-		xfs_rwlock(bdp, locktype);
+		xfs_rwlock(xip, locktype);
+		error2 = xfs_write_sync_logforce(mp, xip);
+		if (!error)
+			error = error2;
 	}
 
  out_unlock_internal:
@@ -875,7 +857,7 @@ retry:
 			xip->i_d.di_size = xip->i_size;
 		xfs_iunlock(xip, XFS_ILOCK_EXCL);
 	}
-	xfs_rwunlock(bdp, locktype);
+	xfs_rwunlock(xip, locktype);
  out_unlock_mutex:
 	if (need_i_mutex)
 		mutex_unlock(&inode->i_mutex);
@@ -914,14 +896,14 @@ xfs_bdstrat_cb(struct xfs_buf *bp)
 
 
 int
-xfs_bmap(bhv_desc_t	*bdp,
+xfs_bmap(
+	xfs_inode_t	*ip,
 	xfs_off_t	offset,
 	ssize_t		count,
 	int		flags,
 	xfs_iomap_t	*iomapp,
 	int		*niomaps)
 {
-	xfs_inode_t	*ip = XFS_BHVTOI(bdp);
 	xfs_iocore_t	*io = &ip->i_iocore;
 
 	ASSERT((ip->i_d.di_mode & S_IFMT) == S_IFREG);

fs/xfs/linux-2.6/xfs_lrw.h

@@ -18,8 +18,6 @@
 #ifndef __XFS_LRW_H__
 #define __XFS_LRW_H__
 
-struct bhv_desc;
-struct bhv_vnode;
 struct xfs_mount;
 struct xfs_iocore;
 struct xfs_inode;
@@ -71,30 +69,11 @@ extern void xfs_inval_cached_trace(struct xfs_iocore *,
 #define xfs_inval_cached_trace(io, offset, len, first, last)
 #endif
 
-/*
- * Maximum count of bmaps used by read and write paths.
- */
-#define	XFS_MAX_RW_NBMAPS	4
-
-extern int xfs_bmap(struct bhv_desc *, xfs_off_t, ssize_t, int,
-			struct xfs_iomap *, int *);
 extern int xfsbdstrat(struct xfs_mount *, struct xfs_buf *);
 extern int xfs_bdstrat_cb(struct xfs_buf *);
 extern int xfs_dev_is_read_only(struct xfs_mount *, char *);
 
-extern int xfs_zero_eof(struct bhv_vnode *, struct xfs_iocore *, xfs_off_t,
+extern int xfs_zero_eof(struct inode *, struct xfs_iocore *, xfs_off_t,
 				xfs_fsize_t);
-extern ssize_t xfs_read(struct bhv_desc *, struct kiocb *,
-				const struct iovec *, unsigned int,
-				loff_t *, int, struct cred *);
-extern ssize_t xfs_write(struct bhv_desc *, struct kiocb *,
-				const struct iovec *, unsigned int,
-				loff_t *, int, struct cred *);
-extern ssize_t xfs_splice_read(struct bhv_desc *, struct file *, loff_t *,
-				struct pipe_inode_info *, size_t, int, int,
-				struct cred *);
-extern ssize_t xfs_splice_write(struct bhv_desc *, struct pipe_inode_info *,
-				struct file *, loff_t *, size_t, int, int,
-				struct cred *);
 
 #endif	/* __XFS_LRW_H__ */

fs/xfs/linux-2.6/xfs_super.c

@@ -46,6 +46,8 @@
 #include "xfs_attr.h"
 #include "xfs_buf_item.h"
 #include "xfs_utils.h"
+#include "xfs_vnodeops.h"
+#include "xfs_vfsops.h"
 #include "xfs_version.h"
 
 #include <linux/namei.h>
@@ -196,23 +198,20 @@ xfs_revalidate_inode(
 		inode->i_flags |= S_NOATIME;
 	else
 		inode->i_flags &= ~S_NOATIME;
-	vp->v_flag &= ~VMODIFIED;
+	xfs_iflags_clear(ip, XFS_IMODIFIED);
 }
 
 void
 xfs_initialize_vnode(
-	bhv_desc_t		*bdp,
+	struct xfs_mount	*mp,
 	bhv_vnode_t		*vp,
-	bhv_desc_t		*inode_bhv,
-	int			unlock)
+	struct xfs_inode	*ip)
 {
-	xfs_inode_t		*ip = XFS_BHVTOI(inode_bhv);
 	struct inode		*inode = vn_to_inode(vp);
 
-	if (!inode_bhv->bd_vobj) {
-		vp->v_vfsp = bhvtovfs(bdp);
-		bhv_desc_init(inode_bhv, ip, vp, &xfs_vnodeops);
-		bhv_insert(VN_BHV_HEAD(vp), inode_bhv);
+	if (!ip->i_vnode) {
+		ip->i_vnode = vp;
+		inode->i_private = ip;
 	}
 
 	/*
@@ -222,8 +221,8 @@ xfs_initialize_vnode(
 	 * second time once the inode is properly set up, and then we can
 	 * finish our work.
 	 */
-	if (ip->i_d.di_mode != 0 && unlock && (inode->i_state & I_NEW)) {
-		xfs_revalidate_inode(XFS_BHVTOM(bdp), vp, ip);
+	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);
@@ -409,19 +408,22 @@ xfs_fs_write_inode(
 	struct inode		*inode,
 	int			sync)
 {
-	bhv_vnode_t		*vp = vn_from_inode(inode);
 	int			error = 0, flags = FLUSH_INODE;
 
-	if (vp) {
-		vn_trace_entry(vp, __FUNCTION__, (inst_t *)__return_address);
-		if (sync) {
-			filemap_fdatawait(inode->i_mapping);
-			flags |= FLUSH_SYNC;
-		}
-		error = bhv_vop_iflush(vp, flags);
-		if (error == EAGAIN)
-			error = sync? bhv_vop_iflush(vp, flags | FLUSH_LOG) : 0;
+	vn_trace_entry(XFS_I(inode), __FUNCTION__,
+			(inst_t *)__return_address);
+	if (sync) {
+		filemap_fdatawait(inode->i_mapping);
+		flags |= FLUSH_SYNC;
 	}
+	error = xfs_inode_flush(XFS_I(inode), flags);
+	/*
+	 * if we failed to write out the inode then mark
+	 * it dirty again so we'll try again later.
+	 */
+	if (error)
+		mark_inode_dirty_sync(inode);
+
 	return -error;
 }
 
@@ -429,35 +431,27 @@ STATIC void
 xfs_fs_clear_inode(
 	struct inode		*inode)
 {
-	bhv_vnode_t		*vp = vn_from_inode(inode);
-
-	vn_trace_entry(vp, __FUNCTION__, (inst_t *)__return_address);
-
-	XFS_STATS_INC(vn_rele);
-	XFS_STATS_INC(vn_remove);
-	XFS_STATS_INC(vn_reclaim);
-	XFS_STATS_DEC(vn_active);
+	xfs_inode_t		*ip = XFS_I(inode);
 
 	/*
-	 * This can happen because xfs_iget_core calls xfs_idestroy if we
+	 * ip can be null when xfs_iget_core calls xfs_idestroy if we
 	 * find an inode with di_mode == 0 but without IGET_CREATE set.
 	 */
-	if (VNHEAD(vp))
-		bhv_vop_inactive(vp, NULL);
-
-	VN_LOCK(vp);
-	vp->v_flag &= ~VMODIFIED;
-	VN_UNLOCK(vp, 0);
-
-	if (VNHEAD(vp))
-		if (bhv_vop_reclaim(vp))
-			panic("%s: cannot reclaim 0x%p\n", __FUNCTION__, vp);
-
-	ASSERT(VNHEAD(vp) == NULL);
+	if (ip) {
+		vn_trace_entry(ip, __FUNCTION__, (inst_t *)__return_address);
+
+		XFS_STATS_INC(vn_rele);
+		XFS_STATS_INC(vn_remove);
+		XFS_STATS_INC(vn_reclaim);
+		XFS_STATS_DEC(vn_active);
+
+		xfs_inactive(ip);
+		xfs_iflags_clear(ip, XFS_IMODIFIED);
+		if (xfs_reclaim(ip))
+			panic("%s: cannot reclaim 0x%p\n", __FUNCTION__, inode);
+	}
 
-#ifdef XFS_VNODE_TRACE
-	ktrace_free(vp->v_trace);
-#endif
+	ASSERT(XFS_I(inode) == NULL);
 }
 
 /*
@@ -469,9 +463,9 @@ xfs_fs_clear_inode(
  */
 STATIC void
 xfs_syncd_queue_work(
-	struct bhv_vfs	*vfs,
+	struct xfs_mount *mp,
 	void		*data,
-	void		(*syncer)(bhv_vfs_t *, void *))
+	void		(*syncer)(struct xfs_mount *, void *))
 {
 	struct bhv_vfs_sync_work *work;
 
@@ -479,11 +473,11 @@ xfs_syncd_queue_work(
 	INIT_LIST_HEAD(&work->w_list);
 	work->w_syncer = syncer;
 	work->w_data = data;
-	work->w_vfs = vfs;
-	spin_lock(&vfs->vfs_sync_lock);
-	list_add_tail(&work->w_list, &vfs->vfs_sync_list);
-	spin_unlock(&vfs->vfs_sync_lock);
-	wake_up_process(vfs->vfs_sync_task);
+	work->w_mount = mp;
+	spin_lock(&mp->m_sync_lock);
+	list_add_tail(&work->w_list, &mp->m_sync_list);
+	spin_unlock(&mp->m_sync_lock);
+	wake_up_process(mp->m_sync_task);
 }
 
 /*
@@ -494,22 +488,22 @@ xfs_syncd_queue_work(
  */
 STATIC void
 xfs_flush_inode_work(
-	bhv_vfs_t	*vfs,
-	void		*inode)
+	struct xfs_mount *mp,
+	void		*arg)
 {
-	filemap_flush(((struct inode *)inode)->i_mapping);
-	iput((struct inode *)inode);
+	struct inode	*inode = arg;
+	filemap_flush(inode->i_mapping);
+	iput(inode);
 }
 
 void
 xfs_flush_inode(
 	xfs_inode_t	*ip)
 {
-	struct inode	*inode = vn_to_inode(XFS_ITOV(ip));
-	struct bhv_vfs	*vfs = XFS_MTOVFS(ip->i_mount);
+	struct inode	*inode = ip->i_vnode;
 
 	igrab(inode);
-	xfs_syncd_queue_work(vfs, inode, xfs_flush_inode_work);
+	xfs_syncd_queue_work(ip->i_mount, inode, xfs_flush_inode_work);
 	delay(msecs_to_jiffies(500));
 }
 
@@ -519,11 +513,12 @@ xfs_flush_inode(
  */
 STATIC void
 xfs_flush_device_work(
-	bhv_vfs_t	*vfs,
-	void		*inode)
+	struct xfs_mount *mp,
+	void		*arg)
 {
-	sync_blockdev(vfs->vfs_super->s_bdev);
-	iput((struct inode *)inode);
+	struct inode	*inode = arg;
+	sync_blockdev(mp->m_super->s_bdev);
+	iput(inode);
 }
 
 void
@@ -531,35 +526,33 @@ xfs_flush_device(
 	xfs_inode_t	*ip)
 {
 	struct inode	*inode = vn_to_inode(XFS_ITOV(ip));
-	struct bhv_vfs	*vfs = XFS_MTOVFS(ip->i_mount);
 
 	igrab(inode);
-	xfs_syncd_queue_work(vfs, inode, xfs_flush_device_work);
+	xfs_syncd_queue_work(ip->i_mount, inode, xfs_flush_device_work);
 	delay(msecs_to_jiffies(500));
 	xfs_log_force(ip->i_mount, (xfs_lsn_t)0, XFS_LOG_FORCE|XFS_LOG_SYNC);
 }
 
 STATIC void
-vfs_sync_worker(
-	bhv_vfs_t	*vfsp,
+xfs_sync_worker(
+	struct xfs_mount *mp,
 	void		*unused)
 {
 	int		error;
 
-	if (!(vfsp->vfs_flag & VFS_RDONLY))
-		error = bhv_vfs_sync(vfsp, SYNC_FSDATA | SYNC_BDFLUSH | \
-					SYNC_ATTR | SYNC_REFCACHE | SYNC_SUPER,
-					NULL);
-	vfsp->vfs_sync_seq++;
-	wake_up(&vfsp->vfs_wait_single_sync_task);
+	if (!(mp->m_flags & XFS_MOUNT_RDONLY))
+		error = xfs_sync(mp, SYNC_FSDATA | SYNC_BDFLUSH | SYNC_ATTR |
+				     SYNC_REFCACHE | SYNC_SUPER);
+	mp->m_sync_seq++;
+	wake_up(&mp->m_wait_single_sync_task);
 }
 
 STATIC int
 xfssyncd(
 	void			*arg)
 {
+	struct xfs_mount	*mp = arg;
 	long			timeleft;
-	bhv_vfs_t		*vfsp = (bhv_vfs_t *) arg;
 	bhv_vfs_sync_work_t	*work, *n;
 	LIST_HEAD		(tmp);
 
@@ -569,31 +562,31 @@ xfssyncd(
 		timeleft = schedule_timeout_interruptible(timeleft);
 		/* swsusp */
 		try_to_freeze();
-		if (kthread_should_stop() && list_empty(&vfsp->vfs_sync_list))
+		if (kthread_should_stop() && list_empty(&mp->m_sync_list))
 			break;
 
-		spin_lock(&vfsp->vfs_sync_lock);
+		spin_lock(&mp->m_sync_lock);
 		/*
 		 * We can get woken by laptop mode, to do a sync -
 		 * that's the (only!) case where the list would be
 		 * empty with time remaining.
 		 */
-		if (!timeleft || list_empty(&vfsp->vfs_sync_list)) {
+		if (!timeleft || list_empty(&mp->m_sync_list)) {
 			if (!timeleft)
 				timeleft = xfs_syncd_centisecs *
 							msecs_to_jiffies(10);
-			INIT_LIST_HEAD(&vfsp->vfs_sync_work.w_list);
-			list_add_tail(&vfsp->vfs_sync_work.w_list,
-					&vfsp->vfs_sync_list);
+			INIT_LIST_HEAD(&mp->m_sync_work.w_list);
+			list_add_tail(&mp->m_sync_work.w_list,
+					&mp->m_sync_list);
 		}
-		list_for_each_entry_safe(work, n, &vfsp->vfs_sync_list, w_list)
+		list_for_each_entry_safe(work, n, &mp->m_sync_list, w_list)
 			list_move(&work->w_list, &tmp);
-		spin_unlock(&vfsp->vfs_sync_lock);
+		spin_unlock(&mp->m_sync_lock);
 
 		list_for_each_entry_safe(work, n, &tmp, w_list) {
-			(*work->w_syncer)(vfsp, work->w_data);
+			(*work->w_syncer)(mp, work->w_data);
 			list_del(&work->w_list);
-			if (work == &vfsp->vfs_sync_work)
+			if (work == &mp->m_sync_work)
 				continue;
 			kmem_free(work, sizeof(struct bhv_vfs_sync_work));
 		}
@@ -602,41 +595,19 @@ xfssyncd(
 	return 0;
 }
 
-STATIC int
-xfs_fs_start_syncd(
-	bhv_vfs_t		*vfsp)
-{
-	vfsp->vfs_sync_work.w_syncer = vfs_sync_worker;
-	vfsp->vfs_sync_work.w_vfs = vfsp;
-	vfsp->vfs_sync_task = kthread_run(xfssyncd, vfsp, "xfssyncd");
-	if (IS_ERR(vfsp->vfs_sync_task))
-		return -PTR_ERR(vfsp->vfs_sync_task);
-	return 0;
-}
-
-STATIC void
-xfs_fs_stop_syncd(
-	bhv_vfs_t		*vfsp)
-{
-	kthread_stop(vfsp->vfs_sync_task);
-}
-
 STATIC void
 xfs_fs_put_super(
 	struct super_block	*sb)
 {
-	bhv_vfs_t		*vfsp = vfs_from_sb(sb);
+	struct xfs_mount	*mp = XFS_M(sb);
 	int			error;
 
-	xfs_fs_stop_syncd(vfsp);
-	bhv_vfs_sync(vfsp, SYNC_ATTR | SYNC_DELWRI, NULL);
-	error = bhv_vfs_unmount(vfsp, 0, NULL);
-	if (error) {
+	kthread_stop(mp->m_sync_task);
+
+	xfs_sync(mp, SYNC_ATTR | SYNC_DELWRI);
+	error = xfs_unmount(mp, 0, NULL);
+	if (error)
 		printk("XFS: unmount got error=%d\n", error);
-		printk("%s: vfs=0x%p left dangling!\n", __FUNCTION__, vfsp);
-	} else {
-		vfs_deallocate(vfsp);
-	}
 }
 
 STATIC void
@@ -644,7 +615,7 @@ xfs_fs_write_super(
 	struct super_block	*sb)
 {
 	if (!(sb->s_flags & MS_RDONLY))
-		bhv_vfs_sync(vfs_from_sb(sb), SYNC_FSDATA, NULL);
+		xfs_sync(XFS_M(sb), SYNC_FSDATA);
 	sb->s_dirt = 0;
 }
 
@@ -653,11 +624,23 @@ xfs_fs_sync_super(
 	struct super_block	*sb,
 	int			wait)
 {
-	bhv_vfs_t		*vfsp = vfs_from_sb(sb);
+	struct xfs_mount	*mp = XFS_M(sb);
 	int			error;
 	int			flags;
 
-	if (unlikely(sb->s_frozen == SB_FREEZE_WRITE)) {
+	/*
+	 * Treat a sync operation like a freeze.  This is to work
+	 * around a race in sync_inodes() which works in two phases
+	 * - an asynchronous flush, which can write out an inode
+	 * without waiting for file size updates to complete, and a
+	 * synchronous flush, which wont do anything because the
+	 * async flush removed the inode's dirty flag.  Also
+	 * sync_inodes() will not see any files that just have
+	 * outstanding transactions to be flushed because we don't
+	 * dirty the Linux inode until after the transaction I/O
+	 * completes.
+	 */
+	if (wait || unlikely(sb->s_frozen == SB_FREEZE_WRITE)) {
 		/*
 		 * First stage of freeze - no more writers will make progress
 		 * now we are here, so we flush delwri and delalloc buffers
@@ -668,28 +651,28 @@ xfs_fs_sync_super(
 		 */
 		flags = SYNC_DATA_QUIESCE;
 	} else
-		flags = SYNC_FSDATA | (wait ? SYNC_WAIT : 0);
+		flags = SYNC_FSDATA;
 
-	error = bhv_vfs_sync(vfsp, flags, NULL);
+	error = xfs_sync(mp, flags);
 	sb->s_dirt = 0;
 
 	if (unlikely(laptop_mode)) {
-		int	prev_sync_seq = vfsp->vfs_sync_seq;
+		int	prev_sync_seq = mp->m_sync_seq;
 
 		/*
 		 * The disk must be active because we're syncing.
 		 * We schedule xfssyncd now (now that the disk is
 		 * active) instead of later (when it might not be).
 		 */
-		wake_up_process(vfsp->vfs_sync_task);
+		wake_up_process(mp->m_sync_task);
 		/*
 		 * We have to wait for the sync iteration to complete.
 		 * If we don't, the disk activity caused by the sync
 		 * will come after the sync is completed, and that
 		 * triggers another sync from laptop mode.
 		 */
-		wait_event(vfsp->vfs_wait_single_sync_task,
-				vfsp->vfs_sync_seq != prev_sync_seq);
+		wait_event(mp->m_wait_single_sync_task,
+				mp->m_sync_seq != prev_sync_seq);
 	}
 
 	return -error;
@@ -700,7 +683,7 @@ xfs_fs_statfs(
 	struct dentry		*dentry,
 	struct kstatfs		*statp)
 {
-	return -bhv_vfs_statvfs(vfs_from_sb(dentry->d_sb), statp,
+	return -xfs_statvfs(XFS_M(dentry->d_sb), statp,
 				vn_from_inode(dentry->d_inode));
 }
 
@@ -710,13 +693,13 @@ xfs_fs_remount(
 	int			*flags,
 	char			*options)
 {
-	bhv_vfs_t		*vfsp = vfs_from_sb(sb);
+	struct xfs_mount	*mp = XFS_M(sb);
 	struct xfs_mount_args	*args = xfs_args_allocate(sb, 0);
 	int			error;
 
-	error = bhv_vfs_parseargs(vfsp, options, args, 1);
+	error = xfs_parseargs(mp, options, args, 1);
 	if (!error)
-		error = bhv_vfs_mntupdate(vfsp, flags, args);
+		error = xfs_mntupdate(mp, flags, args);
 	kmem_free(args, sizeof(*args));
 	return -error;
 }
@@ -725,7 +708,7 @@ STATIC void
 xfs_fs_lockfs(
 	struct super_block	*sb)
 {
-	bhv_vfs_freeze(vfs_from_sb(sb));
+	xfs_freeze(XFS_M(sb));
 }
 
 STATIC int
@@ -733,7 +716,7 @@ xfs_fs_show_options(
 	struct seq_file		*m,
 	struct vfsmount		*mnt)
 {
-	return -bhv_vfs_showargs(vfs_from_sb(mnt->mnt_sb), m);
+	return -xfs_showargs(XFS_M(mnt->mnt_sb), m);
 }
 
 STATIC int
@@ -741,7 +724,7 @@ xfs_fs_quotasync(
 	struct super_block	*sb,
 	int			type)
 {
-	return -bhv_vfs_quotactl(vfs_from_sb(sb), Q_XQUOTASYNC, 0, NULL);
+	return -XFS_QM_QUOTACTL(XFS_M(sb), Q_XQUOTASYNC, 0, NULL);
 }
 
 STATIC int
@@ -749,7 +732,7 @@ xfs_fs_getxstate(
 	struct super_block	*sb,
 	struct fs_quota_stat	*fqs)
 {
-	return -bhv_vfs_quotactl(vfs_from_sb(sb), Q_XGETQSTAT, 0, (caddr_t)fqs);
+	return -XFS_QM_QUOTACTL(XFS_M(sb), Q_XGETQSTAT, 0, (caddr_t)fqs);
 }
 
 STATIC int
@@ -758,7 +741,7 @@ xfs_fs_setxstate(
 	unsigned int		flags,
 	int			op)
 {
-	return -bhv_vfs_quotactl(vfs_from_sb(sb), op, 0, (caddr_t)&flags);
+	return -XFS_QM_QUOTACTL(XFS_M(sb), op, 0, (caddr_t)&flags);
 }
 
 STATIC int
@@ -768,7 +751,7 @@ xfs_fs_getxquota(
 	qid_t			id,
 	struct fs_disk_quota	*fdq)
 {
-	return -bhv_vfs_quotactl(vfs_from_sb(sb),
+	return -XFS_QM_QUOTACTL(XFS_M(sb),
 				 (type == USRQUOTA) ? Q_XGETQUOTA :
 				  ((type == GRPQUOTA) ? Q_XGETGQUOTA :
 				   Q_XGETPQUOTA), id, (caddr_t)fdq);
@@ -781,7 +764,7 @@ xfs_fs_setxquota(
 	qid_t			id,
 	struct fs_disk_quota	*fdq)
 {
-	return -bhv_vfs_quotactl(vfs_from_sb(sb),
+	return -XFS_QM_QUOTACTL(XFS_M(sb),
 				 (type == USRQUOTA) ? Q_XSETQLIM :
 				  ((type == GRPQUOTA) ? Q_XSETGQLIM :
 				   Q_XSETPQLIM), id, (caddr_t)fdq);
@@ -793,32 +776,38 @@ xfs_fs_fill_super(
 	void			*data,
 	int			silent)
 {
-	struct bhv_vnode	*rootvp;
-	struct bhv_vfs		*vfsp = vfs_allocate(sb);
+	struct inode		*rootvp;
+	struct xfs_mount	*mp = NULL;
 	struct xfs_mount_args	*args = xfs_args_allocate(sb, silent);
 	struct kstatfs		statvfs;
 	int			error;
 
-	bhv_insert_all_vfsops(vfsp);
+	mp = xfs_mount_init();
 
-	error = bhv_vfs_parseargs(vfsp, (char *)data, args, 0);
-	if (error) {
-		bhv_remove_all_vfsops(vfsp, 1);
+	INIT_LIST_HEAD(&mp->m_sync_list);
+	spin_lock_init(&mp->m_sync_lock);
+	init_waitqueue_head(&mp->m_wait_single_sync_task);
+
+	mp->m_super = sb;
+	sb->s_fs_info = mp;
+
+	if (sb->s_flags & MS_RDONLY)
+		mp->m_flags |= XFS_MOUNT_RDONLY;
+
+	error = xfs_parseargs(mp, (char *)data, args, 0);
+	if (error)
 		goto fail_vfsop;
-	}
 
 	sb_min_blocksize(sb, BBSIZE);
 	sb->s_export_op = &xfs_export_operations;
 	sb->s_qcop = &xfs_quotactl_operations;
 	sb->s_op = &xfs_super_operations;
 
-	error = bhv_vfs_mount(vfsp, args, NULL);
-	if (error) {
-		bhv_remove_all_vfsops(vfsp, 1);
+	error = xfs_mount(mp, args, NULL);
+	if (error)
 		goto fail_vfsop;
-	}
 
-	error = bhv_vfs_statvfs(vfsp, &statvfs, NULL);
+	error = xfs_statvfs(mp, &statvfs, NULL);
 	if (error)
 		goto fail_unmount;
 
@@ -830,7 +819,7 @@ xfs_fs_fill_super(
 	sb->s_time_gran = 1;
 	set_posix_acl_flag(sb);
 
-	error = bhv_vfs_root(vfsp, &rootvp);
+	error = xfs_root(mp, &rootvp);
 	if (error)
 		goto fail_unmount;
 
@@ -843,9 +832,17 @@ xfs_fs_fill_super(
 		error = EINVAL;
 		goto fail_vnrele;
 	}
-	if ((error = xfs_fs_start_syncd(vfsp)))
+
+	mp->m_sync_work.w_syncer = xfs_sync_worker;
+	mp->m_sync_work.w_mount = mp;
+	mp->m_sync_task = kthread_run(xfssyncd, mp, "xfssyncd");
+	if (IS_ERR(mp->m_sync_task)) {
+		error = -PTR_ERR(mp->m_sync_task);
 		goto fail_vnrele;
-	vn_trace_exit(rootvp, __FUNCTION__, (inst_t *)__return_address);
+	}
+
+	vn_trace_exit(XFS_I(sb->s_root->d_inode), __FUNCTION__,
+			(inst_t *)__return_address);
 
 	kmem_free(args, sizeof(*args));
 	return 0;
@@ -859,10 +856,9 @@ fail_vnrele:
 	}
 
 fail_unmount:
-	bhv_vfs_unmount(vfsp, 0, NULL);
+	xfs_unmount(mp, 0, NULL);
 
 fail_vfsop:
-	vfs_deallocate(vfsp);
 	kmem_free(args, sizeof(*args));
 	return -error;
 }
@@ -914,15 +910,11 @@ STATIC int __init
 init_xfs_fs( void )
 {
 	int			error;
-	struct sysinfo		si;
 	static char		message[] __initdata = KERN_INFO \
 		XFS_VERSION_STRING " with " XFS_BUILD_OPTIONS " enabled\n";
 
 	printk(message);
 
-	si_meminfo(&si);
-	xfs_physmem = si.totalram;
-
 	ktrace_init(64);
 
 	error = xfs_init_zones();

fs/xfs/linux-2.6/xfs_super.h

@@ -107,7 +107,8 @@ struct block_device;
 
 extern __uint64_t xfs_max_file_offset(unsigned int);
 
-extern void xfs_initialize_vnode(bhv_desc_t *, bhv_vnode_t *, bhv_desc_t *, 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 *);
@@ -119,4 +120,6 @@ extern void xfs_blkdev_issue_flush(struct xfs_buftarg *);
 
 extern struct export_operations xfs_export_operations;
 
+#define XFS_M(sb)		((struct xfs_mount *)((sb)->s_fs_info))
+
 #endif	/* __XFS_SUPER_H__ */

fs/xfs/linux-2.6/xfs_vfs.c

@@ -1,327 +0,0 @@
-/*
- * Copyright (c) 2000-2005 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
- */
-#include "xfs.h"
-#include "xfs_fs.h"
-#include "xfs_inum.h"
-#include "xfs_log.h"
-#include "xfs_clnt.h"
-#include "xfs_trans.h"
-#include "xfs_sb.h"
-#include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_imap.h"
-#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
-#include "xfs_mount.h"
-#include "xfs_quota.h"
-
-int
-vfs_mount(
-	struct bhv_desc		*bdp,
-	struct xfs_mount_args	*args,
-	struct cred		*cr)
-{
-	struct bhv_desc		*next = bdp;
-
-	ASSERT(next);
-	while (! (bhvtovfsops(next))->vfs_mount)
-		next = BHV_NEXT(next);
-	return ((*bhvtovfsops(next)->vfs_mount)(next, args, cr));
-}
-
-int
-vfs_parseargs(
-	struct bhv_desc		*bdp,
-	char			*s,
-	struct xfs_mount_args	*args,
-	int			f)
-{
-	struct bhv_desc		*next = bdp;
-
-	ASSERT(next);
-	while (! (bhvtovfsops(next))->vfs_parseargs)
-		next = BHV_NEXT(next);
-	return ((*bhvtovfsops(next)->vfs_parseargs)(next, s, args, f));
-}
-
-int
-vfs_showargs(
-	struct bhv_desc		*bdp,
-	struct seq_file		*m)
-{
-	struct bhv_desc		*next = bdp;
-
-	ASSERT(next);
-	while (! (bhvtovfsops(next))->vfs_showargs)
-		next = BHV_NEXT(next);
-	return ((*bhvtovfsops(next)->vfs_showargs)(next, m));
-}
-
-int
-vfs_unmount(
-	struct bhv_desc		*bdp,
-	int			fl,
-	struct cred		*cr)
-{
-	struct bhv_desc		*next = bdp;
-
-	ASSERT(next);
-	while (! (bhvtovfsops(next))->vfs_unmount)
-		next = BHV_NEXT(next);
-	return ((*bhvtovfsops(next)->vfs_unmount)(next, fl, cr));
-}
-
-int
-vfs_mntupdate(
-	struct bhv_desc		*bdp,
-	int			*fl,
-	struct xfs_mount_args	*args)
-{
-	struct bhv_desc		*next = bdp;
-
-	ASSERT(next);
-	while (! (bhvtovfsops(next))->vfs_mntupdate)
-		next = BHV_NEXT(next);
-	return ((*bhvtovfsops(next)->vfs_mntupdate)(next, fl, args));
-}
-
-int
-vfs_root(
-	struct bhv_desc		*bdp,
-	struct bhv_vnode	**vpp)
-{
-	struct bhv_desc		*next = bdp;
-
-	ASSERT(next);
-	while (! (bhvtovfsops(next))->vfs_root)
-		next = BHV_NEXT(next);
-	return ((*bhvtovfsops(next)->vfs_root)(next, vpp));
-}
-
-int
-vfs_statvfs(
-	struct bhv_desc		*bdp,
-	bhv_statvfs_t		*statp,
-	struct bhv_vnode	*vp)
-{
-	struct bhv_desc		*next = bdp;
-
-	ASSERT(next);
-	while (! (bhvtovfsops(next))->vfs_statvfs)
-		next = BHV_NEXT(next);
-	return ((*bhvtovfsops(next)->vfs_statvfs)(next, statp, vp));
-}
-
-int
-vfs_sync(
-	struct bhv_desc		*bdp,
-	int			fl,
-	struct cred		*cr)
-{
-	struct bhv_desc		*next = bdp;
-
-	ASSERT(next);
-	while (! (bhvtovfsops(next))->vfs_sync)
-		next = BHV_NEXT(next);
-	return ((*bhvtovfsops(next)->vfs_sync)(next, fl, cr));
-}
-
-int
-vfs_vget(
-	struct bhv_desc		*bdp,
-	struct bhv_vnode	**vpp,
-	struct fid		*fidp)
-{
-	struct bhv_desc		*next = bdp;
-
-	ASSERT(next);
-	while (! (bhvtovfsops(next))->vfs_vget)
-		next = BHV_NEXT(next);
-	return ((*bhvtovfsops(next)->vfs_vget)(next, vpp, fidp));
-}
-
-int
-vfs_dmapiops(
-	struct bhv_desc		*bdp,
-	caddr_t			addr)
-{
-	struct bhv_desc		*next = bdp;
-
-	ASSERT(next);
-	while (! (bhvtovfsops(next))->vfs_dmapiops)
-		next = BHV_NEXT(next);
-	return ((*bhvtovfsops(next)->vfs_dmapiops)(next, addr));
-}
-
-int
-vfs_quotactl(
-	struct bhv_desc		*bdp,
-	int			cmd,
-	int			id,
-	caddr_t			addr)
-{
-	struct bhv_desc		*next = bdp;
-
-	ASSERT(next);
-	while (! (bhvtovfsops(next))->vfs_quotactl)
-		next = BHV_NEXT(next);
-	return ((*bhvtovfsops(next)->vfs_quotactl)(next, cmd, id, addr));
-}
-
-void
-vfs_init_vnode(
-	struct bhv_desc		*bdp,
-	struct bhv_vnode	*vp,
-	struct bhv_desc		*bp,
-	int			unlock)
-{
-	struct bhv_desc		*next = bdp;
-
-	ASSERT(next);
-	while (! (bhvtovfsops(next))->vfs_init_vnode)
-		next = BHV_NEXT(next);
-	((*bhvtovfsops(next)->vfs_init_vnode)(next, vp, bp, unlock));
-}
-
-void
-vfs_force_shutdown(
-	struct bhv_desc		*bdp,
-	int			fl,
-	char			*file,
-	int			line)
-{
-	struct bhv_desc		*next = bdp;
-
-	ASSERT(next);
-	while (! (bhvtovfsops(next))->vfs_force_shutdown)
-		next = BHV_NEXT(next);
-	((*bhvtovfsops(next)->vfs_force_shutdown)(next, fl, file, line));
-}
-
-void
-vfs_freeze(
-	struct bhv_desc		*bdp)
-{
-	struct bhv_desc		*next = bdp;
-
-	ASSERT(next);
-	while (! (bhvtovfsops(next))->vfs_freeze)
-		next = BHV_NEXT(next);
-	((*bhvtovfsops(next)->vfs_freeze)(next));
-}
-
-bhv_vfs_t *
-vfs_allocate(
-	struct super_block	*sb)
-{
-	struct bhv_vfs		*vfsp;
-
-	vfsp = kmem_zalloc(sizeof(bhv_vfs_t), KM_SLEEP);
-	bhv_head_init(VFS_BHVHEAD(vfsp), "vfs");
-	INIT_LIST_HEAD(&vfsp->vfs_sync_list);
-	spin_lock_init(&vfsp->vfs_sync_lock);
-	init_waitqueue_head(&vfsp->vfs_wait_single_sync_task);
-
-	vfsp->vfs_super = sb;
-	sb->s_fs_info = vfsp;
-
-	if (sb->s_flags & MS_RDONLY)
-		vfsp->vfs_flag |= VFS_RDONLY;
-
-	return vfsp;
-}
-
-bhv_vfs_t *
-vfs_from_sb(
-	struct super_block	*sb)
-{
-	return (bhv_vfs_t *)sb->s_fs_info;
-}
-
-void
-vfs_deallocate(
-	struct bhv_vfs		*vfsp)
-{
-	bhv_head_destroy(VFS_BHVHEAD(vfsp));
-	kmem_free(vfsp, sizeof(bhv_vfs_t));
-}
-
-void
-vfs_insertops(
-	struct bhv_vfs		*vfsp,
-	struct bhv_module_vfsops *vfsops)
-{
-	struct bhv_desc		*bdp;
-
-	bdp = kmem_alloc(sizeof(struct bhv_desc), KM_SLEEP);
-	bhv_desc_init(bdp, NULL, vfsp, vfsops);
-	bhv_insert(&vfsp->vfs_bh, bdp);
-}
-
-void
-vfs_insertbhv(
-	struct bhv_vfs		*vfsp,
-	struct bhv_desc		*bdp,
-	struct bhv_vfsops	*vfsops,
-	void			*mount)
-{
-	bhv_desc_init(bdp, mount, vfsp, vfsops);
-	bhv_insert_initial(&vfsp->vfs_bh, bdp);
-}
-
-void
-bhv_remove_vfsops(
-	struct bhv_vfs		*vfsp,
-	int			pos)
-{
-	struct bhv_desc		*bhv;
-
-	bhv = bhv_lookup_range(&vfsp->vfs_bh, pos, pos);
-	if (!bhv)
-		return;
-	bhv_remove(&vfsp->vfs_bh, bhv);
-	kmem_free(bhv, sizeof(*bhv));
-}
-
-void
-bhv_remove_all_vfsops(
-	struct bhv_vfs		*vfsp,
-	int			freebase)
-{
-	struct xfs_mount	*mp;
-
-	bhv_remove_vfsops(vfsp, VFS_POSITION_QM);
-	bhv_remove_vfsops(vfsp, VFS_POSITION_DM);
-	if (!freebase)
-		return;
-	mp = XFS_VFSTOM(vfsp);
-	VFS_REMOVEBHV(vfsp, &mp->m_bhv);
-	xfs_mount_free(mp, 0);
-}
-
-void
-bhv_insert_all_vfsops(
-	struct bhv_vfs		*vfsp)
-{
-	struct xfs_mount	*mp;
-
-	mp = xfs_mount_init();
-	vfs_insertbhv(vfsp, &mp->m_bhv, &xfs_vfsops, mp);
-	vfs_insertdmapi(vfsp);
-	vfs_insertquota(vfsp);
-}

fs/xfs/linux-2.6/xfs_vfs.h

@@ -21,68 +21,25 @@
 #include <linux/vfs.h>
 #include "xfs_fs.h"
 
-struct bhv_vfs;
-struct bhv_vnode;
+struct inode;
 
 struct fid;
 struct cred;
 struct seq_file;
 struct super_block;
+struct xfs_inode;
+struct xfs_mount;
 struct xfs_mount_args;
 
 typedef struct kstatfs	bhv_statvfs_t;
 
 typedef struct bhv_vfs_sync_work {
 	struct list_head	w_list;
-	struct bhv_vfs		*w_vfs;
+	struct xfs_mount	*w_mount;
 	void			*w_data;	/* syncer routine argument */
-	void			(*w_syncer)(struct bhv_vfs *, void *);
+	void			(*w_syncer)(struct xfs_mount *, void *);
 } bhv_vfs_sync_work_t;
 
-typedef struct bhv_vfs {
-	u_int			vfs_flag;	/* flags */
-	xfs_fsid_t		vfs_fsid;	/* file system ID */
-	xfs_fsid_t		*vfs_altfsid;	/* An ID fixed for life of FS */
-	bhv_head_t		vfs_bh;		/* head of vfs behavior chain */
-	struct super_block	*vfs_super;	/* generic superblock pointer */
-	struct task_struct	*vfs_sync_task;	/* generalised sync thread */
-	bhv_vfs_sync_work_t	vfs_sync_work;	/* work item for VFS_SYNC */
-	struct list_head	vfs_sync_list;	/* sync thread work item list */
-	spinlock_t		vfs_sync_lock;	/* work item list lock */
-	int			vfs_sync_seq;	/* sync thread generation no. */
-	wait_queue_head_t	vfs_wait_single_sync_task;
-} bhv_vfs_t;
-
-#define bhvtovfs(bdp)		( (struct bhv_vfs *)BHV_VOBJ(bdp) )
-#define bhvtovfsops(bdp)	( (struct bhv_vfsops *)BHV_OPS(bdp) )
-#define VFS_BHVHEAD(vfs)	( &(vfs)->vfs_bh )
-#define VFS_REMOVEBHV(vfs, bdp)	( bhv_remove(VFS_BHVHEAD(vfs), bdp) )
-
-#define VFS_POSITION_BASE	BHV_POSITION_BASE	/* chain bottom */
-#define VFS_POSITION_TOP	BHV_POSITION_TOP	/* chain top */
-#define VFS_POSITION_INVALID	BHV_POSITION_INVALID	/* invalid pos. num */
-
-typedef enum {
-	VFS_BHV_UNKNOWN,	/* not specified */
-	VFS_BHV_XFS,		/* xfs */
-	VFS_BHV_DM,		/* data migration */
-	VFS_BHV_QM,		/* quota manager */
-	VFS_BHV_IO,		/* IO path */
-	VFS_BHV_END		/* housekeeping end-of-range */
-} bhv_vfs_type_t;
-
-#define VFS_POSITION_XFS	(BHV_POSITION_BASE)
-#define VFS_POSITION_DM		(VFS_POSITION_BASE+10)
-#define VFS_POSITION_QM		(VFS_POSITION_BASE+20)
-#define VFS_POSITION_IO		(VFS_POSITION_BASE+30)
-
-#define VFS_RDONLY		0x0001	/* read-only vfs */
-#define VFS_GRPID		0x0002	/* group-ID assigned from directory */
-#define VFS_DMI			0x0004	/* filesystem has the DMI enabled */
-/* ---- VFS_UMOUNT ----		0x0008	-- unneeded, fixed via kthread APIs */
-#define VFS_32BITINODES		0x0010	/* do not use inums above 32 bits */
-#define VFS_END			0x0010	/* max flag */
-
 #define SYNC_ATTR		0x0001	/* sync attributes */
 #define SYNC_CLOSE		0x0002	/* close file system down */
 #define SYNC_DELWRI		0x0004	/* look at delayed writes */
@@ -115,118 +72,7 @@ typedef enum {
 #define SHUTDOWN_REMOTE_REQ	0x0010	/* shutdown came from remote cell */
 #define SHUTDOWN_DEVICE_REQ	0x0020	/* failed all paths to the device */
 
-typedef int	(*vfs_mount_t)(bhv_desc_t *,
-				struct xfs_mount_args *, struct cred *);
-typedef int	(*vfs_parseargs_t)(bhv_desc_t *, char *,
-				struct xfs_mount_args *, int);
-typedef	int	(*vfs_showargs_t)(bhv_desc_t *, struct seq_file *);
-typedef int	(*vfs_unmount_t)(bhv_desc_t *, int, struct cred *);
-typedef int	(*vfs_mntupdate_t)(bhv_desc_t *, int *,
-				struct xfs_mount_args *);
-typedef int	(*vfs_root_t)(bhv_desc_t *, struct bhv_vnode **);
-typedef int	(*vfs_statvfs_t)(bhv_desc_t *, bhv_statvfs_t *,
-				struct bhv_vnode *);
-typedef int	(*vfs_sync_t)(bhv_desc_t *, int, struct cred *);
-typedef int	(*vfs_vget_t)(bhv_desc_t *, struct bhv_vnode **, struct fid *);
-typedef int	(*vfs_dmapiops_t)(bhv_desc_t *, caddr_t);
-typedef int	(*vfs_quotactl_t)(bhv_desc_t *, int, int, caddr_t);
-typedef void	(*vfs_init_vnode_t)(bhv_desc_t *,
-				struct bhv_vnode *, bhv_desc_t *, int);
-typedef void	(*vfs_force_shutdown_t)(bhv_desc_t *, int, char *, int);
-typedef void	(*vfs_freeze_t)(bhv_desc_t *);
-
-typedef struct bhv_vfsops {
-	bhv_position_t		vf_position;	/* behavior chain position */
-	vfs_mount_t		vfs_mount;	/* mount file system */
-	vfs_parseargs_t		vfs_parseargs;	/* parse mount options */
-	vfs_showargs_t		vfs_showargs;	/* unparse mount options */
-	vfs_unmount_t		vfs_unmount;	/* unmount file system */
-	vfs_mntupdate_t		vfs_mntupdate;	/* update file system options */
-	vfs_root_t		vfs_root;	/* get root vnode */
-	vfs_statvfs_t		vfs_statvfs;	/* file system statistics */
-	vfs_sync_t		vfs_sync;	/* flush files */
-	vfs_vget_t		vfs_vget;	/* get vnode from fid */
-	vfs_dmapiops_t		vfs_dmapiops;	/* data migration */
-	vfs_quotactl_t		vfs_quotactl;	/* disk quota */
-	vfs_init_vnode_t	vfs_init_vnode;	/* initialize a new vnode */
-	vfs_force_shutdown_t	vfs_force_shutdown;	/* crash and burn */
-	vfs_freeze_t		vfs_freeze;	/* freeze fs for snapshot */
-} bhv_vfsops_t;
-
-/*
- * Virtual filesystem operations, operating from head bhv.
- */
-#define VFSHEAD(v)			((v)->vfs_bh.bh_first)
-#define bhv_vfs_mount(v, ma,cr)		vfs_mount(VFSHEAD(v), ma,cr)
-#define bhv_vfs_parseargs(v, o,ma,f)	vfs_parseargs(VFSHEAD(v), o,ma,f)
-#define bhv_vfs_showargs(v, m)		vfs_showargs(VFSHEAD(v), m)
-#define bhv_vfs_unmount(v, f,cr)	vfs_unmount(VFSHEAD(v), f,cr)
-#define bhv_vfs_mntupdate(v, fl,args)	vfs_mntupdate(VFSHEAD(v), fl,args)
-#define bhv_vfs_root(v, vpp)		vfs_root(VFSHEAD(v), vpp)
-#define bhv_vfs_statvfs(v, sp,vp)	vfs_statvfs(VFSHEAD(v), sp,vp)
-#define bhv_vfs_sync(v, flag,cr)	vfs_sync(VFSHEAD(v), flag,cr)
-#define bhv_vfs_vget(v, vpp,fidp)	vfs_vget(VFSHEAD(v), vpp,fidp)
-#define bhv_vfs_dmapiops(v, p)		vfs_dmapiops(VFSHEAD(v), p)
-#define bhv_vfs_quotactl(v, c,id,p)	vfs_quotactl(VFSHEAD(v), c,id,p)
-#define bhv_vfs_init_vnode(v, vp,b,ul)	vfs_init_vnode(VFSHEAD(v), vp,b,ul)
-#define bhv_vfs_force_shutdown(v,u,f,l)	vfs_force_shutdown(VFSHEAD(v), u,f,l)
-#define bhv_vfs_freeze(v)		vfs_freeze(VFSHEAD(v))
-
-/*
- * Virtual filesystem operations, operating from next bhv.
- */
-#define bhv_next_vfs_mount(b, ma,cr)		vfs_mount(b, ma,cr)
-#define bhv_next_vfs_parseargs(b, o,ma,f)	vfs_parseargs(b, o,ma,f)
-#define bhv_next_vfs_showargs(b, m)		vfs_showargs(b, m)
-#define bhv_next_vfs_unmount(b, f,cr)		vfs_unmount(b, f,cr)
-#define bhv_next_vfs_mntupdate(b, fl,args)	vfs_mntupdate(b, fl, args)
-#define bhv_next_vfs_root(b, vpp)		vfs_root(b, vpp)
-#define bhv_next_vfs_statvfs(b, sp,vp)		vfs_statvfs(b, sp,vp)
-#define bhv_next_vfs_sync(b, flag,cr)		vfs_sync(b, flag,cr)
-#define bhv_next_vfs_vget(b, vpp,fidp)		vfs_vget(b, vpp,fidp)
-#define bhv_next_vfs_dmapiops(b, p)		vfs_dmapiops(b, p)
-#define bhv_next_vfs_quotactl(b, c,id,p)	vfs_quotactl(b, c,id,p)
-#define bhv_next_vfs_init_vnode(b, vp,b2,ul)	vfs_init_vnode(b, vp,b2,ul)
-#define bhv_next_force_shutdown(b, fl,f,l)	vfs_force_shutdown(b, fl,f,l)
-#define bhv_next_vfs_freeze(b)			vfs_freeze(b)
-
-extern int vfs_mount(bhv_desc_t *, struct xfs_mount_args *, struct cred *);
-extern int vfs_parseargs(bhv_desc_t *, char *, struct xfs_mount_args *, int);
-extern int vfs_showargs(bhv_desc_t *, struct seq_file *);
-extern int vfs_unmount(bhv_desc_t *, int, struct cred *);
-extern int vfs_mntupdate(bhv_desc_t *, int *, struct xfs_mount_args *);
-extern int vfs_root(bhv_desc_t *, struct bhv_vnode **);
-extern int vfs_statvfs(bhv_desc_t *, bhv_statvfs_t *, struct bhv_vnode *);
-extern int vfs_sync(bhv_desc_t *, int, struct cred *);
-extern int vfs_vget(bhv_desc_t *, struct bhv_vnode **, struct fid *);
-extern int vfs_dmapiops(bhv_desc_t *, caddr_t);
-extern int vfs_quotactl(bhv_desc_t *, int, int, caddr_t);
-extern void vfs_init_vnode(bhv_desc_t *, struct bhv_vnode *, bhv_desc_t *, int);
-extern void vfs_force_shutdown(bhv_desc_t *, int, char *, int);
-extern void vfs_freeze(bhv_desc_t *);
-
-#define vfs_test_for_freeze(vfs)	((vfs)->vfs_super->s_frozen)
-#define vfs_wait_for_freeze(vfs,l)	vfs_check_frozen((vfs)->vfs_super, (l))
- 
-typedef struct bhv_module_vfsops {
-	struct bhv_vfsops	bhv_common;
-	void *			bhv_custom;
-} bhv_module_vfsops_t;
-
-#define vfs_bhv_lookup(v, id)	(bhv_lookup_range(&(v)->vfs_bh, (id), (id)))
-#define vfs_bhv_custom(b)	(((bhv_module_vfsops_t*)BHV_OPS(b))->bhv_custom)
-#define vfs_bhv_set_custom(b,o)	((b)->bhv_custom = (void *)(o))
-#define vfs_bhv_clr_custom(b)	((b)->bhv_custom = NULL)
-
-extern bhv_vfs_t *vfs_allocate(struct super_block *);
-extern bhv_vfs_t *vfs_from_sb(struct super_block *);
-extern void vfs_deallocate(bhv_vfs_t *);
-extern void vfs_insertbhv(bhv_vfs_t *, bhv_desc_t *, bhv_vfsops_t *, void *);
-
-extern void vfs_insertops(bhv_vfs_t *, bhv_module_vfsops_t *);
-
-extern void bhv_insert_all_vfsops(struct bhv_vfs *);
-extern void bhv_remove_all_vfsops(struct bhv_vfs *, int);
-extern void bhv_remove_vfsops(struct bhv_vfs *, int);
+#define xfs_test_for_freeze(mp)		((mp)->m_super->s_frozen)
+#define xfs_wait_for_freeze(mp,l)	vfs_check_frozen((mp)->m_super, (l))
 
 #endif	/* __XFS_VFS_H__ */

fs/xfs/linux-2.6/xfs_vnode.c

@@ -16,9 +16,21 @@
  * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
  */
 #include "xfs.h"
+#include "xfs_vnodeops.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_inode.h"
+
+/*
+ * And this gunk is needed for xfs_mount.h"
+ */
+#include "xfs_log.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_dmapi.h"
+#include "xfs_inum.h"
+#include "xfs_ag.h"
+#include "xfs_mount.h"
 
-uint64_t vn_generation;		/* vnode generation number */
-DEFINE_SPINLOCK(vnumber_lock);
 
 /*
  * Dedicated vnode inactive/reclaim sync semaphores.
@@ -39,19 +51,19 @@ vn_init(void)
 
 void
 vn_iowait(
-	bhv_vnode_t	*vp)
+	xfs_inode_t	*ip)
 {
-	wait_queue_head_t *wq = vptosync(vp);
+	wait_queue_head_t *wq = vptosync(ip);
 
-	wait_event(*wq, (atomic_read(&vp->v_iocount) == 0));
+	wait_event(*wq, (atomic_read(&ip->i_iocount) == 0));
 }
 
 void
 vn_iowake(
-	bhv_vnode_t	*vp)
+	xfs_inode_t	*ip)
 {
-	if (atomic_dec_and_test(&vp->v_iocount))
-		wake_up(vptosync(vp));
+	if (atomic_dec_and_test(&ip->i_iocount))
+		wake_up(vptosync(ip));
 }
 
 /*
@@ -61,13 +73,13 @@ vn_iowake(
  */
 void
 vn_ioerror(
-	bhv_vnode_t	*vp,
+	xfs_inode_t	*ip,
 	int		error,
 	char		*f,
 	int		l)
 {
 	if (unlikely(error == -ENODEV))
-		bhv_vfs_force_shutdown(vp->v_vfsp, SHUTDOWN_DEVICE_REQ, f, l);
+		xfs_do_force_shutdown(ip->i_mount, SHUTDOWN_DEVICE_REQ, f, l);
 }
 
 bhv_vnode_t *
@@ -79,27 +91,8 @@ vn_initialize(
 	XFS_STATS_INC(vn_active);
 	XFS_STATS_INC(vn_alloc);
 
-	vp->v_flag = VMODIFIED;
-	spinlock_init(&vp->v_lock, "v_lock");
-
-	spin_lock(&vnumber_lock);
-	if (!++vn_generation)	/* v_number shouldn't be zero */
-		vn_generation++;
-	vp->v_number = vn_generation;
-	spin_unlock(&vnumber_lock);
-
 	ASSERT(VN_CACHED(vp) == 0);
 
-	/* Initialize the first behavior and the behavior chain head. */
-	vn_bhv_head_init(VN_BHV_HEAD(vp), "vnode");
-
-	atomic_set(&vp->v_iocount, 0);
-
-#ifdef	XFS_VNODE_TRACE
-	vp->v_trace = ktrace_alloc(VNODE_TRACE_SIZE, KM_SLEEP);
-#endif	/* XFS_VNODE_TRACE */
-
-	vn_trace_exit(vp, __FUNCTION__, (inst_t *)__return_address);
 	return vp;
 }
 
@@ -150,12 +143,12 @@ __vn_revalidate(
 {
 	int		error;
 
-	vn_trace_entry(vp, __FUNCTION__, (inst_t *)__return_address);
+	vn_trace_entry(xfs_vtoi(vp), __FUNCTION__, (inst_t *)__return_address);
 	vattr->va_mask = XFS_AT_STAT | XFS_AT_XFLAGS;
-	error = bhv_vop_getattr(vp, vattr, 0, NULL);
+	error = xfs_getattr(xfs_vtoi(vp), vattr, 0);
 	if (likely(!error)) {
 		vn_revalidate_core(vp, vattr);
-		VUNMODIFY(vp);
+		xfs_iflags_clear(xfs_vtoi(vp), XFS_IMODIFIED);
 	}
 	return -error;
 }
@@ -180,24 +173,35 @@ vn_hold(
 
 	XFS_STATS_INC(vn_hold);
 
-	VN_LOCK(vp);
 	inode = igrab(vn_to_inode(vp));
 	ASSERT(inode);
-	VN_UNLOCK(vp, 0);
 
 	return vp;
 }
 
 #ifdef	XFS_VNODE_TRACE
 
-#define KTRACE_ENTER(vp, vk, s, line, ra)			\
-	ktrace_enter(	(vp)->v_trace,				\
+/*
+ * Reference count of Linux inode if present, -1 if the xfs_inode
+ * has no associated Linux inode.
+ */
+static inline int xfs_icount(struct xfs_inode *ip)
+{
+	bhv_vnode_t *vp = XFS_ITOV_NULL(ip);
+
+	if (vp)
+		return vn_count(vp);
+	return -1;
+}
+
+#define KTRACE_ENTER(ip, vk, s, line, ra)			\
+	ktrace_enter(	(ip)->i_trace,				\
 /*  0 */		(void *)(__psint_t)(vk),		\
 /*  1 */		(void *)(s),				\
 /*  2 */		(void *)(__psint_t) line,		\
-/*  3 */		(void *)(__psint_t)(vn_count(vp)),	\
+/*  3 */		(void *)(__psint_t)xfs_icount(ip),	\
 /*  4 */		(void *)(ra),				\
-/*  5 */		(void *)(__psunsigned_t)(vp)->v_flag,	\
+/*  5 */		NULL,					\
 /*  6 */		(void *)(__psint_t)current_cpu(),	\
 /*  7 */		(void *)(__psint_t)current_pid(),	\
 /*  8 */		(void *)__return_address,		\
@@ -207,32 +211,32 @@ vn_hold(
  * Vnode tracing code.
  */
 void
-vn_trace_entry(bhv_vnode_t *vp, const char *func, inst_t *ra)
+vn_trace_entry(xfs_inode_t *ip, const char *func, inst_t *ra)
 {
-	KTRACE_ENTER(vp, VNODE_KTRACE_ENTRY, func, 0, ra);
+	KTRACE_ENTER(ip, VNODE_KTRACE_ENTRY, func, 0, ra);
 }
 
 void
-vn_trace_exit(bhv_vnode_t *vp, const char *func, inst_t *ra)
+vn_trace_exit(xfs_inode_t *ip, const char *func, inst_t *ra)
 {
-	KTRACE_ENTER(vp, VNODE_KTRACE_EXIT, func, 0, ra);
+	KTRACE_ENTER(ip, VNODE_KTRACE_EXIT, func, 0, ra);
 }
 
 void
-vn_trace_hold(bhv_vnode_t *vp, char *file, int line, inst_t *ra)
+vn_trace_hold(xfs_inode_t *ip, char *file, int line, inst_t *ra)
 {
-	KTRACE_ENTER(vp, VNODE_KTRACE_HOLD, file, line, ra);
+	KTRACE_ENTER(ip, VNODE_KTRACE_HOLD, file, line, ra);
 }
 
 void
-vn_trace_ref(bhv_vnode_t *vp, char *file, int line, inst_t *ra)
+vn_trace_ref(xfs_inode_t *ip, char *file, int line, inst_t *ra)
 {
-	KTRACE_ENTER(vp, VNODE_KTRACE_REF, file, line, ra);
+	KTRACE_ENTER(ip, VNODE_KTRACE_REF, file, line, ra);
 }
 
 void
-vn_trace_rele(bhv_vnode_t *vp, char *file, int line, inst_t *ra)
+vn_trace_rele(xfs_inode_t *ip, char *file, int line, inst_t *ra)
 {
-	KTRACE_ENTER(vp, VNODE_KTRACE_RELE, file, line, ra);
+	KTRACE_ENTER(ip, VNODE_KTRACE_RELE, file, line, ra);
 }
 #endif	/* XFS_VNODE_TRACE */

fs/xfs/linux-2.6/xfs_vnode.h

@@ -18,84 +18,31 @@
 #ifndef __XFS_VNODE_H__
 #define __XFS_VNODE_H__
 
-struct uio;
 struct file;
-struct bhv_vfs;
 struct bhv_vattr;
 struct xfs_iomap;
 struct attrlist_cursor_kern;
 
 typedef struct dentry	bhv_vname_t;
 typedef __u64		bhv_vnumber_t;
+typedef struct inode	bhv_vnode_t;
 
-typedef enum bhv_vflags {
-	VMODIFIED	= 0x08,	/* XFS inode state possibly differs */
-				/* to the Linux inode state. */
-	VTRUNCATED	= 0x40,	/* truncated down so flush-on-close */
-} bhv_vflags_t;
-
-/*
- * MP locking protocols:
- *	v_flag, v_vfsp				VN_LOCK/VN_UNLOCK
- */
-typedef struct bhv_vnode {
-	bhv_vflags_t	v_flag;			/* vnode flags (see above) */
-	bhv_vfs_t	*v_vfsp;		/* ptr to containing VFS */
-	bhv_vnumber_t	v_number;		/* in-core vnode number */
-	bhv_head_t	v_bh;			/* behavior head */
-	spinlock_t	v_lock;			/* VN_LOCK/VN_UNLOCK */
-	atomic_t	v_iocount;		/* outstanding I/O count */
-#ifdef XFS_VNODE_TRACE
-	struct ktrace	*v_trace;		/* trace header structure    */
-#endif
-	struct inode	v_inode;		/* Linux inode */
-	/* inode MUST be last */
-} bhv_vnode_t;
-
-#define VN_ISLNK(vp)	S_ISLNK((vp)->v_inode.i_mode)
-#define VN_ISREG(vp)	S_ISREG((vp)->v_inode.i_mode)
-#define VN_ISDIR(vp)	S_ISDIR((vp)->v_inode.i_mode)
-#define VN_ISCHR(vp)	S_ISCHR((vp)->v_inode.i_mode)
-#define VN_ISBLK(vp)	S_ISBLK((vp)->v_inode.i_mode)
-
-#define VNODE_POSITION_BASE	BHV_POSITION_BASE	/* chain bottom */
-#define VNODE_POSITION_TOP	BHV_POSITION_TOP	/* chain top */
-#define VNODE_POSITION_INVALID	BHV_POSITION_INVALID	/* invalid pos. num */
-
-typedef enum {
-	VN_BHV_UNKNOWN,		/* not specified */
-	VN_BHV_XFS,		/* xfs */
-	VN_BHV_DM,		/* data migration */
-	VN_BHV_QM,		/* quota manager */
-	VN_BHV_IO,		/* IO path */
-	VN_BHV_END		/* housekeeping end-of-range */
-} vn_bhv_t;
-
-#define VNODE_POSITION_XFS	(VNODE_POSITION_BASE)
-#define VNODE_POSITION_DM	(VNODE_POSITION_BASE+10)
-#define VNODE_POSITION_QM	(VNODE_POSITION_BASE+20)
-#define VNODE_POSITION_IO	(VNODE_POSITION_BASE+30)
-
-/*
- * Macros for dealing with the behavior descriptor inside of the vnode.
- */
-#define BHV_TO_VNODE(bdp)	((bhv_vnode_t *)BHV_VOBJ(bdp))
-#define BHV_TO_VNODE_NULL(bdp)	((bhv_vnode_t *)BHV_VOBJNULL(bdp))
-
-#define VN_BHV_HEAD(vp)			((bhv_head_t *)(&((vp)->v_bh)))
-#define vn_bhv_head_init(bhp,name)	bhv_head_init(bhp,name)
-#define vn_bhv_remove(bhp,bdp)		bhv_remove(bhp,bdp)
+#define VN_ISLNK(vp)	S_ISLNK((vp)->i_mode)
+#define VN_ISREG(vp)	S_ISREG((vp)->i_mode)
+#define VN_ISDIR(vp)	S_ISDIR((vp)->i_mode)
+#define VN_ISCHR(vp)	S_ISCHR((vp)->i_mode)
+#define VN_ISBLK(vp)	S_ISBLK((vp)->i_mode)
 
 /*
  * Vnode to Linux inode mapping.
  */
-static inline struct bhv_vnode *vn_from_inode(struct inode *inode)
+static inline bhv_vnode_t *vn_from_inode(struct inode *inode)
 {
-	return container_of(inode, bhv_vnode_t, v_inode);
+	return inode;
 }
-static inline struct inode *vn_to_inode(struct bhv_vnode *vnode)
+static inline struct inode *vn_to_inode(bhv_vnode_t *vnode)
 {
-	return &vnode->v_inode;
+	return vnode;
 }
 
 /*
@@ -111,200 +58,13 @@ typedef enum bhv_vrwlock {
 } bhv_vrwlock_t;
 
 /*
- * Return values for bhv_vop_inactive.  A return value of
+ * Return values for xfs_inactive.  A return value of
  * VN_INACTIVE_NOCACHE implies that the file system behavior
  * has disassociated its state and bhv_desc_t from the vnode.
  */
 #define	VN_INACTIVE_CACHE	0
 #define	VN_INACTIVE_NOCACHE	1
 
-/*
- * Values for the cmd code given to vop_vnode_change.
- */
-typedef enum bhv_vchange {
-	VCHANGE_FLAGS_FRLOCKS		= 0,
-	VCHANGE_FLAGS_ENF_LOCKING	= 1,
-	VCHANGE_FLAGS_TRUNCATED		= 2,
-	VCHANGE_FLAGS_PAGE_DIRTY	= 3,
-	VCHANGE_FLAGS_IOEXCL_COUNT	= 4
-} bhv_vchange_t;
-
-typedef int	(*vop_open_t)(bhv_desc_t *, struct cred *);
-typedef ssize_t (*vop_read_t)(bhv_desc_t *, struct kiocb *,
-				const struct iovec *, unsigned int,
-				loff_t *, int, struct cred *);
-typedef ssize_t (*vop_write_t)(bhv_desc_t *, struct kiocb *,
-				const struct iovec *, unsigned int,
-				loff_t *, int, struct cred *);
-typedef ssize_t (*vop_splice_read_t)(bhv_desc_t *, struct file *, loff_t *,
-				struct pipe_inode_info *, size_t, int, int,
-				struct cred *);
-typedef ssize_t (*vop_splice_write_t)(bhv_desc_t *, struct pipe_inode_info *,
-				struct file *, loff_t *, size_t, int, int,
-				struct cred *);
-typedef int	(*vop_ioctl_t)(bhv_desc_t *, struct inode *, struct file *,
-				int, unsigned int, void __user *);
-typedef int	(*vop_getattr_t)(bhv_desc_t *, struct bhv_vattr *, int,
-				struct cred *);
-typedef int	(*vop_setattr_t)(bhv_desc_t *, struct bhv_vattr *, int,
-				struct cred *);
-typedef int	(*vop_access_t)(bhv_desc_t *, int, struct cred *);
-typedef int	(*vop_lookup_t)(bhv_desc_t *, bhv_vname_t *, bhv_vnode_t **,
-				int, bhv_vnode_t *, struct cred *);
-typedef int	(*vop_create_t)(bhv_desc_t *, bhv_vname_t *, struct bhv_vattr *,
-				bhv_vnode_t **, struct cred *);
-typedef int	(*vop_remove_t)(bhv_desc_t *, bhv_vname_t *, struct cred *);
-typedef int	(*vop_link_t)(bhv_desc_t *, bhv_vnode_t *, bhv_vname_t *,
-				struct cred *);
-typedef int	(*vop_rename_t)(bhv_desc_t *, bhv_vname_t *, bhv_vnode_t *,
-				bhv_vname_t *, struct cred *);
-typedef int	(*vop_mkdir_t)(bhv_desc_t *, bhv_vname_t *, struct bhv_vattr *,
-				bhv_vnode_t **, struct cred *);
-typedef int	(*vop_rmdir_t)(bhv_desc_t *, bhv_vname_t *, struct cred *);
-typedef int	(*vop_readdir_t)(bhv_desc_t *, struct uio *, struct cred *,
-				int *);
-typedef int	(*vop_symlink_t)(bhv_desc_t *, bhv_vname_t *, struct bhv_vattr*,
-				char *, bhv_vnode_t **, struct cred *);
-typedef int	(*vop_readlink_t)(bhv_desc_t *, struct uio *, int,
-				struct cred *);
-typedef int	(*vop_fsync_t)(bhv_desc_t *, int, struct cred *,
-				xfs_off_t, xfs_off_t);
-typedef int	(*vop_inactive_t)(bhv_desc_t *, struct cred *);
-typedef int	(*vop_fid2_t)(bhv_desc_t *, struct fid *);
-typedef int	(*vop_release_t)(bhv_desc_t *);
-typedef int	(*vop_rwlock_t)(bhv_desc_t *, bhv_vrwlock_t);
-typedef void	(*vop_rwunlock_t)(bhv_desc_t *, bhv_vrwlock_t);
-typedef int	(*vop_bmap_t)(bhv_desc_t *, xfs_off_t, ssize_t, int,
-				struct xfs_iomap *, int *);
-typedef int	(*vop_reclaim_t)(bhv_desc_t *);
-typedef int	(*vop_attr_get_t)(bhv_desc_t *, const char *, char *, int *,
-				int, struct cred *);
-typedef	int	(*vop_attr_set_t)(bhv_desc_t *, const char *, char *, int,
-				int, struct cred *);
-typedef	int	(*vop_attr_remove_t)(bhv_desc_t *, const char *,
-				int, struct cred *);
-typedef	int	(*vop_attr_list_t)(bhv_desc_t *, char *, int, int,
-				struct attrlist_cursor_kern *, struct cred *);
-typedef void	(*vop_link_removed_t)(bhv_desc_t *, bhv_vnode_t *, int);
-typedef void	(*vop_vnode_change_t)(bhv_desc_t *, bhv_vchange_t, __psint_t);
-typedef void	(*vop_ptossvp_t)(bhv_desc_t *, xfs_off_t, xfs_off_t, int);
-typedef int	(*vop_pflushinvalvp_t)(bhv_desc_t *, xfs_off_t, xfs_off_t, int);
-typedef int	(*vop_pflushvp_t)(bhv_desc_t *, xfs_off_t, xfs_off_t,
-				uint64_t, int);
-typedef int	(*vop_iflush_t)(bhv_desc_t *, int);
-
-
-typedef struct bhv_vnodeops {
-	bhv_position_t  vn_position;    /* position within behavior chain */
-	vop_open_t		vop_open;
-	vop_read_t		vop_read;
-	vop_write_t		vop_write;
-	vop_splice_read_t	vop_splice_read;
-	vop_splice_write_t	vop_splice_write;
-	vop_ioctl_t		vop_ioctl;
-	vop_getattr_t		vop_getattr;
-	vop_setattr_t		vop_setattr;
-	vop_access_t		vop_access;
-	vop_lookup_t		vop_lookup;
-	vop_create_t		vop_create;
-	vop_remove_t		vop_remove;
-	vop_link_t		vop_link;
-	vop_rename_t		vop_rename;
-	vop_mkdir_t		vop_mkdir;
-	vop_rmdir_t		vop_rmdir;
-	vop_readdir_t		vop_readdir;
-	vop_symlink_t		vop_symlink;
-	vop_readlink_t		vop_readlink;
-	vop_fsync_t		vop_fsync;
-	vop_inactive_t		vop_inactive;
-	vop_fid2_t		vop_fid2;
-	vop_rwlock_t		vop_rwlock;
-	vop_rwunlock_t		vop_rwunlock;
-	vop_bmap_t		vop_bmap;
-	vop_reclaim_t		vop_reclaim;
-	vop_attr_get_t		vop_attr_get;
-	vop_attr_set_t		vop_attr_set;
-	vop_attr_remove_t	vop_attr_remove;
-	vop_attr_list_t		vop_attr_list;
-	vop_link_removed_t	vop_link_removed;
-	vop_vnode_change_t	vop_vnode_change;
-	vop_ptossvp_t		vop_tosspages;
-	vop_pflushinvalvp_t	vop_flushinval_pages;
-	vop_pflushvp_t		vop_flush_pages;
-	vop_release_t		vop_release;
-	vop_iflush_t		vop_iflush;
-} bhv_vnodeops_t;
-
-/*
- * Virtual node operations, operating from head bhv.
- */
-#define VNHEAD(vp)	((vp)->v_bh.bh_first)
-#define VOP(op, vp)	(*((bhv_vnodeops_t *)VNHEAD(vp)->bd_ops)->op)
-#define bhv_vop_open(vp, cr)		VOP(vop_open, vp)(VNHEAD(vp),cr)
-#define bhv_vop_read(vp,file,iov,segs,offset,ioflags,cr)		\
-		VOP(vop_read, vp)(VNHEAD(vp),file,iov,segs,offset,ioflags,cr)
-#define bhv_vop_write(vp,file,iov,segs,offset,ioflags,cr)		\
-		VOP(vop_write, vp)(VNHEAD(vp),file,iov,segs,offset,ioflags,cr)
-#define bhv_vop_splice_read(vp,f,o,pipe,cnt,fl,iofl,cr)			\
-		VOP(vop_splice_read, vp)(VNHEAD(vp),f,o,pipe,cnt,fl,iofl,cr)
-#define bhv_vop_splice_write(vp,f,o,pipe,cnt,fl,iofl,cr)		\
-		VOP(vop_splice_write, vp)(VNHEAD(vp),f,o,pipe,cnt,fl,iofl,cr)
-#define bhv_vop_bmap(vp,of,sz,rw,b,n)					\
-		VOP(vop_bmap, vp)(VNHEAD(vp),of,sz,rw,b,n)
-#define bhv_vop_getattr(vp, vap,f,cr)					\
-		VOP(vop_getattr, vp)(VNHEAD(vp), vap,f,cr)
-#define	bhv_vop_setattr(vp, vap,f,cr)					\
-		VOP(vop_setattr, vp)(VNHEAD(vp), vap,f,cr)
-#define	bhv_vop_access(vp, mode,cr)	VOP(vop_access, vp)(VNHEAD(vp), mode,cr)
-#define	bhv_vop_lookup(vp,d,vpp,f,rdir,cr)				\
-		VOP(vop_lookup, vp)(VNHEAD(vp),d,vpp,f,rdir,cr)
-#define bhv_vop_create(dvp,d,vap,vpp,cr)				\
-		VOP(vop_create, dvp)(VNHEAD(dvp),d,vap,vpp,cr)
-#define bhv_vop_remove(dvp,d,cr)	VOP(vop_remove, dvp)(VNHEAD(dvp),d,cr)
-#define	bhv_vop_link(dvp,fvp,d,cr)	VOP(vop_link, dvp)(VNHEAD(dvp),fvp,d,cr)
-#define	bhv_vop_rename(fvp,fnm,tdvp,tnm,cr)				\
-		VOP(vop_rename, fvp)(VNHEAD(fvp),fnm,tdvp,tnm,cr)
-#define	bhv_vop_mkdir(dp,d,vap,vpp,cr)					\
-		VOP(vop_mkdir, dp)(VNHEAD(dp),d,vap,vpp,cr)
-#define	bhv_vop_rmdir(dp,d,cr)	 	VOP(vop_rmdir, dp)(VNHEAD(dp),d,cr)
-#define	bhv_vop_readdir(vp,uiop,cr,eofp)				\
-		VOP(vop_readdir, vp)(VNHEAD(vp),uiop,cr,eofp)
-#define	bhv_vop_symlink(dvp,d,vap,tnm,vpp,cr)				\
-		VOP(vop_symlink, dvp)(VNHEAD(dvp),d,vap,tnm,vpp,cr)
-#define	bhv_vop_readlink(vp,uiop,fl,cr)					\
-		VOP(vop_readlink, vp)(VNHEAD(vp),uiop,fl,cr)
-#define	bhv_vop_fsync(vp,f,cr,b,e)	VOP(vop_fsync, vp)(VNHEAD(vp),f,cr,b,e)
-#define bhv_vop_inactive(vp,cr)		VOP(vop_inactive, vp)(VNHEAD(vp),cr)
-#define bhv_vop_release(vp)		VOP(vop_release, vp)(VNHEAD(vp))
-#define bhv_vop_fid2(vp,fidp)		VOP(vop_fid2, vp)(VNHEAD(vp),fidp)
-#define bhv_vop_rwlock(vp,i)		VOP(vop_rwlock, vp)(VNHEAD(vp),i)
-#define bhv_vop_rwlock_try(vp,i)	VOP(vop_rwlock, vp)(VNHEAD(vp),i)
-#define bhv_vop_rwunlock(vp,i)		VOP(vop_rwunlock, vp)(VNHEAD(vp),i)
-#define bhv_vop_frlock(vp,c,fl,flags,offset,fr)				\
-		VOP(vop_frlock, vp)(VNHEAD(vp),c,fl,flags,offset,fr)
-#define bhv_vop_reclaim(vp)		VOP(vop_reclaim, vp)(VNHEAD(vp))
-#define bhv_vop_attr_get(vp, name, val, vallenp, fl, cred)		\
-		VOP(vop_attr_get, vp)(VNHEAD(vp),name,val,vallenp,fl,cred)
-#define	bhv_vop_attr_set(vp, name, val, vallen, fl, cred)		\
-		VOP(vop_attr_set, vp)(VNHEAD(vp),name,val,vallen,fl,cred)
-#define	bhv_vop_attr_remove(vp, name, flags, cred)			\
-		VOP(vop_attr_remove, vp)(VNHEAD(vp),name,flags,cred)
-#define	bhv_vop_attr_list(vp, buf, buflen, fl, cursor, cred)		\
-		VOP(vop_attr_list, vp)(VNHEAD(vp),buf,buflen,fl,cursor,cred)
-#define bhv_vop_link_removed(vp, dvp, linkzero)				\
-		VOP(vop_link_removed, vp)(VNHEAD(vp), dvp, linkzero)
-#define bhv_vop_vnode_change(vp, cmd, val)				\
-		VOP(vop_vnode_change, vp)(VNHEAD(vp), cmd, val)
-#define bhv_vop_toss_pages(vp, first, last, fiopt)			\
-		VOP(vop_tosspages, vp)(VNHEAD(vp), first, last, fiopt)
-#define bhv_vop_flushinval_pages(vp, first, last, fiopt)		\
-		VOP(vop_flushinval_pages, vp)(VNHEAD(vp),first,last,fiopt)
-#define bhv_vop_flush_pages(vp, first, last, flags, fiopt)		\
-		VOP(vop_flush_pages, vp)(VNHEAD(vp),first,last,flags,fiopt)
-#define bhv_vop_ioctl(vp, inode, filp, fl, cmd, arg)			\
-		VOP(vop_ioctl, vp)(VNHEAD(vp),inode,filp,fl,cmd,arg)
-#define bhv_vop_iflush(vp, flags)	VOP(vop_iflush, vp)(VNHEAD(vp), flags)
-
 /*
  * Flags for read/write calls - same values as IRIX
  */
@@ -428,16 +188,19 @@ typedef struct bhv_vattr {
 
 extern void	vn_init(void);
 extern bhv_vnode_t	*vn_initialize(struct inode *);
-extern int	vn_revalidate(struct bhv_vnode *);
-extern int	__vn_revalidate(struct bhv_vnode *, bhv_vattr_t *);
-extern void	vn_revalidate_core(struct bhv_vnode *, bhv_vattr_t *);
-
-extern void	vn_iowait(struct bhv_vnode *vp);
-extern void	vn_iowake(struct bhv_vnode *vp);
+extern int	vn_revalidate(bhv_vnode_t *);
+extern int	__vn_revalidate(bhv_vnode_t *, bhv_vattr_t *);
+extern void	vn_revalidate_core(bhv_vnode_t *, bhv_vattr_t *);
 
-extern void	vn_ioerror(struct bhv_vnode *vp, int error, char *f, int l);
+/*
+ * Yeah, these don't take vnode anymore at all, all this should be
+ * cleaned up at some point.
+ */
+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(struct bhv_vnode *vp)
+static inline int vn_count(bhv_vnode_t *vp)
 {
 	return atomic_read(&vn_to_inode(vp)->i_count);
 }
@@ -445,21 +208,21 @@ static inline int vn_count(struct bhv_vnode *vp)
 /*
  * Vnode reference counting functions (and macros for compatibility).
  */
-extern bhv_vnode_t	*vn_hold(struct bhv_vnode *);
+extern bhv_vnode_t	*vn_hold(bhv_vnode_t *);
 
 #if defined(XFS_VNODE_TRACE)
 #define VN_HOLD(vp)		\
 	((void)vn_hold(vp),	\
-	  vn_trace_hold(vp, __FILE__, __LINE__, (inst_t *)__return_address))
+	  vn_trace_hold(xfs_vtoi(vp), __FILE__, __LINE__, (inst_t *)__return_address))
 #define VN_RELE(vp)		\
-	  (vn_trace_rele(vp, __FILE__, __LINE__, (inst_t *)__return_address), \
+	  (vn_trace_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
 
-static inline struct bhv_vnode *vn_grab(struct bhv_vnode *vp)
+static inline bhv_vnode_t *vn_grab(bhv_vnode_t *vp)
 {
 	struct inode *inode = igrab(vn_to_inode(vp));
 	return inode ? vn_from_inode(inode) : NULL;
@@ -472,44 +235,15 @@ static inline struct bhv_vnode *vn_grab(struct bhv_vnode *vp)
 #define VNAMELEN(dentry)	((dentry)->d_name.len)
 #define VNAME_TO_VNODE(dentry)	(vn_from_inode((dentry)->d_inode))
 
-/*
- * Vnode spinlock manipulation.
- */
-#define VN_LOCK(vp)		mutex_spinlock(&(vp)->v_lock)
-#define VN_UNLOCK(vp, s)	mutex_spinunlock(&(vp)->v_lock, s)
-
-STATIC_INLINE void vn_flagset(struct bhv_vnode *vp, uint flag)
-{
-	spin_lock(&vp->v_lock);
-	vp->v_flag |= flag;
-	spin_unlock(&vp->v_lock);
-}
-
-STATIC_INLINE uint vn_flagclr(struct bhv_vnode *vp, uint flag)
-{
-	uint	cleared;
-
-	spin_lock(&vp->v_lock);
-	cleared = (vp->v_flag & flag);
-	vp->v_flag &= ~flag;
-	spin_unlock(&vp->v_lock);
-	return cleared;
-}
-
-#define VMODIFY(vp)	vn_flagset(vp, VMODIFIED)
-#define VUNMODIFY(vp)	vn_flagclr(vp, VMODIFIED)
-#define VTRUNCATE(vp)	vn_flagset(vp, VTRUNCATED)
-#define VUNTRUNCATE(vp)	vn_flagclr(vp, VTRUNCATED)
-
 /*
  * Dealing with bad inodes
  */
-static inline void vn_mark_bad(struct bhv_vnode *vp)
+static inline void vn_mark_bad(bhv_vnode_t *vp)
 {
 	make_bad_inode(vn_to_inode(vp));
 }
 
-static inline int VN_BAD(struct bhv_vnode *vp)
+static inline int VN_BAD(bhv_vnode_t *vp)
 {
 	return is_bad_inode(vn_to_inode(vp));
 }
@@ -519,18 +253,18 @@ static inline int VN_BAD(struct bhv_vnode *vp)
  */
 static inline void vn_atime_to_bstime(bhv_vnode_t *vp, xfs_bstime_t *bs_atime)
 {
-	bs_atime->tv_sec = vp->v_inode.i_atime.tv_sec;
-	bs_atime->tv_nsec = vp->v_inode.i_atime.tv_nsec;
+	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)
 {
-	*ts = vp->v_inode.i_atime;
+	*ts = vp->i_atime;
 }
 
 static inline void vn_atime_to_time_t(bhv_vnode_t *vp, time_t *tt)
 {
-	*tt = vp->v_inode.i_atime.tv_sec;
+	*tt = vp->i_atime.tv_sec;
 }
 
 /*
@@ -540,7 +274,6 @@ static inline void vn_atime_to_time_t(bhv_vnode_t *vp, time_t *tt)
 #define VN_CACHED(vp)	(vn_to_inode(vp)->i_mapping->nrpages)
 #define VN_DIRTY(vp)	mapping_tagged(vn_to_inode(vp)->i_mapping, \
 					PAGECACHE_TAG_DIRTY)
-#define VN_TRUNC(vp)	((vp)->v_flag & VTRUNCATED)
 
 /*
  * Flags to vop_setattr/getattr.
@@ -572,21 +305,17 @@ static inline void vn_atime_to_time_t(bhv_vnode_t *vp, time_t *tt)
 #define	VNODE_KTRACE_REF	4
 #define	VNODE_KTRACE_RELE	5
 
-extern void vn_trace_entry(struct bhv_vnode *, const char *, inst_t *);
-extern void vn_trace_exit(struct bhv_vnode *, const char *, inst_t *);
-extern void vn_trace_hold(struct bhv_vnode *, char *, int, inst_t *);
-extern void vn_trace_ref(struct bhv_vnode *, char *, int, inst_t *);
-extern void vn_trace_rele(struct bhv_vnode *, char *, int, inst_t *);
-
-#define	VN_TRACE(vp)		\
-	vn_trace_ref(vp, __FILE__, __LINE__, (inst_t *)__return_address)
+extern void vn_trace_entry(struct xfs_inode *, const char *, inst_t *);
+extern void vn_trace_exit(struct xfs_inode *, const char *, inst_t *);
+extern void vn_trace_hold(struct xfs_inode *, char *, int, inst_t *);
+extern void vn_trace_ref(struct xfs_inode *, char *, int, inst_t *);
+extern void vn_trace_rele(struct xfs_inode *, char *, int, inst_t *);
 #else
 #define	vn_trace_entry(a,b,c)
 #define	vn_trace_exit(a,b,c)
 #define	vn_trace_hold(a,b,c,d)
 #define	vn_trace_ref(a,b,c,d)
 #define	vn_trace_rele(a,b,c,d)
-#define	VN_TRACE(vp)
 #endif
 
 #endif	/* __XFS_VNODE_H__ */

fs/xfs/quota/xfs_qm.c

@@ -287,45 +287,6 @@ xfs_qm_rele_quotafs_ref(
 	XFS_QM_UNLOCK(xfs_Gqm);
 }
 
-/*
- * This is called at mount time from xfs_mountfs to initialize the quotainfo
- * structure and start the global quota manager (xfs_Gqm) if it hasn't done
- * so already.	Note that the superblock has not been read in yet.
- */
-void
-xfs_qm_mount_quotainit(
-	xfs_mount_t	*mp,
-	uint		flags)
-{
-	/*
-	 * User, projects or group quotas has to be on.
-	 */
-	ASSERT(flags & (XFSMNT_UQUOTA | XFSMNT_PQUOTA | XFSMNT_GQUOTA));
-
-	/*
-	 * Initialize the flags in the mount structure. From this point
-	 * onwards we look at m_qflags to figure out if quotas's ON/OFF, etc.
-	 * Note that we enforce nothing if accounting is off.
-	 * ie.	XFSMNT_*QUOTA must be ON for XFSMNT_*QUOTAENF.
-	 * It isn't necessary to take the quotaoff lock to do this; this is
-	 * called from mount.
-	 */
-	if (flags & XFSMNT_UQUOTA) {
-		mp->m_qflags |= (XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE);
-		if (flags & XFSMNT_UQUOTAENF)
-			mp->m_qflags |= XFS_UQUOTA_ENFD;
-	}
-	if (flags & XFSMNT_GQUOTA) {
-		mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE);
-		if (flags & XFSMNT_GQUOTAENF)
-			mp->m_qflags |= XFS_OQUOTA_ENFD;
-	} else if (flags & XFSMNT_PQUOTA) {
-		mp->m_qflags |= (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE);
-		if (flags & XFSMNT_PQUOTAENF)
-			mp->m_qflags |= XFS_OQUOTA_ENFD;
-	}
-}
-
 /*
  * Just destroy the quotainfo structure.
  */
@@ -1039,7 +1000,7 @@ xfs_qm_dqdetach(
 int
 xfs_qm_sync(
 	xfs_mount_t	*mp,
-	short		flags)
+	int		flags)
 {
 	int		recl, restarts;
 	xfs_dquot_t	*dqp;
@@ -1717,7 +1678,6 @@ xfs_qm_get_rtblks(
 	xfs_extnum_t	idx;			/* extent record index */
 	xfs_ifork_t	*ifp;			/* inode fork pointer */
 	xfs_extnum_t	nextents;		/* number of extent entries */
-	xfs_bmbt_rec_t	*ep;			/* pointer to an extent entry */
 	int		error;
 
 	ASSERT(XFS_IS_REALTIME_INODE(ip));
@@ -1728,10 +1688,8 @@ xfs_qm_get_rtblks(
 	}
 	rtblks = 0;
 	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	for (idx = 0; idx < nextents; idx++) {
-		ep = xfs_iext_get_ext(ifp, idx);
-		rtblks += xfs_bmbt_get_blockcount(ep);
-	}
+	for (idx = 0; idx < nextents; idx++)
+		rtblks += xfs_bmbt_get_blockcount(xfs_iext_get_ext(ifp, idx));
 	*O_rtblks = (xfs_qcnt_t)rtblks;
 	return 0;
 }
@@ -2459,8 +2417,7 @@ xfs_qm_vop_dqalloc(
 	lockflags = XFS_ILOCK_EXCL;
 	xfs_ilock(ip, lockflags);
 
-	if ((flags & XFS_QMOPT_INHERIT) &&
-	    XFS_INHERIT_GID(ip, XFS_MTOVFS(mp)))
+	if ((flags & XFS_QMOPT_INHERIT) && XFS_INHERIT_GID(ip))
 		gid = ip->i_d.di_gid;
 
 	/*

fs/xfs/quota/xfs_qm.h

@@ -166,12 +166,11 @@ typedef struct xfs_dquot_acct {
 
 extern void		xfs_qm_destroy_quotainfo(xfs_mount_t *);
 extern int		xfs_qm_mount_quotas(xfs_mount_t *, int);
-extern void		xfs_qm_mount_quotainit(xfs_mount_t *, uint);
 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 *);
 extern int		xfs_qm_write_sb_changes(xfs_mount_t *, __int64_t);
-extern int		xfs_qm_sync(xfs_mount_t *, short);
+extern int		xfs_qm_sync(xfs_mount_t *, int);
 
 /* dquot stuff */
 extern boolean_t	xfs_qm_dqalloc_incore(xfs_dquot_t **);
@@ -199,7 +198,8 @@ extern void		xfs_qm_freelist_unlink(xfs_dquot_t *);
 extern int		xfs_qm_freelist_lock_nowait(xfs_qm_t *);
 
 /* system call interface */
-extern int		xfs_qm_quotactl(bhv_desc_t *, int, int, xfs_caddr_t);
+extern int		xfs_qm_quotactl(struct xfs_mount *, int, int,
+				xfs_caddr_t);
 
 #ifdef DEBUG
 extern int		xfs_qm_internalqcheck(xfs_mount_t *);

fs/xfs/quota/xfs_qm_bhv.c

@@ -48,172 +48,13 @@
 #include "xfs_buf_item.h"
 #include "xfs_qm.h"
 
-#define MNTOPT_QUOTA	"quota"		/* disk quotas (user) */
-#define MNTOPT_NOQUOTA	"noquota"	/* no quotas */
-#define MNTOPT_USRQUOTA	"usrquota"	/* user quota enabled */
-#define MNTOPT_GRPQUOTA	"grpquota"	/* group quota enabled */
-#define MNTOPT_PRJQUOTA	"prjquota"	/* project quota enabled */
-#define MNTOPT_UQUOTA	"uquota"	/* user quota (IRIX variant) */
-#define MNTOPT_GQUOTA	"gquota"	/* group quota (IRIX variant) */
-#define MNTOPT_PQUOTA	"pquota"	/* project quota (IRIX variant) */
-#define MNTOPT_UQUOTANOENF "uqnoenforce"/* user quota limit enforcement */
-#define MNTOPT_GQUOTANOENF "gqnoenforce"/* group quota limit enforcement */
-#define MNTOPT_PQUOTANOENF "pqnoenforce"/* project quota limit enforcement */
-#define MNTOPT_QUOTANOENF  "qnoenforce"	/* same as uqnoenforce */
 
-STATIC int
-xfs_qm_parseargs(
-	struct bhv_desc		*bhv,
-	char			*options,
-	struct xfs_mount_args	*args,
-	int			update)
-{
-	size_t			length;
-	char			*local_options = options;
-	char			*this_char;
-	int			error;
-	int			referenced = update;
-
-	while ((this_char = strsep(&local_options, ",")) != NULL) {
-		length = strlen(this_char);
-		if (local_options)
-			length++;
-
-		if (!strcmp(this_char, MNTOPT_NOQUOTA)) {
-			args->flags &= ~(XFSMNT_UQUOTAENF|XFSMNT_UQUOTA);
-			args->flags &= ~(XFSMNT_GQUOTAENF|XFSMNT_GQUOTA);
-			referenced = update;
-		} else if (!strcmp(this_char, MNTOPT_QUOTA) ||
-			   !strcmp(this_char, MNTOPT_UQUOTA) ||
-			   !strcmp(this_char, MNTOPT_USRQUOTA)) {
-			args->flags |= XFSMNT_UQUOTA | XFSMNT_UQUOTAENF;
-			referenced = 1;
-		} else if (!strcmp(this_char, MNTOPT_QUOTANOENF) ||
-			   !strcmp(this_char, MNTOPT_UQUOTANOENF)) {
-			args->flags |= XFSMNT_UQUOTA;
-			args->flags &= ~XFSMNT_UQUOTAENF;
-			referenced = 1;
-		} else if (!strcmp(this_char, MNTOPT_PQUOTA) ||
-			   !strcmp(this_char, MNTOPT_PRJQUOTA)) {
-			args->flags |= XFSMNT_PQUOTA | XFSMNT_PQUOTAENF;
-			referenced = 1;
-		} else if (!strcmp(this_char, MNTOPT_PQUOTANOENF)) {
-			args->flags |= XFSMNT_PQUOTA;
-			args->flags &= ~XFSMNT_PQUOTAENF;
-			referenced = 1;
-		} else if (!strcmp(this_char, MNTOPT_GQUOTA) ||
-			   !strcmp(this_char, MNTOPT_GRPQUOTA)) {
-			args->flags |= XFSMNT_GQUOTA | XFSMNT_GQUOTAENF;
-			referenced = 1;
-		} else if (!strcmp(this_char, MNTOPT_GQUOTANOENF)) {
-			args->flags |= XFSMNT_GQUOTA;
-			args->flags &= ~XFSMNT_GQUOTAENF;
-			referenced = 1;
-		} else {
-			if (local_options)
-				*(local_options-1) = ',';
-			continue;
-		}
-
-		while (length--)
-			*this_char++ = ',';
-	}
-
-	if ((args->flags & XFSMNT_GQUOTA) && (args->flags & XFSMNT_PQUOTA)) {
-		cmn_err(CE_WARN,
-			"XFS: cannot mount with both project and group quota");
-		return XFS_ERROR(EINVAL);
-	}
-
-	error = bhv_next_vfs_parseargs(BHV_NEXT(bhv), options, args, update);
-	if (!error && !referenced)
-		bhv_remove_vfsops(bhvtovfs(bhv), VFS_POSITION_QM);
-	return error;
-}
-
-STATIC int
-xfs_qm_showargs(
-	struct bhv_desc		*bhv,
-	struct seq_file		*m)
-{
-	struct bhv_vfs		*vfsp = bhvtovfs(bhv);
-	struct xfs_mount	*mp = XFS_VFSTOM(vfsp);
-
-	if (mp->m_qflags & XFS_UQUOTA_ACCT) {
-		(mp->m_qflags & XFS_UQUOTA_ENFD) ?
-			seq_puts(m, "," MNTOPT_USRQUOTA) :
-			seq_puts(m, "," MNTOPT_UQUOTANOENF);
-	}
-
-	if (mp->m_qflags & XFS_PQUOTA_ACCT) {
-		(mp->m_qflags & XFS_OQUOTA_ENFD) ?
-			seq_puts(m, "," MNTOPT_PRJQUOTA) :
-			seq_puts(m, "," MNTOPT_PQUOTANOENF);
-	}
-
-	if (mp->m_qflags & XFS_GQUOTA_ACCT) {
-		(mp->m_qflags & XFS_OQUOTA_ENFD) ?
-			seq_puts(m, "," MNTOPT_GRPQUOTA) :
-			seq_puts(m, "," MNTOPT_GQUOTANOENF);
-	}
-
-	if (!(mp->m_qflags & XFS_ALL_QUOTA_ACCT))
-		seq_puts(m, "," MNTOPT_NOQUOTA);
-
-	return bhv_next_vfs_showargs(BHV_NEXT(bhv), m);
-}
-
-STATIC int
-xfs_qm_mount(
-	struct bhv_desc		*bhv,
-	struct xfs_mount_args	*args,
-	struct cred		*cr)
-{
-	struct bhv_vfs		*vfsp = bhvtovfs(bhv);
-	struct xfs_mount	*mp = XFS_VFSTOM(vfsp);
-
-	if (args->flags & (XFSMNT_UQUOTA | XFSMNT_GQUOTA | XFSMNT_PQUOTA))
-		xfs_qm_mount_quotainit(mp, args->flags);
-	return bhv_next_vfs_mount(BHV_NEXT(bhv), args, cr);
-}
-
-/*
- * Directory tree accounting is implemented using project quotas, where
- * the project identifier is inherited from parent directories.
- * A statvfs (df, etc.) of a directory that is using project quota should
- * return a statvfs of the project, not the entire filesystem.
- * This makes such trees appear as if they are filesystems in themselves.
- */
-STATIC int
-xfs_qm_statvfs(
-	struct bhv_desc		*bhv,
+STATIC void
+xfs_fill_statvfs_from_dquot(
 	bhv_statvfs_t		*statp,
-	struct bhv_vnode	*vnode)
+	xfs_disk_dquot_t	*dp)
 {
-	xfs_mount_t		*mp;
-	xfs_inode_t		*ip;
-	xfs_dquot_t		*dqp;
-	xfs_disk_dquot_t	*dp;
 	__uint64_t		limit;
-	int			error;
-
-	error = bhv_next_vfs_statvfs(BHV_NEXT(bhv), statp, vnode);
-	if (error || !vnode)
-		return error;
-
-	mp = xfs_vfstom(bhvtovfs(bhv));
-	ip = xfs_vtoi(vnode);
-
-	if (!(ip->i_d.di_flags & XFS_DIFLAG_PROJINHERIT))
-		return 0;
-	if (!(mp->m_qflags & XFS_PQUOTA_ACCT))
-		return 0;
-	if (!(mp->m_qflags & XFS_OQUOTA_ENFD))
-		return 0;
-
-	if (xfs_qm_dqget(mp, NULL, ip->i_d.di_projid, XFS_DQ_PROJ, 0, &dqp))
-		return 0;
-	dp = &dqp->q_core;
 
 	limit = dp->d_blk_softlimit ?
 		be64_to_cpu(dp->d_blk_softlimit) :
@@ -234,37 +75,35 @@ xfs_qm_statvfs(
 			(statp->f_files > be64_to_cpu(dp->d_icount)) ?
 			 (statp->f_ffree - be64_to_cpu(dp->d_icount)) : 0;
 	}
-
-	xfs_qm_dqput(dqp);
-	return 0;
 }
 
-STATIC int
-xfs_qm_syncall(
-	struct bhv_desc		*bhv,
-	int			flags,
-	cred_t			*credp)
+
+/*
+ * Directory tree accounting is implemented using project quotas, where
+ * the project identifier is inherited from parent directories.
+ * A statvfs (df, etc.) of a directory that is using project quota should
+ * return a statvfs of the project, not the entire filesystem.
+ * This makes such trees appear as if they are filesystems in themselves.
+ */
+STATIC void
+xfs_qm_statvfs(
+	xfs_inode_t		*ip,
+	bhv_statvfs_t		*statp)
 {
-	struct bhv_vfs		*vfsp = bhvtovfs(bhv);
-	struct xfs_mount	*mp = XFS_VFSTOM(vfsp);
-	int			error;
+	xfs_mount_t		*mp = ip->i_mount;
+	xfs_dquot_t		*dqp;
 
-	/*
-	 * Get the Quota Manager to flush the dquots.
-	 */
-	if (XFS_IS_QUOTA_ON(mp)) {
-		if ((error = xfs_qm_sync(mp, flags))) {
-			/*
-			 * If we got an IO error, we will be shutting down.
-			 * So, there's nothing more for us to do here.
-			 */
-			ASSERT(error != EIO || XFS_FORCED_SHUTDOWN(mp));
-			if (XFS_FORCED_SHUTDOWN(mp)) {
-				return XFS_ERROR(error);
-			}
-		}
+	if (!(ip->i_d.di_flags & XFS_DIFLAG_PROJINHERIT) ||
+	    !((mp->m_qflags & (XFS_PQUOTA_ACCT|XFS_OQUOTA_ENFD))) ==
+	    		      (XFS_PQUOTA_ACCT|XFS_OQUOTA_ENFD))
+		return;
+
+	if (!xfs_qm_dqget(mp, NULL, ip->i_d.di_projid, XFS_DQ_PROJ, 0, &dqp)) {
+		xfs_disk_dquot_t	*dp = &dqp->q_core;
+
+		xfs_fill_statvfs_from_dquot(statp, dp);
+		xfs_qm_dqput(dqp);
 	}
-	return bhv_next_vfs_sync(BHV_NEXT(bhv), flags, credp);
 }
 
 STATIC int
@@ -382,7 +221,7 @@ xfs_qm_dqrele_null(
 }
 
 
-static struct xfs_qmops xfs_qmcore_xfs = {
+struct xfs_qmops xfs_qmcore_xfs = {
 	.xfs_qminit		= xfs_qm_newmount,
 	.xfs_qmdone		= xfs_qm_unmount_quotadestroy,
 	.xfs_qmmount		= xfs_qm_endmount,
@@ -396,36 +235,24 @@ static struct xfs_qmops xfs_qmcore_xfs = {
 	.xfs_dqvoprename	= xfs_qm_vop_rename_dqattach,
 	.xfs_dqvopchown		= xfs_qm_vop_chown,
 	.xfs_dqvopchownresv	= xfs_qm_vop_chown_reserve,
+	.xfs_dqstatvfs		= xfs_qm_statvfs,
+	.xfs_dqsync		= xfs_qm_sync,
+	.xfs_quotactl		= xfs_qm_quotactl,
 	.xfs_dqtrxops		= &xfs_trans_dquot_ops,
 };
-
-struct bhv_module_vfsops xfs_qmops = { {
-	BHV_IDENTITY_INIT(VFS_BHV_QM, VFS_POSITION_QM),
-	.vfs_parseargs		= xfs_qm_parseargs,
-	.vfs_showargs		= xfs_qm_showargs,
-	.vfs_mount		= xfs_qm_mount,
-	.vfs_statvfs		= xfs_qm_statvfs,
-	.vfs_sync		= xfs_qm_syncall,
-	.vfs_quotactl		= xfs_qm_quotactl, },
-};
-
+EXPORT_SYMBOL(xfs_qmcore_xfs);
 
 void __init
 xfs_qm_init(void)
 {
-	static char	message[] __initdata =
-		KERN_INFO "SGI XFS Quota Management subsystem\n";
-
-	printk(message);
+	printk(KERN_INFO "SGI XFS Quota Management subsystem\n");
 	mutex_init(&xfs_Gqm_lock);
-	vfs_bhv_set_custom(&xfs_qmops, &xfs_qmcore_xfs);
 	xfs_qm_init_procfs();
 }
 
 void __exit
 xfs_qm_exit(void)
 {
-	vfs_bhv_clr_custom(&xfs_qmops);
 	xfs_qm_cleanup_procfs();
 	if (qm_dqzone)
 		kmem_zone_destroy(qm_dqzone);

fs/xfs/quota/xfs_qm_syscalls.c

@@ -81,18 +81,13 @@ STATIC void	xfs_qm_export_dquot(xfs_mount_t *, xfs_disk_dquot_t *,
  */
 int
 xfs_qm_quotactl(
-	struct bhv_desc *bdp,
+	xfs_mount_t	*mp,
 	int		cmd,
 	int		id,
 	xfs_caddr_t	addr)
 {
-	xfs_mount_t	*mp;
-	bhv_vfs_t	*vfsp;
 	int		error;
 
-	vfsp = bhvtovfs(bdp);
-	mp = XFS_VFSTOM(vfsp);
-
 	ASSERT(addr != NULL || cmd == Q_XQUOTASYNC);
 
 	/*
@@ -105,7 +100,7 @@ xfs_qm_quotactl(
 		 */
 		if (XFS_IS_QUOTA_ON(mp))
 			return XFS_ERROR(EINVAL);
-		if (vfsp->vfs_flag & VFS_RDONLY)
+		if (mp->m_flags & XFS_MOUNT_RDONLY)
 			return XFS_ERROR(EROFS);
 		return (xfs_qm_scall_trunc_qfiles(mp,
 			       xfs_qm_import_qtype_flags(*(uint *)addr)));
@@ -121,13 +116,13 @@ xfs_qm_quotactl(
 		 * QUOTAON - enabling quota enforcement.
 		 * Quota accounting must be turned on at mount time.
 		 */
-		if (vfsp->vfs_flag & VFS_RDONLY)
+		if (mp->m_flags & XFS_MOUNT_RDONLY)
 			return XFS_ERROR(EROFS);
 		return (xfs_qm_scall_quotaon(mp,
 					  xfs_qm_import_flags(*(uint *)addr)));
 
 	case Q_XQUOTAOFF:
-		if (vfsp->vfs_flag & VFS_RDONLY)
+		if (mp->m_flags & XFS_MOUNT_RDONLY)
 			return XFS_ERROR(EROFS);
 		break;
 
@@ -143,7 +138,7 @@ xfs_qm_quotactl(
 
 	switch (cmd) {
 	case Q_XQUOTAOFF:
-		if (vfsp->vfs_flag & VFS_RDONLY)
+		if (mp->m_flags & XFS_MOUNT_RDONLY)
 			return XFS_ERROR(EROFS);
 		error = xfs_qm_scall_quotaoff(mp,
 					    xfs_qm_import_flags(*(uint *)addr),
@@ -164,19 +159,19 @@ xfs_qm_quotactl(
 		break;
 
 	case Q_XSETQLIM:
-		if (vfsp->vfs_flag & VFS_RDONLY)
+		if (mp->m_flags & XFS_MOUNT_RDONLY)
 			return XFS_ERROR(EROFS);
 		error = xfs_qm_scall_setqlim(mp, (xfs_dqid_t)id, XFS_DQ_USER,
 					     (fs_disk_quota_t *)addr);
 		break;
 	case Q_XSETGQLIM:
-		if (vfsp->vfs_flag & VFS_RDONLY)
+		if (mp->m_flags & XFS_MOUNT_RDONLY)
 			return XFS_ERROR(EROFS);
 		error = xfs_qm_scall_setqlim(mp, (xfs_dqid_t)id, XFS_DQ_GROUP,
 					     (fs_disk_quota_t *)addr);
 		break;
 	case Q_XSETPQLIM:
-		if (vfsp->vfs_flag & VFS_RDONLY)
+		if (mp->m_flags & XFS_MOUNT_RDONLY)
 			return XFS_ERROR(EROFS);
 		error = xfs_qm_scall_setqlim(mp, (xfs_dqid_t)id, XFS_DQ_PROJ,
 					     (fs_disk_quota_t *)addr);

fs/xfs/support/move.c

@@ -1,51 +0,0 @@
-/*
- * Copyright (c) 2000-2003,2005 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
- */
-#include <xfs.h>
-
-/* Read from kernel buffer at src to user/kernel buffer defined
- * by the uio structure. Advance the pointer in the uio struct
- * as we go.
- */
-int
-xfs_uio_read(caddr_t src, size_t len, struct uio *uio)
-{
-	size_t	count;
-
-	if (!len || !uio->uio_resid)
-		return 0;
-
-	count = uio->uio_iov->iov_len;
-	if (!count)
-		return 0;
-	if (count > len)
-		count = len;
-
-	if (uio->uio_segflg == UIO_USERSPACE) {
-		if (copy_to_user(uio->uio_iov->iov_base, src, count))
-			return EFAULT;
-	} else {
-		ASSERT(uio->uio_segflg == UIO_SYSSPACE);
-		memcpy(uio->uio_iov->iov_base, src, count);
-	}
-
-	uio->uio_iov->iov_base = (void*)((char*)uio->uio_iov->iov_base + count);
-	uio->uio_iov->iov_len -= count;
-	uio->uio_offset += count;
-	uio->uio_resid -= count;
-	return 0;
-}

fs/xfs/support/move.h

@@ -1,70 +0,0 @@
-/*
- * Copyright (c) 2000-2003,2005 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
- *
- * Portions Copyright (c) 1982, 1986, 1993, 1994
- *	The Regents of the University of California.  All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. Neither the name of the University nor the names of its contributors
- *    may be used to endorse or promote products derived from this software
- *    without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- */
-#ifndef __XFS_SUPPORT_MOVE_H__
-#define __XFS_SUPPORT_MOVE_H__
-
-#include <linux/uio.h>
-#include <asm/uaccess.h>
-
-/* Segment flag values. */
-enum uio_seg {
-	UIO_USERSPACE,          /* from user data space */
-	UIO_SYSSPACE,           /* from system space */
-};
-
-struct uio {
-	struct kvec	*uio_iov;   /* pointer to array of iovecs */
-	int		uio_iovcnt; /* number of iovecs in array */
-	xfs_off_t	uio_offset; /* offset in file this uio corresponds to */
-	int		uio_resid;  /* residual i/o count */
-	enum uio_seg	uio_segflg; /* see above */
-};
-
-typedef struct uio uio_t;
-typedef struct kvec iovec_t;
-
-extern int	xfs_uio_read (caddr_t, size_t, uio_t *);
-
-#endif  /* __XFS_SUPPORT_MOVE_H__ */

fs/xfs/xfs_acl.c

@@ -32,6 +32,7 @@
 #include "xfs_btree.h"
 #include "xfs_acl.h"
 #include "xfs_attr.h"
+#include "xfs_vnodeops.h"
 
 #include <linux/capability.h>
 #include <linux/posix_acl_xattr.h>
@@ -241,7 +242,7 @@ xfs_acl_vget(
 			bhv_vattr_t	va;
 
 			va.va_mask = XFS_AT_MODE;
-			error = bhv_vop_getattr(vp, &va, 0, sys_cred);
+			error = xfs_getattr(xfs_vtoi(vp), &va, 0);
 			if (error)
 				goto out;
 			xfs_acl_sync_mode(va.va_mode, xfs_acl);
@@ -265,9 +266,10 @@ xfs_acl_vremove(
 	VN_HOLD(vp);
 	error = xfs_acl_allow_set(vp, kind);
 	if (!error) {
-		error = bhv_vop_attr_remove(vp, kind == _ACL_TYPE_DEFAULT?
+		error = xfs_attr_remove(xfs_vtoi(vp),
+						kind == _ACL_TYPE_DEFAULT?
 						SGI_ACL_DEFAULT: SGI_ACL_FILE,
-						ATTR_ROOT, sys_cred);
+						ATTR_ROOT);
 		if (error == ENOATTR)
 			error = 0;	/* 'scool */
 	}
@@ -370,17 +372,18 @@ xfs_acl_allow_set(
 	bhv_vnode_t	*vp,
 	int		kind)
 {
+	xfs_inode_t	*ip = xfs_vtoi(vp);
 	bhv_vattr_t	va;
 	int		error;
 
-	if (vp->v_inode.i_flags & (S_IMMUTABLE|S_APPEND))
+	if (vp->i_flags & (S_IMMUTABLE|S_APPEND))
 		return EPERM;
 	if (kind == _ACL_TYPE_DEFAULT && !VN_ISDIR(vp))
 		return ENOTDIR;
-	if (vp->v_vfsp->vfs_flag & VFS_RDONLY)
+	if (vp->i_sb->s_flags & MS_RDONLY)
 		return EROFS;
 	va.va_mask = XFS_AT_UID;
-	error = bhv_vop_getattr(vp, &va, 0, NULL);
+	error = xfs_getattr(ip, &va, 0);
 	if (error)
 		return error;
 	if (va.va_uid != current->fsuid && !capable(CAP_FOWNER))
@@ -613,7 +616,8 @@ xfs_acl_get_attr(
 
 	ASSERT((flags & ATTR_KERNOVAL) ? (aclp == NULL) : 1);
 	flags |= ATTR_ROOT;
-	*error = bhv_vop_attr_get(vp, kind == _ACL_TYPE_ACCESS ?
+	*error = xfs_attr_get(xfs_vtoi(vp),
+					kind == _ACL_TYPE_ACCESS ?
 					SGI_ACL_FILE : SGI_ACL_DEFAULT,
 					(char *)aclp, &len, flags, sys_cred);
 	if (*error || (flags & ATTR_KERNOVAL))
@@ -651,9 +655,10 @@ 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 = bhv_vop_attr_set(vp, kind == _ACL_TYPE_ACCESS ?
+	*error = xfs_attr_set(xfs_vtoi(vp),
+				kind == _ACL_TYPE_ACCESS ?
 				SGI_ACL_FILE: SGI_ACL_DEFAULT,
-				(char *)newacl, len, ATTR_ROOT, sys_cred);
+				(char *)newacl, len, ATTR_ROOT);
 	_ACL_FREE(newacl);
 }
 
@@ -675,7 +680,7 @@ xfs_acl_vtoacl(
 		if (!error) {
 			/* Got the ACL, need the mode... */
 			va.va_mask = XFS_AT_MODE;
-			error = bhv_vop_getattr(vp, &va, 0, sys_cred);
+			error = xfs_getattr(xfs_vtoi(vp), &va, 0);
 		}
 
 		if (error)
@@ -699,7 +704,7 @@ xfs_acl_vtoacl(
 int
 xfs_acl_inherit(
 	bhv_vnode_t	*vp,
-	bhv_vattr_t	*vap,
+	mode_t		mode,
 	xfs_acl_t	*pdaclp)
 {
 	xfs_acl_t	*cacl;
@@ -727,7 +732,7 @@ xfs_acl_inherit(
 		return ENOMEM;
 
 	memcpy(cacl, pdaclp, sizeof(xfs_acl_t));
-	xfs_acl_filter_mode(vap->va_mode, cacl);
+	xfs_acl_filter_mode(mode, cacl);
 	xfs_acl_setmode(vp, cacl, &basicperms);
 
 	/*
@@ -773,7 +778,7 @@ xfs_acl_setmode(
 	 * mode.  The m:: bits take precedence over the g:: bits.
 	 */
 	va.va_mask = XFS_AT_MODE;
-	error = bhv_vop_getattr(vp, &va, 0, sys_cred);
+	error = xfs_getattr(xfs_vtoi(vp), &va, 0);
 	if (error)
 		return error;
 
@@ -807,7 +812,7 @@ xfs_acl_setmode(
 	if (gap && nomask)
 		va.va_mode |= gap->ae_perm << 3;
 
-	return bhv_vop_setattr(vp, &va, 0, sys_cred);
+	return xfs_setattr(xfs_vtoi(vp), &va, 0, sys_cred);
 }
 
 /*

fs/xfs/xfs_acl.h

@@ -50,7 +50,6 @@ typedef struct xfs_acl {
 #ifdef CONFIG_XFS_POSIX_ACL
 
 struct vattr;
-struct bhv_vnode;
 struct xfs_inode;
 
 extern struct kmem_zone *xfs_acl_zone;
@@ -58,20 +57,20 @@ 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(struct bhv_vnode *, struct bhv_vattr *, xfs_acl_t *);
+extern int xfs_acl_inherit(bhv_vnode_t *, mode_t mode, xfs_acl_t *);
 extern int xfs_acl_iaccess(struct xfs_inode *, mode_t, cred_t *);
-extern int xfs_acl_vtoacl(struct bhv_vnode *, xfs_acl_t *, xfs_acl_t *);
-extern int xfs_acl_vhasacl_access(struct bhv_vnode *);
-extern int xfs_acl_vhasacl_default(struct bhv_vnode *);
-extern int xfs_acl_vset(struct bhv_vnode *, void *, size_t, int);
-extern int xfs_acl_vget(struct bhv_vnode *, void *, size_t, int);
-extern int xfs_acl_vremove(struct bhv_vnode *, int);
+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);
 
 #define _ACL_TYPE_ACCESS	1
 #define _ACL_TYPE_DEFAULT	2
 #define _ACL_PERM_INVALID(perm)	((perm) & ~(ACL_READ|ACL_WRITE|ACL_EXECUTE))
 
-#define _ACL_INHERIT(c,v,d)	(xfs_acl_inherit(c,v,d))
+#define _ACL_INHERIT(c,m,d)	(xfs_acl_inherit(c,m,d))
 #define _ACL_GET_ACCESS(pv,pa)	(xfs_acl_vtoacl(pv,pa,NULL) == 0)
 #define _ACL_GET_DEFAULT(pv,pd)	(xfs_acl_vtoacl(pv,NULL,pd) == 0)
 #define _ACL_ACCESS_EXISTS	xfs_acl_vhasacl_access
@@ -91,7 +90,7 @@ extern int xfs_acl_vremove(struct bhv_vnode *, int);
 #define xfs_acl_vhasacl_default(v)	(0)
 #define _ACL_ALLOC(a)		(1)	/* successfully allocate nothing */
 #define _ACL_FREE(a)		((void)0)
-#define _ACL_INHERIT(c,v,d)	(0)
+#define _ACL_INHERIT(c,m,d)	(0)
 #define _ACL_GET_ACCESS(pv,pa)	(0)
 #define _ACL_GET_DEFAULT(pv,pd)	(0)
 #define _ACL_ACCESS_EXISTS	(NULL)

fs/xfs/xfs_ag.h

@@ -197,6 +197,10 @@ typedef struct xfs_perag
 #endif
 	xfs_perag_busy_t *pagb_list;	/* unstable blocks */
 	atomic_t        pagf_fstrms;    /* # of filestreams active in this AG */
+
+	int		pag_ici_init;	/* incore inode cache initialised */
+	rwlock_t	pag_ici_lock;	/* incore inode lock */
+	struct radix_tree_root pag_ici_root;	/* incore inode cache root */
 } xfs_perag_t;
 
 #define	XFS_AG_MAXLEVELS(mp)		((mp)->m_ag_maxlevels)

fs/xfs/xfs_attr.c

@@ -49,6 +49,7 @@
 #include "xfs_trans_space.h"
 #include "xfs_acl.h"
 #include "xfs_rw.h"
+#include "xfs_vnodeops.h"
 
 /*
  * xfs_attr.c
@@ -156,10 +157,14 @@ xfs_attr_fetch(xfs_inode_t *ip, const char *name, int namelen,
 }
 
 int
-xfs_attr_get(bhv_desc_t *bdp, const char *name, char *value, int *valuelenp,
-	     int flags, struct cred *cred)
+xfs_attr_get(
+	xfs_inode_t	*ip,
+	const char	*name,
+	char		*value,
+	int		*valuelenp,
+	int		flags,
+	cred_t		*cred)
 {
-	xfs_inode_t	*ip = XFS_BHVTOI(bdp);
 	int		error, namelen;
 
 	XFS_STATS_INC(xs_attr_get);
@@ -417,10 +422,13 @@ out:
 }
 
 int
-xfs_attr_set(bhv_desc_t *bdp, const char *name, char *value, int valuelen, int flags,
-	     struct cred *cred)
+xfs_attr_set(
+	xfs_inode_t	*dp,
+	const char	*name,
+	char		*value,
+	int		valuelen,
+	int		flags)
 {
-	xfs_inode_t	*dp;
 	int             namelen;
 
 	namelen = strlen(name);
@@ -429,7 +437,6 @@ xfs_attr_set(bhv_desc_t *bdp, const char *name, char *value, int valuelen, int f
 
 	XFS_STATS_INC(xs_attr_set);
 
-	dp = XFS_BHVTOI(bdp);
 	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
 		return (EIO);
 
@@ -563,10 +570,12 @@ out:
 }
 
 int
-xfs_attr_remove(bhv_desc_t *bdp, const char *name, int flags, struct cred *cred)
+xfs_attr_remove(
+	xfs_inode_t	*dp,
+	const char	*name,
+	int		flags)
 {
-	xfs_inode_t         *dp;
-	int                 namelen;
+	int		namelen;
 
 	namelen = strlen(name);
 	if (namelen >= MAXNAMELEN)
@@ -574,7 +583,6 @@ xfs_attr_remove(bhv_desc_t *bdp, const char *name, int flags, struct cred *cred)
 
 	XFS_STATS_INC(xs_attr_remove);
 
-	dp = XFS_BHVTOI(bdp);
 	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
 		return (EIO);
 
@@ -702,11 +710,14 @@ xfs_attr_kern_list_sizes(xfs_attr_list_context_t *context, attrnames_t *namesp,
  * success.
  */
 int
-xfs_attr_list(bhv_desc_t *bdp, char *buffer, int bufsize, int flags,
-		      attrlist_cursor_kern_t *cursor, struct cred *cred)
+xfs_attr_list(
+	xfs_inode_t	*dp,
+	char		*buffer,
+	int		bufsize,
+	int		flags,
+	attrlist_cursor_kern_t *cursor)
 {
 	xfs_attr_list_context_t context;
-	xfs_inode_t *dp;
 	int error;
 
 	XFS_STATS_INC(xs_attr_list);
@@ -731,7 +742,7 @@ xfs_attr_list(bhv_desc_t *bdp, char *buffer, int bufsize, int flags,
 	/*
 	 * Initialize the output buffer.
 	 */
-	context.dp = dp = XFS_BHVTOI(bdp);
+	context.dp = dp;
 	context.cursor = cursor;
 	context.count = 0;
 	context.dupcnt = 0;
@@ -2502,7 +2513,7 @@ STATIC int
 attr_generic_set(
 	bhv_vnode_t *vp, char *name, void *data, size_t size, int xflags)
 {
-	return -bhv_vop_attr_set(vp, name, data, size, xflags, NULL);
+	return -xfs_attr_set(xfs_vtoi(vp), name, data, size, xflags);
 }
 
 STATIC int
@@ -2511,7 +2522,8 @@ attr_generic_get(
 {
 	int	error, asize = size;
 
-	error = bhv_vop_attr_get(vp, name, data, &asize, xflags, NULL);
+	error = xfs_attr_get(xfs_vtoi(vp), name, data,
+				    &asize, xflags, NULL);
 	if (!error)
 		return asize;
 	return -error;
@@ -2521,7 +2533,7 @@ STATIC int
 attr_generic_remove(
 	bhv_vnode_t *vp, char *name, int xflags)
 {
-	return -bhv_vop_attr_remove(vp, name, xflags, NULL);
+	return -xfs_attr_remove(xfs_vtoi(vp), name, xflags);
 }
 
 STATIC int
@@ -2576,7 +2588,7 @@ attr_generic_list(
 	attrlist_cursor_kern_t	cursor = { 0 };
 	int			error;
 
-	error = bhv_vop_attr_list(vp, data, size, xflags, &cursor, NULL);
+	error = xfs_attr_list(xfs_vtoi(vp), data, size, xflags, &cursor);
 	if (error > 0)
 		return -error;
 	*result = -error;

fs/xfs/xfs_attr.h

@@ -36,14 +36,13 @@
  *========================================================================*/
 
 struct cred;
-struct bhv_vnode;
 struct xfs_attr_list_context;
 
-typedef int (*attrset_t)(struct bhv_vnode *, char *, void *, size_t, int);
-typedef int (*attrget_t)(struct bhv_vnode *, char *, void *, size_t, int);
-typedef int (*attrremove_t)(struct bhv_vnode *, char *, int);
-typedef int (*attrexists_t)(struct bhv_vnode *);
-typedef int (*attrcapable_t)(struct bhv_vnode *, struct cred *);
+typedef int (*attrset_t)(bhv_vnode_t *, char *, void *, size_t, int);
+typedef int (*attrget_t)(bhv_vnode_t *, char *, void *, size_t, int);
+typedef int (*attrremove_t)(bhv_vnode_t *, char *, int);
+typedef int (*attrexists_t)(bhv_vnode_t *);
+typedef int (*attrcapable_t)(bhv_vnode_t *, struct cred *);
 
 typedef struct attrnames {
 	char *		attr_name;
@@ -64,7 +63,7 @@ extern struct attrnames attr_trusted;
 extern struct attrnames *attr_namespaces[ATTR_NAMECOUNT];
 
 extern attrnames_t *attr_lookup_namespace(char *, attrnames_t **, int);
-extern int attr_generic_list(struct bhv_vnode *, void *, size_t, int, ssize_t *);
+extern int attr_generic_list(bhv_vnode_t *, void *, size_t, int, ssize_t *);
 
 #define ATTR_DONTFOLLOW	0x0001	/* -- unused, from IRIX -- */
 #define ATTR_ROOT	0x0002	/* use attrs in root (trusted) namespace */
@@ -159,12 +158,8 @@ struct xfs_da_args;
 /*
  * Overall external interface routines.
  */
-int xfs_attr_get(bhv_desc_t *, const char *, char *, int *, int, struct cred *);
-int xfs_attr_set(bhv_desc_t *, const char *, char *, int, int, struct cred *);
 int xfs_attr_set_int(struct xfs_inode *, const char *, int, char *, int, int);
-int xfs_attr_remove(bhv_desc_t *, const char *, int, struct cred *);
 int xfs_attr_remove_int(struct xfs_inode *, const char *, int, int);
-int xfs_attr_list(bhv_desc_t *, char *, int, int, struct attrlist_cursor_kern *, struct cred *);
 int xfs_attr_list_int(struct xfs_attr_list_context *);
 int xfs_attr_inactive(struct xfs_inode *dp);
 

fs/xfs/xfs_behavior.c

@@ -1,183 +0,0 @@
-/*
- * Copyright (c) 2000-2003,2005 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
- */
-#include "xfs.h"
-
-/*
- * Source file used to associate/disassociate behaviors with virtualized
- * objects.  See xfs_behavior.h for more information about behaviors, etc.
- *
- * The implementation is split between functions in this file and macros
- * in xfs_behavior.h.
- */
-
-/*
- * Insert a new behavior descriptor into a behavior chain.
- *
- * The behavior chain is ordered based on the 'position' number which
- * lives in the first field of the ops vector (higher numbers first).
- *
- * Attempts to insert duplicate ops result in an EINVAL return code.
- * Otherwise, return 0 to indicate success.
- */
-int
-bhv_insert(bhv_head_t *bhp, bhv_desc_t *bdp)
-{
-	bhv_desc_t	*curdesc, *prev;
-	int		position;
-
-	/*
-	 * Validate the position value of the new behavior.
-	 */
-	position = BHV_POSITION(bdp);
-	ASSERT(position >= BHV_POSITION_BASE && position <= BHV_POSITION_TOP);
-
-	/*
-	 * Find location to insert behavior.  Check for duplicates.
-	 */
-	prev = NULL;
-	for (curdesc = bhp->bh_first;
-	     curdesc != NULL;
-	     curdesc = curdesc->bd_next) {
-
-		/* Check for duplication. */
-		if (curdesc->bd_ops == bdp->bd_ops) {
-			ASSERT(0);
-			return EINVAL;
-		}
-
-		/* Find correct position */
-		if (position >= BHV_POSITION(curdesc)) {
-			ASSERT(position != BHV_POSITION(curdesc));
-			break;		/* found it */
-		}
-
-		prev = curdesc;
-	}
-
-	if (prev == NULL) {
-		/* insert at front of chain */
-		bdp->bd_next = bhp->bh_first;
-		bhp->bh_first = bdp;
-	} else {
-		/* insert after prev */
-		bdp->bd_next = prev->bd_next;
-		prev->bd_next = bdp;
-	}
-
-	return 0;
-}
-
-/*
- * Remove a behavior descriptor from a position in a behavior chain;
- * the position is guaranteed not to be the first position.
- * Should only be called by the bhv_remove() macro.
- */
-void
-bhv_remove_not_first(bhv_head_t *bhp, bhv_desc_t *bdp)
-{
-	bhv_desc_t	*curdesc, *prev;
-
-	ASSERT(bhp->bh_first != NULL);
-	ASSERT(bhp->bh_first->bd_next != NULL);
-
-	prev = bhp->bh_first;
-	for (curdesc = bhp->bh_first->bd_next;
-	     curdesc != NULL;
-	     curdesc = curdesc->bd_next) {
-
-		if (curdesc == bdp)
-			break;		/* found it */
-		prev = curdesc;
-	}
-
-	ASSERT(curdesc == bdp);
-	prev->bd_next = bdp->bd_next;	/* remove from after prev */
-}
-
-/*
- * Looks for the first behavior within a specified range of positions.
- * Return the associated behavior descriptor.  Or NULL, if none found.
- */
-bhv_desc_t *
-bhv_lookup_range(bhv_head_t *bhp, int low, int high)
-{
-	bhv_desc_t	*curdesc;
-
-	for (curdesc = bhp->bh_first;
-	     curdesc != NULL;
-	     curdesc = curdesc->bd_next) {
-
-		int	position = BHV_POSITION(curdesc);
-
-		if (position <= high) {
-			if (position >= low)
-				return curdesc;
-			return NULL;
-		}
-	}
-
-	return NULL;
-}
-
-/*
- * Return the base behavior in the chain, or NULL if the chain
- * is empty.
- *
- * The caller has not read locked the behavior chain, so acquire the
- * lock before traversing the chain.
- */
-bhv_desc_t *
-bhv_base(bhv_head_t *bhp)
-{
-	bhv_desc_t	*curdesc;
-
-	for (curdesc = bhp->bh_first;
-	     curdesc != NULL;
-	     curdesc = curdesc->bd_next) {
-
-		if (curdesc->bd_next == NULL) {
-			return curdesc;
-		}
-	}
-
-	return NULL;
-}
-
-void
-bhv_head_init(
-	bhv_head_t *bhp,
-	char *name)
-{
-	bhp->bh_first = NULL;
-}
-
-void
-bhv_insert_initial(
-	bhv_head_t *bhp,
-	bhv_desc_t *bdp)
-{
-	ASSERT(bhp->bh_first == NULL);
-	(bhp)->bh_first = bdp;
-}
-
-void
-bhv_head_destroy(
-	bhv_head_t *bhp)
-{
-	ASSERT(bhp->bh_first == NULL);
-}

fs/xfs/xfs_behavior.h

@@ -1,185 +0,0 @@
-/*
- * Copyright (c) 2000-2003,2005 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_BEHAVIOR_H__
-#define __XFS_BEHAVIOR_H__
-
-/*
- * Header file used to associate behaviors with virtualized objects.
- *
- * A virtualized object is an internal, virtualized representation of
- * OS entities such as persistent files, processes, or sockets.  Examples
- * of virtualized objects include vnodes, vprocs, and vsockets.  Often
- * a virtualized object is referred to simply as an "object."
- *
- * A behavior is essentially an implementation layer associated with
- * an object.  Multiple behaviors for an object are chained together,
- * the order of chaining determining the order of invocation.  Each
- * behavior of a given object implements the same set of interfaces
- * (e.g., the VOP interfaces).
- *
- * Behaviors may be dynamically inserted into an object's behavior chain,
- * such that the addition is transparent to consumers that already have
- * references to the object.  Typically, a given behavior will be inserted
- * at a particular location in the behavior chain.  Insertion of new
- * behaviors is synchronized with operations-in-progress (oip's) so that
- * the oip's always see a consistent view of the chain.
- *
- * The term "interposition" is used to refer to the act of inserting
- * a behavior such that it interposes on (i.e., is inserted in front
- * of) a particular other behavior.  A key example of this is when a
- * system implementing distributed single system image wishes to
- * interpose a distribution layer (providing distributed coherency)
- * in front of an object that is otherwise only accessed locally.
- *
- * Note that the traditional vnode/inode combination is simply a virtualized
- * object that has exactly one associated behavior.
- *
- * Behavior synchronization is logic which is necessary under certain
- * circumstances that there is no conflict between ongoing operations
- * traversing the behavior chain and those dynamically modifying the
- * behavior chain.  Because behavior synchronization adds extra overhead
- * to virtual operation invocation, we want to restrict, as much as
- * we can, the requirement for this extra code, to those situations
- * in which it is truly necessary.
- *
- * Behavior synchronization is needed whenever there's at least one class
- * of object in the system for which:
- * 1) multiple behaviors for a given object are supported,
- * -- AND --
- * 2a) insertion of a new behavior can happen dynamically at any time during
- *     the life of an active object,
- *	-- AND --
- *	3a) insertion of a new behavior needs to synchronize with existing
- *	    ops-in-progress.
- *	-- OR --
- *	3b) multiple different behaviors can be dynamically inserted at
- *	    any time during the life of an active object
- *	-- OR --
- *	3c) removal of a behavior can occur at any time during the life of
- *	    an active object.
- * -- OR --
- * 2b) removal of a behavior can occur at any time during the life of an
- *     active object
- *
- */
-
-/*
- * Behavior head.  Head of the chain of behaviors.
- * Contained within each virtualized object data structure.
- */
-typedef struct bhv_head {
-	struct bhv_desc *bh_first;	/* first behavior in chain */
-} bhv_head_t;
-
-/*
- * Behavior descriptor.	 Descriptor associated with each behavior.
- * Contained within the behavior's private data structure.
- */
-typedef struct bhv_desc {
-	void		*bd_pdata;	/* private data for this behavior */
-	void		*bd_vobj;	/* virtual object associated with */
-	void		*bd_ops;	/* ops for this behavior */
-	struct bhv_desc *bd_next;	/* next behavior in chain */
-} bhv_desc_t;
-
-/*
- * Behavior identity field.  A behavior's identity determines the position
- * where it lives within a behavior chain, and it's always the first field
- * of the behavior's ops vector. The optional id field further identifies the
- * subsystem responsible for the behavior.
- */
-typedef struct bhv_identity {
-	__u16	bi_id;		/* owning subsystem id */
-	__u16	bi_position;	/* position in chain */
-} bhv_identity_t;
-
-typedef bhv_identity_t bhv_position_t;
-
-#define BHV_IDENTITY_INIT(id,pos)	{id, pos}
-#define BHV_IDENTITY_INIT_POSITION(pos) BHV_IDENTITY_INIT(0, pos)
-
-/*
- * Define boundaries of position values.
- */
-#define BHV_POSITION_INVALID	0	/* invalid position number */
-#define BHV_POSITION_BASE	1	/* base (last) implementation layer */
-#define BHV_POSITION_TOP	63	/* top (first) implementation layer */
-
-/*
- * Plumbing macros.
- */
-#define BHV_HEAD_FIRST(bhp)	(ASSERT((bhp)->bh_first), (bhp)->bh_first)
-#define BHV_NEXT(bdp)		(ASSERT((bdp)->bd_next), (bdp)->bd_next)
-#define BHV_NEXTNULL(bdp)	((bdp)->bd_next)
-#define BHV_VOBJ(bdp)		(ASSERT((bdp)->bd_vobj), (bdp)->bd_vobj)
-#define BHV_VOBJNULL(bdp)	((bdp)->bd_vobj)
-#define BHV_PDATA(bdp)		(bdp)->bd_pdata
-#define BHV_OPS(bdp)		(bdp)->bd_ops
-#define BHV_IDENTITY(bdp)	((bhv_identity_t *)(bdp)->bd_ops)
-#define BHV_POSITION(bdp)	(BHV_IDENTITY(bdp)->bi_position)
-
-extern void bhv_head_init(bhv_head_t *, char *);
-extern void bhv_head_destroy(bhv_head_t *);
-extern int  bhv_insert(bhv_head_t *, bhv_desc_t *);
-extern void bhv_insert_initial(bhv_head_t *, bhv_desc_t *);
-
-/*
- * Initialize a new behavior descriptor.
- * Arguments:
- *   bdp - pointer to behavior descriptor
- *   pdata - pointer to behavior's private data
- *   vobj - pointer to associated virtual object
- *   ops - pointer to ops for this behavior
- */
-#define bhv_desc_init(bdp, pdata, vobj, ops)		\
- {							\
-	(bdp)->bd_pdata = pdata;			\
-	(bdp)->bd_vobj = vobj;				\
-	(bdp)->bd_ops = ops;				\
-	(bdp)->bd_next = NULL;				\
- }
-
-/*
- * Remove a behavior descriptor from a behavior chain.
- */
-#define bhv_remove(bhp, bdp)				\
- {							\
-	if ((bhp)->bh_first == (bdp)) {			\
-		/*					\
-		* Remove from front of chain.		\
-		* Atomic wrt oip's.			\
-		*/					\
-	       (bhp)->bh_first = (bdp)->bd_next;	\
-	} else {					\
-	       /* remove from non-front of chain */	\
-	       bhv_remove_not_first(bhp, bdp);		\
-	}						\
-	(bdp)->bd_vobj = NULL;				\
- }
-
-/*
- * Behavior module prototypes.
- */
-extern void		bhv_remove_not_first(bhv_head_t *bhp, bhv_desc_t *bdp);
-extern bhv_desc_t *	bhv_lookup_range(bhv_head_t *bhp, int low, int high);
-extern bhv_desc_t *	bhv_base(bhv_head_t *bhp);
-
-/* No bhv locking on Linux */
-#define bhv_base_unlocked	bhv_base
-
-#endif /* __XFS_BEHAVIOR_H__ */

fs/xfs/xfs_bmap.c

@@ -53,6 +53,7 @@
 #include "xfs_trans_space.h"
 #include "xfs_buf_item.h"
 #include "xfs_filestream.h"
+#include "xfs_vnodeops.h"
 
 
 #ifdef DEBUG
@@ -248,7 +249,7 @@ xfs_bmap_local_to_extents(
  * Else, *lastxp will be set to the index of the found
  * entry; *gotp will contain the entry.
  */
-STATIC xfs_bmbt_rec_t *			/* pointer to found extent entry */
+STATIC xfs_bmbt_rec_host_t *		/* pointer to found extent entry */
 xfs_bmap_search_extents(
 	xfs_inode_t	*ip,		/* incore inode pointer */
 	xfs_fileoff_t	bno,		/* block number searched for */
@@ -272,21 +273,6 @@ xfs_bmap_isaeof(
 	char		*aeof);		/* return value */
 
 #ifdef XFS_BMAP_TRACE
-/*
- * Add a bmap trace buffer entry.  Base routine for the others.
- */
-STATIC void
-xfs_bmap_trace_addentry(
-	int		opcode,		/* operation */
-	const char	*fname,		/* function name */
-	char		*desc,		/* operation description */
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_extnum_t	idx,		/* index of entry(ies) */
-	xfs_extnum_t	cnt,		/* count of entries, 1 or 2 */
-	xfs_bmbt_rec_t	*r1,		/* first record */
-	xfs_bmbt_rec_t	*r2,		/* second record or null */
-	int		whichfork);	/* data or attr fork */
-
 /*
  * Add bmap trace entry prior to a call to xfs_iext_remove.
  */
@@ -714,7 +700,7 @@ xfs_bmap_add_extent_delay_real(
 {
 	xfs_btree_cur_t		*cur;	/* btree cursor */
 	int			diff;	/* temp value */
-	xfs_bmbt_rec_t		*ep;	/* extent entry for idx */
+	xfs_bmbt_rec_host_t	*ep;	/* extent entry for idx */
 	int			error;	/* error return value */
 	int			i;	/* temp state */
 	xfs_ifork_t		*ifp;	/* inode fork pointer */
@@ -1270,7 +1256,7 @@ xfs_bmap_add_extent_unwritten_real(
 	xfs_extdelta_t		*delta) /* Change made to incore extents */
 {
 	xfs_btree_cur_t		*cur;	/* btree cursor */
-	xfs_bmbt_rec_t		*ep;	/* extent entry for idx */
+	xfs_bmbt_rec_host_t	*ep;	/* extent entry for idx */
 	int			error;	/* error return value */
 	int			i;	/* temp state */
 	xfs_ifork_t		*ifp;	/* inode fork pointer */
@@ -1823,7 +1809,7 @@ xfs_bmap_add_extent_hole_delay(
 	xfs_extdelta_t		*delta, /* Change made to incore extents */
 	int			rsvd)		/* OK to allocate reserved blocks */
 {
-	xfs_bmbt_rec_t		*ep;	/* extent record for idx */
+	xfs_bmbt_rec_host_t	*ep;	/* extent record for idx */
 	xfs_ifork_t		*ifp;	/* inode fork pointer */
 	xfs_bmbt_irec_t		left;	/* left neighbor extent entry */
 	xfs_filblks_t		newlen=0;	/* new indirect size */
@@ -2012,7 +1998,7 @@ xfs_bmap_add_extent_hole_real(
 	xfs_extdelta_t		*delta, /* Change made to incore extents */
 	int			whichfork) /* data or attr fork */
 {
-	xfs_bmbt_rec_t		*ep;	/* pointer to extent entry ins. point */
+	xfs_bmbt_rec_host_t	*ep;	/* pointer to extent entry ins. point */
 	int			error;	/* error return value */
 	int			i;	/* temp state */
 	xfs_ifork_t		*ifp;	/* inode fork pointer */
@@ -3070,7 +3056,7 @@ xfs_bmap_del_extent(
 	xfs_fileoff_t		del_endoff;	/* first offset past del */
 	int			delay;	/* current block is delayed allocated */
 	int			do_fx;	/* free extent at end of routine */
-	xfs_bmbt_rec_t		*ep;	/* current extent entry pointer */
+	xfs_bmbt_rec_host_t	*ep;	/* current extent entry pointer */
 	int			error;	/* error return value */
 	int			flags;	/* inode logging flags */
 	xfs_bmbt_irec_t		got;	/* current extent entry */
@@ -3418,7 +3404,7 @@ xfs_bmap_extents_to_btree(
 	xfs_bmbt_rec_t		*arp;		/* child record pointer */
 	xfs_bmbt_block_t	*block;		/* btree root block */
 	xfs_btree_cur_t		*cur;		/* bmap btree cursor */
-	xfs_bmbt_rec_t		*ep;		/* extent record pointer */
+	xfs_bmbt_rec_host_t	*ep;		/* extent record pointer */
 	int			error;		/* error return value */
 	xfs_extnum_t		i, cnt;		/* extent record index */
 	xfs_ifork_t		*ifp;		/* inode fork pointer */
@@ -3507,8 +3493,8 @@ xfs_bmap_extents_to_btree(
 	for (cnt = i = 0; i < nextents; i++) {
 		ep = xfs_iext_get_ext(ifp, i);
 		if (!ISNULLSTARTBLOCK(xfs_bmbt_get_startblock(ep))) {
-			arp->l0 = INT_GET(ep->l0, ARCH_CONVERT);
-			arp->l1 = INT_GET(ep->l1, ARCH_CONVERT);
+			arp->l0 = cpu_to_be64(ep->l0);
+			arp->l1 = cpu_to_be64(ep->l1);
 			arp++; cnt++;
 		}
 	}
@@ -3590,7 +3576,7 @@ xfs_bmap_local_to_extents(
 	if (ifp->if_bytes) {
 		xfs_alloc_arg_t	args;	/* allocation arguments */
 		xfs_buf_t	*bp;	/* buffer for extent block */
-		xfs_bmbt_rec_t	*ep;	/* extent record pointer */
+		xfs_bmbt_rec_host_t *ep;/* extent record pointer */
 
 		args.tp = tp;
 		args.mp = ip->i_mount;
@@ -3655,7 +3641,7 @@ done:
  * entry (null if none).  Else, *lastxp will be set to the index
  * of the found entry; *gotp will contain the entry.
  */
-xfs_bmbt_rec_t *			/* pointer to found extent entry */
+xfs_bmbt_rec_host_t *			/* pointer to found extent entry */
 xfs_bmap_search_multi_extents(
 	xfs_ifork_t	*ifp,		/* inode fork pointer */
 	xfs_fileoff_t	bno,		/* block number searched for */
@@ -3664,7 +3650,7 @@ xfs_bmap_search_multi_extents(
 	xfs_bmbt_irec_t	*gotp,		/* out: extent entry found */
 	xfs_bmbt_irec_t	*prevp)		/* out: previous extent entry found */
 {
-	xfs_bmbt_rec_t	*ep;		/* extent record pointer */
+	xfs_bmbt_rec_host_t *ep;		/* extent record pointer */
 	xfs_extnum_t	lastx;		/* last extent index */
 
 	/*
@@ -3706,7 +3692,7 @@ xfs_bmap_search_multi_extents(
  * Else, *lastxp will be set to the index of the found
  * entry; *gotp will contain the entry.
  */
-STATIC xfs_bmbt_rec_t *                 /* pointer to found extent entry */
+STATIC xfs_bmbt_rec_host_t *                 /* pointer to found extent entry */
 xfs_bmap_search_extents(
 	xfs_inode_t     *ip,            /* incore inode pointer */
 	xfs_fileoff_t   bno,            /* block number searched for */
@@ -3717,7 +3703,7 @@ xfs_bmap_search_extents(
 	xfs_bmbt_irec_t *prevp)         /* out: previous extent entry found */
 {
 	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_bmbt_rec_t  *ep;            /* extent record pointer */
+	xfs_bmbt_rec_host_t  *ep;            /* extent record pointer */
 
 	XFS_STATS_INC(xs_look_exlist);
 	ifp = XFS_IFORK_PTR(ip, fork);
@@ -3757,11 +3743,11 @@ xfs_bmap_trace_addentry(
 	xfs_inode_t	*ip,		/* incore inode pointer */
 	xfs_extnum_t	idx,		/* index of entry(ies) */
 	xfs_extnum_t	cnt,		/* count of entries, 1 or 2 */
-	xfs_bmbt_rec_t	*r1,		/* first record */
-	xfs_bmbt_rec_t	*r2,		/* second record or null */
+	xfs_bmbt_rec_host_t *r1,	/* first record */
+	xfs_bmbt_rec_host_t *r2,	/* second record or null */
 	int		whichfork)	/* data or attr fork */
 {
-	xfs_bmbt_rec_t	tr2;
+	xfs_bmbt_rec_host_t tr2;
 
 	ASSERT(cnt == 1 || cnt == 2);
 	ASSERT(r1 != NULL);
@@ -3842,8 +3828,8 @@ xfs_bmap_trace_insert(
 	xfs_bmbt_irec_t	*r2,		/* inserted record 2 or null */
 	int		whichfork)	/* data or attr fork */
 {
-	xfs_bmbt_rec_t	tr1;		/* compressed record 1 */
-	xfs_bmbt_rec_t	tr2;		/* compressed record 2 if needed */
+	xfs_bmbt_rec_host_t tr1;	/* compressed record 1 */
+	xfs_bmbt_rec_host_t tr2;	/* compressed record 2 if needed */
 
 	xfs_bmbt_set_all(&tr1, r1);
 	if (cnt == 2) {
@@ -4316,7 +4302,6 @@ xfs_bmap_first_unused(
 	xfs_fileoff_t	*first_unused,		/* unused block */
 	int		whichfork)		/* data or attr fork */
 {
-	xfs_bmbt_rec_t	*ep;			/* pointer to an extent entry */
 	int		error;			/* error return value */
 	int		idx;			/* extent record index */
 	xfs_ifork_t	*ifp;			/* inode fork pointer */
@@ -4340,7 +4325,7 @@ xfs_bmap_first_unused(
 	lowest = *first_unused;
 	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
 	for (idx = 0, lastaddr = 0, max = lowest; idx < nextents; idx++) {
-		ep = xfs_iext_get_ext(ifp, idx);
+		xfs_bmbt_rec_host_t *ep = xfs_iext_get_ext(ifp, idx);
 		off = xfs_bmbt_get_startoff(ep);
 		/*
 		 * See if the hole before this extent will work.
@@ -4371,7 +4356,7 @@ xfs_bmap_last_before(
 {
 	xfs_fileoff_t	bno;			/* input file offset */
 	int		eof;			/* hit end of file */
-	xfs_bmbt_rec_t	*ep;			/* pointer to last extent */
+	xfs_bmbt_rec_host_t *ep;		/* pointer to last extent */
 	int		error;			/* error return value */
 	xfs_bmbt_irec_t	got;			/* current extent value */
 	xfs_ifork_t	*ifp;			/* inode fork pointer */
@@ -4417,7 +4402,7 @@ xfs_bmap_last_offset(
 	xfs_fileoff_t	*last_block,		/* last block */
 	int		whichfork)		/* data or attr fork */
 {
-	xfs_bmbt_rec_t	*ep;			/* pointer to last extent */
+	xfs_bmbt_rec_host_t *ep;		/* pointer to last extent */
 	int		error;			/* error return value */
 	xfs_ifork_t	*ifp;			/* inode fork pointer */
 	xfs_extnum_t	nextents;		/* number of extent entries */
@@ -4454,7 +4439,7 @@ xfs_bmap_one_block(
 	xfs_inode_t	*ip,		/* incore inode */
 	int		whichfork)	/* data or attr fork */
 {
-	xfs_bmbt_rec_t	*ep;		/* ptr to fork's extent */
+	xfs_bmbt_rec_host_t *ep;	/* ptr to fork's extent */
 	xfs_ifork_t	*ifp;		/* inode fork pointer */
 	int		rval;		/* return value */
 	xfs_bmbt_irec_t	s;		/* internal version of extent */
@@ -4549,7 +4534,7 @@ xfs_bmap_read_extents(
 	 * Loop over all leaf nodes.  Copy information to the extent records.
 	 */
 	for (;;) {
-		xfs_bmbt_rec_t	*frp, *trp;
+		xfs_bmbt_rec_t	*frp;
 		xfs_fsblock_t	nextbno;
 		xfs_extnum_t	num_recs;
 		xfs_extnum_t	start;
@@ -4581,9 +4566,9 @@ xfs_bmap_read_extents(
 		frp = XFS_BTREE_REC_ADDR(xfs_bmbt, block, 1);
 		start = i;
 		for (j = 0; j < num_recs; j++, i++, frp++) {
-			trp = xfs_iext_get_ext(ifp, i);
-			trp->l0 = INT_GET(frp->l0, ARCH_CONVERT);
-			trp->l1 = INT_GET(frp->l1, ARCH_CONVERT);
+			xfs_bmbt_rec_host_t *trp = xfs_iext_get_ext(ifp, i);
+			trp->l0 = be64_to_cpu(frp->l0);
+			trp->l1 = be64_to_cpu(frp->l1);
 		}
 		if (exntf == XFS_EXTFMT_NOSTATE) {
 			/*
@@ -4631,7 +4616,7 @@ xfs_bmap_trace_exlist(
 	xfs_extnum_t	cnt,		/* count of entries in the list */
 	int		whichfork)	/* data or attr fork */
 {
-	xfs_bmbt_rec_t	*ep;		/* current extent record */
+	xfs_bmbt_rec_host_t *ep;	/* current extent record */
 	xfs_extnum_t	idx;		/* extent record index */
 	xfs_ifork_t	*ifp;		/* inode fork pointer */
 	xfs_bmbt_irec_t	s;		/* file extent record */
@@ -4727,7 +4712,7 @@ xfs_bmapi(
 	xfs_btree_cur_t	*cur;		/* bmap btree cursor */
 	xfs_fileoff_t	end;		/* end of mapped file region */
 	int		eof;		/* we've hit the end of extents */
-	xfs_bmbt_rec_t	*ep;		/* extent record pointer */
+	xfs_bmbt_rec_host_t *ep;	/* extent record pointer */
 	int		error;		/* error return */
 	xfs_bmbt_irec_t	got;		/* current file extent record */
 	xfs_ifork_t	*ifp;		/* inode fork pointer */
@@ -5378,7 +5363,7 @@ xfs_bunmapi(
 	xfs_btree_cur_t		*cur;		/* bmap btree cursor */
 	xfs_bmbt_irec_t		del;		/* extent being deleted */
 	int			eof;		/* is deleting at eof */
-	xfs_bmbt_rec_t		*ep;		/* extent record pointer */
+	xfs_bmbt_rec_host_t	*ep;		/* extent record pointer */
 	int			error;		/* error return value */
 	xfs_extnum_t		extno;		/* extent number in list */
 	xfs_bmbt_irec_t		got;		/* current extent record */
@@ -5742,12 +5727,45 @@ error0:
 	return error;
 }
 
+/*
+ * returns 1 for success, 0 if we failed to map the extent.
+ */
+STATIC int
+xfs_getbmapx_fix_eof_hole(
+	xfs_inode_t		*ip,		/* xfs incore inode pointer */
+	struct getbmap		*out,		/* output structure */
+	int			prealloced,	/* this is a file with
+						* preallocated data space */
+	__int64_t		end,		/* last block requested */
+	xfs_fsblock_t		startblock)
+{
+	__int64_t		fixlen;
+	xfs_mount_t		*mp;		/* file system mount point */
+
+	if (startblock == HOLESTARTBLOCK) {
+		mp = ip->i_mount;
+		out->bmv_block = -1;
+		fixlen = XFS_FSB_TO_BB(mp, XFS_B_TO_FSB(mp, ip->i_size));
+		fixlen -= out->bmv_offset;
+		if (prealloced && out->bmv_offset + out->bmv_length == end) {
+			/* Came to hole at EOF. Trim it. */
+			if (fixlen <= 0)
+				return 0;
+			out->bmv_length = fixlen;
+		}
+	} else {
+		out->bmv_block = XFS_FSB_TO_DB(ip, startblock);
+	}
+
+	return 1;
+}
+
 /*
  * Fcntl interface to xfs_bmapi.
  */
 int						/* error code */
 xfs_getbmap(
-	bhv_desc_t		*bdp,		/* XFS behavior descriptor*/
+	xfs_inode_t		*ip,
 	struct getbmap		*bmv,		/* user bmap structure */
 	void			__user *ap,	/* pointer to user's array */
 	int			interface)	/* interface flags */
@@ -5756,7 +5774,6 @@ xfs_getbmap(
 	int			error;		/* return value */
 	__int64_t		fixlen;		/* length for -1 case */
 	int			i;		/* extent number */
-	xfs_inode_t		*ip;		/* xfs incore inode pointer */
 	bhv_vnode_t		*vp;		/* corresponding vnode */
 	int			lock;		/* lock state */
 	xfs_bmbt_irec_t		*map;		/* buffer for user's data */
@@ -5774,8 +5791,7 @@ xfs_getbmap(
 	int			bmapi_flags;	/* flags for xfs_bmapi */
 	__int32_t		oflags;		/* getbmapx bmv_oflags field */
 
-	vp = BHV_TO_VNODE(bdp);
-	ip = XFS_BHVTOI(bdp);
+	vp = XFS_ITOV(ip);
 	mp = ip->i_mount;
 
 	whichfork = interface & BMV_IF_ATTRFORK ? XFS_ATTR_FORK : XFS_DATA_FORK;
@@ -5794,10 +5810,9 @@ xfs_getbmap(
 	 *	could misinterpret holes in a DMAPI file as true holes,
 	 *	when in fact they may represent offline user data.
 	 */
-	if (   (interface & BMV_IF_NO_DMAPI_READ) == 0
-	    && DM_EVENT_ENABLED(vp->v_vfsp, ip, DM_EVENT_READ)
-	    && whichfork == XFS_DATA_FORK) {
-
+	if ((interface & BMV_IF_NO_DMAPI_READ) == 0 &&
+	    DM_EVENT_ENABLED(ip, DM_EVENT_READ) &&
+	    whichfork == XFS_DATA_FORK) {
 		error = XFS_SEND_DATA(mp, DM_EVENT_READ, vp, 0, 0, 0, NULL);
 		if (error)
 			return XFS_ERROR(error);
@@ -5854,7 +5869,8 @@ xfs_getbmap(
 	if (whichfork == XFS_DATA_FORK &&
 		(ip->i_delayed_blks || ip->i_size > ip->i_d.di_size)) {
 		/* xfs_fsize_t last_byte = xfs_file_last_byte(ip); */
-		error = bhv_vop_flush_pages(vp, (xfs_off_t)0, -1, 0, FI_REMAPF);
+		error = xfs_flush_pages(ip, (xfs_off_t)0,
+					       -1, 0, FI_REMAPF);
 	}
 
 	ASSERT(whichfork == XFS_ATTR_FORK || ip->i_delayed_blks == 0);
@@ -5904,18 +5920,15 @@ xfs_getbmap(
 			out.bmv_length = XFS_FSB_TO_BB(mp, map[i].br_blockcount);
 			ASSERT(map[i].br_startblock != DELAYSTARTBLOCK);
                         if (map[i].br_startblock == HOLESTARTBLOCK &&
-                           ((prealloced && out.bmv_offset + out.bmv_length == bmvend) ||
-                             whichfork == XFS_ATTR_FORK )) {
-                                /*
-                                 * came to hole at end of file or the end of
-                                   attribute fork
-                                 */
+			    whichfork == XFS_ATTR_FORK) {
+				/* came to the end of attribute fork */
 				goto unlock_and_return;
 			} else {
-				out.bmv_block =
-				    (map[i].br_startblock == HOLESTARTBLOCK) ?
-					-1 :
-					XFS_FSB_TO_DB(ip, map[i].br_startblock);
+				if (!xfs_getbmapx_fix_eof_hole(ip, &out,
+							prealloced, bmvend,
+							map[i].br_startblock)) {
+					goto unlock_and_return;
+				}
 
 				/* return either getbmap/getbmapx structure. */
 				if (interface & BMV_IF_EXTENDED) {
@@ -5974,7 +5987,7 @@ xfs_bmap_isaeof(
 {
 	int		error;		/* error return value */
 	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_bmbt_rec_t	*lastrec;	/* extent record pointer */
+	xfs_bmbt_rec_host_t *lastrec;	/* extent record pointer */
 	xfs_extnum_t	nextents;	/* number of file extents */
 	xfs_bmbt_irec_t	s;		/* expanded extent record */
 
@@ -6018,7 +6031,7 @@ xfs_bmap_eof(
 	xfs_fsblock_t	blockcount;	/* extent block count */
 	int		error;		/* error return value */
 	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_bmbt_rec_t	*lastrec;	/* extent record pointer */
+	xfs_bmbt_rec_host_t *lastrec;	/* extent record pointer */
 	xfs_extnum_t	nextents;	/* number of file extents */
 	xfs_fileoff_t	startoff;	/* extent starting file offset */
 
@@ -6465,10 +6478,9 @@ xfs_bmap_count_leaves(
 	int			*count)
 {
 	int		b;
-	xfs_bmbt_rec_t	*frp;
 
 	for (b = 0; b < numrecs; b++) {
-		frp = xfs_iext_get_ext(ifp, idx + b);
+		xfs_bmbt_rec_host_t *frp = xfs_iext_get_ext(ifp, idx + b);
 		*count += xfs_bmbt_get_blockcount(frp);
 	}
 	return 0;

fs/xfs/xfs_bmap.h

@@ -335,7 +335,7 @@ xfs_bunmapi(
  */
 int						/* error code */
 xfs_getbmap(
-	bhv_desc_t		*bdp,		/* XFS behavior descriptor*/
+	xfs_inode_t		*ip,
 	struct getbmap		*bmv,		/* user bmap structure */
 	void			__user *ap,	/* pointer to user's array */
 	int			iflags);	/* interface flags */
@@ -378,7 +378,7 @@ xfs_check_nostate_extents(
  * entry (null if none).  Else, *lastxp will be set to the index
  * of the found entry; *gotp will contain the entry.
  */
-xfs_bmbt_rec_t *
+xfs_bmbt_rec_host_t *
 xfs_bmap_search_multi_extents(struct xfs_ifork *, xfs_fileoff_t, int *,
 			xfs_extnum_t *, xfs_bmbt_irec_t *, xfs_bmbt_irec_t *);
 

fs/xfs/xfs_bmap_btree.c

@@ -260,13 +260,14 @@ xfs_bmbt_trace_cursor(
 	char		*s,
 	int		line)
 {
-	xfs_bmbt_rec_t	r;
+	xfs_bmbt_rec_host_t	r;
 
 	xfs_bmbt_set_all(&r, &cur->bc_rec.b);
 	xfs_bmbt_trace_enter(func, cur, s, XFS_BMBT_KTRACE_CUR, line,
 		(cur->bc_nlevels << 24) | (cur->bc_private.b.flags << 16) |
 		cur->bc_private.b.allocated,
-		INT_GET(r.l0, ARCH_CONVERT) >> 32, (int)INT_GET(r.l0, ARCH_CONVERT), INT_GET(r.l1, ARCH_CONVERT) >> 32, (int)INT_GET(r.l1, ARCH_CONVERT),
+		r.l0 >> 32, (int)r.l0,
+		r.l1 >> 32, (int)r.l1,
 		(unsigned long)cur->bc_bufs[0], (unsigned long)cur->bc_bufs[1],
 		(unsigned long)cur->bc_bufs[2], (unsigned long)cur->bc_bufs[3],
 		(cur->bc_ptrs[0] << 16) | cur->bc_ptrs[1],
@@ -383,7 +384,7 @@ xfs_bmbt_delrec(
 		if (ptr < numrecs) {
 			memmove(&kp[ptr - 1], &kp[ptr],
 				(numrecs - ptr) * sizeof(*kp));
-			memmove(&pp[ptr - 1], &pp[ptr], /* INT_: direct copy */
+			memmove(&pp[ptr - 1], &pp[ptr],
 				(numrecs - ptr) * sizeof(*pp));
 			xfs_bmbt_log_ptrs(cur, bp, ptr, numrecs - 1);
 			xfs_bmbt_log_keys(cur, bp, ptr, numrecs - 1);
@@ -815,7 +816,7 @@ xfs_bmbt_insrec(
 #endif
 		memmove(&kp[ptr], &kp[ptr - 1],
 			(numrecs - ptr + 1) * sizeof(*kp));
-		memmove(&pp[ptr], &pp[ptr - 1], /* INT_: direct copy */
+		memmove(&pp[ptr], &pp[ptr - 1],
 			(numrecs - ptr + 1) * sizeof(*pp));
 #ifdef DEBUG
 		if ((error = xfs_btree_check_lptr(cur, *bnop, level))) {
@@ -1250,7 +1251,7 @@ xfs_bmbt_lshift(
 			return error;
 		}
 #endif
-		*lpp = *rpp; /* INT_: direct copy */
+		*lpp = *rpp;
 		xfs_bmbt_log_ptrs(cur, lbp, lrecs, lrecs);
 	} else {
 		lrp = XFS_BMAP_REC_IADDR(left, lrecs, cur);
@@ -1388,7 +1389,7 @@ xfs_bmbt_rshift(
 		}
 #endif
 		*rkp = *lkp;
-		*rpp = *lpp; /* INT_: direct copy */
+		*rpp = *lpp;
 		xfs_bmbt_log_keys(cur, rbp, 1, be16_to_cpu(right->bb_numrecs) + 1);
 		xfs_bmbt_log_ptrs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs) + 1);
 	} else {
@@ -1826,7 +1827,7 @@ __xfs_bmbt_get_all(
 
 void
 xfs_bmbt_get_all(
-	xfs_bmbt_rec_t	*r,
+	xfs_bmbt_rec_host_t *r,
 	xfs_bmbt_irec_t *s)
 {
 	__xfs_bmbt_get_all(r->l0, r->l1, s);
@@ -1862,7 +1863,7 @@ xfs_bmbt_get_block(
  */
 xfs_filblks_t
 xfs_bmbt_get_blockcount(
-	xfs_bmbt_rec_t	*r)
+	xfs_bmbt_rec_host_t	*r)
 {
 	return (xfs_filblks_t)(r->l1 & XFS_MASK64LO(21));
 }
@@ -1872,7 +1873,7 @@ xfs_bmbt_get_blockcount(
  */
 xfs_fsblock_t
 xfs_bmbt_get_startblock(
-	xfs_bmbt_rec_t	*r)
+	xfs_bmbt_rec_host_t	*r)
 {
 #if XFS_BIG_BLKNOS
 	return (((xfs_fsblock_t)r->l0 & XFS_MASK64LO(9)) << 43) |
@@ -1896,7 +1897,7 @@ xfs_bmbt_get_startblock(
  */
 xfs_fileoff_t
 xfs_bmbt_get_startoff(
-	xfs_bmbt_rec_t	*r)
+	xfs_bmbt_rec_host_t	*r)
 {
 	return ((xfs_fileoff_t)r->l0 &
 		 XFS_MASK64LO(64 - BMBT_EXNTFLAG_BITLEN)) >> 9;
@@ -1904,7 +1905,7 @@ xfs_bmbt_get_startoff(
 
 xfs_exntst_t
 xfs_bmbt_get_state(
-	xfs_bmbt_rec_t	*r)
+	xfs_bmbt_rec_host_t	*r)
 {
 	int	ext_flag;
 
@@ -1913,19 +1914,13 @@ xfs_bmbt_get_state(
 				ext_flag);
 }
 
-#ifndef XFS_NATIVE_HOST
 /* Endian flipping versions of the bmbt extraction functions */
 void
 xfs_bmbt_disk_get_all(
 	xfs_bmbt_rec_t	*r,
 	xfs_bmbt_irec_t *s)
 {
-	__uint64_t	l0, l1;
-
-	l0 = INT_GET(r->l0, ARCH_CONVERT);
-	l1 = INT_GET(r->l1, ARCH_CONVERT);
-
-	__xfs_bmbt_get_all(l0, l1, s);
+	__xfs_bmbt_get_all(be64_to_cpu(r->l0), be64_to_cpu(r->l1), s);
 }
 
 /*
@@ -1935,7 +1930,7 @@ xfs_filblks_t
 xfs_bmbt_disk_get_blockcount(
 	xfs_bmbt_rec_t	*r)
 {
-	return (xfs_filblks_t)(INT_GET(r->l1, ARCH_CONVERT) & XFS_MASK64LO(21));
+	return (xfs_filblks_t)(be64_to_cpu(r->l1) & XFS_MASK64LO(21));
 }
 
 /*
@@ -1945,11 +1940,9 @@ xfs_fileoff_t
 xfs_bmbt_disk_get_startoff(
 	xfs_bmbt_rec_t	*r)
 {
-	return ((xfs_fileoff_t)INT_GET(r->l0, ARCH_CONVERT) &
+	return ((xfs_fileoff_t)be64_to_cpu(r->l0) &
 		 XFS_MASK64LO(64 - BMBT_EXNTFLAG_BITLEN)) >> 9;
 }
-#endif /* XFS_NATIVE_HOST */
-
 
 /*
  * Increment cursor by one record at the level.
@@ -2289,186 +2282,132 @@ xfs_bmbt_newroot(
 	return 0;
 }
 
-/*
- * Set all the fields in a bmap extent record from the uncompressed form.
- */
-void
-xfs_bmbt_set_all(
-	xfs_bmbt_rec_t	*r,
-	xfs_bmbt_irec_t	*s)
-{
-	int	extent_flag;
-
-	ASSERT((s->br_state == XFS_EXT_NORM) ||
-		(s->br_state == XFS_EXT_UNWRITTEN));
-	extent_flag = (s->br_state == XFS_EXT_NORM) ? 0 : 1;
-	ASSERT((s->br_startoff & XFS_MASK64HI(9)) == 0);
-	ASSERT((s->br_blockcount & XFS_MASK64HI(43)) == 0);
-#if XFS_BIG_BLKNOS
-	ASSERT((s->br_startblock & XFS_MASK64HI(12)) == 0);
-	r->l0 = ((xfs_bmbt_rec_base_t)extent_flag << 63) |
-		 ((xfs_bmbt_rec_base_t)s->br_startoff << 9) |
-		 ((xfs_bmbt_rec_base_t)s->br_startblock >> 43);
-	r->l1 = ((xfs_bmbt_rec_base_t)s->br_startblock << 21) |
-		 ((xfs_bmbt_rec_base_t)s->br_blockcount &
-		 (xfs_bmbt_rec_base_t)XFS_MASK64LO(21));
-#else	/* !XFS_BIG_BLKNOS */
-	if (ISNULLSTARTBLOCK(s->br_startblock)) {
-		r->l0 = ((xfs_bmbt_rec_base_t)extent_flag << 63) |
-			((xfs_bmbt_rec_base_t)s->br_startoff << 9) |
-			  (xfs_bmbt_rec_base_t)XFS_MASK64LO(9);
-		r->l1 = XFS_MASK64HI(11) |
-			  ((xfs_bmbt_rec_base_t)s->br_startblock << 21) |
-			  ((xfs_bmbt_rec_base_t)s->br_blockcount &
-			   (xfs_bmbt_rec_base_t)XFS_MASK64LO(21));
-	} else {
-		r->l0 = ((xfs_bmbt_rec_base_t)extent_flag << 63) |
-			((xfs_bmbt_rec_base_t)s->br_startoff << 9);
-		r->l1 = ((xfs_bmbt_rec_base_t)s->br_startblock << 21) |
-			  ((xfs_bmbt_rec_base_t)s->br_blockcount &
-			   (xfs_bmbt_rec_base_t)XFS_MASK64LO(21));
-	}
-#endif	/* XFS_BIG_BLKNOS */
-}
-
 /*
  * Set all the fields in a bmap extent record from the arguments.
  */
 void
 xfs_bmbt_set_allf(
-	xfs_bmbt_rec_t	*r,
-	xfs_fileoff_t	o,
-	xfs_fsblock_t	b,
-	xfs_filblks_t	c,
-	xfs_exntst_t	v)
+	xfs_bmbt_rec_host_t	*r,
+	xfs_fileoff_t		startoff,
+	xfs_fsblock_t		startblock,
+	xfs_filblks_t		blockcount,
+	xfs_exntst_t		state)
 {
-	int	extent_flag;
+	int		extent_flag = (state == XFS_EXT_NORM) ? 0 : 1;
+
+	ASSERT(state == XFS_EXT_NORM || state == XFS_EXT_UNWRITTEN);
+	ASSERT((startoff & XFS_MASK64HI(64-BMBT_STARTOFF_BITLEN)) == 0);
+	ASSERT((blockcount & XFS_MASK64HI(64-BMBT_BLOCKCOUNT_BITLEN)) == 0);
 
-	ASSERT((v == XFS_EXT_NORM) || (v == XFS_EXT_UNWRITTEN));
-	extent_flag = (v == XFS_EXT_NORM) ? 0 : 1;
-	ASSERT((o & XFS_MASK64HI(64-BMBT_STARTOFF_BITLEN)) == 0);
-	ASSERT((c & XFS_MASK64HI(64-BMBT_BLOCKCOUNT_BITLEN)) == 0);
 #if XFS_BIG_BLKNOS
-	ASSERT((b & XFS_MASK64HI(64-BMBT_STARTBLOCK_BITLEN)) == 0);
+	ASSERT((startblock & XFS_MASK64HI(64-BMBT_STARTBLOCK_BITLEN)) == 0);
+
 	r->l0 = ((xfs_bmbt_rec_base_t)extent_flag << 63) |
-		((xfs_bmbt_rec_base_t)o << 9) |
-		((xfs_bmbt_rec_base_t)b >> 43);
-	r->l1 = ((xfs_bmbt_rec_base_t)b << 21) |
-		((xfs_bmbt_rec_base_t)c &
+		((xfs_bmbt_rec_base_t)startoff << 9) |
+		((xfs_bmbt_rec_base_t)startblock >> 43);
+	r->l1 = ((xfs_bmbt_rec_base_t)startblock << 21) |
+		((xfs_bmbt_rec_base_t)blockcount &
 		(xfs_bmbt_rec_base_t)XFS_MASK64LO(21));
 #else	/* !XFS_BIG_BLKNOS */
-	if (ISNULLSTARTBLOCK(b)) {
+	if (ISNULLSTARTBLOCK(startblock)) {
 		r->l0 = ((xfs_bmbt_rec_base_t)extent_flag << 63) |
-			((xfs_bmbt_rec_base_t)o << 9) |
+			((xfs_bmbt_rec_base_t)startoff << 9) |
 			 (xfs_bmbt_rec_base_t)XFS_MASK64LO(9);
 		r->l1 = XFS_MASK64HI(11) |
-			  ((xfs_bmbt_rec_base_t)b << 21) |
-			  ((xfs_bmbt_rec_base_t)c &
+			  ((xfs_bmbt_rec_base_t)startblock << 21) |
+			  ((xfs_bmbt_rec_base_t)blockcount &
 			   (xfs_bmbt_rec_base_t)XFS_MASK64LO(21));
 	} else {
 		r->l0 = ((xfs_bmbt_rec_base_t)extent_flag << 63) |
-			((xfs_bmbt_rec_base_t)o << 9);
-		r->l1 = ((xfs_bmbt_rec_base_t)b << 21) |
-			 ((xfs_bmbt_rec_base_t)c &
+			((xfs_bmbt_rec_base_t)startoff << 9);
+		r->l1 = ((xfs_bmbt_rec_base_t)startblock << 21) |
+			 ((xfs_bmbt_rec_base_t)blockcount &
 			 (xfs_bmbt_rec_base_t)XFS_MASK64LO(21));
 	}
 #endif	/* XFS_BIG_BLKNOS */
 }
 
-#ifndef XFS_NATIVE_HOST
 /*
  * Set all the fields in a bmap extent record from the uncompressed form.
  */
 void
-xfs_bmbt_disk_set_all(
-	xfs_bmbt_rec_t	*r,
-	xfs_bmbt_irec_t *s)
+xfs_bmbt_set_all(
+	xfs_bmbt_rec_host_t *r,
+	xfs_bmbt_irec_t	*s)
 {
-	int	extent_flag;
-
-	ASSERT((s->br_state == XFS_EXT_NORM) ||
-		(s->br_state == XFS_EXT_UNWRITTEN));
-	extent_flag = (s->br_state == XFS_EXT_NORM) ? 0 : 1;
-	ASSERT((s->br_startoff & XFS_MASK64HI(9)) == 0);
-	ASSERT((s->br_blockcount & XFS_MASK64HI(43)) == 0);
-#if XFS_BIG_BLKNOS
-	ASSERT((s->br_startblock & XFS_MASK64HI(12)) == 0);
-	INT_SET(r->l0, ARCH_CONVERT, ((xfs_bmbt_rec_base_t)extent_flag << 63) |
-		  ((xfs_bmbt_rec_base_t)s->br_startoff << 9) |
-		  ((xfs_bmbt_rec_base_t)s->br_startblock >> 43));
-	INT_SET(r->l1, ARCH_CONVERT, ((xfs_bmbt_rec_base_t)s->br_startblock << 21) |
-		  ((xfs_bmbt_rec_base_t)s->br_blockcount &
-		   (xfs_bmbt_rec_base_t)XFS_MASK64LO(21)));
-#else	/* !XFS_BIG_BLKNOS */
-	if (ISNULLSTARTBLOCK(s->br_startblock)) {
-		INT_SET(r->l0, ARCH_CONVERT, ((xfs_bmbt_rec_base_t)extent_flag << 63) |
-			((xfs_bmbt_rec_base_t)s->br_startoff << 9) |
-			  (xfs_bmbt_rec_base_t)XFS_MASK64LO(9));
-		INT_SET(r->l1, ARCH_CONVERT, XFS_MASK64HI(11) |
-			  ((xfs_bmbt_rec_base_t)s->br_startblock << 21) |
-			  ((xfs_bmbt_rec_base_t)s->br_blockcount &
-			   (xfs_bmbt_rec_base_t)XFS_MASK64LO(21)));
-	} else {
-		INT_SET(r->l0, ARCH_CONVERT, ((xfs_bmbt_rec_base_t)extent_flag << 63) |
-			((xfs_bmbt_rec_base_t)s->br_startoff << 9));
-		INT_SET(r->l1, ARCH_CONVERT, ((xfs_bmbt_rec_base_t)s->br_startblock << 21) |
-			  ((xfs_bmbt_rec_base_t)s->br_blockcount &
-			   (xfs_bmbt_rec_base_t)XFS_MASK64LO(21)));
-	}
-#endif	/* XFS_BIG_BLKNOS */
+	xfs_bmbt_set_allf(r, s->br_startoff, s->br_startblock,
+			     s->br_blockcount, s->br_state);
 }
 
+
 /*
  * Set all the fields in a disk format bmap extent record from the arguments.
  */
 void
 xfs_bmbt_disk_set_allf(
-	xfs_bmbt_rec_t	*r,
-	xfs_fileoff_t	o,
-	xfs_fsblock_t	b,
-	xfs_filblks_t	c,
-	xfs_exntst_t	v)
+	xfs_bmbt_rec_t		*r,
+	xfs_fileoff_t		startoff,
+	xfs_fsblock_t		startblock,
+	xfs_filblks_t		blockcount,
+	xfs_exntst_t		state)
 {
-	int	extent_flag;
+	int			extent_flag = (state == XFS_EXT_NORM) ? 0 : 1;
+
+	ASSERT(state == XFS_EXT_NORM || state == XFS_EXT_UNWRITTEN);
+	ASSERT((startoff & XFS_MASK64HI(64-BMBT_STARTOFF_BITLEN)) == 0);
+	ASSERT((blockcount & XFS_MASK64HI(64-BMBT_BLOCKCOUNT_BITLEN)) == 0);
 
-	ASSERT((v == XFS_EXT_NORM) || (v == XFS_EXT_UNWRITTEN));
-	extent_flag = (v == XFS_EXT_NORM) ? 0 : 1;
-	ASSERT((o & XFS_MASK64HI(64-BMBT_STARTOFF_BITLEN)) == 0);
-	ASSERT((c & XFS_MASK64HI(64-BMBT_BLOCKCOUNT_BITLEN)) == 0);
 #if XFS_BIG_BLKNOS
-	ASSERT((b & XFS_MASK64HI(64-BMBT_STARTBLOCK_BITLEN)) == 0);
-	INT_SET(r->l0, ARCH_CONVERT, ((xfs_bmbt_rec_base_t)extent_flag << 63) |
-		((xfs_bmbt_rec_base_t)o << 9) |
-		((xfs_bmbt_rec_base_t)b >> 43));
-	INT_SET(r->l1, ARCH_CONVERT, ((xfs_bmbt_rec_base_t)b << 21) |
-		  ((xfs_bmbt_rec_base_t)c &
-		   (xfs_bmbt_rec_base_t)XFS_MASK64LO(21)));
+	ASSERT((startblock & XFS_MASK64HI(64-BMBT_STARTBLOCK_BITLEN)) == 0);
+
+	r->l0 = cpu_to_be64(
+		((xfs_bmbt_rec_base_t)extent_flag << 63) |
+		 ((xfs_bmbt_rec_base_t)startoff << 9) |
+		 ((xfs_bmbt_rec_base_t)startblock >> 43));
+	r->l1 = cpu_to_be64(
+		((xfs_bmbt_rec_base_t)startblock << 21) |
+		 ((xfs_bmbt_rec_base_t)blockcount &
+		  (xfs_bmbt_rec_base_t)XFS_MASK64LO(21)));
 #else	/* !XFS_BIG_BLKNOS */
-	if (ISNULLSTARTBLOCK(b)) {
-		INT_SET(r->l0, ARCH_CONVERT, ((xfs_bmbt_rec_base_t)extent_flag << 63) |
-			((xfs_bmbt_rec_base_t)o << 9) |
-			 (xfs_bmbt_rec_base_t)XFS_MASK64LO(9));
-		INT_SET(r->l1, ARCH_CONVERT, XFS_MASK64HI(11) |
-			  ((xfs_bmbt_rec_base_t)b << 21) |
-			  ((xfs_bmbt_rec_base_t)c &
+	if (ISNULLSTARTBLOCK(startblock)) {
+		r->l0 = cpu_to_be64(
+			((xfs_bmbt_rec_base_t)extent_flag << 63) |
+			 ((xfs_bmbt_rec_base_t)startoff << 9) |
+			  (xfs_bmbt_rec_base_t)XFS_MASK64LO(9));
+		r->l1 = cpu_to_be64(XFS_MASK64HI(11) |
+			  ((xfs_bmbt_rec_base_t)startblock << 21) |
+			  ((xfs_bmbt_rec_base_t)blockcount &
 			   (xfs_bmbt_rec_base_t)XFS_MASK64LO(21)));
 	} else {
-		INT_SET(r->l0, ARCH_CONVERT, ((xfs_bmbt_rec_base_t)extent_flag << 63) |
-			((xfs_bmbt_rec_base_t)o << 9));
-		INT_SET(r->l1, ARCH_CONVERT, ((xfs_bmbt_rec_base_t)b << 21) |
-			  ((xfs_bmbt_rec_base_t)c &
-			   (xfs_bmbt_rec_base_t)XFS_MASK64LO(21)));
+		r->l0 = cpu_to_be64(
+			((xfs_bmbt_rec_base_t)extent_flag << 63) |
+			 ((xfs_bmbt_rec_base_t)startoff << 9));
+		r->l1 = cpu_to_be64(
+			((xfs_bmbt_rec_base_t)startblock << 21) |
+			 ((xfs_bmbt_rec_base_t)blockcount &
+			  (xfs_bmbt_rec_base_t)XFS_MASK64LO(21)));
 	}
 #endif	/* XFS_BIG_BLKNOS */
 }
-#endif /* XFS_NATIVE_HOST */
+
+/*
+ * Set all the fields in a bmap extent record from the uncompressed form.
+ */
+void
+xfs_bmbt_disk_set_all(
+	xfs_bmbt_rec_t	*r,
+	xfs_bmbt_irec_t *s)
+{
+	xfs_bmbt_disk_set_allf(r, s->br_startoff, s->br_startblock,
+				  s->br_blockcount, s->br_state);
+}
 
 /*
  * Set the blockcount field in a bmap extent record.
  */
 void
 xfs_bmbt_set_blockcount(
-	xfs_bmbt_rec_t	*r,
+	xfs_bmbt_rec_host_t *r,
 	xfs_filblks_t	v)
 {
 	ASSERT((v & XFS_MASK64HI(43)) == 0);
@@ -2481,7 +2420,7 @@ xfs_bmbt_set_blockcount(
  */
 void
 xfs_bmbt_set_startblock(
-	xfs_bmbt_rec_t	*r,
+	xfs_bmbt_rec_host_t *r,
 	xfs_fsblock_t	v)
 {
 #if XFS_BIG_BLKNOS
@@ -2509,7 +2448,7 @@ xfs_bmbt_set_startblock(
  */
 void
 xfs_bmbt_set_startoff(
-	xfs_bmbt_rec_t	*r,
+	xfs_bmbt_rec_host_t *r,
 	xfs_fileoff_t	v)
 {
 	ASSERT((v & XFS_MASK64HI(9)) == 0);
@@ -2523,7 +2462,7 @@ xfs_bmbt_set_startoff(
  */
 void
 xfs_bmbt_set_state(
-	xfs_bmbt_rec_t	*r,
+	xfs_bmbt_rec_host_t *r,
 	xfs_exntst_t	v)
 {
 	ASSERT(v == XFS_EXT_NORM || v == XFS_EXT_UNWRITTEN);
@@ -2624,10 +2563,8 @@ xfs_check_nostate_extents(
 	xfs_extnum_t		idx,
 	xfs_extnum_t		num)
 {
-	xfs_bmbt_rec_t		*ep;
-
 	for (; num > 0; num--, idx++) {
-		ep = xfs_iext_get_ext(ifp, idx);
+		xfs_bmbt_rec_host_t *ep = xfs_iext_get_ext(ifp, idx);
 		if ((ep->l0 >>
 		     (64 - BMBT_EXNTFLAG_BITLEN)) != 0) {
 			ASSERT(0);

fs/xfs/xfs_bmap_btree.h

@@ -35,45 +35,16 @@ typedef struct xfs_bmdr_block {
 
 /*
  * Bmap btree record and extent descriptor.
- * For 32-bit kernels,
- *  l0:31 is an extent flag (value 1 indicates non-normal).
- *  l0:0-30 and l1:9-31 are startoff.
- *  l1:0-8, l2:0-31, and l3:21-31 are startblock.
- *  l3:0-20 are blockcount.
- * For 64-bit kernels,
  *  l0:63 is an extent flag (value 1 indicates non-normal).
  *  l0:9-62 are startoff.
  *  l0:0-8 and l1:21-63 are startblock.
  *  l1:0-20 are blockcount.
  */
-
-#ifndef XFS_NATIVE_HOST
-
-#define BMBT_TOTAL_BITLEN	128	/* 128 bits, 16 bytes */
-#define BMBT_EXNTFLAG_BITOFF	0
 #define BMBT_EXNTFLAG_BITLEN	1
-#define BMBT_STARTOFF_BITOFF	(BMBT_EXNTFLAG_BITOFF + BMBT_EXNTFLAG_BITLEN)
 #define BMBT_STARTOFF_BITLEN	54
-#define BMBT_STARTBLOCK_BITOFF	(BMBT_STARTOFF_BITOFF + BMBT_STARTOFF_BITLEN)
 #define BMBT_STARTBLOCK_BITLEN	52
-#define BMBT_BLOCKCOUNT_BITOFF	\
-	(BMBT_STARTBLOCK_BITOFF + BMBT_STARTBLOCK_BITLEN)
-#define BMBT_BLOCKCOUNT_BITLEN	(BMBT_TOTAL_BITLEN - BMBT_BLOCKCOUNT_BITOFF)
-
-#else
-
-#define BMBT_TOTAL_BITLEN	128	/* 128 bits, 16 bytes */
-#define BMBT_EXNTFLAG_BITOFF	63
-#define BMBT_EXNTFLAG_BITLEN	1
-#define BMBT_STARTOFF_BITOFF	(BMBT_EXNTFLAG_BITOFF - BMBT_STARTOFF_BITLEN)
-#define BMBT_STARTOFF_BITLEN	54
-#define BMBT_STARTBLOCK_BITOFF	85 /* 128 - 43 (other 9 is in first word) */
-#define BMBT_STARTBLOCK_BITLEN	52
-#define BMBT_BLOCKCOUNT_BITOFF	64 /* Start of second 64 bit container */
 #define BMBT_BLOCKCOUNT_BITLEN	21
 
-#endif /* XFS_NATIVE_HOST */
-
 
 #define BMBT_USE_64	1
 
@@ -83,12 +54,16 @@ typedef struct xfs_bmbt_rec_32
 } xfs_bmbt_rec_32_t;
 typedef struct xfs_bmbt_rec_64
 {
-	__uint64_t		l0, l1;
+	__be64			l0, l1;
 } xfs_bmbt_rec_64_t;
 
 typedef __uint64_t	xfs_bmbt_rec_base_t;	/* use this for casts */
 typedef xfs_bmbt_rec_64_t xfs_bmbt_rec_t, xfs_bmdr_rec_t;
 
+typedef struct xfs_bmbt_rec_host {
+	__uint64_t		l0, l1;
+} xfs_bmbt_rec_host_t;
+
 /*
  * Values and macros for delayed-allocation startblock fields.
  */
@@ -281,23 +256,17 @@ extern ktrace_t	*xfs_bmbt_trace_buf;
 extern void xfs_bmdr_to_bmbt(xfs_bmdr_block_t *, int, xfs_bmbt_block_t *, int);
 extern int xfs_bmbt_decrement(struct xfs_btree_cur *, int, int *);
 extern int xfs_bmbt_delete(struct xfs_btree_cur *, int *);
-extern void xfs_bmbt_get_all(xfs_bmbt_rec_t *r, xfs_bmbt_irec_t *s);
+extern void xfs_bmbt_get_all(xfs_bmbt_rec_host_t *r, xfs_bmbt_irec_t *s);
 extern xfs_bmbt_block_t *xfs_bmbt_get_block(struct xfs_btree_cur *cur,
 						int, struct xfs_buf **bpp);
-extern xfs_filblks_t xfs_bmbt_get_blockcount(xfs_bmbt_rec_t *r);
-extern xfs_fsblock_t xfs_bmbt_get_startblock(xfs_bmbt_rec_t *r);
-extern xfs_fileoff_t xfs_bmbt_get_startoff(xfs_bmbt_rec_t *r);
-extern xfs_exntst_t xfs_bmbt_get_state(xfs_bmbt_rec_t *r);
+extern xfs_filblks_t xfs_bmbt_get_blockcount(xfs_bmbt_rec_host_t *r);
+extern xfs_fsblock_t xfs_bmbt_get_startblock(xfs_bmbt_rec_host_t *r);
+extern xfs_fileoff_t xfs_bmbt_get_startoff(xfs_bmbt_rec_host_t *r);
+extern xfs_exntst_t xfs_bmbt_get_state(xfs_bmbt_rec_host_t *r);
 
-#ifndef XFS_NATIVE_HOST
 extern void xfs_bmbt_disk_get_all(xfs_bmbt_rec_t *r, xfs_bmbt_irec_t *s);
 extern xfs_filblks_t xfs_bmbt_disk_get_blockcount(xfs_bmbt_rec_t *r);
 extern xfs_fileoff_t xfs_bmbt_disk_get_startoff(xfs_bmbt_rec_t *r);
-#else
-#define xfs_bmbt_disk_get_all(r, s)	xfs_bmbt_get_all(r, s)
-#define xfs_bmbt_disk_get_blockcount(r)	xfs_bmbt_get_blockcount(r)
-#define xfs_bmbt_disk_get_startoff(r)	xfs_bmbt_get_startoff(r)
-#endif /* XFS_NATIVE_HOST */
 
 extern int xfs_bmbt_increment(struct xfs_btree_cur *, int, int *);
 extern int xfs_bmbt_insert(struct xfs_btree_cur *, int *);
@@ -315,22 +284,17 @@ extern int xfs_bmbt_lookup_ge(struct xfs_btree_cur *, xfs_fileoff_t,
  */
 extern int xfs_bmbt_newroot(struct xfs_btree_cur *cur, int *lflags, int *stat);
 
-extern void xfs_bmbt_set_all(xfs_bmbt_rec_t *r, xfs_bmbt_irec_t *s);
-extern void xfs_bmbt_set_allf(xfs_bmbt_rec_t *r, xfs_fileoff_t o,
+extern void xfs_bmbt_set_all(xfs_bmbt_rec_host_t *r, xfs_bmbt_irec_t *s);
+extern void xfs_bmbt_set_allf(xfs_bmbt_rec_host_t *r, xfs_fileoff_t o,
 			xfs_fsblock_t b, xfs_filblks_t c, xfs_exntst_t v);
-extern void xfs_bmbt_set_blockcount(xfs_bmbt_rec_t *r, xfs_filblks_t v);
-extern void xfs_bmbt_set_startblock(xfs_bmbt_rec_t *r, xfs_fsblock_t v);
-extern void xfs_bmbt_set_startoff(xfs_bmbt_rec_t *r, xfs_fileoff_t v);
-extern void xfs_bmbt_set_state(xfs_bmbt_rec_t *r, xfs_exntst_t v);
+extern void xfs_bmbt_set_blockcount(xfs_bmbt_rec_host_t *r, xfs_filblks_t v);
+extern void xfs_bmbt_set_startblock(xfs_bmbt_rec_host_t *r, xfs_fsblock_t v);
+extern void xfs_bmbt_set_startoff(xfs_bmbt_rec_host_t *r, xfs_fileoff_t v);
+extern void xfs_bmbt_set_state(xfs_bmbt_rec_host_t *r, xfs_exntst_t v);
 
-#ifndef XFS_NATIVE_HOST
 extern void xfs_bmbt_disk_set_all(xfs_bmbt_rec_t *r, xfs_bmbt_irec_t *s);
 extern void xfs_bmbt_disk_set_allf(xfs_bmbt_rec_t *r, xfs_fileoff_t o,
 			xfs_fsblock_t b, xfs_filblks_t c, xfs_exntst_t v);
-#else
-#define xfs_bmbt_disk_set_all(r, s)		xfs_bmbt_set_all(r, s)
-#define xfs_bmbt_disk_set_allf(r, o, b, c, v)	xfs_bmbt_set_allf(r, o, b, c, v)
-#endif /* XFS_NATIVE_HOST */
 
 extern void xfs_bmbt_to_bmdr(xfs_bmbt_block_t *, int, xfs_bmdr_block_t *, int);
 extern int xfs_bmbt_update(struct xfs_btree_cur *, xfs_fileoff_t,

fs/xfs/xfs_buf_item.c

@@ -23,6 +23,7 @@
 #include "xfs_inum.h"
 #include "xfs_trans.h"
 #include "xfs_sb.h"
+#include "xfs_ag.h"
 #include "xfs_dmapi.h"
 #include "xfs_mount.h"
 #include "xfs_buf_item.h"

fs/xfs/xfs_clnt.h

@@ -89,7 +89,6 @@ struct xfs_mount_args {
 #define XFSMNT_IDELETE		0x08000000	/* inode cluster delete */
 #define XFSMNT_SWALLOC		0x10000000	/* turn on stripe width
 						 * allocation */
-#define XFSMNT_IHASHSIZE	0x20000000	/* inode hash table size */
 #define XFSMNT_DIRSYNC		0x40000000	/* sync creat,link,unlink,rename
 						 * symlink,mkdir,rmdir,mknod */
 #define XFSMNT_FLAGS2		0x80000000	/* more flags set in flags2 */

fs/xfs/xfs_dfrag.c

@@ -42,6 +42,7 @@
 #include "xfs_dfrag.h"
 #include "xfs_error.h"
 #include "xfs_rw.h"
+#include "xfs_vnodeops.h"
 
 /*
  * Syssgi interface for swapext
@@ -199,7 +200,8 @@ xfs_swap_extents(
 
 	if (VN_CACHED(tvp) != 0) {
 		xfs_inval_cached_trace(&tip->i_iocore, 0, -1, 0, -1);
-		error = bhv_vop_flushinval_pages(tvp, 0, -1, FI_REMAPF_LOCKED);
+		error = xfs_flushinval_pages(tip, 0, -1,
+				FI_REMAPF_LOCKED);
 		if (error)
 			goto error0;
 	}
@@ -265,7 +267,7 @@ xfs_swap_extents(
 	 * fields change.
 	 */
 
-	bhv_vop_toss_pages(vp, 0, -1, FI_REMAPF);
+	xfs_tosspages(ip, 0, -1, FI_REMAPF);
 
 	tp = xfs_trans_alloc(mp, XFS_TRANS_SWAPEXT);
 	if ((error = xfs_trans_reserve(tp, 0,

fs/xfs/xfs_dinode.h

@@ -34,41 +34,41 @@ struct xfs_mount;
  * because we only need the core part in the in-core inode.
  */
 typedef struct xfs_timestamp {
-	__int32_t	t_sec;		/* timestamp seconds */
-	__int32_t	t_nsec;		/* timestamp nanoseconds */
+	__be32		t_sec;		/* timestamp seconds */
+	__be32		t_nsec;		/* timestamp nanoseconds */
 } xfs_timestamp_t;
 
 /*
  * Note: Coordinate changes to this structure with the XFS_DI_* #defines
- * below and the offsets table in xfs_ialloc_log_di().
+ * below, the offsets table in xfs_ialloc_log_di() and struct xfs_icdinode
+ * in xfs_inode.h.
  */
-typedef struct xfs_dinode_core
-{
-	__uint16_t	di_magic;	/* inode magic # = XFS_DINODE_MAGIC */
-	__uint16_t	di_mode;	/* mode and type of file */
-	__int8_t	di_version;	/* inode version */
-	__int8_t	di_format;	/* format of di_c data */
-	__uint16_t	di_onlink;	/* old number of links to file */
-	__uint32_t	di_uid;		/* owner's user id */
-	__uint32_t	di_gid;		/* owner's group id */
-	__uint32_t	di_nlink;	/* number of links to file */
-	__uint16_t	di_projid;	/* owner's project id */
-	__uint8_t	di_pad[8];	/* unused, zeroed space */
-	__uint16_t	di_flushiter;	/* incremented on flush */
+typedef struct xfs_dinode_core {
+	__be16		di_magic;	/* inode magic # = XFS_DINODE_MAGIC */
+	__be16		di_mode;	/* mode and type of file */
+	__u8		di_version;	/* inode version */
+	__u8		di_format;	/* format of di_c data */
+	__be16		di_onlink;	/* old number of links to file */
+	__be32		di_uid;		/* owner's user id */
+	__be32		di_gid;		/* owner's group id */
+	__be32		di_nlink;	/* number of links to file */
+	__be16		di_projid;	/* owner's project id */
+	__u8		di_pad[8];	/* unused, zeroed space */
+	__be16		di_flushiter;	/* incremented on flush */
 	xfs_timestamp_t	di_atime;	/* time last accessed */
 	xfs_timestamp_t	di_mtime;	/* time last modified */
 	xfs_timestamp_t	di_ctime;	/* time created/inode modified */
-	xfs_fsize_t	di_size;	/* number of bytes in file */
-	xfs_drfsbno_t	di_nblocks;	/* # of direct & btree blocks used */
-	xfs_extlen_t	di_extsize;	/* basic/minimum extent size for file */
-	xfs_extnum_t	di_nextents;	/* number of extents in data fork */
-	xfs_aextnum_t	di_anextents;	/* number of extents in attribute fork*/
-	__uint8_t	di_forkoff;	/* attr fork offs, <<3 for 64b align */
-	__int8_t	di_aformat;	/* format of attr fork's data */
-	__uint32_t	di_dmevmask;	/* DMIG event mask */
-	__uint16_t	di_dmstate;	/* DMIG state info */
-	__uint16_t	di_flags;	/* random flags, XFS_DIFLAG_... */
-	__uint32_t	di_gen;		/* generation number */
+	__be64		di_size;	/* number of bytes in file */
+	__be64		di_nblocks;	/* # of direct & btree blocks used */
+	__be32		di_extsize;	/* basic/minimum extent size for file */
+	__be32		di_nextents;	/* number of extents in data fork */
+	__be16		di_anextents;	/* number of extents in attribute fork*/
+	__u8		di_forkoff;	/* attr fork offs, <<3 for 64b align */
+	__s8		di_aformat;	/* format of attr fork's data */
+	__be32		di_dmevmask;	/* DMIG event mask */
+	__be16		di_dmstate;	/* DMIG state info */
+	__be16		di_flags;	/* random flags, XFS_DIFLAG_... */
+	__be32		di_gen;		/* generation number */
 } xfs_dinode_core_t;
 
 #define DI_MAX_FLUSH 0xffff
@@ -81,13 +81,13 @@ typedef struct xfs_dinode
 	 * sure to update the macros like XFS_LITINO below and
 	 * XFS_BMAP_RBLOCK_DSIZE in xfs_bmap_btree.h.
 	 */
-	xfs_agino_t		di_next_unlinked;/* agi unlinked list ptr */
+	__be32			di_next_unlinked;/* agi unlinked list ptr */
 	union {
 		xfs_bmdr_block_t di_bmbt;	/* btree root block */
 		xfs_bmbt_rec_32_t di_bmx[1];	/* extent list */
 		xfs_dir2_sf_t	di_dir2sf;	/* shortform directory v2 */
 		char		di_c[1];	/* local contents */
-		xfs_dev_t	di_dev;		/* device for S_IFCHR/S_IFBLK */
+		__be32		di_dev;		/* device for S_IFCHR/S_IFBLK */
 		uuid_t		di_muuid;	/* mount point value */
 		char		di_symlink[1];	/* local symbolic link */
 	}		di_u;
@@ -175,8 +175,7 @@ typedef enum xfs_dinode_fmt
 #define	XFS_CFORK_Q_DISK(dcp)		    ((dcp)->di_forkoff != 0)
 
 #define XFS_CFORK_BOFF(dcp)                 ((int)((dcp)->di_forkoff << 3))
-#define	XFS_CFORK_BOFF_DISK(dcp) \
-	((int)(INT_GET((dcp)->di_forkoff, ARCH_CONVERT) << 3))
+#define	XFS_CFORK_BOFF_DISK(dcp)	    ((int)((dcp)->di_forkoff << 3))
 
 #define	XFS_CFORK_DSIZE_DISK(dcp,mp) \
 	(XFS_CFORK_Q_DISK(dcp) ? XFS_CFORK_BOFF_DISK(dcp) : XFS_LITINO(mp))
@@ -225,8 +224,8 @@ typedef enum xfs_dinode_fmt
 
 #define	XFS_CFORK_NEXTENTS_DISK(dcp,w) \
 	((w) == XFS_DATA_FORK ? \
-	 	INT_GET((dcp)->di_nextents, ARCH_CONVERT) : \
-	 	INT_GET((dcp)->di_anextents, ARCH_CONVERT))
+	 	be32_to_cpu((dcp)->di_nextents) : \
+	 	be16_to_cpu((dcp)->di_anextents))
 #define XFS_CFORK_NEXTENTS(dcp,w) \
 	((w) == XFS_DATA_FORK ? (dcp)->di_nextents : (dcp)->di_anextents)
 #define	XFS_DFORK_NEXTENTS(dip,w) XFS_CFORK_NEXTENTS_DISK(&(dip)->di_core, w)

fs/xfs/xfs_dir2.c

@@ -43,8 +43,6 @@
 #include "xfs_dir2_trace.h"
 #include "xfs_error.h"
 
-static int	xfs_dir2_put_dirent64_direct(xfs_dir2_put_args_t *pa);
-static int	xfs_dir2_put_dirent64_uio(xfs_dir2_put_args_t *pa);
 
 void
 xfs_dir_mount(
@@ -293,47 +291,33 @@ xfs_dir_removename(
  * Read a directory.
  */
 int
-xfs_dir_getdents(
-	xfs_trans_t	*tp,
+xfs_readdir(
 	xfs_inode_t	*dp,
-	uio_t		*uio,		/* caller's buffer control */
-	int		*eofp)		/* out: eof reached */
+	void		*dirent,
+	size_t		bufsize,
+	xfs_off_t	*offset,
+	filldir_t	filldir)
 {
-	int		alignment;	/* alignment required for ABI */
-	xfs_dirent_t	*dbp;		/* malloc'ed buffer */
-	xfs_dir2_put_t	put;		/* entry formatting routine */
 	int		rval;		/* return value */
 	int		v;		/* type-checking value */
 
+	vn_trace_entry(dp, __FUNCTION__, (inst_t *)__return_address);
+
+	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
+		return XFS_ERROR(EIO);
+
 	ASSERT((dp->i_d.di_mode & S_IFMT) == S_IFDIR);
 	XFS_STATS_INC(xs_dir_getdents);
-	/*
-	 * If our caller has given us a single contiguous aligned memory buffer,
-	 * just work directly within that buffer.  If it's in user memory,
-	 * lock it down first.
-	 */
-	alignment = sizeof(xfs_off_t) - 1;
-	if ((uio->uio_iovcnt == 1) &&
-	    (((__psint_t)uio->uio_iov[0].iov_base & alignment) == 0) &&
-	    ((uio->uio_iov[0].iov_len & alignment) == 0)) {
-		dbp = NULL;
-		put = xfs_dir2_put_dirent64_direct;
-	} else {
-		dbp = kmem_alloc(sizeof(*dbp) + MAXNAMELEN, KM_SLEEP);
-		put = xfs_dir2_put_dirent64_uio;
-	}
 
-	*eofp = 0;
 	if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL)
-		rval = xfs_dir2_sf_getdents(dp, uio, eofp, dbp, put);
-	else if ((rval = xfs_dir2_isblock(tp, dp, &v)))
+		rval = xfs_dir2_sf_getdents(dp, dirent, offset, filldir);
+	else if ((rval = xfs_dir2_isblock(NULL, dp, &v)))
 		;
 	else if (v)
-		rval = xfs_dir2_block_getdents(tp, dp, uio, eofp, dbp, put);
+		rval = xfs_dir2_block_getdents(dp, dirent, offset, filldir);
 	else
-		rval = xfs_dir2_leaf_getdents(tp, dp, uio, eofp, dbp, put);
-	if (dbp != NULL)
-		kmem_free(dbp, sizeof(*dbp) + MAXNAMELEN);
+		rval = xfs_dir2_leaf_getdents(dp, dirent, bufsize, offset,
+					      filldir);
 	return rval;
 }
 
@@ -612,77 +596,6 @@ xfs_dir2_isleaf(
 	return 0;
 }
 
-/*
- * Getdents put routine for 64-bit ABI, direct form.
- */
-static int
-xfs_dir2_put_dirent64_direct(
-	xfs_dir2_put_args_t	*pa)
-{
-	xfs_dirent_t		*idbp;		/* dirent pointer */
-	iovec_t			*iovp;		/* io vector */
-	int			namelen;	/* entry name length */
-	int			reclen;		/* entry total length */
-	uio_t			*uio;		/* I/O control */
-
-	namelen = pa->namelen;
-	reclen = DIRENTSIZE(namelen);
-	uio = pa->uio;
-	/*
-	 * Won't fit in the remaining space.
-	 */
-	if (reclen > uio->uio_resid) {
-		pa->done = 0;
-		return 0;
-	}
-	iovp = uio->uio_iov;
-	idbp = (xfs_dirent_t *)iovp->iov_base;
-	iovp->iov_base = (char *)idbp + reclen;
-	iovp->iov_len -= reclen;
-	uio->uio_resid -= reclen;
-	idbp->d_reclen = reclen;
-	idbp->d_ino = pa->ino;
-	idbp->d_off = pa->cook;
-	idbp->d_name[namelen] = '\0';
-	pa->done = 1;
-	memcpy(idbp->d_name, pa->name, namelen);
-	return 0;
-}
-
-/*
- * Getdents put routine for 64-bit ABI, uio form.
- */
-static int
-xfs_dir2_put_dirent64_uio(
-	xfs_dir2_put_args_t	*pa)
-{
-	xfs_dirent_t		*idbp;		/* dirent pointer */
-	int			namelen;	/* entry name length */
-	int			reclen;		/* entry total length */
-	int			rval;		/* return value */
-	uio_t			*uio;		/* I/O control */
-
-	namelen = pa->namelen;
-	reclen = DIRENTSIZE(namelen);
-	uio = pa->uio;
-	/*
-	 * Won't fit in the remaining space.
-	 */
-	if (reclen > uio->uio_resid) {
-		pa->done = 0;
-		return 0;
-	}
-	idbp = pa->dbp;
-	idbp->d_reclen = reclen;
-	idbp->d_ino = pa->ino;
-	idbp->d_off = pa->cook;
-	idbp->d_name[namelen] = '\0';
-	memcpy(idbp->d_name, pa->name, namelen);
-	rval = xfs_uio_read((caddr_t)idbp, reclen, uio);
-	pa->done = (rval == 0);
-	return rval;
-}
-
 /*
  * Remove the given block from the directory.
  * This routine is used for data and free blocks, leaf/node are done

fs/xfs/xfs_dir2.h

@@ -59,21 +59,6 @@ typedef	__uint32_t	xfs_dir2_db_t;
  */
 typedef	xfs_off_t	xfs_dir2_off_t;
 
-/*
- * For getdents, argument struct for put routines.
- */
-typedef int (*xfs_dir2_put_t)(struct xfs_dir2_put_args *pa);
-typedef struct xfs_dir2_put_args {
-	xfs_off_t	cook;		/* cookie of (next) entry */
-	xfs_intino_t	ino;		/* inode number */
-	xfs_dirent_t	*dbp;		/* buffer pointer */
-	char		*name;		/* directory entry name */
-	int		namelen;	/* length of name */
-	int		done;		/* output: set if value was stored */
-	xfs_dir2_put_t	put;		/* put function ptr (i/o) */
-	struct uio	*uio;		/* uio control structure */
-} xfs_dir2_put_args_t;
-
 /*
  * Generic directory interface routines
  */
@@ -92,8 +77,6 @@ extern int xfs_dir_removename(struct xfs_trans *tp, struct xfs_inode *dp,
 				char *name, int namelen, xfs_ino_t ino,
 				xfs_fsblock_t *first,
 				struct xfs_bmap_free *flist, xfs_extlen_t tot);
-extern int xfs_dir_getdents(struct xfs_trans *tp, struct xfs_inode *dp,
-				uio_t *uio, int *eofp);
 extern int xfs_dir_replace(struct xfs_trans *tp, struct xfs_inode *dp,
 				char *name, int namelen, xfs_ino_t inum,
 				xfs_fsblock_t *first,

fs/xfs/xfs_dir2_block.c

@@ -22,6 +22,7 @@
 #include "xfs_inum.h"
 #include "xfs_trans.h"
 #include "xfs_sb.h"
+#include "xfs_ag.h"
 #include "xfs_dir2.h"
 #include "xfs_dmapi.h"
 #include "xfs_mount.h"
@@ -432,12 +433,10 @@ xfs_dir2_block_addname(
  */
 int						/* error */
 xfs_dir2_block_getdents(
-	xfs_trans_t		*tp,		/* transaction (NULL) */
 	xfs_inode_t		*dp,		/* incore inode */
-	uio_t			*uio,		/* caller's buffer control */
-	int			*eofp,		/* eof reached? (out) */
-	xfs_dirent_t		*dbp,		/* caller's buffer */
-	xfs_dir2_put_t		put)		/* abi's formatting function */
+	void			*dirent,
+	xfs_off_t		*offset,
+	filldir_t		filldir)
 {
 	xfs_dir2_block_t	*block;		/* directory block structure */
 	xfs_dabuf_t		*bp;		/* buffer for block */
@@ -447,31 +446,32 @@ xfs_dir2_block_getdents(
 	char			*endptr;	/* end of the data entries */
 	int			error;		/* error return value */
 	xfs_mount_t		*mp;		/* filesystem mount point */
-	xfs_dir2_put_args_t	p;		/* arg package for put rtn */
 	char			*ptr;		/* current data entry */
 	int			wantoff;	/* starting block offset */
+	xfs_ino_t		ino;
+	xfs_off_t		cook;
 
 	mp = dp->i_mount;
 	/*
 	 * If the block number in the offset is out of range, we're done.
 	 */
-	if (xfs_dir2_dataptr_to_db(mp, uio->uio_offset) > mp->m_dirdatablk) {
-		*eofp = 1;
+	if (xfs_dir2_dataptr_to_db(mp, *offset) > mp->m_dirdatablk) {
 		return 0;
 	}
 	/*
 	 * Can't read the block, give up, else get dabuf in bp.
 	 */
-	if ((error =
-	    xfs_da_read_buf(tp, dp, mp->m_dirdatablk, -1, &bp, XFS_DATA_FORK))) {
+	error = xfs_da_read_buf(NULL, dp, mp->m_dirdatablk, -1,
+				&bp, XFS_DATA_FORK);
+	if (error)
 		return error;
-	}
+
 	ASSERT(bp != NULL);
 	/*
 	 * Extract the byte offset we start at from the seek pointer.
 	 * We'll skip entries before this.
 	 */
-	wantoff = xfs_dir2_dataptr_to_off(mp, uio->uio_offset);
+	wantoff = xfs_dir2_dataptr_to_off(mp, *offset);
 	block = bp->data;
 	xfs_dir2_data_check(dp, bp);
 	/*
@@ -480,9 +480,7 @@ xfs_dir2_block_getdents(
 	btp = xfs_dir2_block_tail_p(mp, block);
 	ptr = (char *)block->u;
 	endptr = (char *)xfs_dir2_block_leaf_p(btp);
-	p.dbp = dbp;
-	p.put = put;
-	p.uio = uio;
+
 	/*
 	 * Loop over the data portion of the block.
 	 * Each object is a real entry (dep) or an unused one (dup).
@@ -508,33 +506,24 @@ xfs_dir2_block_getdents(
 		 */
 		if ((char *)dep - (char *)block < wantoff)
 			continue;
-		/*
-		 * Set up argument structure for put routine.
-		 */
-		p.namelen = dep->namelen;
 
-		p.cook = xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
+		cook = xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
 						    ptr - (char *)block);
-		p.ino = be64_to_cpu(dep->inumber);
+		ino = be64_to_cpu(dep->inumber);
 #if XFS_BIG_INUMS
-		p.ino += mp->m_inoadd;
+		ino += mp->m_inoadd;
 #endif
-		p.name = (char *)dep->name;
-
-		/*
-		 * Put the entry in the caller's buffer.
-		 */
-		error = p.put(&p);
 
 		/*
 		 * If it didn't fit, set the final offset to here & return.
 		 */
-		if (!p.done) {
-			uio->uio_offset =
-				xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
+		if (filldir(dirent, dep->name, dep->namelen, cook,
+			    ino, DT_UNKNOWN)) {
+			*offset = xfs_dir2_db_off_to_dataptr(mp,
+					mp->m_dirdatablk,
 					(char *)dep - (char *)block);
-			xfs_da_brelse(tp, bp);
-			return error;
+			xfs_da_brelse(NULL, bp);
+			return 0;
 		}
 	}
 
@@ -542,13 +531,8 @@ xfs_dir2_block_getdents(
 	 * Reached the end of the block.
 	 * Set the offset to a non-existent block 1 and return.
 	 */
-	*eofp = 1;
-
-	uio->uio_offset =
-		xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk + 1, 0);
-
-	xfs_da_brelse(tp, bp);
-
+	*offset = xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk + 1, 0);
+	xfs_da_brelse(NULL, bp);
 	return 0;
 }
 

fs/xfs/xfs_dir2_block.h

@@ -80,9 +80,8 @@ xfs_dir2_block_leaf_p(xfs_dir2_block_tail_t *btp)
  * Function declarations.
  */
 extern int xfs_dir2_block_addname(struct xfs_da_args *args);
-extern int xfs_dir2_block_getdents(struct xfs_trans *tp, struct xfs_inode *dp,
-				   struct uio *uio, int *eofp,
-				   struct xfs_dirent *dbp, xfs_dir2_put_t put);
+extern int xfs_dir2_block_getdents(struct xfs_inode *dp, void *dirent,
+				   xfs_off_t *offset, filldir_t filldir);
 extern int xfs_dir2_block_lookup(struct xfs_da_args *args);
 extern int xfs_dir2_block_removename(struct xfs_da_args *args);
 extern int xfs_dir2_block_replace(struct xfs_da_args *args);

fs/xfs/xfs_dir2_data.c

@@ -22,6 +22,7 @@
 #include "xfs_inum.h"
 #include "xfs_trans.h"
 #include "xfs_sb.h"
+#include "xfs_ag.h"
 #include "xfs_dir2.h"
 #include "xfs_dmapi.h"
 #include "xfs_mount.h"

fs/xfs/xfs_dir2_leaf.c

@@ -749,12 +749,11 @@ xfs_dir2_leaf_compact_x1(
  */
 int						/* error */
 xfs_dir2_leaf_getdents(
-	xfs_trans_t		*tp,		/* transaction pointer */
 	xfs_inode_t		*dp,		/* incore directory inode */
-	uio_t			*uio,		/* I/O control & vectors */
-	int			*eofp,		/* out: reached end of dir */
-	xfs_dirent_t		*dbp,		/* caller's buffer */
-	xfs_dir2_put_t		put)		/* ABI formatting routine */
+	void			*dirent,
+	size_t			bufsize,
+	xfs_off_t		*offset,
+	filldir_t		filldir)
 {
 	xfs_dabuf_t		*bp;		/* data block buffer */
 	int			byteoff;	/* offset in current block */
@@ -763,7 +762,6 @@ xfs_dir2_leaf_getdents(
 	xfs_dir2_data_t		*data;		/* data block structure */
 	xfs_dir2_data_entry_t	*dep;		/* data entry */
 	xfs_dir2_data_unused_t	*dup;		/* unused entry */
-	int			eof;		/* reached end of directory */
 	int			error = 0;	/* error return value */
 	int			i;		/* temporary loop index */
 	int			j;		/* temporary loop index */
@@ -776,46 +774,38 @@ xfs_dir2_leaf_getdents(
 	xfs_mount_t		*mp;		/* filesystem mount point */
 	xfs_dir2_off_t		newoff;		/* new curoff after new blk */
 	int			nmap;		/* mappings to ask xfs_bmapi */
-	xfs_dir2_put_args_t	*p;		/* formatting arg bundle */
 	char			*ptr = NULL;	/* pointer to current data */
 	int			ra_current;	/* number of read-ahead blks */
 	int			ra_index;	/* *map index for read-ahead */
 	int			ra_offset;	/* map entry offset for ra */
 	int			ra_want;	/* readahead count wanted */
+	xfs_ino_t		ino;
 
 	/*
 	 * If the offset is at or past the largest allowed value,
-	 * give up right away, return eof.
+	 * give up right away.
 	 */
-	if (uio->uio_offset >= XFS_DIR2_MAX_DATAPTR) {
-		*eofp = 1;
+	if (*offset >= XFS_DIR2_MAX_DATAPTR)
 		return 0;
-	}
+
 	mp = dp->i_mount;
-	/*
-	 * Setup formatting arguments.
-	 */
-	p = kmem_alloc(sizeof(*p), KM_SLEEP);
-	p->dbp = dbp;
-	p->put = put;
-	p->uio = uio;
+
 	/*
 	 * Set up to bmap a number of blocks based on the caller's
 	 * buffer size, the directory block size, and the filesystem
 	 * block size.
 	 */
-	map_size =
-		howmany(uio->uio_resid + mp->m_dirblksize,
-			mp->m_sb.sb_blocksize);
+	map_size = howmany(bufsize + mp->m_dirblksize, mp->m_sb.sb_blocksize);
 	map = kmem_alloc(map_size * sizeof(*map), KM_SLEEP);
 	map_valid = ra_index = ra_offset = ra_current = map_blocks = 0;
 	bp = NULL;
-	eof = 1;
+
 	/*
 	 * Inside the loop we keep the main offset value as a byte offset
 	 * in the directory file.
 	 */
-	curoff = xfs_dir2_dataptr_to_byte(mp, uio->uio_offset);
+	curoff = xfs_dir2_dataptr_to_byte(mp, *offset);
+
 	/*
 	 * Force this conversion through db so we truncate the offset
 	 * down to get the start of the data block.
@@ -836,7 +826,7 @@ xfs_dir2_leaf_getdents(
 			 * take it out of the mapping.
 			 */
 			if (bp) {
-				xfs_da_brelse(tp, bp);
+				xfs_da_brelse(NULL, bp);
 				bp = NULL;
 				map_blocks -= mp->m_dirblkfsbs;
 				/*
@@ -862,8 +852,9 @@ xfs_dir2_leaf_getdents(
 			/*
 			 * Recalculate the readahead blocks wanted.
 			 */
-			ra_want = howmany(uio->uio_resid + mp->m_dirblksize,
+			ra_want = howmany(bufsize + mp->m_dirblksize,
 					  mp->m_sb.sb_blocksize) - 1;
+
 			/*
 			 * If we don't have as many as we want, and we haven't
 			 * run out of data blocks, get some more mappings.
@@ -876,7 +867,7 @@ xfs_dir2_leaf_getdents(
 				 * we already have in the table.
 				 */
 				nmap = map_size - map_valid;
-				error = xfs_bmapi(tp, dp,
+				error = xfs_bmapi(NULL, dp,
 					map_off,
 					xfs_dir2_byte_to_da(mp,
 						XFS_DIR2_LEAF_OFFSET) - map_off,
@@ -939,7 +930,7 @@ xfs_dir2_leaf_getdents(
 			 * mapping.
 			 */
 			curdb = xfs_dir2_da_to_db(mp, map->br_startoff);
-			error = xfs_da_read_buf(tp, dp, map->br_startoff,
+			error = xfs_da_read_buf(NULL, dp, map->br_startoff,
 				map->br_blockcount >= mp->m_dirblkfsbs ?
 				    XFS_FSB_TO_DADDR(mp, map->br_startblock) :
 				    -1,
@@ -982,7 +973,7 @@ xfs_dir2_leaf_getdents(
 				 * is a very rare case.
 				 */
 				else if (i > ra_current) {
-					(void)xfs_da_reada_buf(tp, dp,
+					(void)xfs_da_reada_buf(NULL, dp,
 						map[ra_index].br_startoff +
 						ra_offset, XFS_DATA_FORK);
 					ra_current = i;
@@ -1089,46 +1080,39 @@ xfs_dir2_leaf_getdents(
 		 */
 		dep = (xfs_dir2_data_entry_t *)ptr;
 
-		p->namelen = dep->namelen;
-
-		length = xfs_dir2_data_entsize(p->namelen);
-
-		p->cook = xfs_dir2_byte_to_dataptr(mp, curoff + length);
+		length = xfs_dir2_data_entsize(dep->namelen);
 
-		p->ino = be64_to_cpu(dep->inumber);
+		ino = be64_to_cpu(dep->inumber);
 #if XFS_BIG_INUMS
-		p->ino += mp->m_inoadd;
+		ino += mp->m_inoadd;
 #endif
-		p->name = (char *)dep->name;
-
-		error = p->put(p);
 
 		/*
 		 * Won't fit.  Return to caller.
 		 */
-		if (!p->done) {
-			eof = 0;
+		if (filldir(dirent, dep->name, dep->namelen,
+			    xfs_dir2_byte_to_dataptr(mp, curoff + length),
+			    ino, DT_UNKNOWN))
 			break;
-		}
+
 		/*
 		 * Advance to next entry in the block.
 		 */
 		ptr += length;
 		curoff += length;
+		bufsize -= length;
 	}
 
 	/*
 	 * All done.  Set output offset value to current offset.
 	 */
-	*eofp = eof;
 	if (curoff > xfs_dir2_dataptr_to_byte(mp, XFS_DIR2_MAX_DATAPTR))
-		uio->uio_offset = XFS_DIR2_MAX_DATAPTR;
+		*offset = XFS_DIR2_MAX_DATAPTR;
 	else
-		uio->uio_offset = xfs_dir2_byte_to_dataptr(mp, curoff);
+		*offset = xfs_dir2_byte_to_dataptr(mp, curoff);
 	kmem_free(map, map_size * sizeof(*map));
-	kmem_free(p, sizeof(*p));
 	if (bp)
-		xfs_da_brelse(tp, bp);
+		xfs_da_brelse(NULL, bp);
 	return error;
 }
 

fs/xfs/xfs_dir2_leaf.h

@@ -232,9 +232,9 @@ extern void xfs_dir2_leaf_compact(struct xfs_da_args *args,
 extern void xfs_dir2_leaf_compact_x1(struct xfs_dabuf *bp, int *indexp,
 				     int *lowstalep, int *highstalep,
 				     int *lowlogp, int *highlogp);
-extern int xfs_dir2_leaf_getdents(struct xfs_trans *tp, struct xfs_inode *dp,
-				  struct uio *uio, int *eofp,
-				  struct xfs_dirent *dbp, xfs_dir2_put_t put);
+extern int xfs_dir2_leaf_getdents(struct xfs_inode *dp, void *dirent,
+				  size_t bufsize, xfs_off_t *offset,
+				  filldir_t filldir);
 extern int xfs_dir2_leaf_init(struct xfs_da_args *args, xfs_dir2_db_t bno,
 			      struct xfs_dabuf **bpp, int magic);
 extern void xfs_dir2_leaf_log_ents(struct xfs_trans *tp, struct xfs_dabuf *bp,

fs/xfs/xfs_dir2_node.c

@@ -22,6 +22,7 @@
 #include "xfs_inum.h"
 #include "xfs_trans.h"
 #include "xfs_sb.h"
+#include "xfs_ag.h"
 #include "xfs_dir2.h"
 #include "xfs_dmapi.h"
 #include "xfs_mount.h"

fs/xfs/xfs_dir2_sf.c

@@ -22,6 +22,7 @@
 #include "xfs_inum.h"
 #include "xfs_trans.h"
 #include "xfs_sb.h"
+#include "xfs_ag.h"
 #include "xfs_dir2.h"
 #include "xfs_dmapi.h"
 #include "xfs_mount.h"
@@ -695,19 +696,18 @@ xfs_dir2_sf_create(
 int						/* error */
 xfs_dir2_sf_getdents(
 	xfs_inode_t		*dp,		/* incore directory inode */
-	uio_t			*uio,		/* caller's buffer control */
-	int			*eofp,		/* eof reached? (out) */
-	xfs_dirent_t		*dbp,		/* caller's buffer */
-	xfs_dir2_put_t		put)		/* abi's formatting function */
+	void			*dirent,
+	xfs_off_t		*offset,
+	filldir_t		filldir)
 {
-	int			error;		/* error return value */
 	int			i;		/* shortform entry number */
 	xfs_mount_t		*mp;		/* filesystem mount point */
 	xfs_dir2_dataptr_t	off;		/* current entry's offset */
-	xfs_dir2_put_args_t	p;		/* arg package for put rtn */
 	xfs_dir2_sf_entry_t	*sfep;		/* shortform directory entry */
 	xfs_dir2_sf_t		*sfp;		/* shortform structure */
-	xfs_off_t			dir_offset;
+	xfs_dir2_dataptr_t	dot_offset;
+	xfs_dir2_dataptr_t	dotdot_offset;
+	xfs_ino_t		ino;
 
 	mp = dp->i_mount;
 
@@ -720,8 +720,6 @@ xfs_dir2_sf_getdents(
 		return XFS_ERROR(EIO);
 	}
 
-	dir_offset = uio->uio_offset;
-
 	ASSERT(dp->i_df.if_bytes == dp->i_d.di_size);
 	ASSERT(dp->i_df.if_u1.if_data != NULL);
 
@@ -732,108 +730,78 @@ xfs_dir2_sf_getdents(
 	/*
 	 * If the block number in the offset is out of range, we're done.
 	 */
-	if (xfs_dir2_dataptr_to_db(mp, dir_offset) > mp->m_dirdatablk) {
-		*eofp = 1;
+	if (xfs_dir2_dataptr_to_db(mp, *offset) > mp->m_dirdatablk)
 		return 0;
-	}
 
 	/*
-	 * Set up putargs structure.
+	 * Precalculate offsets for . and .. as we will always need them.
+	 *
+	 * XXX(hch): the second argument is sometimes 0 and sometimes
+	 * mp->m_dirdatablk.
 	 */
-	p.dbp = dbp;
-	p.put = put;
-	p.uio = uio;
+	dot_offset = xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
+					     XFS_DIR2_DATA_DOT_OFFSET);
+	dotdot_offset = xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
+						XFS_DIR2_DATA_DOTDOT_OFFSET);
+
 	/*
 	 * Put . entry unless we're starting past it.
 	 */
-	if (dir_offset <=
-		    xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
-					       XFS_DIR2_DATA_DOT_OFFSET)) {
-		p.cook = xfs_dir2_db_off_to_dataptr(mp, 0,
-						XFS_DIR2_DATA_DOTDOT_OFFSET);
-		p.ino = dp->i_ino;
+	if (*offset <= dot_offset) {
+		ino = dp->i_ino;
 #if XFS_BIG_INUMS
-		p.ino += mp->m_inoadd;
+		ino += mp->m_inoadd;
 #endif
-		p.name = ".";
-		p.namelen = 1;
-
-		error = p.put(&p);
-
-		if (!p.done) {
-			uio->uio_offset =
-				xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
-						XFS_DIR2_DATA_DOT_OFFSET);
-			return error;
+		if (filldir(dirent, ".", 1, dotdot_offset, ino, DT_DIR)) {
+			*offset = dot_offset;
+			return 0;
 		}
 	}
 
 	/*
 	 * Put .. entry unless we're starting past it.
 	 */
-	if (dir_offset <=
-		    xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
-					       XFS_DIR2_DATA_DOTDOT_OFFSET)) {
-		p.cook = xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
-						XFS_DIR2_DATA_FIRST_OFFSET);
-		p.ino = xfs_dir2_sf_get_inumber(sfp, &sfp->hdr.parent);
+	if (*offset <= dotdot_offset) {
+		off = xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
+						  XFS_DIR2_DATA_FIRST_OFFSET);
+		ino = xfs_dir2_sf_get_inumber(sfp, &sfp->hdr.parent);
 #if XFS_BIG_INUMS
-		p.ino += mp->m_inoadd;
+		ino += mp->m_inoadd;
 #endif
-		p.name = "..";
-		p.namelen = 2;
-
-		error = p.put(&p);
-
-		if (!p.done) {
-			uio->uio_offset =
-				xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
-					XFS_DIR2_DATA_DOTDOT_OFFSET);
-			return error;
+		if (filldir(dirent, "..", 2, off, ino, DT_DIR)) {
+			*offset = dotdot_offset;
+			return 0;
 		}
 	}
 
 	/*
 	 * Loop while there are more entries and put'ing works.
 	 */
-	for (i = 0, sfep = xfs_dir2_sf_firstentry(sfp);
-		     i < sfp->hdr.count;
-			     i++, sfep = xfs_dir2_sf_nextentry(sfp, sfep)) {
-
+	sfep = xfs_dir2_sf_firstentry(sfp);
+	for (i = 0; i < sfp->hdr.count; i++) {
 		off = xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
 				xfs_dir2_sf_get_offset(sfep));
 
-		if (dir_offset > off)
+		if (*offset > off) {
+			sfep = xfs_dir2_sf_nextentry(sfp, sfep);
 			continue;
+		}
 
-		p.namelen = sfep->namelen;
-
-		p.cook = xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
-			xfs_dir2_sf_get_offset(sfep) +
-			xfs_dir2_data_entsize(p.namelen));
-
-		p.ino = xfs_dir2_sf_get_inumber(sfp, xfs_dir2_sf_inumberp(sfep));
+		ino = xfs_dir2_sf_get_inumber(sfp, xfs_dir2_sf_inumberp(sfep));
 #if XFS_BIG_INUMS
-		p.ino += mp->m_inoadd;
+		ino += mp->m_inoadd;
 #endif
-		p.name = (char *)sfep->name;
-
-		error = p.put(&p);
 
-		if (!p.done) {
-			uio->uio_offset = off;
-			return error;
+		if (filldir(dirent, sfep->name, sfep->namelen,
+			    off + xfs_dir2_data_entsize(sfep->namelen),
+			    ino, DT_UNKNOWN)) {
+			*offset = off;
+			return 0;
 		}
+		sfep = xfs_dir2_sf_nextentry(sfp, sfep);
 	}
 
-	/*
-	 * They all fit.
-	 */
-	*eofp = 1;
-
-	uio->uio_offset =
-		xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk + 1, 0);
-
+	*offset = xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk + 1, 0);
 	return 0;
 }
 

fs/xfs/xfs_dir2_sf.h

@@ -169,9 +169,8 @@ extern int xfs_dir2_block_to_sf(struct xfs_da_args *args, struct xfs_dabuf *bp,
 				int size, xfs_dir2_sf_hdr_t *sfhp);
 extern int xfs_dir2_sf_addname(struct xfs_da_args *args);
 extern int xfs_dir2_sf_create(struct xfs_da_args *args, xfs_ino_t pino);
-extern int xfs_dir2_sf_getdents(struct xfs_inode *dp, struct uio *uio,
-				int *eofp, struct xfs_dirent *dbp,
-				xfs_dir2_put_t put);
+extern int xfs_dir2_sf_getdents(struct xfs_inode *dp, void *dirent,
+				xfs_off_t *offset, filldir_t filldir);
 extern int xfs_dir2_sf_lookup(struct xfs_da_args *args);
 extern int xfs_dir2_sf_removename(struct xfs_da_args *args);
 extern int xfs_dir2_sf_replace(struct xfs_da_args *args);

fs/xfs/xfs_dmapi.h

@@ -67,17 +67,15 @@ typedef enum {
 #define HAVE_DM_RIGHT_T
 
 /* Defines for determining if an event message should be sent. */
-#define	DM_EVENT_ENABLED(vfsp, ip, event) ( \
-	unlikely ((vfsp)->vfs_flag & VFS_DMI) && \
+#ifdef HAVE_DMAPI
+#define	DM_EVENT_ENABLED(ip, event) ( \
+	unlikely ((ip)->i_mount->m_flags & XFS_MOUNT_DMAPI) && \
 		( ((ip)->i_d.di_dmevmask & (1 << event)) || \
 		  ((ip)->i_mount->m_dmevmask & (1 << event)) ) \
 	)
-
-#define	DM_EVENT_ENABLED_IO(vfsp, io, event) ( \
-	unlikely ((vfsp)->vfs_flag & VFS_DMI) && \
-		( ((io)->io_dmevmask & (1 << event)) || \
-		  ((io)->io_mount->m_dmevmask & (1 << event)) ) \
-	)
+#else
+#define DM_EVENT_ENABLED(ip, event)	(0)
+#endif
 
 #define DM_XFS_VALID_FS_EVENTS		( \
 	(1 << DM_EVENT_PREUNMOUNT)	| \
@@ -170,7 +168,4 @@ typedef enum {
 			DM_FLAGS_NDELAY : 0)
 #define AT_DELAY_FLAG(f) ((f&ATTR_NONBLOCK) ? DM_FLAGS_NDELAY : 0)
 
-
-extern struct bhv_module_vfsops xfs_dmops;
-
 #endif  /* __XFS_DMAPI_H__ */

fs/xfs/xfs_dmops.c

@@ -19,18 +19,51 @@
 #include "xfs_fs.h"
 #include "xfs_types.h"
 #include "xfs_log.h"
-#include "xfs_inum.h"
 #include "xfs_trans.h"
 #include "xfs_sb.h"
-#include "xfs_ag.h"
-#include "xfs_dir2.h"
 #include "xfs_dmapi.h"
+#include "xfs_inum.h"
+#include "xfs_ag.h"
 #include "xfs_mount.h"
+#include "xfs_clnt.h"
+
 
-xfs_dmops_t	xfs_dmcore_stub = {
+static struct xfs_dmops xfs_dmcore_stub = {
 	.xfs_send_data		= (xfs_send_data_t)fs_nosys,
 	.xfs_send_mmap		= (xfs_send_mmap_t)fs_noerr,
 	.xfs_send_destroy	= (xfs_send_destroy_t)fs_nosys,
 	.xfs_send_namesp	= (xfs_send_namesp_t)fs_nosys,
-	.xfs_send_unmount	= (xfs_send_unmount_t)fs_noval,
+	.xfs_send_mount		= (xfs_send_mount_t)fs_nosys,
+	.xfs_send_unmount	= (xfs_send_unmount_t)fs_noerr,
 };
+
+int
+xfs_dmops_get(struct xfs_mount *mp, struct xfs_mount_args *args)
+{
+	if (args->flags & XFSMNT_DMAPI) {
+		struct xfs_dmops *ops;
+
+		ops = symbol_get(xfs_dmcore_xfs);
+		if (!ops) {
+			request_module("xfs_dmapi");
+			ops = symbol_get(xfs_dmcore_xfs);
+		}
+
+		if (!ops) {
+			cmn_err(CE_WARN, "XFS: no dmapi support available.");
+			return EINVAL;
+		}
+		mp->m_dm_ops = ops;
+	} else {
+		mp->m_dm_ops = &xfs_dmcore_stub;
+	}
+
+	return 0;
+}
+
+void
+xfs_dmops_put(struct xfs_mount *mp)
+{
+	if (mp->m_dm_ops != &xfs_dmcore_stub)
+		symbol_put(xfs_dmcore_xfs);
+}

fs/xfs/xfs_error.c

@@ -22,6 +22,7 @@
 #include "xfs_inum.h"
 #include "xfs_trans.h"
 #include "xfs_sb.h"
+#include "xfs_ag.h"
 #include "xfs_dir2.h"
 #include "xfs_dmapi.h"
 #include "xfs_mount.h"
@@ -132,10 +133,14 @@ xfs_errortag_add(int error_tag, xfs_mount_t *mp)
 }
 
 int
-xfs_errortag_clearall_umount(int64_t fsid, char *fsname, int loud)
+xfs_errortag_clearall(xfs_mount_t *mp, int loud)
 {
-	int i;
+	int64_t fsid;
 	int cleared = 0;
+	int i;
+
+	memcpy(&fsid, mp->m_fixedfsid, sizeof(xfs_fsid_t));
+
 
 	for (i = 0; i < XFS_NUM_INJECT_ERROR; i++) {
 		if ((fsid == 0LL || xfs_etest_fsid[i] == fsid) &&
@@ -154,20 +159,10 @@ xfs_errortag_clearall_umount(int64_t fsid, char *fsname, int loud)
 	if (loud || cleared)
 		cmn_err(CE_WARN,
 			"Cleared all XFS error tags for filesystem \"%s\"",
-			fsname);
+			mp->m_fsname);
 
 	return 0;
 }
-
-int
-xfs_errortag_clearall(xfs_mount_t *mp)
-{
-	int64_t fsid;
-
-	memcpy(&fsid, mp->m_fixedfsid, sizeof(xfs_fsid_t));
-
-	return xfs_errortag_clearall_umount(fsid, mp->m_fsname, 1);
-}
 #endif /* DEBUG || INDUCE_IO_ERROR */
 
 static void

fs/xfs/xfs_error.h

@@ -144,12 +144,11 @@ extern void xfs_error_test_init(void);
 #endif /* __ANSI_CPP__ */
 
 extern int xfs_errortag_add(int error_tag, xfs_mount_t *mp);
-extern int xfs_errortag_clearall(xfs_mount_t *mp);
-extern int xfs_errortag_clearall_umount(int64_t fsid, char *fsname, int loud);
+extern int xfs_errortag_clearall(xfs_mount_t *mp, int loud);
 #else
 #define XFS_TEST_ERROR(expr, mp, tag, rf)	(expr)
 #define xfs_errortag_add(tag, mp)		(ENOSYS)
-#define xfs_errortag_clearall(mp)		(ENOSYS)
+#define xfs_errortag_clearall(mp, loud)		(ENOSYS)
 #endif /* (DEBUG || INDUCE_IO_ERROR) */
 
 /*

fs/xfs/xfs_extfree_item.c

@@ -23,6 +23,7 @@
 #include "xfs_trans.h"
 #include "xfs_buf_item.h"
 #include "xfs_sb.h"
+#include "xfs_ag.h"
 #include "xfs_dmapi.h"
 #include "xfs_mount.h"
 #include "xfs_trans_priv.h"

fs/xfs/xfs_fsops.c

@@ -136,7 +136,6 @@ xfs_growfs_data_private(
 	xfs_rfsblock_t		nfree;
 	xfs_agnumber_t		oagcount;
 	int			pct;
-	xfs_sb_t		*sbp;
 	xfs_trans_t		*tp;
 
 	nb = in->newblocks;
@@ -175,7 +174,7 @@ xfs_growfs_data_private(
 		memset(&mp->m_perag[oagcount], 0,
 			(nagcount - oagcount) * sizeof(xfs_perag_t));
 		mp->m_flags |= XFS_MOUNT_32BITINODES;
-		nagimax = xfs_initialize_perag(XFS_MTOVFS(mp), mp, nagcount);
+		nagimax = xfs_initialize_perag(mp, nagcount);
 		up_write(&mp->m_peraglock);
 	}
 	tp = xfs_trans_alloc(mp, XFS_TRANS_GROWFS);
@@ -377,8 +376,7 @@ xfs_growfs_data_private(
 				error, agno);
 			break;
 		}
-		sbp = XFS_BUF_TO_SBP(bp);
-		xfs_xlatesb(sbp, &mp->m_sb, -1, XFS_SB_ALL_BITS);
+		xfs_sb_to_disk(XFS_BUF_TO_SBP(bp), &mp->m_sb, XFS_SB_ALL_BITS);
 		/*
 		 * If we get an error writing out the alternate superblocks,
 		 * just issue a warning and continue.  The real work is
@@ -435,10 +433,10 @@ xfs_growfs_data(
 	xfs_growfs_data_t	*in)
 {
 	int error;
-	if (!cpsema(&mp->m_growlock))
+	if (!mutex_trylock(&mp->m_growlock))
 		return XFS_ERROR(EWOULDBLOCK);
 	error = xfs_growfs_data_private(mp, in);
-	vsema(&mp->m_growlock);
+	mutex_unlock(&mp->m_growlock);
 	return error;
 }
 
@@ -448,10 +446,10 @@ xfs_growfs_log(
 	xfs_growfs_log_t	*in)
 {
 	int error;
-	if (!cpsema(&mp->m_growlock))
+	if (!mutex_trylock(&mp->m_growlock))
 		return XFS_ERROR(EWOULDBLOCK);
 	error = xfs_growfs_log_private(mp, in);
-	vsema(&mp->m_growlock);
+	mutex_unlock(&mp->m_growlock);
 	return error;
 }
 
@@ -628,8 +626,7 @@ xfs_fs_goingdown(
 {
 	switch (inflags) {
 	case XFS_FSOP_GOING_FLAGS_DEFAULT: {
-		struct bhv_vfs *vfsp = XFS_MTOVFS(mp);
-		struct super_block *sb = freeze_bdev(vfsp->vfs_super->s_bdev);
+		struct super_block *sb = freeze_bdev(mp->m_super->s_bdev);
 
 		if (sb && !IS_ERR(sb)) {
 			xfs_force_shutdown(mp, SHUTDOWN_FORCE_UMOUNT);

fs/xfs/xfs_ialloc.c

@@ -293,9 +293,9 @@ xfs_ialloc_ag_alloc(
 		xfs_biozero(fbuf, 0, ninodes << args.mp->m_sb.sb_inodelog);
 		for (i = 0; i < ninodes; i++) {
 			free = XFS_MAKE_IPTR(args.mp, fbuf, i);
-			INT_SET(free->di_core.di_magic, ARCH_CONVERT, XFS_DINODE_MAGIC);
-			INT_SET(free->di_core.di_version, ARCH_CONVERT, version);
-			INT_SET(free->di_next_unlinked, ARCH_CONVERT, NULLAGINO);
+			free->di_core.di_magic = cpu_to_be16(XFS_DINODE_MAGIC);
+			free->di_core.di_version = version;
+			free->di_next_unlinked = cpu_to_be32(NULLAGINO);
 			xfs_ialloc_log_di(tp, fbuf, i,
 				XFS_DI_CORE_BITS | XFS_DI_NEXT_UNLINKED);
 		}

fs/xfs/xfs_ialloc.h

@@ -30,14 +30,9 @@ struct xfs_trans;
 #define	XFS_IALLOC_BLOCKS(mp)	(mp)->m_ialloc_blks
 
 /*
- * For small block file systems, move inodes in clusters of this size.
- * When we don't have a lot of memory, however, we go a bit smaller
- * to reduce the number of AGI and ialloc btree blocks we need to keep
- * around for xfs_dilocate().  We choose which one to use in
- * xfs_mount_int().
+ * Move inodes in clusters of this size.
  */
 #define	XFS_INODE_BIG_CLUSTER_SIZE	8192
-#define	XFS_INODE_SMALL_CLUSTER_SIZE	4096
 #define	XFS_INODE_CLUSTER_SIZE(mp)	(mp)->m_inode_cluster_size
 
 /*

fs/xfs/xfs_iget.c

@@ -39,132 +39,14 @@
 #include "xfs_quota.h"
 #include "xfs_utils.h"
 
-/*
- * Initialize the inode hash table for the newly mounted file system.
- * Choose an initial table size based on user specified value, else
- * use a simple algorithm using the maximum number of inodes as an
- * indicator for table size, and clamp it between one and some large
- * number of pages.
- */
-void
-xfs_ihash_init(xfs_mount_t *mp)
-{
-	__uint64_t	icount;
-	uint		i;
-
-	if (!mp->m_ihsize) {
-		icount = mp->m_maxicount ? mp->m_maxicount :
-			 (mp->m_sb.sb_dblocks << mp->m_sb.sb_inopblog);
-		mp->m_ihsize = 1 << max_t(uint, 8,
-					(xfs_highbit64(icount) + 1) / 2);
-		mp->m_ihsize = min_t(uint, mp->m_ihsize,
-					(64 * NBPP) / sizeof(xfs_ihash_t));
-	}
-
-	mp->m_ihash = kmem_zalloc_greedy(&mp->m_ihsize,
-					 NBPC * sizeof(xfs_ihash_t),
-					 mp->m_ihsize * sizeof(xfs_ihash_t),
-					 KM_SLEEP | KM_MAYFAIL | KM_LARGE);
-	mp->m_ihsize /= sizeof(xfs_ihash_t);
-	for (i = 0; i < mp->m_ihsize; i++)
-		rwlock_init(&(mp->m_ihash[i].ih_lock));
-}
-
-/*
- * Free up structures allocated by xfs_ihash_init, at unmount time.
- */
-void
-xfs_ihash_free(xfs_mount_t *mp)
-{
-	kmem_free(mp->m_ihash, mp->m_ihsize * sizeof(xfs_ihash_t));
-	mp->m_ihash = NULL;
-}
-
-/*
- * Initialize the inode cluster hash table for the newly mounted file system.
- * Its size is derived from the ihash table size.
- */
-void
-xfs_chash_init(xfs_mount_t *mp)
-{
-	uint	i;
-
-	mp->m_chsize = max_t(uint, 1, mp->m_ihsize /
-			 (XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_inodelog));
-	mp->m_chsize = min_t(uint, mp->m_chsize, mp->m_ihsize);
-	mp->m_chash = (xfs_chash_t *)kmem_zalloc(mp->m_chsize
-						 * sizeof(xfs_chash_t),
-						 KM_SLEEP | KM_LARGE);
-	for (i = 0; i < mp->m_chsize; i++) {
-		spinlock_init(&mp->m_chash[i].ch_lock,"xfshash");
-	}
-}
-
-/*
- * Free up structures allocated by xfs_chash_init, at unmount time.
- */
-void
-xfs_chash_free(xfs_mount_t *mp)
-{
-	int	i;
-
-	for (i = 0; i < mp->m_chsize; i++) {
-		spinlock_destroy(&mp->m_chash[i].ch_lock);
-	}
-
-	kmem_free(mp->m_chash, mp->m_chsize*sizeof(xfs_chash_t));
-	mp->m_chash = NULL;
-}
-
-/*
- * Try to move an inode to the front of its hash list if possible
- * (and if its not there already).  Called right after obtaining
- * the list version number and then dropping the read_lock on the
- * hash list in question (which is done right after looking up the
- * inode in question...).
- */
-STATIC void
-xfs_ihash_promote(
-	xfs_ihash_t	*ih,
-	xfs_inode_t	*ip,
-	ulong		version)
-{
-	xfs_inode_t	*iq;
-
-	if ((ip->i_prevp != &ih->ih_next) && write_trylock(&ih->ih_lock)) {
-		if (likely(version == ih->ih_version)) {
-			/* remove from list */
-			if ((iq = ip->i_next)) {
-				iq->i_prevp = ip->i_prevp;
-			}
-			*ip->i_prevp = iq;
-
-			/* insert at list head */
-			iq = ih->ih_next;
-			iq->i_prevp = &ip->i_next;
-			ip->i_next = iq;
-			ip->i_prevp = &ih->ih_next;
-			ih->ih_next = ip;
-		}
-		write_unlock(&ih->ih_lock);
-	}
-}
-
 /*
  * Look up an inode by number in the given file system.
- * The inode is looked up in the hash table for the file system
- * represented by the mount point parameter mp.  Each bucket of
- * the hash table is guarded by an individual semaphore.
- *
- * If the inode is found in the hash table, its corresponding vnode
- * is obtained with a call to vn_get().  This call takes care of
- * coordination with the reclamation of the inode and vnode.  Note
- * that the vmap structure is filled in while holding the hash lock.
- * This gives us the state of the inode/vnode when we found it and
- * is used for coordination in vn_get().
+ * The inode is looked up in the cache held in each AG.
+ * If the inode is found in the cache, attach it to the provided
+ * vnode.
  *
- * If it is not in core, read it in from the file system's device and
- * add the inode into the hash table.
+ * If it is not in core, read it in from the file system's device,
+ * add it to the cache and attach the provided vnode.
  *
  * The inode is locked according to the value of the lock_flags parameter.
  * This flag parameter indicates how and if the inode's IO lock and inode lock
@@ -192,274 +74,241 @@ xfs_iget_core(
 	xfs_inode_t	**ipp,
 	xfs_daddr_t	bno)
 {
-	xfs_ihash_t	*ih;
 	xfs_inode_t	*ip;
 	xfs_inode_t	*iq;
 	bhv_vnode_t	*inode_vp;
-	ulong		version;
 	int		error;
-	/* REFERENCED */
-	xfs_chash_t	*ch;
-	xfs_chashlist_t	*chl, *chlnew;
-	SPLDECL(s);
+	xfs_icluster_t	*icl, *new_icl = NULL;
+	unsigned long	first_index, mask;
+	xfs_perag_t	*pag;
+	xfs_agino_t	agino;
+
+	/* the radix tree exists only in inode capable AGs */
+	if (XFS_INO_TO_AGNO(mp, ino) >= mp->m_maxagi)
+		return EINVAL;
+
+	/* get the perag structure and ensure that it's inode capable */
+	pag = xfs_get_perag(mp, ino);
+	if (!pag->pagi_inodeok)
+		return EINVAL;
+	ASSERT(pag->pag_ici_init);
+	agino = XFS_INO_TO_AGINO(mp, ino);
 
+again:
+	read_lock(&pag->pag_ici_lock);
+	ip = radix_tree_lookup(&pag->pag_ici_root, agino);
 
-	ih = XFS_IHASH(mp, ino);
+	if (ip != NULL) {
+		/*
+		 * If INEW is set this inode is being set up
+		 * we need to pause and try again.
+		 */
+		if (xfs_iflags_test(ip, XFS_INEW)) {
+			read_unlock(&pag->pag_ici_lock);
+			delay(1);
+			XFS_STATS_INC(xs_ig_frecycle);
 
-again:
-	read_lock(&ih->ih_lock);
+			goto again;
+		}
 
-	for (ip = ih->ih_next; ip != NULL; ip = ip->i_next) {
-		if (ip->i_ino == ino) {
+		inode_vp = XFS_ITOV_NULL(ip);
+		if (inode_vp == NULL) {
 			/*
-			 * If INEW is set this inode is being set up
+			 * If IRECLAIM is set this inode is
+			 * on its way out of the system,
 			 * we need to pause and try again.
 			 */
-			if (xfs_iflags_test(ip, XFS_INEW)) {
-				read_unlock(&ih->ih_lock);
+			if (xfs_iflags_test(ip, XFS_IRECLAIM)) {
+				read_unlock(&pag->pag_ici_lock);
 				delay(1);
 				XFS_STATS_INC(xs_ig_frecycle);
 
 				goto again;
 			}
+			ASSERT(xfs_iflags_test(ip, XFS_IRECLAIMABLE));
 
-			inode_vp = XFS_ITOV_NULL(ip);
-			if (inode_vp == NULL) {
-				/*
-				 * If IRECLAIM is set this inode is
-				 * on its way out of the system,
-				 * we need to pause and try again.
-				 */
-				if (xfs_iflags_test(ip, XFS_IRECLAIM)) {
-					read_unlock(&ih->ih_lock);
-					delay(1);
-					XFS_STATS_INC(xs_ig_frecycle);
-
-					goto again;
-				}
-				ASSERT(xfs_iflags_test(ip, XFS_IRECLAIMABLE));
-
-				/*
-				 * If lookup is racing with unlink, then we
-				 * should return an error immediately so we
-				 * don't remove it from the reclaim list and
-				 * potentially leak the inode.
-				 */
-				if ((ip->i_d.di_mode == 0) &&
-				    !(flags & XFS_IGET_CREATE)) {
-					read_unlock(&ih->ih_lock);
-					return ENOENT;
-				}
-
-				/*
-				 * There may be transactions sitting in the
-				 * incore log buffers or being flushed to disk
-				 * at this time.  We can't clear the
-				 * XFS_IRECLAIMABLE flag until these
-				 * transactions have hit the disk, otherwise we
-				 * will void the guarantee the flag provides
-				 * xfs_iunpin()
-				 */
-				if (xfs_ipincount(ip)) {
-					read_unlock(&ih->ih_lock);
-					xfs_log_force(mp, 0,
-						XFS_LOG_FORCE|XFS_LOG_SYNC);
-					XFS_STATS_INC(xs_ig_frecycle);
-					goto again;
-				}
-
-				vn_trace_exit(vp, "xfs_iget.alloc",
-					(inst_t *)__return_address);
+			/*
+			 * If lookup is racing with unlink, then we
+			 * should return an error immediately so we
+			 * don't remove it from the reclaim list and
+			 * potentially leak the inode.
+			 */
+			if ((ip->i_d.di_mode == 0) &&
+			    !(flags & XFS_IGET_CREATE)) {
+				read_unlock(&pag->pag_ici_lock);
+				xfs_put_perag(mp, pag);
+				return ENOENT;
+			}
+
+			/*
+			 * There may be transactions sitting in the
+			 * incore log buffers or being flushed to disk
+			 * at this time.  We can't clear the
+			 * XFS_IRECLAIMABLE flag until these
+			 * transactions have hit the disk, otherwise we
+			 * will void the guarantee the flag provides
+			 * xfs_iunpin()
+			 */
+			if (xfs_ipincount(ip)) {
+				read_unlock(&pag->pag_ici_lock);
+				xfs_log_force(mp, 0,
+					XFS_LOG_FORCE|XFS_LOG_SYNC);
+				XFS_STATS_INC(xs_ig_frecycle);
+				goto again;
+			}
 
-				XFS_STATS_INC(xs_ig_found);
+			vn_trace_exit(ip, "xfs_iget.alloc",
+				(inst_t *)__return_address);
 
-				xfs_iflags_clear(ip, XFS_IRECLAIMABLE);
-				version = ih->ih_version;
-				read_unlock(&ih->ih_lock);
-				xfs_ihash_promote(ih, ip, version);
+			XFS_STATS_INC(xs_ig_found);
 
-				XFS_MOUNT_ILOCK(mp);
-				list_del_init(&ip->i_reclaim);
-				XFS_MOUNT_IUNLOCK(mp);
+			xfs_iflags_clear(ip, XFS_IRECLAIMABLE);
+			read_unlock(&pag->pag_ici_lock);
 
-				goto finish_inode;
+			XFS_MOUNT_ILOCK(mp);
+			list_del_init(&ip->i_reclaim);
+			XFS_MOUNT_IUNLOCK(mp);
 
-			} else if (vp != inode_vp) {
-				struct inode *inode = vn_to_inode(inode_vp);
+			goto finish_inode;
 
-				/* The inode is being torn down, pause and
-				 * try again.
-				 */
-				if (inode->i_state & (I_FREEING | I_CLEAR)) {
-					read_unlock(&ih->ih_lock);
-					delay(1);
-					XFS_STATS_INC(xs_ig_frecycle);
+		} else if (vp != inode_vp) {
+			struct inode *inode = vn_to_inode(inode_vp);
 
-					goto again;
-				}
-/* Chances are the other vnode (the one in the inode) is being torn
- * down right now, and we landed on top of it. Question is, what do
- * we do? Unhook the old inode and hook up the new one?
- */
-				cmn_err(CE_PANIC,
-			"xfs_iget_core: ambiguous vns: vp/0x%p, invp/0x%p",
-						inode_vp, vp);
+			/* The inode is being torn down, pause and
+			 * try again.
+			 */
+			if (inode->i_state & (I_FREEING | I_CLEAR)) {
+				read_unlock(&pag->pag_ici_lock);
+				delay(1);
+				XFS_STATS_INC(xs_ig_frecycle);
+
+				goto again;
 			}
+/* Chances are the other vnode (the one in the inode) is being torn
+* down right now, and we landed on top of it. Question is, what do
+* we do? Unhook the old inode and hook up the new one?
+*/
+			cmn_err(CE_PANIC,
+		"xfs_iget_core: ambiguous vns: vp/0x%p, invp/0x%p",
+					inode_vp, vp);
+		}
 
-			/*
-			 * Inode cache hit: if ip is not at the front of
-			 * its hash chain, move it there now.
-			 * Do this with the lock held for update, but
-			 * do statistics after releasing the lock.
-			 */
-			version = ih->ih_version;
-			read_unlock(&ih->ih_lock);
-			xfs_ihash_promote(ih, ip, version);
-			XFS_STATS_INC(xs_ig_found);
+		/*
+		 * Inode cache hit
+		 */
+		read_unlock(&pag->pag_ici_lock);
+		XFS_STATS_INC(xs_ig_found);
 
 finish_inode:
-			if (ip->i_d.di_mode == 0) {
-				if (!(flags & XFS_IGET_CREATE))
-					return ENOENT;
-				xfs_iocore_inode_reinit(ip);
+		if (ip->i_d.di_mode == 0) {
+			if (!(flags & XFS_IGET_CREATE)) {
+				xfs_put_perag(mp, pag);
+				return ENOENT;
 			}
+			xfs_iocore_inode_reinit(ip);
+		}
 
-			if (lock_flags != 0)
-				xfs_ilock(ip, lock_flags);
+		if (lock_flags != 0)
+			xfs_ilock(ip, lock_flags);
 
-			xfs_iflags_clear(ip, XFS_ISTALE);
-			vn_trace_exit(vp, "xfs_iget.found",
-						(inst_t *)__return_address);
-			goto return_ip;
-		}
+		xfs_iflags_clear(ip, XFS_ISTALE);
+		vn_trace_exit(ip, "xfs_iget.found",
+					(inst_t *)__return_address);
+		goto return_ip;
 	}
 
 	/*
-	 * Inode cache miss: save the hash chain version stamp and unlock
-	 * the chain, so we don't deadlock in vn_alloc.
+	 * Inode cache miss
 	 */
+	read_unlock(&pag->pag_ici_lock);
 	XFS_STATS_INC(xs_ig_missed);
 
-	version = ih->ih_version;
-
-	read_unlock(&ih->ih_lock);
-
 	/*
 	 * Read the disk inode attributes into a new inode structure and get
 	 * a new vnode for it. This should also initialize i_ino and i_mount.
 	 */
 	error = xfs_iread(mp, tp, ino, &ip, bno,
 			  (flags & XFS_IGET_BULKSTAT) ? XFS_IMAP_BULKSTAT : 0);
-	if (error)
+	if (error) {
+		xfs_put_perag(mp, pag);
 		return error;
+	}
 
-	vn_trace_exit(vp, "xfs_iget.alloc", (inst_t *)__return_address);
+	vn_trace_exit(ip, "xfs_iget.alloc", (inst_t *)__return_address);
 
 	xfs_inode_lock_init(ip, vp);
 	xfs_iocore_inode_init(ip);
-
 	if (lock_flags)
 		xfs_ilock(ip, lock_flags);
 
 	if ((ip->i_d.di_mode == 0) && !(flags & XFS_IGET_CREATE)) {
 		xfs_idestroy(ip);
+		xfs_put_perag(mp, pag);
 		return ENOENT;
 	}
 
 	/*
-	 * Put ip on its hash chain, unless someone else hashed a duplicate
-	 * after we released the hash lock.
+	 * This is a bit messy - we preallocate everything we _might_
+	 * need before we pick up the ici lock. That way we don't have to
+	 * juggle locks and go all the way back to the start.
 	 */
-	write_lock(&ih->ih_lock);
+	new_icl = kmem_zone_alloc(xfs_icluster_zone, KM_SLEEP);
+	if (radix_tree_preload(GFP_KERNEL)) {
+		delay(1);
+		goto again;
+	}
+	mask = ~(((XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_inodelog)) - 1);
+	first_index = agino & mask;
+	write_lock(&pag->pag_ici_lock);
 
-	if (ih->ih_version != version) {
-		for (iq = ih->ih_next; iq != NULL; iq = iq->i_next) {
-			if (iq->i_ino == ino) {
-				write_unlock(&ih->ih_lock);
-				xfs_idestroy(ip);
+	/*
+	 * Find the cluster if it exists
+	 */
+	icl = NULL;
+	if (radix_tree_gang_lookup(&pag->pag_ici_root, (void**)&iq,
+							first_index, 1)) {
+		if ((iq->i_ino & mask) == first_index)
+			icl = iq->i_cluster;
+	}
 
-				XFS_STATS_INC(xs_ig_dup);
-				goto again;
-			}
-		}
+	/*
+	 * insert the new inode
+	 */
+	error = radix_tree_insert(&pag->pag_ici_root, agino, ip);
+	if (unlikely(error)) {
+		BUG_ON(error != -EEXIST);
+		write_unlock(&pag->pag_ici_lock);
+		radix_tree_preload_end();
+		xfs_idestroy(ip);
+		XFS_STATS_INC(xs_ig_dup);
+		goto again;
 	}
 
 	/*
 	 * These values _must_ be set before releasing ihlock!
 	 */
-	ip->i_hash = ih;
-	if ((iq = ih->ih_next)) {
-		iq->i_prevp = &ip->i_next;
-	}
-	ip->i_next = iq;
-	ip->i_prevp = &ih->ih_next;
-	ih->ih_next = ip;
 	ip->i_udquot = ip->i_gdquot = NULL;
-	ih->ih_version++;
 	xfs_iflags_set(ip, XFS_INEW);
-	write_unlock(&ih->ih_lock);
 
-	/*
-	 * put ip on its cluster's hash chain
-	 */
-	ASSERT(ip->i_chash == NULL && ip->i_cprev == NULL &&
-	       ip->i_cnext == NULL);
-
-	chlnew = NULL;
-	ch = XFS_CHASH(mp, ip->i_blkno);
- chlredo:
-	s = mutex_spinlock(&ch->ch_lock);
-	for (chl = ch->ch_list; chl != NULL; chl = chl->chl_next) {
-		if (chl->chl_blkno == ip->i_blkno) {
-
-			/* insert this inode into the doubly-linked list
-			 * where chl points */
-			if ((iq = chl->chl_ip)) {
-				ip->i_cprev = iq->i_cprev;
-				iq->i_cprev->i_cnext = ip;
-				iq->i_cprev = ip;
-				ip->i_cnext = iq;
-			} else {
-				ip->i_cnext = ip;
-				ip->i_cprev = ip;
-			}
-			chl->chl_ip = ip;
-			ip->i_chash = chl;
-			break;
-		}
-	}
+	ASSERT(ip->i_cluster == NULL);
 
-	/* no hash list found for this block; add a new hash list */
-	if (chl == NULL)  {
-		if (chlnew == NULL) {
-			mutex_spinunlock(&ch->ch_lock, s);
-			ASSERT(xfs_chashlist_zone != NULL);
-			chlnew = (xfs_chashlist_t *)
-					kmem_zone_alloc(xfs_chashlist_zone,
-						KM_SLEEP);
-			ASSERT(chlnew != NULL);
-			goto chlredo;
-		} else {
-			ip->i_cnext = ip;
-			ip->i_cprev = ip;
-			ip->i_chash = chlnew;
-			chlnew->chl_ip = ip;
-			chlnew->chl_blkno = ip->i_blkno;
-			if (ch->ch_list)
-				ch->ch_list->chl_prev = chlnew;
-			chlnew->chl_next = ch->ch_list;
-			chlnew->chl_prev = NULL;
-			ch->ch_list = chlnew;
-			chlnew = NULL;
-		}
+	if (!icl) {
+		spin_lock_init(&new_icl->icl_lock);
+		INIT_HLIST_HEAD(&new_icl->icl_inodes);
+		icl = new_icl;
+		new_icl = NULL;
 	} else {
-		if (chlnew != NULL) {
-			kmem_zone_free(xfs_chashlist_zone, chlnew);
-		}
+		ASSERT(!hlist_empty(&icl->icl_inodes));
 	}
+	spin_lock(&icl->icl_lock);
+	hlist_add_head(&ip->i_cnode, &icl->icl_inodes);
+	ip->i_cluster = icl;
+	spin_unlock(&icl->icl_lock);
 
-	mutex_spinunlock(&ch->ch_lock, s);
-
+	write_unlock(&pag->pag_ici_lock);
+	radix_tree_preload_end();
+	if (new_icl)
+		kmem_zone_free(xfs_icluster_zone, new_icl);
 
 	/*
 	 * Link ip to its mount and thread it on the mount's inode list.
@@ -478,6 +327,7 @@ finish_inode:
 	mp->m_inodes = ip;
 
 	XFS_MOUNT_IUNLOCK(mp);
+	xfs_put_perag(mp, pag);
 
  return_ip:
 	ASSERT(ip->i_df.if_ext_max ==
@@ -486,14 +336,14 @@ finish_inode:
 	ASSERT(((ip->i_d.di_flags & XFS_DIFLAG_REALTIME) != 0) ==
 	       ((ip->i_iocore.io_flags & XFS_IOCORE_RT) != 0));
 
+	xfs_iflags_set(ip, XFS_IMODIFIED);
 	*ipp = 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.
 	 */
-	bhv_vfs_init_vnode(XFS_MTOVFS(mp), vp, XFS_ITOBHV(ip), 1);
-
+	xfs_initialize_vnode(mp, vp, ip);
 	return 0;
 }
 
@@ -519,7 +369,8 @@ xfs_iget(
 	XFS_STATS_INC(xs_ig_attempts);
 
 retry:
-	if ((inode = iget_locked(XFS_MTOVFS(mp)->vfs_super, ino))) {
+	inode = iget_locked(mp->m_super, ino);
+	if (inode) {
 		xfs_inode_t	*ip;
 
 		vp = vn_from_inode(inode);
@@ -570,8 +421,8 @@ xfs_inode_lock_init(
 	bhv_vnode_t	*vp)
 {
 	mrlock_init(&ip->i_lock, MRLOCK_ALLOW_EQUAL_PRI|MRLOCK_BARRIER,
-		     "xfsino", (long)vp->v_number);
-	mrlock_init(&ip->i_iolock, MRLOCK_BARRIER, "xfsio", vp->v_number);
+		     "xfsino", ip->i_ino);
+	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");
@@ -587,32 +438,19 @@ xfs_inode_incore(xfs_mount_t	*mp,
 		 xfs_ino_t	ino,
 		 xfs_trans_t	*tp)
 {
-	xfs_ihash_t	*ih;
 	xfs_inode_t	*ip;
-	ulong		version;
-
-	ih = XFS_IHASH(mp, ino);
-	read_lock(&ih->ih_lock);
-	for (ip = ih->ih_next; ip != NULL; ip = ip->i_next) {
-		if (ip->i_ino == ino) {
-			/*
-			 * If we find it and tp matches, return it.
-			 * Also move it to the front of the hash list
-			 * if we find it and it is not already there.
-			 * Otherwise break from the loop and return
-			 * NULL.
-			 */
-			if (ip->i_transp == tp) {
-				version = ih->ih_version;
-				read_unlock(&ih->ih_lock);
-				xfs_ihash_promote(ih, ip, version);
-				return (ip);
-			}
-			break;
-		}
-	}
-	read_unlock(&ih->ih_lock);
-	return (NULL);
+	xfs_perag_t	*pag;
+
+	pag = xfs_get_perag(mp, ino);
+	read_lock(&pag->pag_ici_lock);
+	ip = radix_tree_lookup(&pag->pag_ici_root, XFS_INO_TO_AGINO(mp, ino));
+	read_unlock(&pag->pag_ici_lock);
+	xfs_put_perag(mp, pag);
+
+	/* the returned inode must match the transaction */
+	if (ip && (ip->i_transp != tp))
+		return NULL;
+	return ip;
 }
 
 /*
@@ -629,7 +467,7 @@ xfs_iput(xfs_inode_t	*ip,
 {
 	bhv_vnode_t	*vp = XFS_ITOV(ip);
 
-	vn_trace_entry(vp, "xfs_iput", (inst_t *)__return_address);
+	vn_trace_entry(ip, "xfs_iput", (inst_t *)__return_address);
 	xfs_iunlock(ip, lock_flags);
 	VN_RELE(vp);
 }
@@ -644,7 +482,7 @@ xfs_iput_new(xfs_inode_t	*ip,
 	bhv_vnode_t	*vp = XFS_ITOV(ip);
 	struct inode	*inode = vn_to_inode(vp);
 
-	vn_trace_entry(vp, "xfs_iput_new", (inst_t *)__return_address);
+	vn_trace_entry(ip, "xfs_iput_new", (inst_t *)__return_address);
 
 	if ((ip->i_d.di_mode == 0)) {
 		ASSERT(!xfs_iflags_test(ip, XFS_IRECLAIMABLE));
@@ -699,7 +537,8 @@ xfs_ireclaim(xfs_inode_t *ip)
 	 */
 	vp = XFS_ITOV_NULL(ip);
 	if (vp) {
-		vn_bhv_remove(VN_BHV_HEAD(vp), XFS_ITOBHV(ip));
+		vn_to_inode(vp)->i_private = NULL;
+		ip->i_vnode = NULL;
 	}
 
 	/*
@@ -718,58 +557,26 @@ void
 xfs_iextract(
 	xfs_inode_t	*ip)
 {
-	xfs_ihash_t	*ih;
+	xfs_mount_t	*mp = ip->i_mount;
+	xfs_perag_t	*pag = xfs_get_perag(mp, ip->i_ino);
 	xfs_inode_t	*iq;
-	xfs_mount_t	*mp;
-	xfs_chash_t	*ch;
-	xfs_chashlist_t *chl, *chm;
-	SPLDECL(s);
-
-	ih = ip->i_hash;
-	write_lock(&ih->ih_lock);
-	if ((iq = ip->i_next)) {
-		iq->i_prevp = ip->i_prevp;
-	}
-	*ip->i_prevp = iq;
-	ih->ih_version++;
-	write_unlock(&ih->ih_lock);
+
+	write_lock(&pag->pag_ici_lock);
+	radix_tree_delete(&pag->pag_ici_root, XFS_INO_TO_AGINO(mp, ip->i_ino));
+	write_unlock(&pag->pag_ici_lock);
+	xfs_put_perag(mp, pag);
 
 	/*
-	 * Remove from cluster hash list
-	 *   1) delete the chashlist if this is the last inode on the chashlist
-	 *   2) unchain from list of inodes
-	 *   3) point chashlist->chl_ip to 'chl_next' if to this inode.
+	 * Remove from cluster list
 	 */
 	mp = ip->i_mount;
-	ch = XFS_CHASH(mp, ip->i_blkno);
-	s = mutex_spinlock(&ch->ch_lock);
-
-	if (ip->i_cnext == ip) {
-		/* Last inode on chashlist */
-		ASSERT(ip->i_cnext == ip && ip->i_cprev == ip);
-		ASSERT(ip->i_chash != NULL);
-		chm=NULL;
-		chl = ip->i_chash;
-		if (chl->chl_prev)
-			chl->chl_prev->chl_next = chl->chl_next;
-		else
-			ch->ch_list = chl->chl_next;
-		if (chl->chl_next)
-			chl->chl_next->chl_prev = chl->chl_prev;
-		kmem_zone_free(xfs_chashlist_zone, chl);
-	} else {
-		/* delete one inode from a non-empty list */
-		iq = ip->i_cnext;
-		iq->i_cprev = ip->i_cprev;
-		ip->i_cprev->i_cnext = iq;
-		if (ip->i_chash->chl_ip == ip) {
-			ip->i_chash->chl_ip = iq;
-		}
-		ip->i_chash = __return_address;
-		ip->i_cprev = __return_address;
-		ip->i_cnext = __return_address;
-	}
-	mutex_spinunlock(&ch->ch_lock, s);
+	spin_lock(&ip->i_cluster->icl_lock);
+	hlist_del(&ip->i_cnode);
+	spin_unlock(&ip->i_cluster->icl_lock);
+
+	/* was last inode in cluster? */
+	if (hlist_empty(&ip->i_cluster->icl_inodes))
+		kmem_zone_free(xfs_icluster_zone, ip->i_cluster);
 
 	/*
 	 * Remove from mount's inode list.

fs/xfs/xfs_inode.c

@@ -49,12 +49,11 @@
 #include "xfs_quota.h"
 #include "xfs_acl.h"
 #include "xfs_filestream.h"
-
-#include <linux/log2.h>
+#include "xfs_vnodeops.h"
 
 kmem_zone_t *xfs_ifork_zone;
 kmem_zone_t *xfs_inode_zone;
-kmem_zone_t *xfs_chashlist_zone;
+kmem_zone_t *xfs_icluster_zone;
 
 /*
  * Used in xfs_itruncate().  This is the maximum number of extents
@@ -67,7 +66,6 @@ STATIC int xfs_iformat_local(xfs_inode_t *, xfs_dinode_t *, int, int);
 STATIC int xfs_iformat_extents(xfs_inode_t *, xfs_dinode_t *, int);
 STATIC int xfs_iformat_btree(xfs_inode_t *, xfs_dinode_t *, int);
 
-
 #ifdef DEBUG
 /*
  * Make sure that the extents in the given memory buffer
@@ -77,28 +75,23 @@ STATIC void
 xfs_validate_extents(
 	xfs_ifork_t		*ifp,
 	int			nrecs,
-	int			disk,
 	xfs_exntfmt_t		fmt)
 {
-	xfs_bmbt_rec_t		*ep;
 	xfs_bmbt_irec_t		irec;
-	xfs_bmbt_rec_t		rec;
+	xfs_bmbt_rec_host_t	rec;
 	int			i;
 
 	for (i = 0; i < nrecs; i++) {
-		ep = xfs_iext_get_ext(ifp, i);
-		rec.l0 = get_unaligned((__uint64_t*)&ep->l0);
-		rec.l1 = get_unaligned((__uint64_t*)&ep->l1);
-		if (disk)
-			xfs_bmbt_disk_get_all(&rec, &irec);
-		else
-			xfs_bmbt_get_all(&rec, &irec);
+		xfs_bmbt_rec_host_t *ep = xfs_iext_get_ext(ifp, i);
+		rec.l0 = get_unaligned(&ep->l0);
+		rec.l1 = get_unaligned(&ep->l1);
+		xfs_bmbt_get_all(&rec, &irec);
 		if (fmt == XFS_EXTFMT_NOSTATE)
 			ASSERT(irec.br_state == XFS_EXT_NORM);
 	}
 }
 #else /* DEBUG */
-#define xfs_validate_extents(ifp, nrecs, disk, fmt)
+#define xfs_validate_extents(ifp, nrecs, fmt)
 #endif /* DEBUG */
 
 /*
@@ -201,8 +194,8 @@ xfs_inotobp(
 	}
 	dip = (xfs_dinode_t *)xfs_buf_offset(bp, 0);
 	di_ok =
-		INT_GET(dip->di_core.di_magic, ARCH_CONVERT) == XFS_DINODE_MAGIC &&
-		XFS_DINODE_GOOD_VERSION(INT_GET(dip->di_core.di_version, ARCH_CONVERT));
+		be16_to_cpu(dip->di_core.di_magic) == XFS_DINODE_MAGIC &&
+		XFS_DINODE_GOOD_VERSION(dip->di_core.di_version);
 	if (unlikely(XFS_TEST_ERROR(!di_ok, mp, XFS_ERRTAG_ITOBP_INOTOBP,
 			XFS_RANDOM_ITOBP_INOTOBP))) {
 		XFS_CORRUPTION_ERROR("xfs_inotobp", XFS_ERRLEVEL_LOW, mp, dip);
@@ -346,8 +339,8 @@ xfs_itobp(
 
 		dip = (xfs_dinode_t *)xfs_buf_offset(bp,
 					(i << mp->m_sb.sb_inodelog));
-		di_ok = INT_GET(dip->di_core.di_magic, ARCH_CONVERT) == XFS_DINODE_MAGIC &&
-			    XFS_DINODE_GOOD_VERSION(INT_GET(dip->di_core.di_version, ARCH_CONVERT));
+		di_ok = be16_to_cpu(dip->di_core.di_magic) == XFS_DINODE_MAGIC &&
+			    XFS_DINODE_GOOD_VERSION(dip->di_core.di_version);
 		if (unlikely(XFS_TEST_ERROR(!di_ok, mp,
 						XFS_ERRTAG_ITOBP_INOTOBP,
 						XFS_RANDOM_ITOBP_INOTOBP))) {
@@ -361,7 +354,7 @@ xfs_itobp(
 					"daddr %lld #%d (magic=%x)",
 				XFS_BUFTARG_NAME(mp->m_ddev_targp),
 				(unsigned long long)imap.im_blkno, i,
-				INT_GET(dip->di_core.di_magic, ARCH_CONVERT));
+				be16_to_cpu(dip->di_core.di_magic));
 #endif
 			XFS_CORRUPTION_ERROR("xfs_itobp", XFS_ERRLEVEL_HIGH,
 					     mp, dip);
@@ -407,27 +400,26 @@ xfs_iformat(
 		XFS_IFORK_DSIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t);
 	error = 0;
 
-	if (unlikely(
-	    INT_GET(dip->di_core.di_nextents, ARCH_CONVERT) +
-		INT_GET(dip->di_core.di_anextents, ARCH_CONVERT) >
-	    INT_GET(dip->di_core.di_nblocks, ARCH_CONVERT))) {
+	if (unlikely(be32_to_cpu(dip->di_core.di_nextents) +
+		     be16_to_cpu(dip->di_core.di_anextents) >
+		     be64_to_cpu(dip->di_core.di_nblocks))) {
 		xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount,
 			"corrupt dinode %Lu, extent total = %d, nblocks = %Lu.",
 			(unsigned long long)ip->i_ino,
-			(int)(INT_GET(dip->di_core.di_nextents, ARCH_CONVERT)
-			    + INT_GET(dip->di_core.di_anextents, ARCH_CONVERT)),
+			(int)(be32_to_cpu(dip->di_core.di_nextents) +
+			      be16_to_cpu(dip->di_core.di_anextents)),
 			(unsigned long long)
-			INT_GET(dip->di_core.di_nblocks, ARCH_CONVERT));
+				be64_to_cpu(dip->di_core.di_nblocks));
 		XFS_CORRUPTION_ERROR("xfs_iformat(1)", XFS_ERRLEVEL_LOW,
 				     ip->i_mount, dip);
 		return XFS_ERROR(EFSCORRUPTED);
 	}
 
-	if (unlikely(INT_GET(dip->di_core.di_forkoff, ARCH_CONVERT) > ip->i_mount->m_sb.sb_inodesize)) {
+	if (unlikely(dip->di_core.di_forkoff > ip->i_mount->m_sb.sb_inodesize)) {
 		xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount,
 			"corrupt dinode %Lu, forkoff = 0x%x.",
 			(unsigned long long)ip->i_ino,
-			(int)(INT_GET(dip->di_core.di_forkoff, ARCH_CONVERT)));
+			dip->di_core.di_forkoff);
 		XFS_CORRUPTION_ERROR("xfs_iformat(2)", XFS_ERRLEVEL_LOW,
 				     ip->i_mount, dip);
 		return XFS_ERROR(EFSCORRUPTED);
@@ -438,25 +430,25 @@ xfs_iformat(
 	case S_IFCHR:
 	case S_IFBLK:
 	case S_IFSOCK:
-		if (unlikely(INT_GET(dip->di_core.di_format, ARCH_CONVERT) != XFS_DINODE_FMT_DEV)) {
+		if (unlikely(dip->di_core.di_format != XFS_DINODE_FMT_DEV)) {
 			XFS_CORRUPTION_ERROR("xfs_iformat(3)", XFS_ERRLEVEL_LOW,
 					      ip->i_mount, dip);
 			return XFS_ERROR(EFSCORRUPTED);
 		}
 		ip->i_d.di_size = 0;
 		ip->i_size = 0;
-		ip->i_df.if_u2.if_rdev = INT_GET(dip->di_u.di_dev, ARCH_CONVERT);
+		ip->i_df.if_u2.if_rdev = be32_to_cpu(dip->di_u.di_dev);
 		break;
 
 	case S_IFREG:
 	case S_IFLNK:
 	case S_IFDIR:
-		switch (INT_GET(dip->di_core.di_format, ARCH_CONVERT)) {
+		switch (dip->di_core.di_format) {
 		case XFS_DINODE_FMT_LOCAL:
 			/*
 			 * no local regular files yet
 			 */
-			if (unlikely((INT_GET(dip->di_core.di_mode, ARCH_CONVERT) & S_IFMT) == S_IFREG)) {
+			if (unlikely((be16_to_cpu(dip->di_core.di_mode) & S_IFMT) == S_IFREG)) {
 				xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount,
 					"corrupt inode %Lu "
 					"(local format for regular file).",
@@ -467,7 +459,7 @@ xfs_iformat(
 				return XFS_ERROR(EFSCORRUPTED);
 			}
 
-			di_size = INT_GET(dip->di_core.di_size, ARCH_CONVERT);
+			di_size = be64_to_cpu(dip->di_core.di_size);
 			if (unlikely(di_size > XFS_DFORK_DSIZE(dip, ip->i_mount))) {
 				xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount,
 					"corrupt inode %Lu "
@@ -509,7 +501,7 @@ xfs_iformat(
 	ip->i_afp = kmem_zone_zalloc(xfs_ifork_zone, KM_SLEEP);
 	ip->i_afp->if_ext_max =
 		XFS_IFORK_ASIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t);
-	switch (INT_GET(dip->di_core.di_aformat, ARCH_CONVERT)) {
+	switch (dip->di_core.di_aformat) {
 	case XFS_DINODE_FMT_LOCAL:
 		atp = (xfs_attr_shortform_t *)XFS_DFORK_APTR(dip);
 		size = be16_to_cpu(atp->hdr.totsize);
@@ -602,7 +594,7 @@ xfs_iformat_extents(
 	xfs_dinode_t	*dip,
 	int		whichfork)
 {
-	xfs_bmbt_rec_t	*ep, *dp;
+	xfs_bmbt_rec_t	*dp;
 	xfs_ifork_t	*ifp;
 	int		nex;
 	int		size;
@@ -637,13 +629,11 @@ xfs_iformat_extents(
 	ifp->if_bytes = size;
 	if (size) {
 		dp = (xfs_bmbt_rec_t *) XFS_DFORK_PTR(dip, whichfork);
-		xfs_validate_extents(ifp, nex, 1, XFS_EXTFMT_INODE(ip));
+		xfs_validate_extents(ifp, nex, XFS_EXTFMT_INODE(ip));
 		for (i = 0; i < nex; i++, dp++) {
-			ep = xfs_iext_get_ext(ifp, i);
-			ep->l0 = INT_GET(get_unaligned((__uint64_t*)&dp->l0),
-								ARCH_CONVERT);
-			ep->l1 = INT_GET(get_unaligned((__uint64_t*)&dp->l1),
-								ARCH_CONVERT);
+			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));
 		}
 		XFS_BMAP_TRACE_EXLIST(ip, nex, whichfork);
 		if (whichfork != XFS_DATA_FORK ||
@@ -719,70 +709,74 @@ xfs_iformat_btree(
 	return 0;
 }
 
-/*
- * xfs_xlate_dinode_core - translate an xfs_inode_core_t between ondisk
- * and native format
- *
- * buf  = on-disk representation
- * dip  = native representation
- * dir  = direction - +ve -> disk to native
- *                    -ve -> native to disk
- */
 void
-xfs_xlate_dinode_core(
-	xfs_caddr_t		buf,
-	xfs_dinode_core_t	*dip,
-	int			dir)
+xfs_dinode_from_disk(
+	xfs_icdinode_t		*to,
+	xfs_dinode_core_t	*from)
 {
-	xfs_dinode_core_t	*buf_core = (xfs_dinode_core_t *)buf;
-	xfs_dinode_core_t	*mem_core = (xfs_dinode_core_t *)dip;
-	xfs_arch_t		arch = ARCH_CONVERT;
-
-	ASSERT(dir);
-
-	INT_XLATE(buf_core->di_magic, mem_core->di_magic, dir, arch);
-	INT_XLATE(buf_core->di_mode, mem_core->di_mode, dir, arch);
-	INT_XLATE(buf_core->di_version,	mem_core->di_version, dir, arch);
-	INT_XLATE(buf_core->di_format, mem_core->di_format, dir, arch);
-	INT_XLATE(buf_core->di_onlink, mem_core->di_onlink, dir, arch);
-	INT_XLATE(buf_core->di_uid, mem_core->di_uid, dir, arch);
-	INT_XLATE(buf_core->di_gid, mem_core->di_gid, dir, arch);
-	INT_XLATE(buf_core->di_nlink, mem_core->di_nlink, dir, arch);
-	INT_XLATE(buf_core->di_projid, mem_core->di_projid, dir, arch);
-
-	if (dir > 0) {
-		memcpy(mem_core->di_pad, buf_core->di_pad,
-			sizeof(buf_core->di_pad));
-	} else {
-		memcpy(buf_core->di_pad, mem_core->di_pad,
-			sizeof(buf_core->di_pad));
-	}
-
-	INT_XLATE(buf_core->di_flushiter, mem_core->di_flushiter, dir, arch);
-
-	INT_XLATE(buf_core->di_atime.t_sec, mem_core->di_atime.t_sec,
-			dir, arch);
-	INT_XLATE(buf_core->di_atime.t_nsec, mem_core->di_atime.t_nsec,
-			dir, arch);
-	INT_XLATE(buf_core->di_mtime.t_sec, mem_core->di_mtime.t_sec,
-			dir, arch);
-	INT_XLATE(buf_core->di_mtime.t_nsec, mem_core->di_mtime.t_nsec,
-			dir, arch);
-	INT_XLATE(buf_core->di_ctime.t_sec, mem_core->di_ctime.t_sec,
-			dir, arch);
-	INT_XLATE(buf_core->di_ctime.t_nsec, mem_core->di_ctime.t_nsec,
-			dir, arch);
-	INT_XLATE(buf_core->di_size, mem_core->di_size, dir, arch);
-	INT_XLATE(buf_core->di_nblocks, mem_core->di_nblocks, dir, arch);
-	INT_XLATE(buf_core->di_extsize, mem_core->di_extsize, dir, arch);
-	INT_XLATE(buf_core->di_nextents, mem_core->di_nextents, dir, arch);
-	INT_XLATE(buf_core->di_anextents, mem_core->di_anextents, dir, arch);
-	INT_XLATE(buf_core->di_forkoff, mem_core->di_forkoff, dir, arch);
-	INT_XLATE(buf_core->di_aformat, mem_core->di_aformat, dir, arch);
-	INT_XLATE(buf_core->di_dmevmask, mem_core->di_dmevmask, dir, arch);
-	INT_XLATE(buf_core->di_dmstate, mem_core->di_dmstate, dir, arch);
-	INT_XLATE(buf_core->di_flags, mem_core->di_flags, dir, arch);
-	INT_XLATE(buf_core->di_gen, mem_core->di_gen, dir, arch);
+	to->di_magic = be16_to_cpu(from->di_magic);
+	to->di_mode = be16_to_cpu(from->di_mode);
+	to->di_version = from ->di_version;
+	to->di_format = from->di_format;
+	to->di_onlink = be16_to_cpu(from->di_onlink);
+	to->di_uid = be32_to_cpu(from->di_uid);
+	to->di_gid = be32_to_cpu(from->di_gid);
+	to->di_nlink = be32_to_cpu(from->di_nlink);
+	to->di_projid = be16_to_cpu(from->di_projid);
+	memcpy(to->di_pad, from->di_pad, sizeof(to->di_pad));
+	to->di_flushiter = be16_to_cpu(from->di_flushiter);
+	to->di_atime.t_sec = be32_to_cpu(from->di_atime.t_sec);
+	to->di_atime.t_nsec = be32_to_cpu(from->di_atime.t_nsec);
+	to->di_mtime.t_sec = be32_to_cpu(from->di_mtime.t_sec);
+	to->di_mtime.t_nsec = be32_to_cpu(from->di_mtime.t_nsec);
+	to->di_ctime.t_sec = be32_to_cpu(from->di_ctime.t_sec);
+	to->di_ctime.t_nsec = be32_to_cpu(from->di_ctime.t_nsec);
+	to->di_size = be64_to_cpu(from->di_size);
+	to->di_nblocks = be64_to_cpu(from->di_nblocks);
+	to->di_extsize = be32_to_cpu(from->di_extsize);
+	to->di_nextents = be32_to_cpu(from->di_nextents);
+	to->di_anextents = be16_to_cpu(from->di_anextents);
+	to->di_forkoff = from->di_forkoff;
+	to->di_aformat	= from->di_aformat;
+	to->di_dmevmask	= be32_to_cpu(from->di_dmevmask);
+	to->di_dmstate	= be16_to_cpu(from->di_dmstate);
+	to->di_flags	= be16_to_cpu(from->di_flags);
+	to->di_gen	= be32_to_cpu(from->di_gen);
+}
+
+void
+xfs_dinode_to_disk(
+	xfs_dinode_core_t	*to,
+	xfs_icdinode_t		*from)
+{
+	to->di_magic = cpu_to_be16(from->di_magic);
+	to->di_mode = cpu_to_be16(from->di_mode);
+	to->di_version = from ->di_version;
+	to->di_format = from->di_format;
+	to->di_onlink = cpu_to_be16(from->di_onlink);
+	to->di_uid = cpu_to_be32(from->di_uid);
+	to->di_gid = cpu_to_be32(from->di_gid);
+	to->di_nlink = cpu_to_be32(from->di_nlink);
+	to->di_projid = cpu_to_be16(from->di_projid);
+	memcpy(to->di_pad, from->di_pad, sizeof(to->di_pad));
+	to->di_flushiter = cpu_to_be16(from->di_flushiter);
+	to->di_atime.t_sec = cpu_to_be32(from->di_atime.t_sec);
+	to->di_atime.t_nsec = cpu_to_be32(from->di_atime.t_nsec);
+	to->di_mtime.t_sec = cpu_to_be32(from->di_mtime.t_sec);
+	to->di_mtime.t_nsec = cpu_to_be32(from->di_mtime.t_nsec);
+	to->di_ctime.t_sec = cpu_to_be32(from->di_ctime.t_sec);
+	to->di_ctime.t_nsec = cpu_to_be32(from->di_ctime.t_nsec);
+	to->di_size = cpu_to_be64(from->di_size);
+	to->di_nblocks = cpu_to_be64(from->di_nblocks);
+	to->di_extsize = cpu_to_be32(from->di_extsize);
+	to->di_nextents = cpu_to_be32(from->di_nextents);
+	to->di_anextents = cpu_to_be16(from->di_anextents);
+	to->di_forkoff = from->di_forkoff;
+	to->di_aformat = from->di_aformat;
+	to->di_dmevmask = cpu_to_be32(from->di_dmevmask);
+	to->di_dmstate = cpu_to_be16(from->di_dmstate);
+	to->di_flags = cpu_to_be16(from->di_flags);
+	to->di_gen = cpu_to_be32(from->di_gen);
 }
 
 STATIC uint
@@ -829,7 +823,7 @@ uint
 xfs_ip2xflags(
 	xfs_inode_t		*ip)
 {
-	xfs_dinode_core_t	*dic = &ip->i_d;
+	xfs_icdinode_t		*dic = &ip->i_d;
 
 	return _xfs_dic2xflags(dic->di_flags) |
 				(XFS_CFORK_Q(dic) ? XFS_XFLAG_HASATTR : 0);
@@ -839,7 +833,7 @@ uint
 xfs_dic2xflags(
 	xfs_dinode_core_t	*dic)
 {
-	return _xfs_dic2xflags(INT_GET(dic->di_flags, ARCH_CONVERT)) |
+	return _xfs_dic2xflags(be16_to_cpu(dic->di_flags)) |
 				(XFS_CFORK_Q_DISK(dic) ? XFS_XFLAG_HASATTR : 0);
 }
 
@@ -870,6 +864,7 @@ xfs_iread(
 	ip = kmem_zone_zalloc(xfs_inode_zone, KM_SLEEP);
 	ip->i_ino = ino;
 	ip->i_mount = mp;
+	atomic_set(&ip->i_iocount, 0);
 	spin_lock_init(&ip->i_flags_lock);
 
 	/*
@@ -889,6 +884,9 @@ xfs_iread(
 	 * Initialize inode's trace buffers.
 	 * Do this before xfs_iformat in case it adds entries.
 	 */
+#ifdef	XFS_VNODE_TRACE
+	ip->i_trace = ktrace_alloc(VNODE_TRACE_SIZE, KM_SLEEP);
+#endif
 #ifdef XFS_BMAP_TRACE
 	ip->i_xtrace = ktrace_alloc(XFS_BMAP_KTRACE_SIZE, KM_SLEEP);
 #endif
@@ -909,14 +907,14 @@ xfs_iread(
 	 * If we got something that isn't an inode it means someone
 	 * (nfs or dmi) has a stale handle.
 	 */
-	if (INT_GET(dip->di_core.di_magic, ARCH_CONVERT) != XFS_DINODE_MAGIC) {
+	if (be16_to_cpu(dip->di_core.di_magic) != XFS_DINODE_MAGIC) {
 		kmem_zone_free(xfs_inode_zone, ip);
 		xfs_trans_brelse(tp, bp);
 #ifdef DEBUG
 		xfs_fs_cmn_err(CE_ALERT, mp, "xfs_iread: "
 				"dip->di_core.di_magic (0x%x) != "
 				"XFS_DINODE_MAGIC (0x%x)",
-				INT_GET(dip->di_core.di_magic, ARCH_CONVERT),
+				be16_to_cpu(dip->di_core.di_magic),
 				XFS_DINODE_MAGIC);
 #endif /* DEBUG */
 		return XFS_ERROR(EINVAL);
@@ -930,8 +928,7 @@ xfs_iread(
 	 * Otherwise, just get the truly permanent information.
 	 */
 	if (dip->di_core.di_mode) {
-		xfs_xlate_dinode_core((xfs_caddr_t)&dip->di_core,
-		     &(ip->i_d), 1);
+		xfs_dinode_from_disk(&ip->i_d, &dip->di_core);
 		error = xfs_iformat(ip, dip);
 		if (error)  {
 			kmem_zone_free(xfs_inode_zone, ip);
@@ -944,10 +941,10 @@ xfs_iread(
 			return error;
 		}
 	} else {
-		ip->i_d.di_magic = INT_GET(dip->di_core.di_magic, ARCH_CONVERT);
-		ip->i_d.di_version = INT_GET(dip->di_core.di_version, ARCH_CONVERT);
-		ip->i_d.di_gen = INT_GET(dip->di_core.di_gen, ARCH_CONVERT);
-		ip->i_d.di_flushiter = INT_GET(dip->di_core.di_flushiter, ARCH_CONVERT);
+		ip->i_d.di_magic = be16_to_cpu(dip->di_core.di_magic);
+		ip->i_d.di_version = dip->di_core.di_version;
+		ip->i_d.di_gen = be32_to_cpu(dip->di_core.di_gen);
+		ip->i_d.di_flushiter = be16_to_cpu(dip->di_core.di_flushiter);
 		/*
 		 * Make sure to pull in the mode here as well in
 		 * case the inode is released without being used.
@@ -1048,7 +1045,7 @@ xfs_iread_extents(
 		ifp->if_flags &= ~XFS_IFEXTENTS;
 		return error;
 	}
-	xfs_validate_extents(ifp, nextents, 0, XFS_EXTFMT_INODE(ip));
+	xfs_validate_extents(ifp, nextents, XFS_EXTFMT_INODE(ip));
 	return 0;
 }
 
@@ -1161,7 +1158,7 @@ xfs_ialloc(
 	if ((prid != 0) && (ip->i_d.di_version == XFS_DINODE_VERSION_1))
 		xfs_bump_ino_vers2(tp, ip);
 
-	if (pip && XFS_INHERIT_GID(pip, vp->v_vfsp)) {
+	if (pip && XFS_INHERIT_GID(pip)) {
 		ip->i_d.di_gid = pip->i_d.di_gid;
 		if ((pip->i_d.di_mode & S_ISGID) && (mode & S_IFMT) == S_IFDIR) {
 			ip->i_d.di_mode |= S_ISGID;
@@ -1275,7 +1272,7 @@ xfs_ialloc(
 	xfs_trans_log_inode(tp, ip, flags);
 
 	/* now that we have an i_mode we can setup inode ops and unlock */
-	bhv_vfs_init_vnode(XFS_MTOVFS(tp->t_mountp), vp, XFS_ITOBHV(ip), 1);
+	xfs_initialize_vnode(tp->t_mountp, vp, ip);
 
 	*ipp = ip;
 	return 0;
@@ -1462,7 +1459,7 @@ xfs_itruncate_start(
 	mp = ip->i_mount;
 	vp = XFS_ITOV(ip);
 
-	vn_iowait(vp);  /* wait for the completion of any pending DIOs */
+	vn_iowait(ip);  /* wait for the completion of any pending DIOs */
 	
 	/*
 	 * Call toss_pages or flushinval_pages to get rid of pages
@@ -1497,9 +1494,11 @@ xfs_itruncate_start(
 			 last_byte);
 	if (last_byte > toss_start) {
 		if (flags & XFS_ITRUNC_DEFINITE) {
-			bhv_vop_toss_pages(vp, toss_start, -1, FI_REMAPF_LOCKED);
+			xfs_tosspages(ip, toss_start,
+					-1, FI_REMAPF_LOCKED);
 		} else {
-			error = bhv_vop_flushinval_pages(vp, toss_start, -1, FI_REMAPF_LOCKED);
+			error = xfs_flushinval_pages(ip, toss_start,
+					-1, FI_REMAPF_LOCKED);
 		}
 	}
 
@@ -1932,9 +1931,9 @@ xfs_iunlink(
 	 */
 	error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, agdaddr,
 				   XFS_FSS_TO_BB(mp, 1), 0, &agibp);
-	if (error) {
+	if (error)
 		return error;
-	}
+
 	/*
 	 * Validate the magic number of the agi block.
 	 */
@@ -1958,6 +1957,24 @@ xfs_iunlink(
 	ASSERT(agi->agi_unlinked[bucket_index]);
 	ASSERT(be32_to_cpu(agi->agi_unlinked[bucket_index]) != agino);
 
+	error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0);
+	if (error)
+		return error;
+
+	/*
+	 * Clear the on-disk di_nlink. This is to prevent xfs_bulkstat
+	 * from picking up this inode when it is reclaimed (its incore state
+	 * initialzed but not flushed to disk yet). The in-core di_nlink is
+	 * already cleared in xfs_droplink() and a corresponding transaction
+	 * logged. The hack here just synchronizes the in-core to on-disk
+	 * di_nlink value in advance before the actual inode sync to disk.
+	 * This is OK because the inode is already unlinked and would never
+	 * change its di_nlink again for this inode generation.
+	 * This is a temporary hack that would require a proper fix
+	 * in the future.
+	 */
+	dip->di_core.di_nlink = 0;
+
 	if (be32_to_cpu(agi->agi_unlinked[bucket_index]) != NULLAGINO) {
 		/*
 		 * There is already another inode in the bucket we need
@@ -1965,12 +1982,7 @@ xfs_iunlink(
 		 * Here we put the head pointer into our next pointer,
 		 * and then we fall through to point the head at us.
 		 */
-		error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0);
-		if (error) {
-			return error;
-		}
-		ASSERT(INT_GET(dip->di_next_unlinked, ARCH_CONVERT) == NULLAGINO);
-		ASSERT(dip->di_next_unlinked);
+		ASSERT(be32_to_cpu(dip->di_next_unlinked) == NULLAGINO);
 		/* both on-disk, don't endian flip twice */
 		dip->di_next_unlinked = agi->agi_unlinked[bucket_index];
 		offset = ip->i_boffset +
@@ -2081,10 +2093,10 @@ xfs_iunlink_remove(
 				error, mp->m_fsname);
 			return error;
 		}
-		next_agino = INT_GET(dip->di_next_unlinked, ARCH_CONVERT);
+		next_agino = be32_to_cpu(dip->di_next_unlinked);
 		ASSERT(next_agino != 0);
 		if (next_agino != NULLAGINO) {
-			INT_SET(dip->di_next_unlinked, ARCH_CONVERT, NULLAGINO);
+			dip->di_next_unlinked = cpu_to_be32(NULLAGINO);
 			offset = ip->i_boffset +
 				offsetof(xfs_dinode_t, di_next_unlinked);
 			xfs_trans_inode_buf(tp, ibp);
@@ -2128,7 +2140,7 @@ xfs_iunlink_remove(
 					error, mp->m_fsname);
 				return error;
 			}
-			next_agino = INT_GET(last_dip->di_next_unlinked, ARCH_CONVERT);
+			next_agino = be32_to_cpu(last_dip->di_next_unlinked);
 			ASSERT(next_agino != NULLAGINO);
 			ASSERT(next_agino != 0);
 		}
@@ -2143,11 +2155,11 @@ xfs_iunlink_remove(
 				error, mp->m_fsname);
 			return error;
 		}
-		next_agino = INT_GET(dip->di_next_unlinked, ARCH_CONVERT);
+		next_agino = be32_to_cpu(dip->di_next_unlinked);
 		ASSERT(next_agino != 0);
 		ASSERT(next_agino != agino);
 		if (next_agino != NULLAGINO) {
-			INT_SET(dip->di_next_unlinked, ARCH_CONVERT, NULLAGINO);
+			dip->di_next_unlinked = cpu_to_be32(NULLAGINO);
 			offset = ip->i_boffset +
 				offsetof(xfs_dinode_t, di_next_unlinked);
 			xfs_trans_inode_buf(tp, ibp);
@@ -2160,7 +2172,7 @@ xfs_iunlink_remove(
 		/*
 		 * Point the previous inode on the list to the next inode.
 		 */
-		INT_SET(last_dip->di_next_unlinked, ARCH_CONVERT, next_agino);
+		last_dip->di_next_unlinked = cpu_to_be32(next_agino);
 		ASSERT(next_agino != 0);
 		offset = last_offset + offsetof(xfs_dinode_t, di_next_unlinked);
 		xfs_trans_inode_buf(tp, last_ibp);
@@ -2191,10 +2203,10 @@ xfs_ifree_cluster(
 	int			i, j, found, pre_flushed;
 	xfs_daddr_t		blkno;
 	xfs_buf_t		*bp;
-	xfs_ihash_t		*ih;
 	xfs_inode_t		*ip, **ip_found;
 	xfs_inode_log_item_t	*iip;
 	xfs_log_item_t		*lip;
+	xfs_perag_t		*pag = xfs_get_perag(mp, inum);
 	SPLDECL(s);
 
 	if (mp->m_sb.sb_blocksize >= XFS_INODE_CLUSTER_SIZE(mp)) {
@@ -2229,23 +2241,20 @@ xfs_ifree_cluster(
 		 */
 		found = 0;
 		for (i = 0; i < ninodes; i++) {
-			ih = XFS_IHASH(mp, inum + i);
-			read_lock(&ih->ih_lock);
-			for (ip = ih->ih_next; ip != NULL; ip = ip->i_next) {
-				if (ip->i_ino == inum + i)
-					break;
-			}
+			read_lock(&pag->pag_ici_lock);
+			ip = radix_tree_lookup(&pag->pag_ici_root,
+					XFS_INO_TO_AGINO(mp, (inum + i)));
 
 			/* Inode not in memory or we found it already,
 			 * nothing to do
 			 */
 			if (!ip || xfs_iflags_test(ip, XFS_ISTALE)) {
-				read_unlock(&ih->ih_lock);
+				read_unlock(&pag->pag_ici_lock);
 				continue;
 			}
 
 			if (xfs_inode_clean(ip)) {
-				read_unlock(&ih->ih_lock);
+				read_unlock(&pag->pag_ici_lock);
 				continue;
 			}
 
@@ -2268,7 +2277,7 @@ xfs_ifree_cluster(
 						ip_found[found++] = ip;
 					}
 				}
-				read_unlock(&ih->ih_lock);
+				read_unlock(&pag->pag_ici_lock);
 				continue;
 			}
 
@@ -2286,8 +2295,7 @@ xfs_ifree_cluster(
 					xfs_iunlock(ip, XFS_ILOCK_EXCL);
 				}
 			}
-
-			read_unlock(&ih->ih_lock);
+			read_unlock(&pag->pag_ici_lock);
 		}
 
 		bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, blkno, 
@@ -2342,6 +2350,7 @@ xfs_ifree_cluster(
 	}
 
 	kmem_free(ip_found, ninodes * sizeof(xfs_inode_t *));
+	xfs_put_perag(mp, pag);
 }
 
 /*
@@ -2737,6 +2746,10 @@ xfs_idestroy(
 	mrfree(&ip->i_lock);
 	mrfree(&ip->i_iolock);
 	freesema(&ip->i_flock);
+
+#ifdef XFS_VNODE_TRACE
+	ktrace_free(ip->i_trace);
+#endif
 #ifdef XFS_BMAP_TRACE
 	ktrace_free(ip->i_xtrace);
 #endif
@@ -2887,12 +2900,10 @@ xfs_iunpin_wait(
 int
 xfs_iextents_copy(
 	xfs_inode_t		*ip,
-	xfs_bmbt_rec_t		*buffer,
+	xfs_bmbt_rec_t		*dp,
 	int			whichfork)
 {
 	int			copied;
-	xfs_bmbt_rec_t		*dest_ep;
-	xfs_bmbt_rec_t		*ep;
 	int			i;
 	xfs_ifork_t		*ifp;
 	int			nrecs;
@@ -2912,10 +2923,9 @@ xfs_iextents_copy(
 	 * the delayed ones.  There must be at least one
 	 * non-delayed extent.
 	 */
-	dest_ep = buffer;
 	copied = 0;
 	for (i = 0; i < nrecs; i++) {
-		ep = xfs_iext_get_ext(ifp, i);
+		xfs_bmbt_rec_host_t *ep = xfs_iext_get_ext(ifp, i);
 		start_block = xfs_bmbt_get_startblock(ep);
 		if (ISNULLSTARTBLOCK(start_block)) {
 			/*
@@ -2925,15 +2935,13 @@ xfs_iextents_copy(
 		}
 
 		/* Translate to on disk format */
-		put_unaligned(INT_GET(ep->l0, ARCH_CONVERT),
-			      (__uint64_t*)&dest_ep->l0);
-		put_unaligned(INT_GET(ep->l1, ARCH_CONVERT),
-			      (__uint64_t*)&dest_ep->l1);
-		dest_ep++;
+		put_unaligned(cpu_to_be64(ep->l0), &dp->l0);
+		put_unaligned(cpu_to_be64(ep->l1), &dp->l1);
+		dp++;
 		copied++;
 	}
 	ASSERT(copied != 0);
-	xfs_validate_extents(ifp, copied, 1, XFS_EXTFMT_INODE(ip));
+	xfs_validate_extents(ifp, copied, XFS_EXTFMT_INODE(ip));
 
 	return (copied * (uint)sizeof(xfs_bmbt_rec_t));
 }
@@ -3024,7 +3032,7 @@ xfs_iflush_fork(
 	case XFS_DINODE_FMT_DEV:
 		if (iip->ili_format.ilf_fields & XFS_ILOG_DEV) {
 			ASSERT(whichfork == XFS_DATA_FORK);
-			INT_SET(dip->di_u.di_dev, ARCH_CONVERT, ip->i_df.if_u2.if_rdev);
+			dip->di_u.di_dev = cpu_to_be32(ip->i_df.if_u2.if_rdev);
 		}
 		break;
 
@@ -3064,12 +3072,11 @@ xfs_iflush(
 	xfs_mount_t		*mp;
 	int			error;
 	/* REFERENCED */
-	xfs_chash_t		*ch;
 	xfs_inode_t		*iq;
 	int			clcount;	/* count of inodes clustered */
 	int			bufwasdelwri;
+	struct hlist_node	*entry;
 	enum { INT_DELWRI = (1 << 0), INT_ASYNC = (1 << 1) };
-	SPLDECL(s);
 
 	XFS_STATS_INC(xs_iflush_count);
 
@@ -3183,14 +3190,14 @@ xfs_iflush(
 	 * inode clustering:
 	 * see if other inodes can be gathered into this write
 	 */
-
-	ip->i_chash->chl_buf = bp;
-
-	ch = XFS_CHASH(mp, ip->i_blkno);
-	s = mutex_spinlock(&ch->ch_lock);
+	spin_lock(&ip->i_cluster->icl_lock);
+	ip->i_cluster->icl_buf = bp;
 
 	clcount = 0;
-	for (iq = ip->i_cnext; iq != ip; iq = iq->i_cnext) {
+	hlist_for_each_entry(iq, entry, &ip->i_cluster->icl_inodes, i_cnode) {
+		if (iq == ip)
+			continue;
+
 		/*
 		 * Do an un-protected check to see if the inode is dirty and
 		 * is a candidate for flushing.  These checks will be repeated
@@ -3241,7 +3248,7 @@ xfs_iflush(
 			xfs_iunlock(iq, XFS_ILOCK_SHARED);
 		}
 	}
-	mutex_spinunlock(&ch->ch_lock, s);
+	spin_unlock(&ip->i_cluster->icl_lock);
 
 	if (clcount) {
 		XFS_STATS_INC(xs_icluster_flushcnt);
@@ -3278,7 +3285,7 @@ cluster_corrupt_out:
 	/* Corruption detected in the clustering loop.  Invalidate the
 	 * inode buffer and shut down the filesystem.
 	 */
-	mutex_spinunlock(&ch->ch_lock, s);
+	spin_unlock(&ip->i_cluster->icl_lock);
 
 	/*
 	 * Clean up the buffer.  If it was B_DELWRI, just release it --
@@ -3373,11 +3380,11 @@ xfs_iflush_int(
 	 */
 	xfs_synchronize_atime(ip);
 
-	if (XFS_TEST_ERROR(INT_GET(dip->di_core.di_magic,ARCH_CONVERT) != XFS_DINODE_MAGIC,
+	if (XFS_TEST_ERROR(be16_to_cpu(dip->di_core.di_magic) != XFS_DINODE_MAGIC,
 			       mp, XFS_ERRTAG_IFLUSH_1, XFS_RANDOM_IFLUSH_1)) {
 		xfs_cmn_err(XFS_PTAG_IFLUSH, CE_ALERT, mp,
 		    "xfs_iflush: Bad inode %Lu magic number 0x%x, ptr 0x%p",
-			ip->i_ino, (int) INT_GET(dip->di_core.di_magic, ARCH_CONVERT), dip);
+			ip->i_ino, be16_to_cpu(dip->di_core.di_magic), dip);
 		goto corrupt_out;
 	}
 	if (XFS_TEST_ERROR(ip->i_d.di_magic != XFS_DINODE_MAGIC,
@@ -3440,7 +3447,7 @@ xfs_iflush_int(
 	 * because if the inode is dirty at all the core must
 	 * be.
 	 */
-	xfs_xlate_dinode_core((xfs_caddr_t)&(dip->di_core), &(ip->i_d), -1);
+	xfs_dinode_to_disk(&dip->di_core, &ip->i_d);
 
 	/* Wrap, we never let the log put out DI_MAX_FLUSH */
 	if (ip->i_d.di_flushiter == DI_MAX_FLUSH)
@@ -3460,7 +3467,7 @@ xfs_iflush_int(
 			 * Convert it back.
 			 */
 			ASSERT(ip->i_d.di_nlink <= XFS_MAXLINK_1);
-			INT_SET(dip->di_core.di_onlink, ARCH_CONVERT, ip->i_d.di_nlink);
+			dip->di_core.di_onlink = cpu_to_be16(ip->i_d.di_nlink);
 		} else {
 			/*
 			 * The superblock version has already been bumped,
@@ -3468,7 +3475,7 @@ xfs_iflush_int(
 			 * format permanent.
 			 */
 			ip->i_d.di_version = XFS_DINODE_VERSION_2;
-			INT_SET(dip->di_core.di_version, ARCH_CONVERT, XFS_DINODE_VERSION_2);
+			dip->di_core.di_version =  XFS_DINODE_VERSION_2;
 			ip->i_d.di_onlink = 0;
 			dip->di_core.di_onlink = 0;
 			memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
@@ -3711,7 +3718,7 @@ xfs_ilock_trace(xfs_inode_t *ip, int lock, unsigned int lockflags, inst_t *ra)
 /*
  * Return a pointer to the extent record at file index idx.
  */
-xfs_bmbt_rec_t *
+xfs_bmbt_rec_host_t *
 xfs_iext_get_ext(
 	xfs_ifork_t	*ifp,		/* inode fork pointer */
 	xfs_extnum_t	idx)		/* index of target extent */
@@ -3744,15 +3751,12 @@ xfs_iext_insert(
 	xfs_extnum_t	count,		/* number of inserted items */
 	xfs_bmbt_irec_t	*new)		/* items to insert */
 {
-	xfs_bmbt_rec_t	*ep;		/* extent record pointer */
 	xfs_extnum_t	i;		/* extent record index */
 
 	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
 	xfs_iext_add(ifp, idx, count);
-	for (i = idx; i < idx + count; i++, new++) {
-		ep = xfs_iext_get_ext(ifp, i);
-		xfs_bmbt_set_all(ep, new);
-	}
+	for (i = idx; i < idx + count; i++, new++)
+		xfs_bmbt_set_all(xfs_iext_get_ext(ifp, i), new);
 }
 
 /*
@@ -4203,7 +4207,7 @@ xfs_iext_realloc_direct(
 			rnew_size = xfs_iroundup(new_size);
 		}
 		if (rnew_size != ifp->if_real_bytes) {
-			ifp->if_u1.if_extents = (xfs_bmbt_rec_t *)
+			ifp->if_u1.if_extents =
 				kmem_realloc(ifp->if_u1.if_extents,
 						rnew_size,
 						ifp->if_real_bytes,
@@ -4266,8 +4270,7 @@ 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 = (xfs_bmbt_rec_t *)
-		kmem_alloc(new_size, KM_SLEEP);
+	ifp->if_u1.if_extents = kmem_alloc(new_size, KM_SLEEP);
 	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,
@@ -4310,7 +4313,7 @@ void
 xfs_iext_indirect_to_direct(
 	 xfs_ifork_t	*ifp)		/* inode fork pointer */
 {
-	xfs_bmbt_rec_t	*ep;		/* extent record pointer */
+	xfs_bmbt_rec_host_t *ep;	/* extent record pointer */
 	xfs_extnum_t	nextents;	/* number of extents in file */
 	int		size;		/* size of file extents */
 
@@ -4362,15 +4365,15 @@ xfs_iext_destroy(
 /*
  * Return a pointer to the extent record for file system block bno.
  */
-xfs_bmbt_rec_t *			/* pointer to found extent record */
+xfs_bmbt_rec_host_t *			/* pointer to found extent record */
 xfs_iext_bno_to_ext(
 	xfs_ifork_t	*ifp,		/* inode fork pointer */
 	xfs_fileoff_t	bno,		/* block number to search for */
 	xfs_extnum_t	*idxp)		/* index of target extent */
 {
-	xfs_bmbt_rec_t	*base;		/* pointer to first extent */
+	xfs_bmbt_rec_host_t *base;	/* pointer to first extent */
 	xfs_filblks_t	blockcount = 0;	/* number of blocks in extent */
-	xfs_bmbt_rec_t	*ep = NULL;	/* pointer to target extent */
+	xfs_bmbt_rec_host_t *ep = NULL;	/* pointer to target extent */
 	xfs_ext_irec_t	*erp = NULL;	/* indirection array pointer */
 	int		high;		/* upper boundary in search */
 	xfs_extnum_t	idx = 0;	/* index of target extent */
@@ -4545,8 +4548,7 @@ xfs_iext_irec_init(
 		kmem_alloc(sizeof(xfs_ext_irec_t), KM_SLEEP);
 
 	if (nextents == 0) {
-		ifp->if_u1.if_extents = (xfs_bmbt_rec_t *)
-			kmem_alloc(XFS_IEXT_BUFSZ, KM_SLEEP);
+		ifp->if_u1.if_extents = kmem_alloc(XFS_IEXT_BUFSZ, KM_SLEEP);
 	} else if (!ifp->if_real_bytes) {
 		xfs_iext_inline_to_direct(ifp, XFS_IEXT_BUFSZ);
 	} else if (ifp->if_real_bytes < XFS_IEXT_BUFSZ) {
@@ -4594,8 +4596,7 @@ xfs_iext_irec_new(
 
 	/* Initialize new extent record */
 	erp = ifp->if_u1.if_ext_irec;
-	erp[erp_idx].er_extbuf = (xfs_bmbt_rec_t *)
-		kmem_alloc(XFS_IEXT_BUFSZ, KM_SLEEP);
+	erp[erp_idx].er_extbuf = kmem_alloc(XFS_IEXT_BUFSZ, KM_SLEEP);
 	ifp->if_real_bytes = nlists * XFS_IEXT_BUFSZ;
 	memset(erp[erp_idx].er_extbuf, 0, XFS_IEXT_BUFSZ);
 	erp[erp_idx].er_extcount = 0;
@@ -4727,7 +4728,7 @@ void
 xfs_iext_irec_compact_full(
 	xfs_ifork_t	*ifp)			/* inode fork pointer */
 {
-	xfs_bmbt_rec_t	*ep, *ep_next;		/* extent record pointers */
+	xfs_bmbt_rec_host_t *ep, *ep_next;	/* extent record pointers */
 	xfs_ext_irec_t	*erp, *erp_next;	/* extent irec pointers */
 	int		erp_idx = 0;		/* extent irec index */
 	int		ext_avail;		/* empty entries in ex list */

fs/xfs/xfs_inode.h

@@ -18,6 +18,10 @@
 #ifndef	__XFS_INODE_H__
 #define	__XFS_INODE_H__
 
+struct xfs_dinode;
+struct xfs_dinode_core;
+
+
 /*
  * Fork identifiers.
  */
@@ -44,7 +48,7 @@
  * it is possible that a third level of indirection may be required.
  */
 typedef struct xfs_ext_irec {
-	xfs_bmbt_rec_t	*er_extbuf;	/* block of extent records */
+	xfs_bmbt_rec_host_t *er_extbuf;	/* block of extent records */
 	xfs_extnum_t	er_extoff;	/* extent offset in file */
 	xfs_extnum_t	er_extcount;	/* number of extents in page/block */
 } xfs_ext_irec_t;
@@ -65,12 +69,12 @@ typedef struct xfs_ifork {
 	unsigned char		if_ext_max;	/* max # of extent records */
 	xfs_extnum_t		if_lastex;	/* last if_extents used */
 	union {
-		xfs_bmbt_rec_t	*if_extents;	/* linear map file exts */
+		xfs_bmbt_rec_host_t *if_extents;/* linear map file exts */
 		xfs_ext_irec_t	*if_ext_irec;	/* irec map file exts */
 		char		*if_data;	/* inline file data */
 	} if_u1;
 	union {
-		xfs_bmbt_rec_t	if_inline_ext[XFS_INLINE_EXTS];
+		xfs_bmbt_rec_host_t if_inline_ext[XFS_INLINE_EXTS];
 						/* very small file extents */
 		char		if_inline_data[XFS_INLINE_DATA];
 						/* very small file data */
@@ -102,7 +106,6 @@ typedef struct xfs_ifork {
 
 #ifdef __KERNEL__
 struct bhv_desc;
-struct bhv_vnode;
 struct cred;
 struct ktrace;
 struct xfs_buf;
@@ -168,41 +171,18 @@ typedef struct xfs_iocore {
 extern void xfs_iocore_inode_init(struct xfs_inode *);
 extern void xfs_iocore_inode_reinit(struct xfs_inode *);
 
-
-/*
- * This is the type used in the xfs inode hash table.
- * An array of these is allocated for each mounted
- * file system to hash the inodes for that file system.
- */
-typedef struct xfs_ihash {
-	struct xfs_inode	*ih_next;
-	rwlock_t		ih_lock;
-	uint			ih_version;
-} xfs_ihash_t;
-
-#define XFS_IHASH(mp,ino) ((mp)->m_ihash + (((uint)(ino)) % (mp)->m_ihsize))
-
 /*
- * This is the xfs inode cluster hash.  This hash is used by xfs_iflush to
- * find inodes that share a cluster and can be flushed to disk at the same
- * time.
+ * This is the xfs inode cluster structure.  This structure is used by
+ * xfs_iflush to find inodes that share a cluster and can be flushed to disk at
+ * the same time.
  */
-typedef struct xfs_chashlist {
-	struct xfs_chashlist	*chl_next;
-	struct xfs_chashlist	*chl_prev;
-	struct xfs_inode	*chl_ip;
-	xfs_daddr_t		chl_blkno;	/* starting block number of
+typedef struct xfs_icluster {
+	struct hlist_head	icl_inodes;	/* list of inodes on cluster */
+	xfs_daddr_t		icl_blkno;	/* starting block number of
 						 * the cluster */
-	struct xfs_buf		*chl_buf;	/* the inode buffer */
-} xfs_chashlist_t;
-
-typedef struct xfs_chash {
-	xfs_chashlist_t		*ch_list;
-	lock_t			ch_lock;
-} xfs_chash_t;
-
-#define XFS_CHASH(mp,blk) ((mp)->m_chash + (((uint)blk) % (mp)->m_chsize))
-
+	struct xfs_buf		*icl_buf;	/* the inode buffer */
+	lock_t			icl_lock;	/* inode list lock */
+} xfs_icluster_t;
 
 /*
  * This is the xfs in-core inode structure.
@@ -227,25 +207,56 @@ typedef struct xfs_chash {
  * chain off the mount structure by xfs_sync calls.
  */
 
+typedef struct xfs_ictimestamp {
+	__int32_t	t_sec;		/* timestamp seconds */
+	__int32_t	t_nsec;		/* timestamp nanoseconds */
+} xfs_ictimestamp_t;
+
+/*
+ * NOTE:  This structure must be kept identical to struct xfs_dinode_core
+ * 	  in xfs_dinode.h except for the endianess annotations.
+ */
+typedef struct xfs_icdinode {
+	__uint16_t	di_magic;	/* inode magic # = XFS_DINODE_MAGIC */
+	__uint16_t	di_mode;	/* mode and type of file */
+	__int8_t	di_version;	/* inode version */
+	__int8_t	di_format;	/* format of di_c data */
+	__uint16_t	di_onlink;	/* old number of links to file */
+	__uint32_t	di_uid;		/* owner's user id */
+	__uint32_t	di_gid;		/* owner's group id */
+	__uint32_t	di_nlink;	/* number of links to file */
+	__uint16_t	di_projid;	/* owner's project id */
+	__uint8_t	di_pad[8];	/* unused, zeroed space */
+	__uint16_t	di_flushiter;	/* incremented on flush */
+	xfs_ictimestamp_t di_atime;	/* time last accessed */
+	xfs_ictimestamp_t di_mtime;	/* time last modified */
+	xfs_ictimestamp_t di_ctime;	/* time created/inode modified */
+	xfs_fsize_t	di_size;	/* number of bytes in file */
+	xfs_drfsbno_t	di_nblocks;	/* # of direct & btree blocks used */
+	xfs_extlen_t	di_extsize;	/* basic/minimum extent size for file */
+	xfs_extnum_t	di_nextents;	/* number of extents in data fork */
+	xfs_aextnum_t	di_anextents;	/* number of extents in attribute fork*/
+	__uint8_t	di_forkoff;	/* attr fork offs, <<3 for 64b align */
+	__int8_t	di_aformat;	/* format of attr fork's data */
+	__uint32_t	di_dmevmask;	/* DMIG event mask */
+	__uint16_t	di_dmstate;	/* DMIG state info */
+	__uint16_t	di_flags;	/* random flags, XFS_DIFLAG_... */
+	__uint32_t	di_gen;		/* generation number */
+} xfs_icdinode_t;
+
 typedef struct {
-	struct xfs_ihash	*ip_hash;	/* pointer to hash header */
-	struct xfs_inode	*ip_next;	/* inode hash link forw */
 	struct xfs_inode	*ip_mnext;	/* next inode in mount list */
 	struct xfs_inode	*ip_mprev;	/* ptr to prev inode */
-	struct xfs_inode	**ip_prevp;	/* ptr to prev i_next */
 	struct xfs_mount	*ip_mount;	/* fs mount struct ptr */
 } xfs_iptr_t;
 
 typedef struct xfs_inode {
 	/* Inode linking and identification information. */
-	struct xfs_ihash	*i_hash;	/* pointer to hash header */
-	struct xfs_inode	*i_next;	/* inode hash link forw */
 	struct xfs_inode	*i_mnext;	/* next inode in mount list */
 	struct xfs_inode	*i_mprev;	/* ptr to prev inode */
-	struct xfs_inode	**i_prevp;	/* ptr to prev i_next */
 	struct xfs_mount	*i_mount;	/* fs mount struct ptr */
 	struct list_head	i_reclaim;	/* reclaim list */
-	struct bhv_desc		i_bhv_desc;	/* inode behavior descriptor*/
+	bhv_vnode_t		*i_vnode;	/* vnode backpointer */
 	struct xfs_dquot	*i_udquot;	/* user dquot */
 	struct xfs_dquot	*i_gdquot;	/* group dquot */
 
@@ -282,13 +293,16 @@ typedef struct xfs_inode {
 	unsigned int		i_gen;		/* generation count */
 	unsigned int		i_delayed_blks;	/* count of delay alloc blks */
 
-	xfs_dinode_core_t	i_d;		/* most of ondisk inode */
-	xfs_chashlist_t		*i_chash;	/* cluster hash list header */
-	struct xfs_inode	*i_cnext;	/* cluster hash link forward */
-	struct xfs_inode	*i_cprev;	/* cluster hash link backward */
+	xfs_icdinode_t		i_d;		/* most of ondisk inode */
+	xfs_icluster_t		*i_cluster;	/* cluster list header */
+	struct hlist_node	i_cnode;	/* cluster link node */
 
 	xfs_fsize_t		i_size;		/* in-memory size */
+	atomic_t		i_iocount;	/* outstanding I/O count */
 	/* Trace buffers per inode. */
+#ifdef XFS_VNODE_TRACE
+	struct ktrace		*i_trace;	/* general inode trace */
+#endif
 #ifdef XFS_BMAP_TRACE
 	struct ktrace		*i_xtrace;	/* inode extent list trace */
 #endif
@@ -349,6 +363,19 @@ xfs_iflags_test(xfs_inode_t *ip, unsigned short flags)
 	spin_unlock(&ip->i_flags_lock);
 	return ret;
 }
+
+static inline int
+xfs_iflags_test_and_clear(xfs_inode_t *ip, unsigned short flags)
+{
+	int ret;
+
+	spin_lock(&ip->i_flags_lock);
+	ret = ip->i_flags & flags;
+	if (ret)
+		ip->i_flags &= ~flags;
+	spin_unlock(&ip->i_flags_lock);
+	return ret;
+}
 #endif	/* __KERNEL__ */
 
 
@@ -380,6 +407,9 @@ xfs_iflags_test(xfs_inode_t *ip, unsigned short flags)
 #define XFS_IRECLAIMABLE 0x0020 /* inode can be reclaimed */
 #define XFS_INEW	0x0040
 #define XFS_IFILESTREAM	0x0080	/* inode is in a filestream directory */
+#define XFS_IMODIFIED	0x0100	/* XFS inode state possibly differs */
+				/* to the Linux inode state. */
+#define XFS_ITRUNCATED	0x0200	/* truncated down so flush-on-close */
 
 /*
  * Flags for inode locking.
@@ -454,20 +484,17 @@ xfs_iflags_test(xfs_inode_t *ip, unsigned short flags)
 #define	XFS_ITRUNC_DEFINITE	0x1
 #define	XFS_ITRUNC_MAYBE	0x2
 
-#define	XFS_ITOV(ip)		BHV_TO_VNODE(XFS_ITOBHV(ip))
-#define	XFS_ITOV_NULL(ip)	BHV_TO_VNODE_NULL(XFS_ITOBHV(ip))
-#define	XFS_ITOBHV(ip)		((struct bhv_desc *)(&((ip)->i_bhv_desc)))
-#define	XFS_BHVTOI(bhvp)	((xfs_inode_t *)((char *)(bhvp) - \
-				(char *)&(((xfs_inode_t *)0)->i_bhv_desc)))
-#define BHV_IS_XFS(bdp)		(BHV_OPS(bdp) == &xfs_vnodeops)
+#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
  * new subdirectory gets S_ISGID bit from parent.
  */
-#define XFS_INHERIT_GID(pip, vfsp)	\
-	(((vfsp)->vfs_flag & VFS_GRPID) || ((pip)->i_d.di_mode & S_ISGID))
+#define XFS_INHERIT_GID(pip)	\
+	(((pip)->i_mount->m_flags & XFS_MOUNT_GRPID) || \
+	 ((pip)->i_d.di_mode & S_ISGID))
 
 /*
  * Flags for xfs_iget()
@@ -480,11 +507,9 @@ xfs_iflags_test(xfs_inode_t *ip, unsigned short flags)
  */
 void		xfs_ihash_init(struct xfs_mount *);
 void		xfs_ihash_free(struct xfs_mount *);
-void		xfs_chash_init(struct xfs_mount *);
-void		xfs_chash_free(struct xfs_mount *);
 xfs_inode_t	*xfs_inode_incore(struct xfs_mount *, xfs_ino_t,
 				  struct xfs_trans *);
-void            xfs_inode_lock_init(xfs_inode_t *, struct bhv_vnode *);
+void            xfs_inode_lock_init(xfs_inode_t *, bhv_vnode_t *);
 int		xfs_iget(struct xfs_mount *, struct xfs_trans *, xfs_ino_t,
 			 uint, uint, xfs_inode_t **, xfs_daddr_t);
 void		xfs_iput(xfs_inode_t *, uint);
@@ -506,7 +531,7 @@ int		xfs_finish_reclaim_all(struct xfs_mount *, int);
  * xfs_inode.c prototypes.
  */
 int		xfs_itobp(struct xfs_mount *, struct xfs_trans *,
-			  xfs_inode_t *, xfs_dinode_t **, struct xfs_buf **,
+			  xfs_inode_t *, struct xfs_dinode **, struct xfs_buf **,
 			  xfs_daddr_t, uint);
 int		xfs_iread(struct xfs_mount *, struct xfs_trans *, xfs_ino_t,
 			  xfs_inode_t **, xfs_daddr_t, uint);
@@ -514,8 +539,11 @@ int		xfs_iread_extents(struct xfs_trans *, xfs_inode_t *, int);
 int		xfs_ialloc(struct xfs_trans *, xfs_inode_t *, mode_t,
 			   xfs_nlink_t, xfs_dev_t, struct cred *, xfs_prid_t,
 			   int, struct xfs_buf **, boolean_t *, xfs_inode_t **);
-void		xfs_xlate_dinode_core(xfs_caddr_t, struct xfs_dinode_core *,
-					int);
+void		xfs_dinode_from_disk(struct xfs_icdinode *,
+				     struct xfs_dinode_core *);
+void		xfs_dinode_to_disk(struct xfs_dinode_core *,
+				   struct xfs_icdinode *);
+
 uint		xfs_ip2xflags(struct xfs_inode *);
 uint		xfs_dic2xflags(struct xfs_dinode_core *);
 int		xfs_ifree(struct xfs_trans *, xfs_inode_t *,
@@ -545,11 +573,9 @@ 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, int, uint);
 
-xfs_inode_t	*xfs_vtoi(struct bhv_vnode *vp);
-
 void		xfs_synchronize_atime(xfs_inode_t *);
 
-xfs_bmbt_rec_t	*xfs_iext_get_ext(xfs_ifork_t *, xfs_extnum_t);
+xfs_bmbt_rec_host_t *xfs_iext_get_ext(xfs_ifork_t *, xfs_extnum_t);
 void		xfs_iext_insert(xfs_ifork_t *, xfs_extnum_t, xfs_extnum_t,
 				xfs_bmbt_irec_t *);
 void		xfs_iext_add(xfs_ifork_t *, xfs_extnum_t, int);
@@ -564,7 +590,7 @@ void		xfs_iext_indirect_to_direct(xfs_ifork_t *);
 void		xfs_iext_direct_to_inline(xfs_ifork_t *, xfs_extnum_t);
 void		xfs_iext_inline_to_direct(xfs_ifork_t *, int);
 void		xfs_iext_destroy(xfs_ifork_t *);
-xfs_bmbt_rec_t	*xfs_iext_bno_to_ext(xfs_ifork_t *, xfs_fileoff_t, int *);
+xfs_bmbt_rec_host_t *xfs_iext_bno_to_ext(xfs_ifork_t *, xfs_fileoff_t, int *);
 xfs_ext_irec_t	*xfs_iext_bno_to_irec(xfs_ifork_t *, xfs_fileoff_t, int *);
 xfs_ext_irec_t	*xfs_iext_idx_to_irec(xfs_ifork_t *, xfs_extnum_t *, int *, int);
 void		xfs_iext_irec_init(xfs_ifork_t *);
@@ -589,7 +615,7 @@ void		xfs_inobp_check(struct xfs_mount *, struct xfs_buf *);
 #define	xfs_inobp_check(mp, bp)
 #endif /* DEBUG */
 
-extern struct kmem_zone	*xfs_chashlist_zone;
+extern struct kmem_zone	*xfs_icluster_zone;
 extern struct kmem_zone	*xfs_ifork_zone;
 extern struct kmem_zone	*xfs_inode_zone;
 extern struct kmem_zone	*xfs_ili_zone;

fs/xfs/xfs_iocore.c

@@ -57,11 +57,11 @@ xfs_size_fn(
 
 STATIC int
 xfs_ioinit(
-	struct bhv_vfs		*vfsp,
+	struct xfs_mount	*mp,
 	struct xfs_mount_args	*mntargs,
 	int			flags)
 {
-	return xfs_mountfs(vfsp, XFS_VFSTOM(vfsp), flags);
+	return xfs_mountfs(mp, flags);
 }
 
 xfs_ioops_t	xfs_iocore_xfs = {

fs/xfs/xfs_iomap.c

@@ -135,14 +135,10 @@ xfs_imap_to_bmap(
 	int		flags)
 {
 	xfs_mount_t	*mp;
-	xfs_fsize_t	nisize;
 	int		pbm;
 	xfs_fsblock_t	start_block;
 
 	mp = io->io_mount;
-	nisize = XFS_SIZE(mp, io);
-	if (io->io_new_size > nisize)
-		nisize = io->io_new_size;
 
 	for (pbm = 0; imaps && pbm < iomaps; imaps--, iomapp++, imap++, pbm++) {
 		iomapp->iomap_offset = XFS_FSB_TO_B(mp, imap->br_startoff);
@@ -169,10 +165,6 @@ xfs_imap_to_bmap(
 				iomapp->iomap_flags |= IOMAP_UNWRITTEN;
 		}
 
-		if ((iomapp->iomap_offset + iomapp->iomap_bsize) >= nisize) {
-			iomapp->iomap_flags |= IOMAP_EOF;
-		}
-
 		offset += iomapp->iomap_bsize - iomapp->iomap_delta;
 	}
 	return pbm;	/* Return the number filled */

fs/xfs/xfs_iomap.h

@@ -23,7 +23,6 @@
 
 typedef enum {				/* iomap_flags values */
 	IOMAP_READ =		0,	/* mapping for a read */
-	IOMAP_EOF =		0x01,	/* mapping contains EOF   */
 	IOMAP_HOLE =		0x02,	/* mapping covers a hole  */
 	IOMAP_DELAY =		0x04,	/* mapping covers delalloc region  */
 	IOMAP_REALTIME =	0x10,	/* mapping on the realtime device  */

fs/xfs/xfs_itable.c

@@ -57,7 +57,7 @@ xfs_bulkstat_one_iget(
 	xfs_bstat_t	*buf,		/* return buffer */
 	int		*stat)		/* BULKSTAT_RV_... */
 {
-	xfs_dinode_core_t *dic;		/* dinode core info pointer */
+	xfs_icdinode_t	*dic;	/* dinode core info pointer */
 	xfs_inode_t	*ip;		/* incore inode pointer */
 	bhv_vnode_t	*vp;
 	int		error;
@@ -151,37 +151,37 @@ xfs_bulkstat_one_dinode(
 	 * the new format. We don't change the version number so that we
 	 * can distinguish this from a real new format inode.
 	 */
-	if (INT_GET(dic->di_version, ARCH_CONVERT) == XFS_DINODE_VERSION_1) {
-		buf->bs_nlink = INT_GET(dic->di_onlink, ARCH_CONVERT);
+	if (dic->di_version == XFS_DINODE_VERSION_1) {
+		buf->bs_nlink = be16_to_cpu(dic->di_onlink);
 		buf->bs_projid = 0;
 	} else {
-		buf->bs_nlink = INT_GET(dic->di_nlink, ARCH_CONVERT);
-		buf->bs_projid = INT_GET(dic->di_projid, ARCH_CONVERT);
+		buf->bs_nlink = be32_to_cpu(dic->di_nlink);
+		buf->bs_projid = be16_to_cpu(dic->di_projid);
 	}
 
 	buf->bs_ino = ino;
-	buf->bs_mode = INT_GET(dic->di_mode, ARCH_CONVERT);
-	buf->bs_uid = INT_GET(dic->di_uid, ARCH_CONVERT);
-	buf->bs_gid = INT_GET(dic->di_gid, ARCH_CONVERT);
-	buf->bs_size = INT_GET(dic->di_size, ARCH_CONVERT);
-	buf->bs_atime.tv_sec = INT_GET(dic->di_atime.t_sec, ARCH_CONVERT);
-	buf->bs_atime.tv_nsec = INT_GET(dic->di_atime.t_nsec, ARCH_CONVERT);
-	buf->bs_mtime.tv_sec = INT_GET(dic->di_mtime.t_sec, ARCH_CONVERT);
-	buf->bs_mtime.tv_nsec = INT_GET(dic->di_mtime.t_nsec, ARCH_CONVERT);
-	buf->bs_ctime.tv_sec = INT_GET(dic->di_ctime.t_sec, ARCH_CONVERT);
-	buf->bs_ctime.tv_nsec = INT_GET(dic->di_ctime.t_nsec, ARCH_CONVERT);
+	buf->bs_mode = be16_to_cpu(dic->di_mode);
+	buf->bs_uid = be32_to_cpu(dic->di_uid);
+	buf->bs_gid = be32_to_cpu(dic->di_gid);
+	buf->bs_size = be64_to_cpu(dic->di_size);
+	buf->bs_atime.tv_sec = be32_to_cpu(dic->di_atime.t_sec);
+	buf->bs_atime.tv_nsec = be32_to_cpu(dic->di_atime.t_nsec);
+	buf->bs_mtime.tv_sec = be32_to_cpu(dic->di_mtime.t_sec);
+	buf->bs_mtime.tv_nsec = be32_to_cpu(dic->di_mtime.t_nsec);
+	buf->bs_ctime.tv_sec = be32_to_cpu(dic->di_ctime.t_sec);
+	buf->bs_ctime.tv_nsec = be32_to_cpu(dic->di_ctime.t_nsec);
 	buf->bs_xflags = xfs_dic2xflags(dic);
-	buf->bs_extsize = INT_GET(dic->di_extsize, ARCH_CONVERT) << mp->m_sb.sb_blocklog;
-	buf->bs_extents = INT_GET(dic->di_nextents, ARCH_CONVERT);
-	buf->bs_gen = INT_GET(dic->di_gen, ARCH_CONVERT);
+	buf->bs_extsize = be32_to_cpu(dic->di_extsize) << mp->m_sb.sb_blocklog;
+	buf->bs_extents = be32_to_cpu(dic->di_nextents);
+	buf->bs_gen = be32_to_cpu(dic->di_gen);
 	memset(buf->bs_pad, 0, sizeof(buf->bs_pad));
-	buf->bs_dmevmask = INT_GET(dic->di_dmevmask, ARCH_CONVERT);
-	buf->bs_dmstate = INT_GET(dic->di_dmstate, ARCH_CONVERT);
-	buf->bs_aextents = INT_GET(dic->di_anextents, ARCH_CONVERT);
+	buf->bs_dmevmask = be32_to_cpu(dic->di_dmevmask);
+	buf->bs_dmstate = be16_to_cpu(dic->di_dmstate);
+	buf->bs_aextents = be16_to_cpu(dic->di_anextents);
 
-	switch (INT_GET(dic->di_format, ARCH_CONVERT)) {
+	switch (dic->di_format) {
 	case XFS_DINODE_FMT_DEV:
-		buf->bs_rdev = INT_GET(dip->di_u.di_dev, ARCH_CONVERT);
+		buf->bs_rdev = be32_to_cpu(dip->di_u.di_dev);
 		buf->bs_blksize = BLKDEV_IOSIZE;
 		buf->bs_blocks = 0;
 		break;
@@ -195,7 +195,7 @@ xfs_bulkstat_one_dinode(
 	case XFS_DINODE_FMT_BTREE:
 		buf->bs_rdev = 0;
 		buf->bs_blksize = mp->m_sb.sb_blocksize;
-		buf->bs_blocks = INT_GET(dic->di_nblocks, ARCH_CONVERT);
+		buf->bs_blocks = be64_to_cpu(dic->di_nblocks);
 		break;
 	}
 
@@ -290,16 +290,23 @@ xfs_bulkstat_use_dinode(
 		return 1;
 	dip = (xfs_dinode_t *)
 			xfs_buf_offset(bp, clustidx << mp->m_sb.sb_inodelog);
-	if (INT_GET(dip->di_core.di_magic, ARCH_CONVERT) != XFS_DINODE_MAGIC ||
-	    !XFS_DINODE_GOOD_VERSION(
-			INT_GET(dip->di_core.di_version, ARCH_CONVERT)))
+	/*
+	 * Check the buffer containing the on-disk inode for di_nlink == 0.
+	 * This is to prevent xfs_bulkstat from picking up just reclaimed
+	 * inodes that have their in-core state initialized but not flushed
+	 * to disk yet. This is a temporary hack that would require a proper
+	 * fix in the future.
+	 */
+	if (be16_to_cpu(dip->di_core.di_magic) != XFS_DINODE_MAGIC ||
+	    !XFS_DINODE_GOOD_VERSION(dip->di_core.di_version) ||
+	    !dip->di_core.di_nlink)
 		return 0;
 	if (flags & BULKSTAT_FG_QUICK) {
 		*dipp = dip;
 		return 1;
 	}
 	/* BULKSTAT_FG_INLINE: if attr fork is local, or not there, use it */
-	aformat = INT_GET(dip->di_core.di_aformat, ARCH_CONVERT);
+	aformat = dip->di_core.di_aformat;
 	if ((XFS_CFORK_Q(&dip->di_core) == 0) ||
 	    (aformat == XFS_DINODE_FMT_LOCAL) ||
 	    (aformat == XFS_DINODE_FMT_EXTENTS && !dip->di_core.di_anextents)) {
@@ -612,21 +619,25 @@ xfs_bulkstat(
 						}
 					}
 				}
+				ino = XFS_AGINO_TO_INO(mp, agno, agino);
+				bno = XFS_AGB_TO_DADDR(mp, agno, agbno);
 				/*
 				 * Skip if this inode is free.
 				 */
-				if (XFS_INOBT_MASK(chunkidx) & irbp->ir_free)
+				if (XFS_INOBT_MASK(chunkidx) & irbp->ir_free) {
+					lastino = ino;
 					continue;
+				}
 				/*
 				 * Count used inodes as free so we can tell
 				 * when the chunk is used up.
 				 */
 				irbp->ir_freecount++;
-				ino = XFS_AGINO_TO_INO(mp, agno, agino);
-				bno = XFS_AGB_TO_DADDR(mp, agno, agbno);
 				if (!xfs_bulkstat_use_dinode(mp, flags, bp,
-							     clustidx, &dip))
+							     clustidx, &dip)) {
+					lastino = ino;
 					continue;
+				}
 				/*
 				 * If we need to do an iget, cannot hold bp.
 				 * Drop it, until starting the next cluster.
@@ -687,8 +698,7 @@ xfs_bulkstat(
 			if (end_of_ag) {
 				agno++;
 				agino = 0;
-			} else
-				agino = XFS_INO_TO_AGINO(mp, lastino);
+			}
 		} else
 			break;
 	}

fs/xfs/xfs_log.c

@@ -252,6 +252,29 @@ xlog_grant_add_space(struct log *log, int bytes)
 	xlog_grant_add_space_reserve(log, bytes);
 }
 
+static void
+xlog_tic_reset_res(xlog_ticket_t *tic)
+{
+	tic->t_res_num = 0;
+	tic->t_res_arr_sum = 0;
+	tic->t_res_num_ophdrs = 0;
+}
+
+static void
+xlog_tic_add_region(xlog_ticket_t *tic, uint len, uint type)
+{
+	if (tic->t_res_num == XLOG_TIC_LEN_MAX) {
+		/* add to overflow and start again */
+		tic->t_res_o_flow += tic->t_res_arr_sum;
+		tic->t_res_num = 0;
+		tic->t_res_arr_sum = 0;
+	}
+
+	tic->t_res_arr[tic->t_res_num].r_len = len;
+	tic->t_res_arr[tic->t_res_num].r_type = type;
+	tic->t_res_arr_sum += len;
+	tic->t_res_num++;
+}
 
 /*
  * NOTES:
@@ -486,7 +509,7 @@ xfs_log_mount(xfs_mount_t	*mp,
 		cmn_err(CE_NOTE,
 			"!Mounting filesystem \"%s\" in no-recovery mode.  Filesystem will be inconsistent.",
 			mp->m_fsname);
-		ASSERT(XFS_MTOVFS(mp)->vfs_flag & VFS_RDONLY);
+		ASSERT(mp->m_flags & XFS_MOUNT_RDONLY);
 	}
 
 	mp->m_log = xlog_alloc_log(mp, log_target, blk_offset, num_bblks);
@@ -496,16 +519,15 @@ xfs_log_mount(xfs_mount_t	*mp,
 	 * just worked.
 	 */
 	if (!(mp->m_flags & XFS_MOUNT_NORECOVERY)) {
-		bhv_vfs_t	*vfsp = XFS_MTOVFS(mp);
-		int		error, readonly = (vfsp->vfs_flag & VFS_RDONLY);
+		int		error, readonly = (mp->m_flags & XFS_MOUNT_RDONLY);
 
 		if (readonly)
-			vfsp->vfs_flag &= ~VFS_RDONLY;
+			mp->m_flags &= ~XFS_MOUNT_RDONLY;
 
 		error = xlog_recover(mp->m_log);
 
 		if (readonly)
-			vfsp->vfs_flag |= VFS_RDONLY;
+			mp->m_flags |= XFS_MOUNT_RDONLY;
 		if (error) {
 			cmn_err(CE_WARN, "XFS: log mount/recovery failed: error %d", error);
 			xlog_dealloc_log(mp->m_log);
@@ -537,7 +559,7 @@ xfs_log_mount_finish(xfs_mount_t *mp, int mfsi_flags)
 		error = xlog_recover_finish(mp->m_log, mfsi_flags);
 	else {
 		error = 0;
-		ASSERT(XFS_MTOVFS(mp)->vfs_flag & VFS_RDONLY);
+		ASSERT(mp->m_flags & XFS_MOUNT_RDONLY);
 	}
 
 	return error;
@@ -597,7 +619,7 @@ xfs_log_unmount_write(xfs_mount_t *mp)
 	 * Don't write out unmount record on read-only mounts.
 	 * Or, if we are doing a forced umount (typically because of IO errors).
 	 */
-	if (XFS_MTOVFS(mp)->vfs_flag & VFS_RDONLY)
+	if (mp->m_flags & XFS_MOUNT_RDONLY)
 		return 0;
 
 	xfs_log_force(mp, 0, XFS_LOG_FORCE|XFS_LOG_SYNC);
@@ -948,6 +970,19 @@ xlog_iodone(xfs_buf_t *bp)
 	 */
 	l = iclog->ic_log;
 
+	/*
+	 * If the ordered flag has been removed by a lower
+	 * layer, it means the underlyin device no longer supports
+	 * barrier I/O. Warn loudly and turn off barriers.
+	 */
+	if ((l->l_mp->m_flags & XFS_MOUNT_BARRIER) && !XFS_BUF_ORDERED(bp)) {
+		l->l_mp->m_flags &= ~XFS_MOUNT_BARRIER;
+		xfs_fs_cmn_err(CE_WARN, l->l_mp,
+				"xlog_iodone: Barriers are no longer supported"
+				" by device. Disabling barriers\n");
+		xfs_buftrace("XLOG_IODONE BARRIERS OFF", bp);
+	}
+
 	/*
 	 * Race to shutdown the filesystem if we see an error.
 	 */
@@ -1012,10 +1047,7 @@ xlog_bdstrat_cb(struct xfs_buf *bp)
 /*
  * Return size of each in-core log record buffer.
  *
- * Low memory machines only get 2 16KB buffers.  We don't want to waste
- * memory here.  However, all other machines get at least 2 32KB buffers.
- * The number is hard coded because we don't care about the minimum
- * memory size, just 32MB systems.
+ * All machines get 8 x 32KB buffers by default, unless tuned otherwise.
  *
  * If the filesystem blocksize is too large, we may need to choose a
  * larger size since the directory code currently logs entire blocks.
@@ -1028,17 +1060,10 @@ xlog_get_iclog_buffer_size(xfs_mount_t	*mp,
 	int size;
 	int xhdrs;
 
-	if (mp->m_logbufs <= 0) {
-		if (xfs_physmem <= btoc(128*1024*1024)) {
-			log->l_iclog_bufs = XLOG_MIN_ICLOGS;
-		} else if (xfs_physmem <= btoc(400*1024*1024)) {
-			log->l_iclog_bufs = XLOG_MED_ICLOGS;
-		} else {	/* 256K with 32K bufs */
-			log->l_iclog_bufs = XLOG_MAX_ICLOGS;
-		}
-	} else {
+	if (mp->m_logbufs <= 0)
+		log->l_iclog_bufs = XLOG_MAX_ICLOGS;
+	else
 		log->l_iclog_bufs = mp->m_logbufs;
-	}
 
 	/*
 	 * Buffer size passed in from mount system call.
@@ -1069,18 +1094,9 @@ xlog_get_iclog_buffer_size(xfs_mount_t	*mp,
 		goto done;
 	}
 
-	/*
-	 * Special case machines that have less than 32MB of memory.
-	 * All machines with more memory use 32KB buffers.
-	 */
-	if (xfs_physmem <= btoc(32*1024*1024)) {
-		/* Don't change; min configuration */
-		log->l_iclog_size = XLOG_RECORD_BSIZE;		/* 16k */
-		log->l_iclog_size_log = XLOG_RECORD_BSHIFT;
-	} else {
-		log->l_iclog_size = XLOG_BIG_RECORD_BSIZE;	/* 32k */
-		log->l_iclog_size_log = XLOG_BIG_RECORD_BSHIFT;
-	}
+	/* All machines use 32KB buffers by default. */
+	log->l_iclog_size = XLOG_BIG_RECORD_BSIZE;
+	log->l_iclog_size_log = XLOG_BIG_RECORD_BSHIFT;
 
 	/* the default log size is 16k or 32k which is one header sector */
 	log->l_iclog_hsize = BBSIZE;
@@ -1771,14 +1787,14 @@ xlog_write(xfs_mount_t *	mp,
     len = 0;
     if (ticket->t_flags & XLOG_TIC_INITED) {    /* acct for start rec of xact */
 	len += sizeof(xlog_op_header_t);
-	XLOG_TIC_ADD_OPHDR(ticket);
+	ticket->t_res_num_ophdrs++;
     }
 
     for (index = 0; index < nentries; index++) {
 	len += sizeof(xlog_op_header_t);	    /* each region gets >= 1 */
-	XLOG_TIC_ADD_OPHDR(ticket);
+	ticket->t_res_num_ophdrs++;
 	len += reg[index].i_len;
-	XLOG_TIC_ADD_REGION(ticket, reg[index].i_len, reg[index].i_type);
+	xlog_tic_add_region(ticket, reg[index].i_len, reg[index].i_type);
     }
     contwr = *start_lsn = 0;
 
@@ -1887,7 +1903,7 @@ xlog_write(xfs_mount_t *	mp,
 		len += sizeof(xlog_op_header_t); /* from splitting of region */
 		/* account for new log op header */
 		ticket->t_curr_res -= sizeof(xlog_op_header_t);
-		XLOG_TIC_ADD_OPHDR(ticket);
+		ticket->t_res_num_ophdrs++;
 	    }
 	    xlog_verify_dest_ptr(log, ptr);
 
@@ -2385,7 +2401,7 @@ restart:
 	 */
 	if (log_offset == 0) {
 		ticket->t_curr_res -= log->l_iclog_hsize;
-		XLOG_TIC_ADD_REGION(ticket,
+		xlog_tic_add_region(ticket,
 				    log->l_iclog_hsize,
 				    XLOG_REG_TYPE_LRHEADER);
 		INT_SET(head->h_cycle, ARCH_CONVERT, log->l_curr_cycle);
@@ -2573,7 +2589,7 @@ xlog_regrant_write_log_space(xlog_t	   *log,
 #endif
 
 	tic->t_curr_res = tic->t_unit_res;
-	XLOG_TIC_RESET_RES(tic);
+	xlog_tic_reset_res(tic);
 
 	if (tic->t_cnt > 0)
 		return 0;
@@ -2714,7 +2730,7 @@ xlog_regrant_reserve_log_space(xlog_t	     *log,
 	s = GRANT_LOCK(log);
 	xlog_grant_sub_space(log, ticket->t_curr_res);
 	ticket->t_curr_res = ticket->t_unit_res;
-	XLOG_TIC_RESET_RES(ticket);
+	xlog_tic_reset_res(ticket);
 	xlog_trace_loggrant(log, ticket,
 			    "xlog_regrant_reserve_log_space: sub current res");
 	xlog_verify_grant_head(log, 1);
@@ -2731,7 +2747,7 @@ xlog_regrant_reserve_log_space(xlog_t	     *log,
 	xlog_verify_grant_head(log, 0);
 	GRANT_UNLOCK(log, s);
 	ticket->t_curr_res = ticket->t_unit_res;
-	XLOG_TIC_RESET_RES(ticket);
+	xlog_tic_reset_res(ticket);
 }	/* xlog_regrant_reserve_log_space */
 
 
@@ -3354,7 +3370,7 @@ xlog_ticket_get(xlog_t		*log,
 		tic->t_flags |= XLOG_TIC_PERM_RESERV;
 	sv_init(&(tic->t_sema), SV_DEFAULT, "logtick");
 
-	XLOG_TIC_RESET_RES(tic);
+	xlog_tic_reset_res(tic);
 
 	return tic;
 }	/* xlog_ticket_get */

fs/xfs/xfs_log_priv.h

@@ -30,17 +30,16 @@ struct xfs_mount;
  */
 
 #define XLOG_MIN_ICLOGS		2
-#define XLOG_MED_ICLOGS		4
 #define XLOG_MAX_ICLOGS		8
 #define XLOG_HEADER_MAGIC_NUM	0xFEEDbabe	/* Invalid cycle number */
 #define XLOG_VERSION_1		1
 #define XLOG_VERSION_2		2		/* Large IClogs, Log sunit */
 #define XLOG_VERSION_OKBITS	(XLOG_VERSION_1 | XLOG_VERSION_2)
-#define XLOG_RECORD_BSIZE	(16*1024)	/* eventually 32k */
+#define XLOG_MIN_RECORD_BSIZE	(16*1024)	/* eventually 32k */
 #define XLOG_BIG_RECORD_BSIZE	(32*1024)	/* 32k buffers */
 #define XLOG_MAX_RECORD_BSIZE	(256*1024)
 #define XLOG_HEADER_CYCLE_SIZE	(32*1024)	/* cycle data in header */
-#define XLOG_RECORD_BSHIFT	14		/* 16384 == 1 << 14 */
+#define XLOG_MIN_RECORD_BSHIFT	14		/* 16384 == 1 << 14 */
 #define XLOG_BIG_RECORD_BSHIFT	15		/* 32k == 1 << 15 */
 #define XLOG_MAX_RECORD_BSHIFT	18		/* 256k == 1 << 18 */
 #define XLOG_BTOLSUNIT(log, b)  (((b)+(log)->l_mp->m_sb.sb_logsunit-1) / \
@@ -250,22 +249,6 @@ typedef __uint32_t xlog_tid_t;
 
 /* Ticket reservation region accounting */ 
 #define XLOG_TIC_LEN_MAX	15
-#define XLOG_TIC_RESET_RES(t) ((t)->t_res_num = \
-				(t)->t_res_arr_sum = (t)->t_res_num_ophdrs = 0)
-#define XLOG_TIC_ADD_OPHDR(t) ((t)->t_res_num_ophdrs++)
-#define XLOG_TIC_ADD_REGION(t, len, type)				\
-	do {								\
-		if ((t)->t_res_num == XLOG_TIC_LEN_MAX) { 		\
-			/* add to overflow and start again */		\
-			(t)->t_res_o_flow += (t)->t_res_arr_sum;	\
-			(t)->t_res_num = 0;				\
-			(t)->t_res_arr_sum = 0;				\
-		}							\
-		(t)->t_res_arr[(t)->t_res_num].r_len = (len);		\
-		(t)->t_res_arr[(t)->t_res_num].r_type = (type);		\
-		(t)->t_res_arr_sum += (len);				\
-		(t)->t_res_num++;					\
-	} while (0)
 
 /*
  * Reservation region

fs/xfs/xfs_log_recover.c

@@ -2245,7 +2245,7 @@ xlog_recover_do_inode_trans(
 	int			error;
 	int			attr_index;
 	uint			fields;
-	xfs_dinode_core_t	*dicp;
+	xfs_icdinode_t		*dicp;
 	int			need_free = 0;
 
 	if (pass == XLOG_RECOVER_PASS1) {
@@ -2309,7 +2309,7 @@ xlog_recover_do_inode_trans(
 	 * Make sure the place we're flushing out to really looks
 	 * like an inode!
 	 */
-	if (unlikely(INT_GET(dip->di_core.di_magic, ARCH_CONVERT) != XFS_DINODE_MAGIC)) {
+	if (unlikely(be16_to_cpu(dip->di_core.di_magic) != XFS_DINODE_MAGIC)) {
 		xfs_buf_relse(bp);
 		xfs_fs_cmn_err(CE_ALERT, mp,
 			"xfs_inode_recover: Bad inode magic number, dino ptr = 0x%p, dino bp = 0x%p, ino = %Ld",
@@ -2319,7 +2319,7 @@ xlog_recover_do_inode_trans(
 		error = EFSCORRUPTED;
 		goto error;
 	}
-	dicp = (xfs_dinode_core_t*)(item->ri_buf[1].i_addr);
+	dicp = (xfs_icdinode_t *)(item->ri_buf[1].i_addr);
 	if (unlikely(dicp->di_magic != XFS_DINODE_MAGIC)) {
 		xfs_buf_relse(bp);
 		xfs_fs_cmn_err(CE_ALERT, mp,
@@ -2332,15 +2332,13 @@ xlog_recover_do_inode_trans(
 	}
 
 	/* Skip replay when the on disk inode is newer than the log one */
-	if (dicp->di_flushiter <
-	    INT_GET(dip->di_core.di_flushiter, ARCH_CONVERT)) {
+	if (dicp->di_flushiter < be16_to_cpu(dip->di_core.di_flushiter)) {
 		/*
 		 * Deal with the wrap case, DI_MAX_FLUSH is less
 		 * than smaller numbers
 		 */
-		if ((INT_GET(dip->di_core.di_flushiter, ARCH_CONVERT)
-							== DI_MAX_FLUSH) &&
-		    (dicp->di_flushiter < (DI_MAX_FLUSH>>1))) {
+		if (be16_to_cpu(dip->di_core.di_flushiter) == DI_MAX_FLUSH &&
+		    dicp->di_flushiter < (DI_MAX_FLUSH >> 1)) {
 			/* do nothing */
 		} else {
 			xfs_buf_relse(bp);
@@ -2411,8 +2409,8 @@ xlog_recover_do_inode_trans(
 	}
 
 	/* The core is in in-core format */
-	xfs_xlate_dinode_core((xfs_caddr_t)&dip->di_core,
-			      (xfs_dinode_core_t*)item->ri_buf[1].i_addr, -1);
+	xfs_dinode_to_disk(&dip->di_core,
+		(xfs_icdinode_t *)item->ri_buf[1].i_addr);
 
 	/* the rest is in on-disk format */
 	if (item->ri_buf[1].i_len > sizeof(xfs_dinode_core_t)) {
@@ -2424,8 +2422,7 @@ xlog_recover_do_inode_trans(
 	fields = in_f->ilf_fields;
 	switch (fields & (XFS_ILOG_DEV | XFS_ILOG_UUID)) {
 	case XFS_ILOG_DEV:
-		INT_SET(dip->di_u.di_dev, ARCH_CONVERT, in_f->ilf_u.ilfu_rdev);
-
+		dip->di_u.di_dev = cpu_to_be32(in_f->ilf_u.ilfu_rdev);
 		break;
 	case XFS_ILOG_UUID:
 		dip->di_u.di_muuid = in_f->ilf_u.ilfu_uuid;
@@ -3234,8 +3231,8 @@ xlog_recover_process_iunlinks(
 					ASSERT(ip->i_d.di_nlink == 0);
 
 					/* setup for the next pass */
-					agino = INT_GET(dip->di_next_unlinked,
-							ARCH_CONVERT);
+					agino = be32_to_cpu(
+							dip->di_next_unlinked);
 					xfs_buf_relse(ibp);
 					/*
 					 * Prevent any DMAPI event from
@@ -3837,7 +3834,10 @@ xlog_do_recover(
 	 */
 	bp = xfs_getsb(log->l_mp, 0);
 	XFS_BUF_UNDONE(bp);
+	ASSERT(!(XFS_BUF_ISWRITE(bp)));
+	ASSERT(!(XFS_BUF_ISDELAYWRITE(bp)));
 	XFS_BUF_READ(bp);
+	XFS_BUF_UNASYNC(bp);
 	xfsbdstrat(log->l_mp, bp);
 	if ((error = xfs_iowait(bp))) {
 		xfs_ioerror_alert("xlog_do_recover",
@@ -3849,7 +3849,7 @@ xlog_do_recover(
 
 	/* Convert superblock from on-disk format */
 	sbp = &log->l_mp->m_sb;
-	xfs_xlatesb(XFS_BUF_TO_SBP(bp), sbp, 1, XFS_SB_ALL_BITS);
+	xfs_sb_from_disk(sbp, XFS_BUF_TO_SBP(bp));
 	ASSERT(sbp->sb_magicnum == XFS_SB_MAGIC);
 	ASSERT(XFS_SB_GOOD_VERSION(sbp));
 	xfs_buf_relse(bp);
@@ -4027,7 +4027,7 @@ xlog_recover_check_summary(
 	sbbp = xfs_getsb(mp, 0);
 #ifdef XFS_LOUD_RECOVERY
 	sbp = &mp->m_sb;
-	xfs_xlatesb(XFS_BUF_TO_SBP(sbbp), sbp, 1, XFS_SB_ALL_BITS);
+	xfs_sb_from_disk(sbp, XFS_BUF_TO_SBP(sbbp));
 	cmn_err(CE_NOTE,
 		"xlog_recover_check_summary: sb_icount %Lu itotal %Lu",
 		sbp->sb_icount, itotal);

fs/xfs/xfs_mount.c

@@ -139,7 +139,7 @@ xfs_mount_init(void)
 	AIL_LOCKINIT(&mp->m_ail_lock, "xfs_ail");
 	spinlock_init(&mp->m_sb_lock, "xfs_sb");
 	mutex_init(&mp->m_ilock);
-	initnsema(&mp->m_growlock, 1, "xfs_grow");
+	mutex_init(&mp->m_growlock);
 	/*
 	 * Initialize the AIL.
 	 */
@@ -157,14 +157,8 @@ xfs_mount_init(void)
  */
 void
 xfs_mount_free(
-	xfs_mount_t	*mp,
-	int		remove_bhv)
+	xfs_mount_t	*mp)
 {
-	if (mp->m_ihash)
-		xfs_ihash_free(mp);
-	if (mp->m_chash)
-		xfs_chash_free(mp);
-
 	if (mp->m_perag) {
 		int	agno;
 
@@ -180,7 +174,7 @@ xfs_mount_free(
 	AIL_LOCK_DESTROY(&mp->m_ail_lock);
 	spinlock_destroy(&mp->m_sb_lock);
 	mutex_destroy(&mp->m_ilock);
-	freesema(&mp->m_growlock);
+	mutex_destroy(&mp->m_growlock);
 	if (mp->m_quotainfo)
 		XFS_QM_DONE(mp);
 
@@ -191,15 +185,7 @@ xfs_mount_free(
 	if (mp->m_logname != NULL)
 		kmem_free(mp->m_logname, strlen(mp->m_logname) + 1);
 
-	if (remove_bhv) {
-		struct bhv_vfs	*vfsp = XFS_MTOVFS(mp);
-
-		bhv_remove_all_vfsops(vfsp, 0);
-		VFS_REMOVEBHV(vfsp, &mp->m_bhv);
-	}
-
 	xfs_icsb_destroy_counters(mp);
-	kmem_free(mp, sizeof(xfs_mount_t));
 }
 
 /*
@@ -342,9 +328,19 @@ xfs_mount_validate_sb(
 	return 0;
 }
 
+STATIC void
+xfs_initialize_perag_icache(
+	xfs_perag_t	*pag)
+{
+	if (!pag->pag_ici_init) {
+		rwlock_init(&pag->pag_ici_lock);
+		INIT_RADIX_TREE(&pag->pag_ici_root, GFP_ATOMIC);
+		pag->pag_ici_init = 1;
+	}
+}
+
 xfs_agnumber_t
 xfs_initialize_perag(
-	bhv_vfs_t	*vfs,
 	xfs_mount_t	*mp,
 	xfs_agnumber_t	agcount)
 {
@@ -362,7 +358,7 @@ xfs_initialize_perag(
 	/* Clear the mount flag if no inode can overflow 32 bits
 	 * on this filesystem, or if specifically requested..
 	 */
-	if ((vfs->vfs_flag & VFS_32BITINODES) && ino > max_inum) {
+	if ((mp->m_flags & XFS_MOUNT_SMALL_INUMS) && ino > max_inum) {
 		mp->m_flags |= XFS_MOUNT_32BITINODES;
 	} else {
 		mp->m_flags &= ~XFS_MOUNT_32BITINODES;
@@ -396,48 +392,92 @@ xfs_initialize_perag(
 			pag->pagi_inodeok = 1;
 			if (index < max_metadata)
 				pag->pagf_metadata = 1;
+			xfs_initialize_perag_icache(pag);
 		}
 	} else {
 		/* Setup default behavior for smaller filesystems */
 		for (index = 0; index < agcount; index++) {
 			pag = &mp->m_perag[index];
 			pag->pagi_inodeok = 1;
+			xfs_initialize_perag_icache(pag);
 		}
 	}
 	return index;
 }
 
+void
+xfs_sb_from_disk(
+	xfs_sb_t	*to,
+	xfs_dsb_t	*from)
+{
+	to->sb_magicnum = be32_to_cpu(from->sb_magicnum);
+	to->sb_blocksize = be32_to_cpu(from->sb_blocksize);
+	to->sb_dblocks = be64_to_cpu(from->sb_dblocks);
+	to->sb_rblocks = be64_to_cpu(from->sb_rblocks);
+	to->sb_rextents = be64_to_cpu(from->sb_rextents);
+	memcpy(&to->sb_uuid, &from->sb_uuid, sizeof(to->sb_uuid));
+	to->sb_logstart = be64_to_cpu(from->sb_logstart);
+	to->sb_rootino = be64_to_cpu(from->sb_rootino);
+	to->sb_rbmino = be64_to_cpu(from->sb_rbmino);
+	to->sb_rsumino = be64_to_cpu(from->sb_rsumino);
+	to->sb_rextsize = be32_to_cpu(from->sb_rextsize);
+	to->sb_agblocks = be32_to_cpu(from->sb_agblocks);
+	to->sb_agcount = be32_to_cpu(from->sb_agcount);
+	to->sb_rbmblocks = be32_to_cpu(from->sb_rbmblocks);
+	to->sb_logblocks = be32_to_cpu(from->sb_logblocks);
+	to->sb_versionnum = be16_to_cpu(from->sb_versionnum);
+	to->sb_sectsize = be16_to_cpu(from->sb_sectsize);
+	to->sb_inodesize = be16_to_cpu(from->sb_inodesize);
+	to->sb_inopblock = be16_to_cpu(from->sb_inopblock);
+	memcpy(&to->sb_fname, &from->sb_fname, sizeof(to->sb_fname));
+	to->sb_blocklog = from->sb_blocklog;
+	to->sb_sectlog = from->sb_sectlog;
+	to->sb_inodelog = from->sb_inodelog;
+	to->sb_inopblog = from->sb_inopblog;
+	to->sb_agblklog = from->sb_agblklog;
+	to->sb_rextslog = from->sb_rextslog;
+	to->sb_inprogress = from->sb_inprogress;
+	to->sb_imax_pct = from->sb_imax_pct;
+	to->sb_icount = be64_to_cpu(from->sb_icount);
+	to->sb_ifree = be64_to_cpu(from->sb_ifree);
+	to->sb_fdblocks = be64_to_cpu(from->sb_fdblocks);
+	to->sb_frextents = be64_to_cpu(from->sb_frextents);
+	to->sb_uquotino = be64_to_cpu(from->sb_uquotino);
+	to->sb_gquotino = be64_to_cpu(from->sb_gquotino);
+	to->sb_qflags = be16_to_cpu(from->sb_qflags);
+	to->sb_flags = from->sb_flags;
+	to->sb_shared_vn = from->sb_shared_vn;
+	to->sb_inoalignmt = be32_to_cpu(from->sb_inoalignmt);
+	to->sb_unit = be32_to_cpu(from->sb_unit);
+	to->sb_width = be32_to_cpu(from->sb_width);
+	to->sb_dirblklog = from->sb_dirblklog;
+	to->sb_logsectlog = from->sb_logsectlog;
+	to->sb_logsectsize = be16_to_cpu(from->sb_logsectsize);
+	to->sb_logsunit = be32_to_cpu(from->sb_logsunit);
+	to->sb_features2 = be32_to_cpu(from->sb_features2);
+}
+
 /*
- * xfs_xlatesb
+ * Copy in core superblock to ondisk one.
  *
- *     data       - on disk version of sb
- *     sb         - a superblock
- *     dir        - conversion direction: <0 - convert sb to buf
- *                                        >0 - convert buf to sb
- *     fields     - which fields to copy (bitmask)
+ * The fields argument is mask of superblock fields to copy.
  */
 void
-xfs_xlatesb(
-	void		*data,
-	xfs_sb_t	*sb,
-	int		dir,
+xfs_sb_to_disk(
+	xfs_dsb_t	*to,
+	xfs_sb_t	*from,
 	__int64_t	fields)
 {
-	xfs_caddr_t	buf_ptr;
-	xfs_caddr_t	mem_ptr;
+	xfs_caddr_t	to_ptr = (xfs_caddr_t)to;
+	xfs_caddr_t	from_ptr = (xfs_caddr_t)from;
 	xfs_sb_field_t	f;
 	int		first;
 	int		size;
 
-	ASSERT(dir);
 	ASSERT(fields);
-
 	if (!fields)
 		return;
 
-	buf_ptr = (xfs_caddr_t)data;
-	mem_ptr = (xfs_caddr_t)sb;
-
 	while (fields) {
 		f = (xfs_sb_field_t)xfs_lowbit64((__uint64_t)fields);
 		first = xfs_sb_info[f].offset;
@@ -446,26 +486,20 @@ xfs_xlatesb(
 		ASSERT(xfs_sb_info[f].type == 0 || xfs_sb_info[f].type == 1);
 
 		if (size == 1 || xfs_sb_info[f].type == 1) {
-			if (dir > 0) {
-				memcpy(mem_ptr + first, buf_ptr + first, size);
-			} else {
-				memcpy(buf_ptr + first, mem_ptr + first, size);
-			}
+			memcpy(to_ptr + first, from_ptr + first, size);
 		} else {
 			switch (size) {
 			case 2:
-				INT_XLATE(*(__uint16_t*)(buf_ptr+first),
-					  *(__uint16_t*)(mem_ptr+first),
-					  dir, ARCH_CONVERT);
+				*(__be16 *)(to_ptr + first) =
+					cpu_to_be16(*(__u16 *)(from_ptr + first));
 				break;
 			case 4:
-				INT_XLATE(*(__uint32_t*)(buf_ptr+first),
-					  *(__uint32_t*)(mem_ptr+first),
-					  dir, ARCH_CONVERT);
+				*(__be32 *)(to_ptr + first) =
+					cpu_to_be32(*(__u32 *)(from_ptr + first));
 				break;
 			case 8:
-				INT_XLATE(*(__uint64_t*)(buf_ptr+first),
-					  *(__uint64_t*)(mem_ptr+first), dir, ARCH_CONVERT);
+				*(__be64 *)(to_ptr + first) =
+					cpu_to_be64(*(__u64 *)(from_ptr + first));
 				break;
 			default:
 				ASSERT(0);
@@ -487,7 +521,6 @@ xfs_readsb(xfs_mount_t *mp, int flags)
 	unsigned int	sector_size;
 	unsigned int	extra_flags;
 	xfs_buf_t	*bp;
-	xfs_sb_t	*sbp;
 	int		error;
 
 	ASSERT(mp->m_sb_bp == NULL);
@@ -515,8 +548,7 @@ xfs_readsb(xfs_mount_t *mp, int flags)
 	 * Initialize the mount structure from the superblock.
 	 * But first do some basic consistency checking.
 	 */
-	sbp = XFS_BUF_TO_SBP(bp);
-	xfs_xlatesb(XFS_BUF_PTR(bp), &(mp->m_sb), 1, XFS_SB_ALL_BITS);
+	xfs_sb_from_disk(&mp->m_sb, XFS_BUF_TO_SBP(bp));
 
 	error = xfs_mount_validate_sb(mp, &(mp->m_sb), flags);
 	if (error) {
@@ -715,7 +747,6 @@ xfs_initialize_perag_data(xfs_mount_t *mp, xfs_agnumber_t agcount)
  */
 int
 xfs_mountfs(
-	bhv_vfs_t	*vfsp,
 	xfs_mount_t	*mp,
 	int		mfsi_flags)
 {
@@ -842,14 +873,11 @@ xfs_mountfs(
 	 */
 	if ((mfsi_flags & XFS_MFSI_SECOND) == 0 &&
 	    (mp->m_flags & XFS_MOUNT_NOUUID) == 0) {
-		__uint64_t	ret64;
 		if (xfs_uuid_mount(mp)) {
 			error = XFS_ERROR(EINVAL);
 			goto error1;
 		}
 		uuid_mounted=1;
-		ret64 = uuid_hash64(&sbp->sb_uuid);
-		memcpy(&vfsp->vfs_fsid, &ret64, sizeof(ret64));
 	}
 
 	/*
@@ -871,16 +899,6 @@ xfs_mountfs(
 		writeio_log = mp->m_writeio_log;
 	}
 
-	/*
-	 * Set the number of readahead buffers to use based on
-	 * physical memory size.
-	 */
-	if (xfs_physmem <= 4096)		/* <= 16MB */
-		mp->m_nreadaheads = XFS_RW_NREADAHEAD_16MB;
-	else if (xfs_physmem <= 8192)	/* <= 32MB */
-		mp->m_nreadaheads = XFS_RW_NREADAHEAD_32MB;
-	else
-		mp->m_nreadaheads = XFS_RW_NREADAHEAD_K32;
 	if (sbp->sb_blocklog > readio_log) {
 		mp->m_readio_log = sbp->sb_blocklog;
 	} else {
@@ -895,15 +913,12 @@ xfs_mountfs(
 	mp->m_writeio_blocks = 1 << (mp->m_writeio_log - sbp->sb_blocklog);
 
 	/*
-	 * Set the inode cluster size based on the physical memory
-	 * size.  This may still be overridden by the file system
+	 * Set the inode cluster size.
+	 * This may still be overridden by the file system
 	 * block size if it is larger than the chosen cluster size.
 	 */
-	if (xfs_physmem <= btoc(32 * 1024 * 1024)) { /* <= 32 MB */
-		mp->m_inode_cluster_size = XFS_INODE_SMALL_CLUSTER_SIZE;
-	} else {
-		mp->m_inode_cluster_size = XFS_INODE_BIG_CLUSTER_SIZE;
-	}
+	mp->m_inode_cluster_size = XFS_INODE_BIG_CLUSTER_SIZE;
+
 	/*
 	 * Set whether we're using inode alignment.
 	 */
@@ -987,16 +1002,6 @@ xfs_mountfs(
 	 */
 	uuid_getnodeuniq(&sbp->sb_uuid, mp->m_fixedfsid);
 
-	/*
-	 *  The vfs structure needs to have a file system independent
-	 *  way of checking for the invariant file system ID.  Since it
-	 *  can't look at mount structures it has a pointer to the data
-	 *  in the mount structure.
-	 *
-	 *  File systems that don't support user level file handles (i.e.
-	 *  all of them except for XFS) will leave vfs_altfsid as NULL.
-	 */
-	vfsp->vfs_altfsid = (xfs_fsid_t *)mp->m_fixedfsid;
 	mp->m_dmevmask = 0;	/* not persistent; set after each mount */
 
 	xfs_dir_mount(mp);
@@ -1011,13 +1016,6 @@ xfs_mountfs(
 	 */
 	xfs_trans_init(mp);
 
-	/*
-	 * Allocate and initialize the inode hash table for this
-	 * file system.
-	 */
-	xfs_ihash_init(mp);
-	xfs_chash_init(mp);
-
 	/*
 	 * Allocate and initialize the per-ag data.
 	 */
@@ -1025,7 +1023,7 @@ xfs_mountfs(
 	mp->m_perag =
 		kmem_zalloc(sbp->sb_agcount * sizeof(xfs_perag_t), KM_SLEEP);
 
-	mp->m_maxagi = xfs_initialize_perag(vfsp, mp, sbp->sb_agcount);
+	mp->m_maxagi = xfs_initialize_perag(mp, sbp->sb_agcount);
 
 	/*
 	 * log's mount-time initialization. Perform 1st part recovery if needed
@@ -1116,7 +1114,7 @@ xfs_mountfs(
 	 * If fs is not mounted readonly, then update the superblock
 	 * unit and width changes.
 	 */
-	if (update_flags && !(vfsp->vfs_flag & VFS_RDONLY))
+	if (update_flags && !(mp->m_flags & XFS_MOUNT_RDONLY))
 		xfs_mount_log_sbunit(mp, update_flags);
 
 	/*
@@ -1169,8 +1167,6 @@ xfs_mountfs(
  error3:
 	xfs_log_unmount_dealloc(mp);
  error2:
-	xfs_ihash_free(mp);
-	xfs_chash_free(mp);
 	for (agno = 0; agno < sbp->sb_agcount; agno++)
 		if (mp->m_perag[agno].pagb_list)
 			kmem_free(mp->m_perag[agno].pagb_list,
@@ -1194,10 +1190,6 @@ xfs_mountfs(
 int
 xfs_unmountfs(xfs_mount_t *mp, struct cred *cr)
 {
-	struct bhv_vfs	*vfsp = XFS_MTOVFS(mp);
-#if defined(DEBUG) || defined(INDUCE_IO_ERROR)
-	int64_t		fsid;
-#endif
 	__uint64_t	resblks;
 
 	/*
@@ -1261,21 +1253,17 @@ xfs_unmountfs(xfs_mount_t *mp, struct cred *cr)
 		xfs_uuid_unmount(mp);
 
 #if defined(DEBUG) || defined(INDUCE_IO_ERROR)
-	/*
-	 * clear all error tags on this filesystem
-	 */
-	memcpy(&fsid, &vfsp->vfs_fsid, sizeof(int64_t));
-	xfs_errortag_clearall_umount(fsid, mp->m_fsname, 0);
+	xfs_errortag_clearall(mp, 0);
 #endif
-	XFS_IODONE(vfsp);
-	xfs_mount_free(mp, 1);
+	XFS_IODONE(mp);
+	xfs_mount_free(mp);
 	return 0;
 }
 
 void
 xfs_unmountfs_close(xfs_mount_t *mp, struct cred *cr)
 {
-	if (mp->m_logdev_targp != mp->m_ddev_targp)
+	if (mp->m_logdev_targp && mp->m_logdev_targp != mp->m_ddev_targp)
 		xfs_free_buftarg(mp->m_logdev_targp, 1);
 	if (mp->m_rtdev_targp)
 		xfs_free_buftarg(mp->m_rtdev_targp, 1);
@@ -1295,10 +1283,8 @@ xfs_unmountfs_wait(xfs_mount_t *mp)
 int
 xfs_fs_writable(xfs_mount_t *mp)
 {
-	bhv_vfs_t	*vfsp = XFS_MTOVFS(mp);
-
-	return !(vfs_test_for_freeze(vfsp) || XFS_FORCED_SHUTDOWN(mp) ||
-		(vfsp->vfs_flag & VFS_RDONLY));
+	return !(xfs_test_for_freeze(mp) || XFS_FORCED_SHUTDOWN(mp) ||
+		(mp->m_flags & XFS_MOUNT_RDONLY));
 }
 
 /*
@@ -1348,34 +1334,44 @@ xfs_log_sbcount(
 	return 0;
 }
 
+STATIC void
+xfs_mark_shared_ro(
+	xfs_mount_t	*mp,
+	xfs_buf_t	*bp)
+{
+	xfs_dsb_t	*sb = XFS_BUF_TO_SBP(bp);
+	__uint16_t	version;
+
+	if (!(sb->sb_flags & XFS_SBF_READONLY))
+		sb->sb_flags |= XFS_SBF_READONLY;
+
+	version = be16_to_cpu(sb->sb_versionnum);
+	if ((version & XFS_SB_VERSION_NUMBITS) != XFS_SB_VERSION_4 ||
+	    !(version & XFS_SB_VERSION_SHAREDBIT))
+		version |= XFS_SB_VERSION_SHAREDBIT;
+	sb->sb_versionnum = cpu_to_be16(version);
+}
+
 int
 xfs_unmountfs_writesb(xfs_mount_t *mp)
 {
 	xfs_buf_t	*sbp;
-	xfs_sb_t	*sb;
 	int		error = 0;
 
 	/*
 	 * skip superblock write if fs is read-only, or
 	 * if we are doing a forced umount.
 	 */
-	if (!(XFS_MTOVFS(mp)->vfs_flag & VFS_RDONLY ||
+	if (!((mp->m_flags & XFS_MOUNT_RDONLY) ||
 		XFS_FORCED_SHUTDOWN(mp))) {
 
 		sbp = xfs_getsb(mp, 0);
- 		sb = XFS_BUF_TO_SBP(sbp);
 
 		/*
 		 * mark shared-readonly if desired
 		 */
-		if (mp->m_mk_sharedro) {
-			if (!(sb->sb_flags & XFS_SBF_READONLY))
-				sb->sb_flags |= XFS_SBF_READONLY;
-			if (!XFS_SB_VERSION_HASSHARED(sb))
-				XFS_SB_VERSION_ADDSHARED(sb);
-			xfs_fs_cmn_err(CE_NOTE, mp,
-				"Unmounting, marking shared read-only");
-		}
+		if (mp->m_mk_sharedro)
+			xfs_mark_shared_ro(mp, sbp);
 
 		XFS_BUF_UNDONE(sbp);
 		XFS_BUF_UNREAD(sbp);
@@ -1410,7 +1406,6 @@ xfs_mod_sb(xfs_trans_t *tp, __int64_t fields)
 	int		first;
 	int		last;
 	xfs_mount_t	*mp;
-	xfs_sb_t	*sbp;
 	xfs_sb_field_t	f;
 
 	ASSERT(fields);
@@ -1418,13 +1413,12 @@ xfs_mod_sb(xfs_trans_t *tp, __int64_t fields)
 		return;
 	mp = tp->t_mountp;
 	bp = xfs_trans_getsb(tp, mp, 0);
-	sbp = XFS_BUF_TO_SBP(bp);
 	first = sizeof(xfs_sb_t);
 	last = 0;
 
 	/* translate/copy */
 
-	xfs_xlatesb(XFS_BUF_PTR(bp), &(mp->m_sb), -1, fields);
+	xfs_sb_to_disk(XFS_BUF_TO_SBP(bp), &mp->m_sb, fields);
 
 	/* find modified range */
 

fs/xfs/xfs_mount.h

@@ -54,13 +54,8 @@ typedef struct xfs_trans_reservations {
 #else
 struct cred;
 struct log;
-struct bhv_vfs;
-struct bhv_vnode;
 struct xfs_mount_args;
-struct xfs_ihash;
-struct xfs_chash;
 struct xfs_inode;
-struct xfs_perag;
 struct xfs_iocore;
 struct xfs_bmbt_irec;
 struct xfs_bmap_free;
@@ -68,9 +63,6 @@ struct xfs_extdelta;
 struct xfs_swapext;
 struct xfs_mru_cache;
 
-extern struct bhv_vfsops xfs_vfsops;
-extern struct bhv_vnodeops xfs_vnodeops;
-
 #define	AIL_LOCK_T		lock_t
 #define	AIL_LOCKINIT(x,y)	spinlock_init(x,y)
 #define	AIL_LOCK_DESTROY(x)	spinlock_destroy(x)
@@ -82,15 +74,17 @@ extern struct bhv_vnodeops xfs_vnodeops;
  * Prototypes and functions for the Data Migration subsystem.
  */
 
-typedef int	(*xfs_send_data_t)(int, struct bhv_vnode *,
+typedef int	(*xfs_send_data_t)(int, bhv_vnode_t *,
 			xfs_off_t, size_t, int, bhv_vrwlock_t *);
 typedef int	(*xfs_send_mmap_t)(struct vm_area_struct *, uint);
-typedef int	(*xfs_send_destroy_t)(struct bhv_vnode *, dm_right_t);
-typedef int	(*xfs_send_namesp_t)(dm_eventtype_t, struct bhv_vfs *,
-			struct bhv_vnode *,
-			dm_right_t, struct bhv_vnode *, dm_right_t,
+typedef int	(*xfs_send_destroy_t)(bhv_vnode_t *, dm_right_t);
+typedef int	(*xfs_send_namesp_t)(dm_eventtype_t, struct xfs_mount *,
+			bhv_vnode_t *,
+			dm_right_t, bhv_vnode_t *, dm_right_t,
 			char *, char *, mode_t, int, int);
-typedef void	(*xfs_send_unmount_t)(struct bhv_vfs *, struct bhv_vnode *,
+typedef int	(*xfs_send_mount_t)(struct xfs_mount *, dm_right_t,
+			char *, char *);
+typedef void	(*xfs_send_unmount_t)(struct xfs_mount *, bhv_vnode_t *,
 			dm_right_t, mode_t, int, int);
 
 typedef struct xfs_dmops {
@@ -98,21 +92,24 @@ typedef struct xfs_dmops {
 	xfs_send_mmap_t		xfs_send_mmap;
 	xfs_send_destroy_t	xfs_send_destroy;
 	xfs_send_namesp_t	xfs_send_namesp;
+	xfs_send_mount_t	xfs_send_mount;
 	xfs_send_unmount_t	xfs_send_unmount;
 } xfs_dmops_t;
 
 #define XFS_SEND_DATA(mp, ev,vp,off,len,fl,lock) \
-	(*(mp)->m_dm_ops.xfs_send_data)(ev,vp,off,len,fl,lock)
+	(*(mp)->m_dm_ops->xfs_send_data)(ev,vp,off,len,fl,lock)
 #define XFS_SEND_MMAP(mp, vma,fl) \
-	(*(mp)->m_dm_ops.xfs_send_mmap)(vma,fl)
+	(*(mp)->m_dm_ops->xfs_send_mmap)(vma,fl)
 #define XFS_SEND_DESTROY(mp, vp,right) \
-	(*(mp)->m_dm_ops.xfs_send_destroy)(vp,right)
+	(*(mp)->m_dm_ops->xfs_send_destroy)(vp,right)
 #define XFS_SEND_NAMESP(mp, ev,b1,r1,b2,r2,n1,n2,mode,rval,fl) \
-	(*(mp)->m_dm_ops.xfs_send_namesp)(ev,NULL,b1,r1,b2,r2,n1,n2,mode,rval,fl)
-#define XFS_SEND_PREUNMOUNT(mp, vfs,b1,r1,b2,r2,n1,n2,mode,rval,fl) \
-	(*(mp)->m_dm_ops.xfs_send_namesp)(DM_EVENT_PREUNMOUNT,vfs,b1,r1,b2,r2,n1,n2,mode,rval,fl)
-#define XFS_SEND_UNMOUNT(mp, vfsp,vp,right,mode,rval,fl) \
-	(*(mp)->m_dm_ops.xfs_send_unmount)(vfsp,vp,right,mode,rval,fl)
+	(*(mp)->m_dm_ops->xfs_send_namesp)(ev,NULL,b1,r1,b2,r2,n1,n2,mode,rval,fl)
+#define XFS_SEND_PREUNMOUNT(mp,b1,r1,b2,r2,n1,n2,mode,rval,fl) \
+	(*(mp)->m_dm_ops->xfs_send_namesp)(DM_EVENT_PREUNMOUNT,mp,b1,r1,b2,r2,n1,n2,mode,rval,fl)
+#define XFS_SEND_MOUNT(mp,right,path,name) \
+	(*(mp)->m_dm_ops->xfs_send_mount)(mp,right,path,name)
+#define XFS_SEND_UNMOUNT(mp, vp,right,mode,rval,fl) \
+	(*(mp)->m_dm_ops->xfs_send_unmount)(mp,vp,right,mode,rval,fl)
 
 
 /*
@@ -142,6 +139,9 @@ typedef struct xfs_dquot * (*xfs_dqvopchown_t)(
 			struct xfs_dquot **, struct xfs_dquot *);
 typedef int	(*xfs_dqvopchownresv_t)(struct xfs_trans *, struct xfs_inode *,
 			struct xfs_dquot *, struct xfs_dquot *, uint);
+typedef void	(*xfs_dqstatvfs_t)(struct xfs_inode *, bhv_statvfs_t *);
+typedef int	(*xfs_dqsync_t)(struct xfs_mount *, int flags);
+typedef int	(*xfs_quotactl_t)(struct xfs_mount *, int, int, xfs_caddr_t);
 
 typedef struct xfs_qmops {
 	xfs_qminit_t		xfs_qminit;
@@ -157,42 +157,51 @@ typedef struct xfs_qmops {
 	xfs_dqvoprename_t	xfs_dqvoprename;
 	xfs_dqvopchown_t	xfs_dqvopchown;
 	xfs_dqvopchownresv_t	xfs_dqvopchownresv;
+	xfs_dqstatvfs_t		xfs_dqstatvfs;
+	xfs_dqsync_t		xfs_dqsync;
+	xfs_quotactl_t		xfs_quotactl;
 	struct xfs_dqtrxops	*xfs_dqtrxops;
 } xfs_qmops_t;
 
 #define XFS_QM_INIT(mp, mnt, fl) \
-	(*(mp)->m_qm_ops.xfs_qminit)(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)
+	(*(mp)->m_qm_ops->xfs_qmmount)(mp, mnt, fl, mfsi_flags)
 #define XFS_QM_UNMOUNT(mp) \
-	(*(mp)->m_qm_ops.xfs_qmunmount)(mp)
+	(*(mp)->m_qm_ops->xfs_qmunmount)(mp)
 #define XFS_QM_DONE(mp) \
-	(*(mp)->m_qm_ops.xfs_qmdone)(mp)
+	(*(mp)->m_qm_ops->xfs_qmdone)(mp)
 #define XFS_QM_DQRELE(mp, dq) \
-	(*(mp)->m_qm_ops.xfs_dqrele)(dq)
+	(*(mp)->m_qm_ops->xfs_dqrele)(dq)
 #define XFS_QM_DQATTACH(mp, ip, fl) \
-	(*(mp)->m_qm_ops.xfs_dqattach)(ip, fl)
+	(*(mp)->m_qm_ops->xfs_dqattach)(ip, fl)
 #define XFS_QM_DQDETACH(mp, ip) \
-	(*(mp)->m_qm_ops.xfs_dqdetach)(ip)
+	(*(mp)->m_qm_ops->xfs_dqdetach)(ip)
 #define XFS_QM_DQPURGEALL(mp, fl) \
-	(*(mp)->m_qm_ops.xfs_dqpurgeall)(mp, fl)
+	(*(mp)->m_qm_ops->xfs_dqpurgeall)(mp, fl)
 #define XFS_QM_DQVOPALLOC(mp, ip, uid, gid, prid, fl, dq1, dq2) \
-	(*(mp)->m_qm_ops.xfs_dqvopalloc)(mp, ip, uid, gid, prid, fl, dq1, dq2)
+	(*(mp)->m_qm_ops->xfs_dqvopalloc)(mp, ip, uid, gid, prid, fl, dq1, dq2)
 #define XFS_QM_DQVOPCREATE(mp, tp, ip, dq1, dq2) \
-	(*(mp)->m_qm_ops.xfs_dqvopcreate)(tp, ip, dq1, dq2)
+	(*(mp)->m_qm_ops->xfs_dqvopcreate)(tp, ip, dq1, dq2)
 #define XFS_QM_DQVOPRENAME(mp, ip) \
-	(*(mp)->m_qm_ops.xfs_dqvoprename)(ip)
+	(*(mp)->m_qm_ops->xfs_dqvoprename)(ip)
 #define XFS_QM_DQVOPCHOWN(mp, tp, ip, dqp, dq) \
-	(*(mp)->m_qm_ops.xfs_dqvopchown)(tp, ip, dqp, dq)
+	(*(mp)->m_qm_ops->xfs_dqvopchown)(tp, ip, dqp, dq)
 #define XFS_QM_DQVOPCHOWNRESV(mp, tp, ip, dq1, dq2, fl) \
-	(*(mp)->m_qm_ops.xfs_dqvopchownresv)(tp, ip, dq1, dq2, fl)
+	(*(mp)->m_qm_ops->xfs_dqvopchownresv)(tp, ip, dq1, dq2, fl)
+#define XFS_QM_DQSTATVFS(ip, statp) \
+	(*(ip)->i_mount->m_qm_ops->xfs_dqstatvfs)(ip, statp)
+#define XFS_QM_DQSYNC(mp, flags) \
+	(*(mp)->m_qm_ops->xfs_dqsync)(mp, flags)
+#define XFS_QM_QUOTACTL(mp, cmd, id, addr) \
+	(*(mp)->m_qm_ops->xfs_quotactl)(mp, cmd, id, addr)
 
 
 /*
  * Prototypes and functions for I/O core modularization.
  */
 
-typedef int		(*xfs_ioinit_t)(struct bhv_vfs *,
+typedef int		(*xfs_ioinit_t)(struct xfs_mount *,
 				struct xfs_mount_args *, int);
 typedef int		(*xfs_bmapi_t)(struct xfs_trans *, void *,
 				xfs_fileoff_t, xfs_filblks_t, int,
@@ -222,7 +231,7 @@ typedef void		(*xfs_lock_demote_t)(void *, uint);
 typedef int		(*xfs_lock_nowait_t)(void *, uint);
 typedef void		(*xfs_unlk_t)(void *, unsigned int);
 typedef xfs_fsize_t	(*xfs_size_t)(void *);
-typedef xfs_fsize_t	(*xfs_iodone_t)(struct bhv_vfs *);
+typedef xfs_fsize_t	(*xfs_iodone_t)(struct xfs_mount *);
 typedef int		(*xfs_swap_extents_t)(void *, void *,
 				struct xfs_swapext*);
 
@@ -245,8 +254,8 @@ typedef struct xfs_ioops {
 	xfs_swap_extents_t		xfs_swap_extents_func;
 } xfs_ioops_t;
 
-#define XFS_IOINIT(vfsp, args, flags) \
-	(*(mp)->m_io_ops.xfs_ioinit)(vfsp, args, flags)
+#define XFS_IOINIT(mp, args, flags) \
+	(*(mp)->m_io_ops.xfs_ioinit)(mp, args, flags)
 #define XFS_BMAPI(mp, trans,io,bno,len,f,first,tot,mval,nmap,flist,delta) \
 	(*(mp)->m_io_ops.xfs_bmapi_func) \
 		(trans,(io)->io_obj,bno,len,f,first,tot,mval,nmap,flist,delta)
@@ -280,8 +289,8 @@ typedef struct xfs_ioops {
 	(*(mp)->m_io_ops.xfs_ilock_demote)((io)->io_obj, mode)
 #define XFS_SIZE(mp, io) \
 	(*(mp)->m_io_ops.xfs_size_func)((io)->io_obj)
-#define XFS_IODONE(vfsp) \
-	(*(mp)->m_io_ops.xfs_iodone)(vfsp)
+#define XFS_IODONE(mp) \
+	(*(mp)->m_io_ops.xfs_iodone)(mp)
 #define XFS_SWAP_EXTENTS(mp, io, tio, sxp) \
 	(*(mp)->m_io_ops.xfs_swap_extents_func) \
 		((io)->io_obj, (tio)->io_obj, sxp)
@@ -318,7 +327,7 @@ extern void	xfs_icsb_sync_counters_flags(struct xfs_mount *, int);
 #endif
 
 typedef struct xfs_mount {
-	bhv_desc_t		m_bhv;		/* vfs xfs behavior */
+	struct super_block	*m_super;
 	xfs_tid_t		m_tid;		/* next unused tid for fs */
 	AIL_LOCK_T		m_ail_lock;	/* fs AIL mutex */
 	xfs_ail_entry_t		m_ail;		/* fs active log item list */
@@ -335,8 +344,6 @@ typedef struct xfs_mount {
 	xfs_agnumber_t		m_agirotor;	/* last ag dir inode alloced */
 	lock_t			m_agirotor_lock;/* .. and lock protecting it */
 	xfs_agnumber_t		m_maxagi;	/* highest inode alloc group */
-	size_t			m_ihsize;	/* size of next field */
-	struct xfs_ihash	*m_ihash;	/* fs private inode hash table*/
 	struct xfs_inode	*m_inodes;	/* active inode list */
 	struct list_head	m_del_inodes;	/* inodes to reclaim */
 	mutex_t			m_ilock;	/* inode list mutex */
@@ -362,7 +369,6 @@ typedef struct xfs_mount {
 	__uint8_t		m_blkbb_log;	/* blocklog - BBSHIFT */
 	__uint8_t		m_agno_log;	/* log #ag's */
 	__uint8_t		m_agino_log;	/* #bits for agino in inum */
-	__uint8_t		m_nreadaheads;	/* #readahead buffers */
 	__uint16_t		m_inode_cluster_size;/* min inode buf size */
 	uint			m_blockmask;	/* sb_blocksize-1 */
 	uint			m_blockwsize;	/* sb_blocksize in words */
@@ -378,7 +384,7 @@ typedef struct xfs_mount {
 	uint			m_in_maxlevels;	/* XFS_IN_MAXLEVELS */
 	struct xfs_perag	*m_perag;	/* per-ag accounting info */
 	struct rw_semaphore	m_peraglock;	/* lock for m_perag (pointer) */
-	sema_t			m_growlock;	/* growfs mutex */
+	struct mutex		m_growlock;	/* growfs mutex */
 	int			m_fixedfsid[2];	/* unchanged for life of FS */
 	uint			m_dmevmask;	/* DMI events for this FS */
 	__uint64_t		m_flags;	/* global mount flags */
@@ -415,8 +421,8 @@ typedef struct xfs_mount {
 	uint			m_chsize;	/* size of next field */
 	struct xfs_chash	*m_chash;	/* fs private inode per-cluster
 						 * hash table */
-	struct xfs_dmops	m_dm_ops;	/* vector of DMI ops */
-	struct xfs_qmops	m_qm_ops;	/* vector of XQM ops */
+	struct xfs_dmops	*m_dm_ops;	/* vector of DMI ops */
+	struct xfs_qmops	*m_qm_ops;	/* vector of XQM ops */
 	struct xfs_ioops	m_io_ops;	/* vector of I/O ops */
 	atomic_t		m_active_trans;	/* number trans frozen */
 #ifdef HAVE_PERCPU_SB
@@ -426,6 +432,12 @@ typedef struct xfs_mount {
 	struct mutex		m_icsb_mutex;	/* balancer sync lock */
 #endif
 	struct xfs_mru_cache	*m_filestream;  /* per-mount filestream data */
+	struct task_struct	*m_sync_task;	/* generalised sync thread */
+	bhv_vfs_sync_work_t	m_sync_work;	/* work item for VFS_SYNC */
+	struct list_head	m_sync_list;	/* sync thread work item list */
+	spinlock_t		m_sync_lock;	/* work item list lock */
+	int			m_sync_seq;	/* sync thread generation no. */
+	wait_queue_head_t	m_wait_single_sync_task;
 } xfs_mount_t;
 
 /*
@@ -435,7 +447,7 @@ typedef struct xfs_mount {
 						   must be synchronous except
 						   for space allocations */
 #define XFS_MOUNT_INO64		(1ULL << 1)
-			     /* (1ULL << 2)	-- currently unused */
+#define XFS_MOUNT_DMAPI		(1ULL << 2)	/* dmapi is enabled */
 #define XFS_MOUNT_WAS_CLEAN	(1ULL << 3)
 #define XFS_MOUNT_FS_SHUTDOWN	(1ULL << 4)	/* atomic stop of all filesystem
 						   operations, typically for
@@ -445,7 +457,7 @@ typedef struct xfs_mount {
 #define XFS_MOUNT_NOALIGN	(1ULL << 7)	/* turn off stripe alignment
 						   allocations */
 #define XFS_MOUNT_ATTR2		(1ULL << 8)	/* allow use of attr2 format */
-			     /*	(1ULL << 9)	-- currently unused */
+#define XFS_MOUNT_GRPID		(1ULL << 9)	/* group-ID assigned from directory */
 #define XFS_MOUNT_NORECOVERY	(1ULL << 10)	/* no recovery - dirty fs */
 #define XFS_MOUNT_SHARED	(1ULL << 11)	/* shared mount */
 #define XFS_MOUNT_DFLT_IOSIZE	(1ULL << 12)	/* set default i/o size */
@@ -453,13 +465,13 @@ typedef struct xfs_mount {
 						/* osyncisdsync is now default*/
 #define XFS_MOUNT_32BITINODES	(1ULL << 14)	/* do not create inodes above
 						 * 32 bits in size */
-			     /* (1ULL << 15)	-- currently unused */
+#define XFS_MOUNT_SMALL_INUMS	(1ULL << 15)	/* users wants 32bit inodes */
 #define XFS_MOUNT_NOUUID	(1ULL << 16)	/* ignore uuid during mount */
 #define XFS_MOUNT_BARRIER	(1ULL << 17)
 #define XFS_MOUNT_IDELETE	(1ULL << 18)	/* delete empty inode clusters*/
 #define XFS_MOUNT_SWALLOC	(1ULL << 19)	/* turn on stripe width
 						 * allocation */
-#define XFS_MOUNT_IHASHSIZE	(1ULL << 20)	/* inode hash table size */
+#define XFS_MOUNT_RDONLY	(1ULL << 20)	/* read-only fs */
 #define XFS_MOUNT_DIRSYNC	(1ULL << 21)	/* synchronous directory ops */
 #define XFS_MOUNT_COMPAT_IOSIZE	(1ULL << 22)	/* don't report large preferred
 						 * I/O size in stat() */
@@ -518,8 +530,10 @@ xfs_preferred_iosize(xfs_mount_t *mp)
 #define XFS_LAST_UNMOUNT_WAS_CLEAN(mp)	\
 				((mp)->m_flags & XFS_MOUNT_WAS_CLEAN)
 #define XFS_FORCED_SHUTDOWN(mp)	((mp)->m_flags & XFS_MOUNT_FS_SHUTDOWN)
+void xfs_do_force_shutdown(struct xfs_mount *mp, int flags, char *fname,
+		int lnnum);
 #define xfs_force_shutdown(m,f)	\
-	bhv_vfs_force_shutdown((XFS_MTOVFS(m)), f, __FILE__, __LINE__)
+	xfs_do_force_shutdown(m, f, __FILE__, __LINE__)
 
 /*
  * Flags for xfs_mountfs
@@ -533,28 +547,6 @@ xfs_preferred_iosize(xfs_mount_t *mp)
 /*	XFS_MFSI_CONVERT_SUNIT	*/
 #define XFS_MFSI_QUIET		0x40	/* Be silent if mount errors found */
 
-/*
- * Macros for getting from mount to vfs and back.
- */
-#define	XFS_MTOVFS(mp)		xfs_mtovfs(mp)
-static inline struct bhv_vfs *xfs_mtovfs(xfs_mount_t *mp)
-{
-	return bhvtovfs(&mp->m_bhv);
-}
-
-#define	XFS_BHVTOM(bdp)	xfs_bhvtom(bdp)
-static inline xfs_mount_t *xfs_bhvtom(bhv_desc_t *bdp)
-{
-	return (xfs_mount_t *)BHV_PDATA(bdp);
-}
-
-#define XFS_VFSTOM(vfs) xfs_vfstom(vfs)
-static inline xfs_mount_t *xfs_vfstom(bhv_vfs_t *vfs)
-{
-	return XFS_BHVTOM(bhv_lookup_range(VFS_BHVHEAD(vfs),
-				VFS_POSITION_XFS, VFS_POSITION_XFS));
-}
-
 #define XFS_DADDR_TO_AGNO(mp,d)         xfs_daddr_to_agno(mp,d)
 static inline xfs_agnumber_t
 xfs_daddr_to_agno(struct xfs_mount *mp, xfs_daddr_t d)
@@ -572,6 +564,21 @@ xfs_daddr_to_agbno(struct xfs_mount *mp, xfs_daddr_t d)
 	return (xfs_agblock_t) do_div(ld, mp->m_sb.sb_agblocks);
 }
 
+/*
+ * perag get/put wrappers for eventual ref counting
+ */
+static inline xfs_perag_t *
+xfs_get_perag(struct xfs_mount *mp, xfs_ino_t ino)
+{
+	return &mp->m_perag[XFS_INO_TO_AGNO(mp, ino)];
+}
+
+static inline void
+xfs_put_perag(struct xfs_mount *mp, xfs_perag_t *pag)
+{
+	/* nothing to see here, move along */
+}
+
 /*
  * Per-cpu superblock locking functions
  */
@@ -609,8 +616,8 @@ typedef struct xfs_mod_sb {
 extern xfs_mount_t *xfs_mount_init(void);
 extern void	xfs_mod_sb(xfs_trans_t *, __int64_t);
 extern int	xfs_log_sbcount(xfs_mount_t *, uint);
-extern void	xfs_mount_free(xfs_mount_t *mp, int remove_bhv);
-extern int	xfs_mountfs(struct bhv_vfs *, xfs_mount_t *mp, int);
+extern void	xfs_mount_free(xfs_mount_t *mp);
+extern int	xfs_mountfs(xfs_mount_t *mp, int);
 extern void	xfs_mountfs_check_barriers(xfs_mount_t *mp);
 
 extern int	xfs_unmountfs(xfs_mount_t *, struct cred *);
@@ -626,16 +633,19 @@ extern struct xfs_buf *xfs_getsb(xfs_mount_t *, int);
 extern int	xfs_readsb(xfs_mount_t *, int);
 extern void	xfs_freesb(xfs_mount_t *);
 extern int	xfs_fs_writable(xfs_mount_t *);
-extern void	xfs_do_force_shutdown(bhv_desc_t *, int, char *, int);
 extern int	xfs_syncsub(xfs_mount_t *, int, int *);
 extern int	xfs_sync_inodes(xfs_mount_t *, int, int *);
-extern xfs_agnumber_t	xfs_initialize_perag(struct bhv_vfs *, xfs_mount_t *,
-						xfs_agnumber_t);
-extern void	xfs_xlatesb(void *, struct xfs_sb *, int, __int64_t);
+extern xfs_agnumber_t	xfs_initialize_perag(xfs_mount_t *, xfs_agnumber_t);
+extern void	xfs_sb_from_disk(struct xfs_sb *, struct xfs_dsb *);
+extern void	xfs_sb_to_disk(struct xfs_dsb *, struct xfs_sb *, __int64_t);
 extern int	xfs_sb_validate_fsb_count(struct xfs_sb *, __uint64_t);
 
-extern struct xfs_dmops xfs_dmcore_stub;
-extern struct xfs_qmops xfs_qmcore_stub;
+extern int	xfs_dmops_get(struct xfs_mount *, struct xfs_mount_args *);
+extern void	xfs_dmops_put(struct xfs_mount *);
+extern int	xfs_qmops_get(struct xfs_mount *, struct xfs_mount_args *);
+extern void	xfs_qmops_put(struct xfs_mount *);
+
+extern struct xfs_dmops xfs_dmcore_xfs;
 extern struct xfs_ioops xfs_iocore_xfs;
 
 extern int	xfs_init(void);

fs/xfs/xfs_qmops.c

@@ -28,6 +28,8 @@
 #include "xfs_mount.h"
 #include "xfs_quota.h"
 #include "xfs_error.h"
+#include "xfs_clnt.h"
+
 
 STATIC struct xfs_dquot *
 xfs_dqvopchown_default(
@@ -64,7 +66,7 @@ xfs_mount_reset_sbqflags(xfs_mount_t *mp)
 	 * if the fs is readonly, let the incore superblock run
 	 * with quotas off but don't flush the update out to disk
 	 */
-	if (XFS_MTOVFS(mp)->vfs_flag & VFS_RDONLY)
+	if (mp->m_flags & XFS_MOUNT_RDONLY)
 		return 0;
 #ifdef QUOTADEBUG
 	xfs_fs_cmn_err(CE_NOTE, mp, "Writing superblock quota changes");
@@ -110,7 +112,7 @@ xfs_noquota_init(
 	return error;
 }
 
-xfs_qmops_t	xfs_qmcore_stub = {
+static struct xfs_qmops xfs_qmcore_stub = {
 	.xfs_qminit		= (xfs_qminit_t) xfs_noquota_init,
 	.xfs_qmdone		= (xfs_qmdone_t) fs_noerr,
 	.xfs_qmmount		= (xfs_qmmount_t) fs_noerr,
@@ -124,4 +126,38 @@ xfs_qmops_t	xfs_qmcore_stub = {
 	.xfs_dqvoprename	= (xfs_dqvoprename_t) fs_noerr,
 	.xfs_dqvopchown		= xfs_dqvopchown_default,
 	.xfs_dqvopchownresv	= (xfs_dqvopchownresv_t) fs_noerr,
+	.xfs_dqstatvfs		= (xfs_dqstatvfs_t) fs_noval,
+	.xfs_dqsync		= (xfs_dqsync_t) fs_noerr,
+	.xfs_quotactl		= (xfs_quotactl_t) fs_nosys,
 };
+
+int
+xfs_qmops_get(struct xfs_mount *mp, struct xfs_mount_args *args)
+{
+	if (args->flags & (XFSMNT_UQUOTA | XFSMNT_PQUOTA | XFSMNT_GQUOTA)) {
+		struct xfs_qmops *ops;
+
+		ops = symbol_get(xfs_qmcore_xfs);
+		if (!ops) {
+			request_module("xfs_quota");
+			ops = symbol_get(xfs_qmcore_xfs);
+		}
+
+		if (!ops) {
+			cmn_err(CE_WARN, "XFS: no quota support available.");
+			return EINVAL;
+		}
+		mp->m_qm_ops = ops;
+	} else {
+		mp->m_qm_ops = &xfs_qmcore_stub;
+	}
+
+	return 0;
+}
+
+void
+xfs_qmops_put(struct xfs_mount *mp)
+{
+	if (mp->m_qm_ops != &xfs_qmcore_stub)
+		symbol_put(xfs_qmcore_xfs);
+}

fs/xfs/xfs_quota.h

@@ -330,12 +330,12 @@ typedef struct xfs_dqtrxops {
 } xfs_dqtrxops_t;
 
 #define XFS_DQTRXOP(mp, tp, op, args...) \
-		((mp)->m_qm_ops.xfs_dqtrxops ? \
-		((mp)->m_qm_ops.xfs_dqtrxops->op)(tp, ## args) : 0)
+		((mp)->m_qm_ops->xfs_dqtrxops ? \
+		((mp)->m_qm_ops->xfs_dqtrxops->op)(tp, ## args) : 0)
 
 #define XFS_DQTRXOP_VOID(mp, tp, op, args...) \
-		((mp)->m_qm_ops.xfs_dqtrxops ? \
-		((mp)->m_qm_ops.xfs_dqtrxops->op)(tp, ## args) : (void)0)
+		((mp)->m_qm_ops->xfs_dqtrxops ? \
+		((mp)->m_qm_ops->xfs_dqtrxops->op)(tp, ## args) : (void)0)
 
 #define XFS_TRANS_DUP_DQINFO(mp, otp, ntp) \
 	XFS_DQTRXOP_VOID(mp, otp, qo_dup_dqinfo, ntp)
@@ -364,7 +364,7 @@ typedef struct xfs_dqtrxops {
 extern int xfs_qm_dqcheck(xfs_disk_dquot_t *, xfs_dqid_t, uint, uint, char *);
 extern int xfs_mount_reset_sbqflags(struct xfs_mount *);
 
-extern struct bhv_module_vfsops xfs_qmops;
+extern struct xfs_qmops xfs_qmcore_xfs;
 
 #endif	/* __KERNEL__ */
 

fs/xfs/xfs_rename.c

@@ -22,6 +22,7 @@
 #include "xfs_inum.h"
 #include "xfs_trans.h"
 #include "xfs_sb.h"
+#include "xfs_ag.h"
 #include "xfs_dir2.h"
 #include "xfs_dmapi.h"
 #include "xfs_mount.h"
@@ -128,8 +129,7 @@ xfs_lock_for_rename(
 		lock_mode = xfs_ilock_map_shared(dp2);
 	}
 
-	error = xfs_dir_lookup_int(XFS_ITOBHV(dp2), lock_mode,
-				   vname2, &inum2, &ip2);
+	error = xfs_dir_lookup_int(dp2, lock_mode, vname2, &inum2, &ip2);
 	if (error == ENOENT) {		/* target does not need to exist. */
 		inum2 = 0;
 	} else if (error) {
@@ -221,15 +221,15 @@ xfs_lock_for_rename(
  */
 int
 xfs_rename(
-	bhv_desc_t	*src_dir_bdp,
+	xfs_inode_t	*src_dp,
 	bhv_vname_t	*src_vname,
 	bhv_vnode_t	*target_dir_vp,
-	bhv_vname_t	*target_vname,
-	cred_t		*credp)
+	bhv_vname_t	*target_vname)
 {
+	bhv_vnode_t	*src_dir_vp = XFS_ITOV(src_dp);
 	xfs_trans_t	*tp;
-	xfs_inode_t	*src_dp, *target_dp, *src_ip, *target_ip;
-	xfs_mount_t	*mp;
+	xfs_inode_t	*target_dp, *src_ip, *target_ip;
+	xfs_mount_t	*mp = src_dp->i_mount;
 	int		new_parent;		/* moving to a new dir */
 	int		src_is_directory;	/* src_name is a directory */
 	int		error;
@@ -239,7 +239,6 @@ xfs_rename(
 	int		committed;
 	xfs_inode_t	*inodes[4];
 	int		target_ip_dropped = 0;	/* dropped target_ip link? */
-	bhv_vnode_t	*src_dir_vp;
 	int		spaceres;
 	int		target_link_zero = 0;
 	int		num_inodes;
@@ -248,9 +247,8 @@ xfs_rename(
 	int		src_namelen = VNAMELEN(src_vname);
 	int		target_namelen = VNAMELEN(target_vname);
 
-	src_dir_vp = BHV_TO_VNODE(src_dir_bdp);
-	vn_trace_entry(src_dir_vp, "xfs_rename", (inst_t *)__return_address);
-	vn_trace_entry(target_dir_vp, "xfs_rename", (inst_t *)__return_address);
+	vn_trace_entry(src_dp, "xfs_rename", (inst_t *)__return_address);
+	vn_trace_entry(xfs_vtoi(target_dir_vp), "xfs_rename", (inst_t *)__return_address);
 
 	/*
 	 * Find the XFS behavior descriptor for the target directory
@@ -261,12 +259,8 @@ xfs_rename(
 		return XFS_ERROR(EXDEV);
 	}
 
-	src_dp = XFS_BHVTOI(src_dir_bdp);
-	mp = src_dp->i_mount;
-
-	if (DM_EVENT_ENABLED(src_dir_vp->v_vfsp, src_dp, DM_EVENT_RENAME) ||
-	    DM_EVENT_ENABLED(target_dir_vp->v_vfsp,
-				target_dp, DM_EVENT_RENAME)) {
+	if (DM_EVENT_ENABLED(src_dp, DM_EVENT_RENAME) ||
+	    DM_EVENT_ENABLED(target_dp, DM_EVENT_RENAME)) {
 		error = XFS_SEND_NAMESP(mp, DM_EVENT_RENAME,
 					src_dir_vp, DM_RIGHT_NULL,
 					target_dir_vp, DM_RIGHT_NULL,
@@ -592,20 +586,16 @@ xfs_rename(
 	/*
 	 * Let interposed file systems know about removed links.
 	 */
-	if (target_ip_dropped) {
-		bhv_vop_link_removed(XFS_ITOV(target_ip), target_dir_vp,
-					target_link_zero);
+	if (target_ip_dropped)
 		IRELE(target_ip);
-	}
 
 	IRELE(src_ip);
 
 	/* Fall through to std_return with error = 0 or errno from
 	 * xfs_trans_commit	 */
 std_return:
-	if (DM_EVENT_ENABLED(src_dir_vp->v_vfsp, src_dp, DM_EVENT_POSTRENAME) ||
-	    DM_EVENT_ENABLED(target_dir_vp->v_vfsp,
-				target_dp, DM_EVENT_POSTRENAME)) {
+	if (DM_EVENT_ENABLED(src_dp, DM_EVENT_POSTRENAME) ||
+	    DM_EVENT_ENABLED(target_dp, DM_EVENT_POSTRENAME)) {
 		(void) XFS_SEND_NAMESP (mp, DM_EVENT_POSTRENAME,
 					src_dir_vp, DM_RIGHT_NULL,
 					target_dir_vp, DM_RIGHT_NULL,

fs/xfs/xfs_rw.c

@@ -178,18 +178,15 @@ xfs_write_sync_logforce(
  * the shop, make sure that absolutely nothing persistent happens to
  * this filesystem after this point.
  */
-
 void
 xfs_do_force_shutdown(
-	bhv_desc_t	*bdp,
+	xfs_mount_t	*mp,
 	int		flags,
 	char		*fname,
 	int		lnnum)
 {
 	int		logerror;
-	xfs_mount_t	*mp;
 
-	mp = XFS_BHVTOM(bdp);
 	logerror = flags & SHUTDOWN_LOG_IO_ERROR;
 
 	if (!(flags & SHUTDOWN_FORCE_UMOUNT)) {

fs/xfs/xfs_rw.h

@@ -22,32 +22,6 @@ struct xfs_buf;
 struct xfs_inode;
 struct xfs_mount;
 
-/*
- * Maximum count of bmaps used by read and write paths.
- */
-#define	XFS_MAX_RW_NBMAPS	4
-
-/*
- * Counts of readahead buffers to use based on physical memory size.
- * None of these should be more than XFS_MAX_RW_NBMAPS.
- */
-#define	XFS_RW_NREADAHEAD_16MB	2
-#define	XFS_RW_NREADAHEAD_32MB	3
-#define	XFS_RW_NREADAHEAD_K32	4
-#define	XFS_RW_NREADAHEAD_K64	4
-
-/*
- * Maximum size of a buffer that we\'ll map.  Making this
- * too big will degrade performance due to the number of
- * pages which need to be gathered.  Making it too small
- * will prevent us from doing large I/O\'s to hardware that
- * needs it.
- *
- * This is currently set to 512 KB.
- */
-#define	XFS_MAX_BMAP_LEN_BB	1024
-#define	XFS_MAX_BMAP_LEN_BYTES	524288
-
 /*
  * Convert the given file system block to a disk block.
  * We have to treat it differently based on whether the
@@ -116,14 +90,6 @@ extern void xfs_ioerror_alert(char *func, struct xfs_mount *mp,
 /*
  * Prototypes for functions in xfs_vnodeops.c.
  */
-extern int xfs_rwlock(bhv_desc_t *bdp, bhv_vrwlock_t write_lock);
-extern void xfs_rwunlock(bhv_desc_t *bdp, bhv_vrwlock_t write_lock);
-extern int xfs_setattr(bhv_desc_t *, bhv_vattr_t *vap, int flags,
-			cred_t *credp);
-extern int xfs_change_file_space(bhv_desc_t *bdp, int cmd, xfs_flock64_t *bf,
-			xfs_off_t offset, cred_t *credp, int flags);
-extern int xfs_set_dmattrs(bhv_desc_t *bdp, u_int evmask, u_int16_t state,
-			cred_t *credp);
 extern int xfs_free_eofblocks(struct xfs_mount *mp, struct xfs_inode *ip,
 			int flags);
 

fs/xfs/xfs_sb.h

@@ -87,8 +87,10 @@ struct xfs_mount;
 	(XFS_SB_VERSION2_OKREALFBITS |	\
 	 XFS_SB_VERSION2_OKSASHFBITS )
 
-typedef struct xfs_sb
-{
+/*
+ * Superblock - in core version.  Must match the ondisk version below.
+ */
+typedef struct xfs_sb {
 	__uint32_t	sb_magicnum;	/* magic number == XFS_SB_MAGIC */
 	__uint32_t	sb_blocksize;	/* logical block size, bytes */
 	xfs_drfsbno_t	sb_dblocks;	/* number of data blocks */
@@ -145,6 +147,66 @@ typedef struct xfs_sb
 	__uint32_t	sb_features2;	/* additional feature bits */
 } xfs_sb_t;
 
+/*
+ * Superblock - on disk version.  Must match the in core version below.
+ */
+typedef struct xfs_dsb {
+	__be32		sb_magicnum;	/* magic number == XFS_SB_MAGIC */
+	__be32		sb_blocksize;	/* logical block size, bytes */
+	__be64		sb_dblocks;	/* number of data blocks */
+	__be64		sb_rblocks;	/* number of realtime blocks */
+	__be64		sb_rextents;	/* number of realtime extents */
+	uuid_t		sb_uuid;	/* file system unique id */
+	__be64		sb_logstart;	/* starting block of log if internal */
+	__be64		sb_rootino;	/* root inode number */
+	__be64		sb_rbmino;	/* bitmap inode for realtime extents */
+	__be64		sb_rsumino;	/* summary inode for rt bitmap */
+	__be32		sb_rextsize;	/* realtime extent size, blocks */
+	__be32		sb_agblocks;	/* size of an allocation group */
+	__be32		sb_agcount;	/* number of allocation groups */
+	__be32		sb_rbmblocks;	/* number of rt bitmap blocks */
+	__be32		sb_logblocks;	/* number of log blocks */
+	__be16		sb_versionnum;	/* header version == XFS_SB_VERSION */
+	__be16		sb_sectsize;	/* volume sector size, bytes */
+	__be16		sb_inodesize;	/* inode size, bytes */
+	__be16		sb_inopblock;	/* inodes per block */
+	char		sb_fname[12];	/* file system name */
+	__u8		sb_blocklog;	/* log2 of sb_blocksize */
+	__u8		sb_sectlog;	/* log2 of sb_sectsize */
+	__u8		sb_inodelog;	/* log2 of sb_inodesize */
+	__u8		sb_inopblog;	/* log2 of sb_inopblock */
+	__u8		sb_agblklog;	/* log2 of sb_agblocks (rounded up) */
+	__u8		sb_rextslog;	/* log2 of sb_rextents */
+	__u8		sb_inprogress;	/* mkfs is in progress, don't mount */
+	__u8		sb_imax_pct;	/* max % of fs for inode space */
+					/* statistics */
+	/*
+	 * These fields must remain contiguous.  If you really
+	 * want to change their layout, make sure you fix the
+	 * code in xfs_trans_apply_sb_deltas().
+	 */
+	__be64		sb_icount;	/* allocated inodes */
+	__be64		sb_ifree;	/* free inodes */
+	__be64		sb_fdblocks;	/* free data blocks */
+	__be64		sb_frextents;	/* free realtime extents */
+	/*
+	 * End contiguous fields.
+	 */
+	__be64		sb_uquotino;	/* user quota inode */
+	__be64		sb_gquotino;	/* group quota inode */
+	__be16		sb_qflags;	/* quota flags */
+	__u8		sb_flags;	/* misc. flags */
+	__u8		sb_shared_vn;	/* shared version number */
+	__be32		sb_inoalignmt;	/* inode chunk alignment, fsblocks */
+	__be32		sb_unit;	/* stripe or raid unit */
+	__be32		sb_width;	/* stripe or raid width */
+	__u8		sb_dirblklog;	/* log2 of dir block size (fsbs) */
+	__u8		sb_logsectlog;	/* log2 of the log sector size */
+	__be16		sb_logsectsize;	/* sector size for the log, bytes */
+	__be32		sb_logsunit;	/* stripe unit size for the log */
+	__be32		sb_features2;	/* additional feature bits */
+} xfs_dsb_t;
+
 /*
  * Sequence number values for the fields.
  */
@@ -446,7 +508,7 @@ static inline void xfs_sb_version_addattr2(xfs_sb_t *sbp)
 
 #define XFS_SB_DADDR		((xfs_daddr_t)0) /* daddr in filesystem/ag */
 #define	XFS_SB_BLOCK(mp)	XFS_HDR_BLOCK(mp, XFS_SB_DADDR)
-#define XFS_BUF_TO_SBP(bp)	((xfs_sb_t *)XFS_BUF_PTR(bp))
+#define XFS_BUF_TO_SBP(bp)	((xfs_dsb_t *)XFS_BUF_PTR(bp))
 
 #define	XFS_HDR_BLOCK(mp,d)	((xfs_agblock_t)XFS_BB_TO_FSBT(mp,d))
 #define	XFS_DADDR_TO_FSB(mp,d)	XFS_AGB_TO_FSB(mp, \

fs/xfs/xfs_trans.c

@@ -234,7 +234,7 @@ xfs_trans_alloc(
 	xfs_mount_t	*mp,
 	uint		type)
 {
-	vfs_wait_for_freeze(XFS_MTOVFS(mp), SB_FREEZE_TRANS);
+	xfs_wait_for_freeze(mp, SB_FREEZE_TRANS);
 	return _xfs_trans_alloc(mp, type);
 }
 
@@ -548,7 +548,7 @@ STATIC void
 xfs_trans_apply_sb_deltas(
 	xfs_trans_t	*tp)
 {
-	xfs_sb_t	*sbp;
+	xfs_dsb_t	*sbp;
 	xfs_buf_t	*bp;
 	int		whole = 0;
 
@@ -566,57 +566,51 @@ xfs_trans_apply_sb_deltas(
 	 * Only update the superblock counters if we are logging them
 	 */
 	if (!xfs_sb_version_haslazysbcount(&(tp->t_mountp->m_sb))) {
-		if (tp->t_icount_delta != 0) {
-			INT_MOD(sbp->sb_icount, ARCH_CONVERT, tp->t_icount_delta);
-		}
-		if (tp->t_ifree_delta != 0) {
-			INT_MOD(sbp->sb_ifree, ARCH_CONVERT, tp->t_ifree_delta);
-		}
-
-		if (tp->t_fdblocks_delta != 0) {
-			INT_MOD(sbp->sb_fdblocks, ARCH_CONVERT, tp->t_fdblocks_delta);
-		}
-		if (tp->t_res_fdblocks_delta != 0) {
-			INT_MOD(sbp->sb_fdblocks, ARCH_CONVERT, tp->t_res_fdblocks_delta);
-		}
+		if (tp->t_icount_delta)
+			be64_add(&sbp->sb_icount, tp->t_icount_delta);
+		if (tp->t_ifree_delta)
+			be64_add(&sbp->sb_ifree, tp->t_ifree_delta);
+		if (tp->t_fdblocks_delta)
+			be64_add(&sbp->sb_fdblocks, tp->t_fdblocks_delta);
+		if (tp->t_res_fdblocks_delta)
+			be64_add(&sbp->sb_fdblocks, tp->t_res_fdblocks_delta);
 	}
 
-	if (tp->t_frextents_delta != 0) {
-		INT_MOD(sbp->sb_frextents, ARCH_CONVERT, tp->t_frextents_delta);
-	}
-	if (tp->t_res_frextents_delta != 0) {
-		INT_MOD(sbp->sb_frextents, ARCH_CONVERT, tp->t_res_frextents_delta);
-	}
-	if (tp->t_dblocks_delta != 0) {
-		INT_MOD(sbp->sb_dblocks, ARCH_CONVERT, tp->t_dblocks_delta);
+	if (tp->t_frextents_delta)
+		be64_add(&sbp->sb_frextents, tp->t_frextents_delta);
+	if (tp->t_res_frextents_delta)
+		be64_add(&sbp->sb_frextents, tp->t_res_frextents_delta);
+
+	if (tp->t_dblocks_delta) {
+		be64_add(&sbp->sb_dblocks, tp->t_dblocks_delta);
 		whole = 1;
 	}
-	if (tp->t_agcount_delta != 0) {
-		INT_MOD(sbp->sb_agcount, ARCH_CONVERT, tp->t_agcount_delta);
+	if (tp->t_agcount_delta) {
+		be32_add(&sbp->sb_agcount, tp->t_agcount_delta);
 		whole = 1;
 	}
-	if (tp->t_imaxpct_delta != 0) {
-		INT_MOD(sbp->sb_imax_pct, ARCH_CONVERT, tp->t_imaxpct_delta);
+	if (tp->t_imaxpct_delta) {
+		sbp->sb_imax_pct += tp->t_imaxpct_delta;
 		whole = 1;
 	}
-	if (tp->t_rextsize_delta != 0) {
-		INT_MOD(sbp->sb_rextsize, ARCH_CONVERT, tp->t_rextsize_delta);
+	if (tp->t_rextsize_delta) {
+		be32_add(&sbp->sb_rextsize, tp->t_rextsize_delta);
 		whole = 1;
 	}
-	if (tp->t_rbmblocks_delta != 0) {
-		INT_MOD(sbp->sb_rbmblocks, ARCH_CONVERT, tp->t_rbmblocks_delta);
+	if (tp->t_rbmblocks_delta) {
+		be32_add(&sbp->sb_rbmblocks, tp->t_rbmblocks_delta);
 		whole = 1;
 	}
-	if (tp->t_rblocks_delta != 0) {
-		INT_MOD(sbp->sb_rblocks, ARCH_CONVERT, tp->t_rblocks_delta);
+	if (tp->t_rblocks_delta) {
+		be64_add(&sbp->sb_rblocks, tp->t_rblocks_delta);
 		whole = 1;
 	}
-	if (tp->t_rextents_delta != 0) {
-		INT_MOD(sbp->sb_rextents, ARCH_CONVERT, tp->t_rextents_delta);
+	if (tp->t_rextents_delta) {
+		be64_add(&sbp->sb_rextents, tp->t_rextents_delta);
 		whole = 1;
 	}
-	if (tp->t_rextslog_delta != 0) {
-		INT_MOD(sbp->sb_rextslog, ARCH_CONVERT, tp->t_rextslog_delta);
+	if (tp->t_rextslog_delta) {
+		sbp->sb_rextslog += tp->t_rextslog_delta;
 		whole = 1;
 	}
 
@@ -624,17 +618,17 @@ xfs_trans_apply_sb_deltas(
 		/*
 		 * Log the whole thing, the fields are noncontiguous.
 		 */
-		xfs_trans_log_buf(tp, bp, 0, sizeof(xfs_sb_t) - 1);
+		xfs_trans_log_buf(tp, bp, 0, sizeof(xfs_dsb_t) - 1);
 	else
 		/*
 		 * Since all the modifiable fields are contiguous, we
 		 * can get away with this.
 		 */
-		xfs_trans_log_buf(tp, bp, offsetof(xfs_sb_t, sb_icount),
-				  offsetof(xfs_sb_t, sb_frextents) +
+		xfs_trans_log_buf(tp, bp, offsetof(xfs_dsb_t, sb_icount),
+				  offsetof(xfs_dsb_t, sb_frextents) +
 				  sizeof(sbp->sb_frextents) - 1);
 
-	XFS_MTOVFS(tp->t_mountp)->vfs_super->s_dirt = 1;
+	tp->t_mountp->m_super->s_dirt = 1;
 }
 
 /*

fs/xfs/xfs_trans_ail.c

@@ -22,6 +22,7 @@
 #include "xfs_inum.h"
 #include "xfs_trans.h"
 #include "xfs_sb.h"
+#include "xfs_ag.h"
 #include "xfs_dmapi.h"
 #include "xfs_mount.h"
 #include "xfs_trans_priv.h"

fs/xfs/xfs_trans_extfree.c

@@ -22,6 +22,7 @@
 #include "xfs_inum.h"
 #include "xfs_trans.h"
 #include "xfs_sb.h"
+#include "xfs_ag.h"
 #include "xfs_dmapi.h"
 #include "xfs_mount.h"
 #include "xfs_trans_priv.h"

fs/xfs/xfs_types.h

@@ -151,18 +151,6 @@ typedef __uint8_t	xfs_arch_t;	/* architecture of an xfs fs */
  */
 #define MAXNAMELEN	256
 
-typedef struct xfs_dirent {		/* data from readdir() */
-	xfs_ino_t	d_ino;		/* inode number of entry */
-	xfs_off_t	d_off;		/* offset of disk directory entry */
-	unsigned short	d_reclen;	/* length of this record */
-	char		d_name[1];	/* name of file */
-} xfs_dirent_t;
-
-#define DIRENTBASESIZE		(((xfs_dirent_t *)0)->d_name - (char *)0)
-#define DIRENTSIZE(namelen)	\
-	((DIRENTBASESIZE + (namelen) + \
-		sizeof(xfs_off_t)) & ~(sizeof(xfs_off_t) - 1))
-
 typedef enum {
 	XFS_LOOKUP_EQi, XFS_LOOKUP_LEi, XFS_LOOKUP_GEi
 } xfs_lookup_t;

fs/xfs/xfs_utils.c

@@ -65,20 +65,15 @@ xfs_get_dir_entry(
 
 int
 xfs_dir_lookup_int(
-	bhv_desc_t	*dir_bdp,
+	xfs_inode_t	*dp,
 	uint		lock_mode,
 	bhv_vname_t	*dentry,
 	xfs_ino_t	*inum,
 	xfs_inode_t	**ipp)
 {
-	bhv_vnode_t	*dir_vp;
-	xfs_inode_t	*dp;
 	int		error;
 
-	dir_vp = BHV_TO_VNODE(dir_bdp);
-	vn_trace_entry(dir_vp, __FUNCTION__, (inst_t *)__return_address);
-
-	dp = XFS_BHVTOI(dir_bdp);
+	vn_trace_entry(dp, __FUNCTION__, (inst_t *)__return_address);
 
 	error = xfs_dir_lookup(NULL, dp, VNAME(dentry), VNAMELEN(dentry), inum);
 	if (!error) {

fs/xfs/xfs_utils.h

@@ -20,13 +20,11 @@
 
 #define IRELE(ip)	VN_RELE(XFS_ITOV(ip))
 #define IHOLD(ip)	VN_HOLD(XFS_ITOV(ip))
-#define	ITRACE(ip)	vn_trace_ref(XFS_ITOV(ip), __FILE__, __LINE__, \
+#define	ITRACE(ip)	vn_trace_ref(ip, __FILE__, __LINE__, \
 				(inst_t *)__return_address)
 
-extern int xfs_rename (bhv_desc_t *, bhv_vname_t *, bhv_vnode_t *,
-			bhv_vname_t *, cred_t *);
 extern int xfs_get_dir_entry (bhv_vname_t *, xfs_inode_t **);
-extern int xfs_dir_lookup_int (bhv_desc_t *, uint, bhv_vname_t *, xfs_ino_t *,
+extern int xfs_dir_lookup_int (xfs_inode_t *, uint, bhv_vname_t *, xfs_ino_t *,
 				xfs_inode_t **);
 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,

fs/xfs/xfs_vfsops.c

@@ -54,8 +54,9 @@
 #include "xfs_mru_cache.h"
 #include "xfs_filestream.h"
 #include "xfs_fsops.h"
+#include "xfs_vnodeops.h"
+#include "xfs_vfsops.h"
 
-STATIC int	xfs_sync(bhv_desc_t *, int, cred_t *);
 
 int
 xfs_init(void)
@@ -117,8 +118,8 @@ xfs_init(void)
 	xfs_ili_zone =
 		kmem_zone_init_flags(sizeof(xfs_inode_log_item_t), "xfs_ili",
 					KM_ZONE_SPREAD, NULL);
-	xfs_chashlist_zone =
-		kmem_zone_init_flags(sizeof(xfs_chashlist_t), "xfs_chashlist",
+	xfs_icluster_zone =
+		kmem_zone_init_flags(sizeof(xfs_icluster_t), "xfs_icluster",
 					KM_ZONE_SPREAD, NULL);
 
 	/*
@@ -163,7 +164,7 @@ xfs_cleanup(void)
 	extern kmem_zone_t	*xfs_efd_zone;
 	extern kmem_zone_t	*xfs_efi_zone;
 	extern kmem_zone_t	*xfs_buf_item_zone;
-	extern kmem_zone_t	*xfs_chashlist_zone;
+	extern kmem_zone_t	*xfs_icluster_zone;
 
 	xfs_cleanup_procfs();
 	xfs_sysctl_unregister();
@@ -199,7 +200,7 @@ xfs_cleanup(void)
 	kmem_zone_destroy(xfs_efi_zone);
 	kmem_zone_destroy(xfs_ifork_zone);
 	kmem_zone_destroy(xfs_ili_zone);
-	kmem_zone_destroy(xfs_chashlist_zone);
+	kmem_zone_destroy(xfs_icluster_zone);
 }
 
 /*
@@ -210,7 +211,6 @@ xfs_cleanup(void)
  */
 STATIC int
 xfs_start_flags(
-	struct bhv_vfs		*vfs,
 	struct xfs_mount_args	*ap,
 	struct xfs_mount	*mp)
 {
@@ -238,17 +238,14 @@ xfs_start_flags(
 	mp->m_logbufs = ap->logbufs;
 	if (ap->logbufsize != -1 &&
 	    ap->logbufsize !=  0 &&
-	    ap->logbufsize != 16 * 1024 &&
-	    ap->logbufsize != 32 * 1024 &&
-	    ap->logbufsize != 64 * 1024 &&
-	    ap->logbufsize != 128 * 1024 &&
-	    ap->logbufsize != 256 * 1024) {
+	    (ap->logbufsize < XLOG_MIN_RECORD_BSIZE ||
+	     ap->logbufsize > XLOG_MAX_RECORD_BSIZE ||
+	     !is_power_of_2(ap->logbufsize))) {
 		cmn_err(CE_WARN,
 	"XFS: invalid logbufsize: %d [not 16k,32k,64k,128k or 256k]",
 			ap->logbufsize);
 		return XFS_ERROR(EINVAL);
 	}
-	mp->m_ihsize = ap->ihashsize;
 	mp->m_logbsize = ap->logbufsize;
 	mp->m_fsname_len = strlen(ap->fsname) + 1;
 	mp->m_fsname = kmem_alloc(mp->m_fsname_len, KM_SLEEP);
@@ -295,8 +292,6 @@ xfs_start_flags(
 		mp->m_readio_log = mp->m_writeio_log = ap->iosizelog;
 	}
 
-	if (ap->flags & XFSMNT_IHASHSIZE)
-		mp->m_flags |= XFS_MOUNT_IHASHSIZE;
 	if (ap->flags & XFSMNT_IDELETE)
 		mp->m_flags |= XFS_MOUNT_IDELETE;
 	if (ap->flags & XFSMNT_DIRSYNC)
@@ -311,7 +306,7 @@ xfs_start_flags(
 	 * no recovery flag requires a read-only mount
 	 */
 	if (ap->flags & XFSMNT_NORECOVERY) {
-		if (!(vfs->vfs_flag & VFS_RDONLY)) {
+		if (!(mp->m_flags & XFS_MOUNT_RDONLY)) {
 			cmn_err(CE_WARN,
 	"XFS: tried to mount a FS read-write without recovery!");
 			return XFS_ERROR(EINVAL);
@@ -329,6 +324,8 @@ xfs_start_flags(
 	if (ap->flags2 & XFSMNT2_FILESTREAMS)
 		mp->m_flags |= XFS_MOUNT_FILESTREAMS;
 
+	if (ap->flags & XFSMNT_DMAPI)
+		mp->m_flags |= XFS_MOUNT_DMAPI;
 	return 0;
 }
 
@@ -338,11 +335,10 @@ xfs_start_flags(
  */
 STATIC int
 xfs_finish_flags(
-	struct bhv_vfs		*vfs,
 	struct xfs_mount_args	*ap,
 	struct xfs_mount	*mp)
 {
-	int			ronly = (vfs->vfs_flag & VFS_RDONLY);
+	int			ronly = (mp->m_flags & XFS_MOUNT_RDONLY);
 
 	/* Fail a mount where the logbuf is smaller then the log stripe */
 	if (XFS_SB_VERSION_HASLOGV2(&mp->m_sb)) {
@@ -403,6 +399,22 @@ xfs_finish_flags(
 			return XFS_ERROR(EINVAL);
 	}
 
+	if (ap->flags & XFSMNT_UQUOTA) {
+		mp->m_qflags |= (XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE);
+		if (ap->flags & XFSMNT_UQUOTAENF)
+			mp->m_qflags |= XFS_UQUOTA_ENFD;
+	}
+
+	if (ap->flags & XFSMNT_GQUOTA) {
+		mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE);
+		if (ap->flags & XFSMNT_GQUOTAENF)
+			mp->m_qflags |= XFS_OQUOTA_ENFD;
+	} else if (ap->flags & XFSMNT_PQUOTA) {
+		mp->m_qflags |= (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE);
+		if (ap->flags & XFSMNT_PQUOTAENF)
+			mp->m_qflags |= XFS_OQUOTA_ENFD;
+	}
+
 	return 0;
 }
 
@@ -418,30 +430,26 @@ xfs_finish_flags(
  * they are present.  The data subvolume has already been opened by
  * get_sb_bdev() and is stored in vfsp->vfs_super->s_bdev.
  */
-STATIC int
+int
 xfs_mount(
-	struct bhv_desc		*bhvp,
+	struct xfs_mount	*mp,
 	struct xfs_mount_args	*args,
 	cred_t			*credp)
 {
-	struct bhv_vfs		*vfsp = bhvtovfs(bhvp);
-	struct bhv_desc		*p;
-	struct xfs_mount	*mp = XFS_BHVTOM(bhvp);
 	struct block_device	*ddev, *logdev, *rtdev;
 	int			flags = 0, error;
 
-	ddev = vfsp->vfs_super->s_bdev;
+	ddev = mp->m_super->s_bdev;
 	logdev = rtdev = NULL;
 
-	/*
-	 * Setup xfs_mount function vectors from available behaviors
-	 */
-	p = vfs_bhv_lookup(vfsp, VFS_POSITION_DM);
-	mp->m_dm_ops = p ? *(xfs_dmops_t *) vfs_bhv_custom(p) : xfs_dmcore_stub;
-	p = vfs_bhv_lookup(vfsp, VFS_POSITION_QM);
-	mp->m_qm_ops = p ? *(xfs_qmops_t *) vfs_bhv_custom(p) : xfs_qmcore_stub;
-	p = vfs_bhv_lookup(vfsp, VFS_POSITION_IO);
-	mp->m_io_ops = p ? *(xfs_ioops_t *) vfs_bhv_custom(p) : xfs_iocore_xfs;
+	error = xfs_dmops_get(mp, args);
+	if (error)
+		return error;
+	error = xfs_qmops_get(mp, args);
+	if (error)
+		return error;
+
+	mp->m_io_ops = xfs_iocore_xfs;
 
 	if (args->flags & XFSMNT_QUIET)
 		flags |= XFS_MFSI_QUIET;
@@ -482,24 +490,30 @@ xfs_mount(
 	}
 	if (rtdev) {
 		mp->m_rtdev_targp = xfs_alloc_buftarg(rtdev, 1);
-		if (!mp->m_rtdev_targp)
+		if (!mp->m_rtdev_targp) {
+			xfs_blkdev_put(logdev);
+			xfs_blkdev_put(rtdev);
 			goto error0;
+		}
 	}
 	mp->m_logdev_targp = (logdev && logdev != ddev) ?
 				xfs_alloc_buftarg(logdev, 1) : mp->m_ddev_targp;
-	if (!mp->m_logdev_targp)
+	if (!mp->m_logdev_targp) {
+		xfs_blkdev_put(logdev);
+		xfs_blkdev_put(rtdev);
 		goto error0;
+	}
 
 	/*
 	 * Setup flags based on mount(2) options and then the superblock
 	 */
-	error = xfs_start_flags(vfsp, args, mp);
+	error = xfs_start_flags(args, mp);
 	if (error)
 		goto error1;
 	error = xfs_readsb(mp, flags);
 	if (error)
 		goto error1;
-	error = xfs_finish_flags(vfsp, args, mp);
+	error = xfs_finish_flags(args, mp);
 	if (error)
 		goto error2;
 
@@ -530,10 +544,12 @@ xfs_mount(
 	if ((error = xfs_filestream_mount(mp)))
 		goto error2;
 
-	error = XFS_IOINIT(vfsp, args, flags);
+	error = XFS_IOINIT(mp, args, flags);
 	if (error)
 		goto error2;
 
+	XFS_SEND_MOUNT(mp, DM_RIGHT_NULL, args->mtpt, args->fsname);
+
 	return 0;
 
 error2:
@@ -547,17 +563,17 @@ error1:
 		xfs_binval(mp->m_rtdev_targp);
 error0:
 	xfs_unmountfs_close(mp, credp);
+	xfs_qmops_put(mp);
+	xfs_dmops_put(mp);
 	return error;
 }
 
-STATIC int
+int
 xfs_unmount(
-	bhv_desc_t	*bdp,
+	xfs_mount_t	*mp,
 	int		flags,
 	cred_t		*credp)
 {
-	bhv_vfs_t	*vfsp = bhvtovfs(bdp);
-	xfs_mount_t	*mp = XFS_BHVTOM(bdp);
 	xfs_inode_t	*rip;
 	bhv_vnode_t	*rvp;
 	int		unmount_event_wanted = 0;
@@ -568,8 +584,9 @@ xfs_unmount(
 	rip = mp->m_rootip;
 	rvp = XFS_ITOV(rip);
 
-	if (vfsp->vfs_flag & VFS_DMI) {
-		error = XFS_SEND_PREUNMOUNT(mp, vfsp,
+#ifdef HAVE_DMAPI
+	if (mp->m_flags & XFS_MOUNT_DMAPI) {
+		error = XFS_SEND_PREUNMOUNT(mp,
 				rvp, DM_RIGHT_NULL, rvp, DM_RIGHT_NULL,
 				NULL, NULL, 0, 0,
 				(mp->m_dmevmask & (1<<DM_EVENT_PREUNMOUNT))?
@@ -580,7 +597,7 @@ xfs_unmount(
 		unmount_event_flags = (mp->m_dmevmask & (1<<DM_EVENT_UNMOUNT))?
 					0 : DM_FLAGS_UNWANTED;
 	}
-
+#endif
 	/*
 	 * First blow any referenced inode from this file system
 	 * out of the reference cache, and delete the timer.
@@ -612,8 +629,7 @@ xfs_unmount(
 	 * referenced vnodes as well.
 	 */
 	if (XFS_FORCED_SHUTDOWN(mp)) {
-		error = xfs_sync(&mp->m_bhv,
-			 (SYNC_WAIT | SYNC_CLOSE), credp);
+		error = xfs_sync(mp, SYNC_WAIT | SYNC_CLOSE);
 		ASSERT(error != EFSCORRUPTED);
 	}
 	xfs_unmountfs_needed = 1;
@@ -627,7 +643,7 @@ out:
 		/* Note: mp structure must still exist for
 		 * XFS_SEND_UNMOUNT() call.
 		 */
-		XFS_SEND_UNMOUNT(mp, vfsp, error == 0 ? rvp : NULL,
+		XFS_SEND_UNMOUNT(mp, error == 0 ? rvp : NULL,
 			DM_RIGHT_NULL, 0, error, unmount_event_flags);
 	}
 	if (xfs_unmountfs_needed) {
@@ -636,6 +652,9 @@ out:
 		 * and free the super block buffer & mount structures.
 		 */
 		xfs_unmountfs(mp, credp);
+		xfs_qmops_put(mp);
+		xfs_dmops_put(mp);
+		kmem_free(mp, sizeof(xfs_mount_t));
 	}
 
 	return XFS_ERROR(error);
@@ -694,29 +713,26 @@ xfs_attr_quiesce(
 	xfs_unmountfs_writesb(mp);
 }
 
-STATIC int
+int
 xfs_mntupdate(
-	bhv_desc_t			*bdp,
+	struct xfs_mount		*mp,
 	int				*flags,
 	struct xfs_mount_args		*args)
 {
-	bhv_vfs_t	*vfsp = bhvtovfs(bdp);
-	xfs_mount_t	*mp = XFS_BHVTOM(bdp);
-
 	if (!(*flags & MS_RDONLY)) {			/* rw/ro -> rw */
-		if (vfsp->vfs_flag & VFS_RDONLY)
-			vfsp->vfs_flag &= ~VFS_RDONLY;
+		if (mp->m_flags & XFS_MOUNT_RDONLY)
+			mp->m_flags &= ~XFS_MOUNT_RDONLY;
 		if (args->flags & XFSMNT_BARRIER) {
 			mp->m_flags |= XFS_MOUNT_BARRIER;
 			xfs_mountfs_check_barriers(mp);
 		} else {
 			mp->m_flags &= ~XFS_MOUNT_BARRIER;
 		}
-	} else if (!(vfsp->vfs_flag & VFS_RDONLY)) {	/* rw -> ro */
+	} else if (!(mp->m_flags & XFS_MOUNT_RDONLY)) {	/* rw -> ro */
 		xfs_filestream_flush(mp);
-		bhv_vfs_sync(vfsp, SYNC_DATA_QUIESCE, NULL);
+		xfs_sync(mp, SYNC_DATA_QUIESCE);
 		xfs_attr_quiesce(mp);
-		vfsp->vfs_flag |= VFS_RDONLY;
+		mp->m_flags |= XFS_MOUNT_RDONLY;
 	}
 	return 0;
 }
@@ -811,14 +827,14 @@ fscorrupt_out2:
  * vpp  -- address of the caller's vnode pointer which should be
  *         set to the desired fs root vnode
  */
-STATIC int
+int
 xfs_root(
-	bhv_desc_t	*bdp,
+	xfs_mount_t	*mp,
 	bhv_vnode_t	**vpp)
 {
 	bhv_vnode_t	*vp;
 
-	vp = XFS_ITOV((XFS_BHVTOM(bdp))->m_rootip);
+	vp = XFS_ITOV(mp->m_rootip);
 	VN_HOLD(vp);
 	*vpp = vp;
 	return 0;
@@ -831,19 +847,17 @@ xfs_root(
  * the superblock lock in the mount structure to ensure a consistent
  * snapshot of the counters returned.
  */
-STATIC int
+int
 xfs_statvfs(
-	bhv_desc_t	*bdp,
+	xfs_mount_t	*mp,
 	bhv_statvfs_t	*statp,
 	bhv_vnode_t	*vp)
 {
 	__uint64_t	fakeinos;
 	xfs_extlen_t	lsize;
-	xfs_mount_t	*mp;
 	xfs_sb_t	*sbp;
 	unsigned long	s;
 
-	mp = XFS_BHVTOM(bdp);
 	sbp = &(mp->m_sb);
 
 	statp->f_type = XFS_SB_MAGIC;
@@ -874,6 +888,8 @@ xfs_statvfs(
 	xfs_statvfs_fsid(statp, mp);
 	statp->f_namelen = MAXNAMELEN - 1;
 
+	if (vp)
+		XFS_QM_DQSTATVFS(xfs_vtoi(vp), statp);
 	return 0;
 }
 
@@ -920,14 +936,30 @@ xfs_statvfs(
  *		       filesystem.
  *
  */
-/*ARGSUSED*/
-STATIC int
+int
 xfs_sync(
-	bhv_desc_t	*bdp,
-	int		flags,
-	cred_t		*credp)
+	xfs_mount_t	*mp,
+	int		flags)
 {
-	xfs_mount_t	*mp = XFS_BHVTOM(bdp);
+	int		error;
+
+	/*
+	 * Get the Quota Manager to flush the dquots.
+	 *
+	 * If XFS quota support is not enabled or this filesystem
+	 * instance does not use quotas XFS_QM_DQSYNC will always
+	 * return zero.
+	 */
+	error = XFS_QM_DQSYNC(mp, flags);
+	if (error) {
+		/*
+		 * If we got an IO error, we will be shutting down.
+		 * So, there's nothing more for us to do here.
+		 */
+		ASSERT(error != EIO || XFS_FORCED_SHUTDOWN(mp));
+		if (XFS_FORCED_SHUTDOWN(mp))
+			return XFS_ERROR(error);
+	}
 
 	if (flags & SYNC_IOWAIT)
 		xfs_filestream_flush(mp);
@@ -1015,7 +1047,7 @@ xfs_sync_inodes(
 
 	if (bypassed)
 		*bypassed = 0;
-	if (XFS_MTOVFS(mp)->vfs_flag & VFS_RDONLY)
+	if (mp->m_flags & XFS_MOUNT_RDONLY)
 		return 0;
 	error = 0;
 	last_error = 0;
@@ -1189,12 +1221,13 @@ xfs_sync_inodes(
 			if (flags & SYNC_CLOSE) {
 				/* Shutdown case. Flush and invalidate. */
 				if (XFS_FORCED_SHUTDOWN(mp))
-					bhv_vop_toss_pages(vp, 0, -1, FI_REMAPF);
+					xfs_tosspages(ip, 0, -1,
+							     FI_REMAPF);
 				else
-					error = bhv_vop_flushinval_pages(vp, 0,
-								-1, FI_REMAPF);
+					error = xfs_flushinval_pages(ip,
+							0, -1, FI_REMAPF);
 			} else if ((flags & SYNC_DELWRI) && VN_DIRTY(vp)) {
-				error = bhv_vop_flush_pages(vp, (xfs_off_t)0,
+				error = xfs_flush_pages(ip, 0,
 							-1, fflag, FI_NONE);
 			}
 
@@ -1204,7 +1237,7 @@ xfs_sync_inodes(
 			 * place after this point
 			 */
 			if (flags & SYNC_IOWAIT)
-				vn_iowait(vp);
+				vn_iowait(ip);
 
 			xfs_ilock(ip, XFS_ILOCK_SHARED);
 		}
@@ -1598,13 +1631,12 @@ xfs_syncsub(
 /*
  * xfs_vget - called by DMAPI and NFSD to get vnode from file handle
  */
-STATIC int
+int
 xfs_vget(
-	bhv_desc_t	*bdp,
+	xfs_mount_t	*mp,
 	bhv_vnode_t	**vpp,
 	fid_t		*fidp)
 {
-	xfs_mount_t	*mp = XFS_BHVTOM(bdp);
 	xfs_fid_t	*xfid = (struct xfs_fid *)fidp;
 	xfs_inode_t	*ip;
 	int		error;
@@ -1668,7 +1700,6 @@ xfs_vget(
 #define MNTOPT_BSDGROUPS    "bsdgroups"    /* group-ID from parent directory */
 #define MNTOPT_SYSVGROUPS   "sysvgroups"   /* group-ID from current process */
 #define MNTOPT_ALLOCSIZE    "allocsize"    /* preferred allocation size */
-#define MNTOPT_IHASHSIZE    "ihashsize"    /* size of inode hash table */
 #define MNTOPT_NORECOVERY   "norecovery"   /* don't run XFS recovery */
 #define MNTOPT_BARRIER	"barrier"	/* use writer barriers for log write and
 					 * unwritten extent conversion */
@@ -1683,6 +1714,21 @@ xfs_vget(
 #define MNTOPT_ATTR2	"attr2"		/* do use attr2 attribute format */
 #define MNTOPT_NOATTR2	"noattr2"	/* do not use attr2 attribute format */
 #define MNTOPT_FILESTREAM  "filestreams" /* use filestreams allocator */
+#define MNTOPT_QUOTA	"quota"		/* disk quotas (user) */
+#define MNTOPT_NOQUOTA	"noquota"	/* no quotas */
+#define MNTOPT_USRQUOTA	"usrquota"	/* user quota enabled */
+#define MNTOPT_GRPQUOTA	"grpquota"	/* group quota enabled */
+#define MNTOPT_PRJQUOTA	"prjquota"	/* project quota enabled */
+#define MNTOPT_UQUOTA	"uquota"	/* user quota (IRIX variant) */
+#define MNTOPT_GQUOTA	"gquota"	/* group quota (IRIX variant) */
+#define MNTOPT_PQUOTA	"pquota"	/* project quota (IRIX variant) */
+#define MNTOPT_UQUOTANOENF "uqnoenforce"/* user quota limit enforcement */
+#define MNTOPT_GQUOTANOENF "gqnoenforce"/* group quota limit enforcement */
+#define MNTOPT_PQUOTANOENF "pqnoenforce"/* project quota limit enforcement */
+#define MNTOPT_QUOTANOENF  "qnoenforce"	/* same as uqnoenforce */
+#define MNTOPT_DMAPI	"dmapi"		/* DMI enabled (DMAPI / XDSM) */
+#define MNTOPT_XDSM	"xdsm"		/* DMI enabled (DMAPI / XDSM) */
+#define MNTOPT_DMI	"dmi"		/* DMI enabled (DMAPI / XDSM) */
 
 STATIC unsigned long
 suffix_strtoul(char *s, char **endp, unsigned int base)
@@ -1707,19 +1753,18 @@ suffix_strtoul(char *s, char **endp, unsigned int base)
 	return simple_strtoul((const char *)s, endp, base) << shift_left_factor;
 }
 
-STATIC int
+int
 xfs_parseargs(
-	struct bhv_desc		*bhv,
+	struct xfs_mount	*mp,
 	char			*options,
 	struct xfs_mount_args	*args,
 	int			update)
 {
-	bhv_vfs_t		*vfsp = bhvtovfs(bhv);
 	char			*this_char, *value, *eov;
 	int			dsunit, dswidth, vol_dsunit, vol_dswidth;
 	int			iosize;
+	int			ikeep = 0;
 
-	args->flags |= XFSMNT_IDELETE;
 	args->flags |= XFSMNT_BARRIER;
 	args->flags2 |= XFSMNT2_COMPAT_IOSIZE;
 
@@ -1794,21 +1839,12 @@ xfs_parseargs(
 			iosize = suffix_strtoul(value, &eov, 10);
 			args->flags |= XFSMNT_IOSIZE;
 			args->iosizelog = ffs(iosize) - 1;
-		} else if (!strcmp(this_char, MNTOPT_IHASHSIZE)) {
-			if (!value || !*value) {
-				cmn_err(CE_WARN,
-					"XFS: %s option requires an argument",
-					this_char);
-				return EINVAL;
-			}
-			args->flags |= XFSMNT_IHASHSIZE;
-			args->ihashsize = simple_strtoul(value, &eov, 10);
 		} else if (!strcmp(this_char, MNTOPT_GRPID) ||
 			   !strcmp(this_char, MNTOPT_BSDGROUPS)) {
-			vfsp->vfs_flag |= VFS_GRPID;
+			mp->m_flags |= XFS_MOUNT_GRPID;
 		} else if (!strcmp(this_char, MNTOPT_NOGRPID) ||
 			   !strcmp(this_char, MNTOPT_SYSVGROUPS)) {
-			vfsp->vfs_flag &= ~VFS_GRPID;
+			mp->m_flags &= ~XFS_MOUNT_GRPID;
 		} else if (!strcmp(this_char, MNTOPT_WSYNC)) {
 			args->flags |= XFSMNT_WSYNC;
 		} else if (!strcmp(this_char, MNTOPT_OSYNCISOSYNC)) {
@@ -1858,6 +1894,7 @@ xfs_parseargs(
 		} else if (!strcmp(this_char, MNTOPT_NOBARRIER)) {
 			args->flags &= ~XFSMNT_BARRIER;
 		} else if (!strcmp(this_char, MNTOPT_IKEEP)) {
+			ikeep = 1;
 			args->flags &= ~XFSMNT_IDELETE;
 		} else if (!strcmp(this_char, MNTOPT_NOIKEEP)) {
 			args->flags |= XFSMNT_IDELETE;
@@ -1871,6 +1908,38 @@ xfs_parseargs(
 			args->flags &= ~XFSMNT_ATTR2;
 		} else if (!strcmp(this_char, MNTOPT_FILESTREAM)) {
 			args->flags2 |= XFSMNT2_FILESTREAMS;
+		} else if (!strcmp(this_char, MNTOPT_NOQUOTA)) {
+			args->flags &= ~(XFSMNT_UQUOTAENF|XFSMNT_UQUOTA);
+			args->flags &= ~(XFSMNT_GQUOTAENF|XFSMNT_GQUOTA);
+		} else if (!strcmp(this_char, MNTOPT_QUOTA) ||
+			   !strcmp(this_char, MNTOPT_UQUOTA) ||
+			   !strcmp(this_char, MNTOPT_USRQUOTA)) {
+			args->flags |= XFSMNT_UQUOTA | XFSMNT_UQUOTAENF;
+		} else if (!strcmp(this_char, MNTOPT_QUOTANOENF) ||
+			   !strcmp(this_char, MNTOPT_UQUOTANOENF)) {
+			args->flags |= XFSMNT_UQUOTA;
+			args->flags &= ~XFSMNT_UQUOTAENF;
+		} else if (!strcmp(this_char, MNTOPT_PQUOTA) ||
+			   !strcmp(this_char, MNTOPT_PRJQUOTA)) {
+			args->flags |= XFSMNT_PQUOTA | XFSMNT_PQUOTAENF;
+		} else if (!strcmp(this_char, MNTOPT_PQUOTANOENF)) {
+			args->flags |= XFSMNT_PQUOTA;
+			args->flags &= ~XFSMNT_PQUOTAENF;
+		} else if (!strcmp(this_char, MNTOPT_GQUOTA) ||
+			   !strcmp(this_char, MNTOPT_GRPQUOTA)) {
+			args->flags |= XFSMNT_GQUOTA | XFSMNT_GQUOTAENF;
+		} else if (!strcmp(this_char, MNTOPT_GQUOTANOENF)) {
+			args->flags |= XFSMNT_GQUOTA;
+			args->flags &= ~XFSMNT_GQUOTAENF;
+		} else if (!strcmp(this_char, MNTOPT_DMAPI)) {
+			args->flags |= XFSMNT_DMAPI;
+		} else if (!strcmp(this_char, MNTOPT_XDSM)) {
+			args->flags |= XFSMNT_DMAPI;
+		} else if (!strcmp(this_char, MNTOPT_DMI)) {
+			args->flags |= XFSMNT_DMAPI;
+		} else if (!strcmp(this_char, "ihashsize")) {
+			cmn_err(CE_WARN,
+	"XFS: ihashsize no longer used, option is deprecated.");
 		} else if (!strcmp(this_char, "osyncisdsync")) {
 			/* no-op, this is now the default */
 			cmn_err(CE_WARN,
@@ -1886,7 +1955,7 @@ xfs_parseargs(
 	}
 
 	if (args->flags & XFSMNT_NORECOVERY) {
-		if ((vfsp->vfs_flag & VFS_RDONLY) == 0) {
+		if ((mp->m_flags & XFS_MOUNT_RDONLY) == 0) {
 			cmn_err(CE_WARN,
 				"XFS: no-recovery mounts must be read-only.");
 			return EINVAL;
@@ -1899,6 +1968,18 @@ xfs_parseargs(
 		return EINVAL;
 	}
 
+	if ((args->flags & XFSMNT_GQUOTA) && (args->flags & XFSMNT_PQUOTA)) {
+		cmn_err(CE_WARN,
+			"XFS: cannot mount with both project and group quota");
+		return EINVAL;
+	}
+
+	if ((args->flags & XFSMNT_DMAPI) && *args->mtpt == '\0') {
+		printk("XFS: %s option needs the mount point option as well\n",
+			MNTOPT_DMAPI);
+		return EINVAL;
+	}
+
 	if ((dsunit && !dswidth) || (!dsunit && dswidth)) {
 		cmn_err(CE_WARN,
 			"XFS: sunit and swidth must be specified together");
@@ -1912,6 +1993,18 @@ xfs_parseargs(
 		return EINVAL;
 	}
 
+	/*
+	 * Applications using DMI filesystems often expect the
+	 * inode generation number to be monotonically increasing.
+	 * If we delete inode chunks we break this assumption, so
+	 * keep unused inode chunks on disk for DMI filesystems
+	 * until we come up with a better solution.
+	 * Note that if "ikeep" or "noikeep" mount options are
+	 * supplied, then they are honored.
+	 */
+	if (!(args->flags & XFSMNT_DMAPI) && !ikeep)
+		args->flags |= XFSMNT_IDELETE;
+
 	if ((args->flags & XFSMNT_NOALIGN) != XFSMNT_NOALIGN) {
 		if (dsunit) {
 			args->sunit = dsunit;
@@ -1927,15 +2020,15 @@ xfs_parseargs(
 
 done:
 	if (args->flags & XFSMNT_32BITINODES)
-		vfsp->vfs_flag |= VFS_32BITINODES;
+		mp->m_flags |= XFS_MOUNT_SMALL_INUMS;
 	if (args->flags2)
 		args->flags |= XFSMNT_FLAGS2;
 	return 0;
 }
 
-STATIC int
+int
 xfs_showargs(
-	struct bhv_desc		*bhv,
+	struct xfs_mount	*mp,
 	struct seq_file		*m)
 {
 	static struct proc_xfs_info {
@@ -1953,17 +2046,12 @@ xfs_showargs(
 		{ 0, NULL }
 	};
 	struct proc_xfs_info	*xfs_infop;
-	struct xfs_mount	*mp = XFS_BHVTOM(bhv);
-	struct bhv_vfs		*vfsp = XFS_MTOVFS(mp);
 
 	for (xfs_infop = xfs_info; xfs_infop->flag; xfs_infop++) {
 		if (mp->m_flags & xfs_infop->flag)
 			seq_puts(m, xfs_infop->str);
 	}
 
-	if (mp->m_flags & XFS_MOUNT_IHASHSIZE)
-		seq_printf(m, "," MNTOPT_IHASHSIZE "=%d", (int)mp->m_ihsize);
-
 	if (mp->m_flags & XFS_MOUNT_DFLT_IOSIZE)
 		seq_printf(m, "," MNTOPT_ALLOCSIZE "=%dk",
 				(int)(1 << mp->m_writeio_log) >> 10);
@@ -1990,11 +2078,37 @@ xfs_showargs(
 	if (!(mp->m_flags & XFS_MOUNT_COMPAT_IOSIZE))
 		seq_printf(m, "," MNTOPT_LARGEIO);
 
-	if (!(vfsp->vfs_flag & VFS_32BITINODES))
+	if (!(mp->m_flags & XFS_MOUNT_SMALL_INUMS))
 		seq_printf(m, "," MNTOPT_64BITINODE);
-	if (vfsp->vfs_flag & VFS_GRPID)
+	if (mp->m_flags & XFS_MOUNT_GRPID)
 		seq_printf(m, "," MNTOPT_GRPID);
 
+	if (mp->m_qflags & XFS_UQUOTA_ACCT) {
+		if (mp->m_qflags & XFS_UQUOTA_ENFD)
+			seq_puts(m, "," MNTOPT_USRQUOTA);
+		else
+			seq_puts(m, "," MNTOPT_UQUOTANOENF);
+	}
+
+	if (mp->m_qflags & XFS_PQUOTA_ACCT) {
+		if (mp->m_qflags & XFS_OQUOTA_ENFD)
+			seq_puts(m, "," MNTOPT_PRJQUOTA);
+		else
+			seq_puts(m, "," MNTOPT_PQUOTANOENF);
+	}
+
+	if (mp->m_qflags & XFS_GQUOTA_ACCT) {
+		if (mp->m_qflags & XFS_OQUOTA_ENFD)
+			seq_puts(m, "," MNTOPT_GRPQUOTA);
+		else
+			seq_puts(m, "," MNTOPT_GQUOTANOENF);
+	}
+
+	if (!(mp->m_qflags & XFS_ALL_QUOTA_ACCT))
+		seq_puts(m, "," MNTOPT_NOQUOTA);
+
+	if (mp->m_flags & XFS_MOUNT_DMAPI)
+		seq_puts(m, "," MNTOPT_DMAPI);
 	return 0;
 }
 
@@ -2003,31 +2117,10 @@ xfs_showargs(
  * need to take care of themetadata. Once that's done write a dummy
  * record to dirty the log in case of a crash while frozen.
  */
-STATIC void
+void
 xfs_freeze(
-	bhv_desc_t	*bdp)
+	xfs_mount_t	*mp)
 {
-	xfs_mount_t	*mp = XFS_BHVTOM(bdp);
-
 	xfs_attr_quiesce(mp);
 	xfs_fs_log_dummy(mp);
 }
-
-
-bhv_vfsops_t xfs_vfsops = {
-	BHV_IDENTITY_INIT(VFS_BHV_XFS,VFS_POSITION_XFS),
-	.vfs_parseargs		= xfs_parseargs,
-	.vfs_showargs		= xfs_showargs,
-	.vfs_mount		= xfs_mount,
-	.vfs_unmount		= xfs_unmount,
-	.vfs_mntupdate		= xfs_mntupdate,
-	.vfs_root		= xfs_root,
-	.vfs_statvfs		= xfs_statvfs,
-	.vfs_sync		= xfs_sync,
-	.vfs_vget		= xfs_vget,
-	.vfs_dmapiops		= (vfs_dmapiops_t)fs_nosys,
-	.vfs_quotactl		= (vfs_quotactl_t)fs_nosys,
-	.vfs_init_vnode		= xfs_initialize_vnode,
-	.vfs_force_shutdown	= xfs_do_force_shutdown,
-	.vfs_freeze		= xfs_freeze,
-};

fs/xfs/xfs_vfsops.h

@@ -0,0 +1,28 @@
+#ifndef _XFS_VFSOPS_H
+#define _XFS_VFSOPS_H 1
+
+struct cred;
+struct fid;
+struct inode;
+struct kstatfs;
+struct xfs_mount;
+struct xfs_mount_args;
+
+int xfs_mount(struct xfs_mount *mp, struct xfs_mount_args *args,
+		struct cred *credp);
+int xfs_unmount(struct xfs_mount *mp, int flags, struct cred *credp);
+int xfs_mntupdate(struct xfs_mount *mp, int *flags,
+		struct xfs_mount_args *args);
+int xfs_root(struct xfs_mount *mp, bhv_vnode_t **vpp);
+int xfs_statvfs(struct xfs_mount *mp, struct kstatfs *statp,
+		bhv_vnode_t *vp);
+int xfs_sync(struct xfs_mount *mp, int flags);
+int xfs_vget(struct xfs_mount *mp, bhv_vnode_t **vpp, struct fid *fidp);
+int xfs_parseargs(struct xfs_mount *mp, char *options,
+		struct xfs_mount_args *args, int update);
+int xfs_showargs(struct xfs_mount *mp, struct seq_file *m);
+void xfs_freeze(struct xfs_mount *mp);
+void xfs_do_force_shutdown(struct xfs_mount *mp, int flags, char *fname,
+		int lnnum);
+
+#endif /* _XFS_VFSOPS_H */

fs/xfs/xfs_vnodeops.h

@@ -0,0 +1,86 @@
+#ifndef _XFS_VNODEOPS_H
+#define _XFS_VNODEOPS_H 1
+
+struct attrlist_cursor_kern;
+struct bhv_vattr;
+struct cred;
+struct file;
+struct inode;
+struct iovec;
+struct kiocb;
+struct pipe_inode_info;
+struct uio;
+struct xfs_inode;
+struct xfs_iomap;
+
+
+int xfs_open(struct xfs_inode *ip);
+int xfs_getattr(struct xfs_inode *ip, struct bhv_vattr *vap, int flags);
+int xfs_setattr(struct xfs_inode *ip, struct bhv_vattr *vap, int flags,
+		struct cred *credp);
+int xfs_access(struct xfs_inode *ip, int mode, struct cred *credp);
+int xfs_readlink(struct xfs_inode *ip, char *link);
+int xfs_fsync(struct xfs_inode *ip, int flag, xfs_off_t start,
+		xfs_off_t stop);
+int xfs_release(struct xfs_inode *ip);
+int xfs_inactive(struct xfs_inode *ip);
+int xfs_lookup(struct xfs_inode *dp, bhv_vname_t *dentry,
+		bhv_vnode_t **vpp);
+int xfs_create(struct xfs_inode *dp, bhv_vname_t *dentry, mode_t mode,
+		xfs_dev_t rdev, bhv_vnode_t **vpp, struct cred *credp);
+int xfs_remove(struct xfs_inode *dp, bhv_vname_t	*dentry);
+int xfs_link(struct xfs_inode *tdp, bhv_vnode_t *src_vp,
+		bhv_vname_t *dentry);
+int xfs_mkdir(struct xfs_inode *dp, bhv_vname_t *dentry,
+		mode_t mode, bhv_vnode_t **vpp, struct cred *credp);
+int xfs_rmdir(struct xfs_inode *dp, bhv_vname_t *dentry);
+int xfs_readdir(struct xfs_inode	*dp, void *dirent, size_t bufsize,
+		       xfs_off_t *offset, filldir_t filldir);
+int xfs_symlink(struct xfs_inode *dp, bhv_vname_t *dentry,
+		char *target_path, mode_t mode, bhv_vnode_t **vpp,
+		struct cred *credp);
+int xfs_fid2(struct xfs_inode *ip, fid_t	*fidp);
+int xfs_rwlock(struct xfs_inode *ip, bhv_vrwlock_t locktype);
+void xfs_rwunlock(struct xfs_inode *ip, bhv_vrwlock_t locktype);
+int xfs_inode_flush(struct xfs_inode *ip, int flags);
+int xfs_set_dmattrs(struct xfs_inode *ip, u_int evmask, u_int16_t state);
+int xfs_reclaim(struct xfs_inode *ip);
+int xfs_change_file_space(struct xfs_inode *ip, int cmd,
+		xfs_flock64_t *bf, xfs_off_t offset,
+		struct cred *credp, int	attr_flags);
+int xfs_rename(struct xfs_inode *src_dp, bhv_vname_t *src_vname,
+		bhv_vnode_t *target_dir_vp, bhv_vname_t *target_vname);
+int xfs_attr_get(struct xfs_inode *ip, const char *name, char *value,
+		int *valuelenp, int flags, cred_t *cred);
+int xfs_attr_set(struct xfs_inode *dp, const char *name, char *value,
+		int valuelen, int flags);
+int xfs_attr_remove(struct xfs_inode *dp, const char *name, int flags);
+int xfs_attr_list(struct xfs_inode *dp, char *buffer, int bufsize,
+		int flags, struct attrlist_cursor_kern *cursor);
+int xfs_ioctl(struct xfs_inode *ip, struct file *filp,
+		int ioflags, unsigned int cmd, void __user *arg);
+ssize_t xfs_read(struct xfs_inode *ip, struct kiocb *iocb,
+		const struct iovec *iovp, unsigned int segs,
+		loff_t *offset, int ioflags);
+ssize_t xfs_sendfile(struct xfs_inode *ip, struct file *filp,
+		loff_t *offset, int ioflags, size_t count,
+		read_actor_t actor, void *target);
+ssize_t xfs_splice_read(struct xfs_inode *ip, struct file *infilp,
+		loff_t *ppos, struct pipe_inode_info *pipe, size_t count,
+		int flags, int ioflags);
+ssize_t xfs_splice_write(struct xfs_inode *ip,
+		struct pipe_inode_info *pipe, struct file *outfilp,
+		loff_t *ppos, size_t count, int flags, int ioflags);
+ssize_t xfs_write(struct xfs_inode *xip, struct kiocb *iocb,
+		const struct iovec *iovp, unsigned int nsegs,
+		loff_t *offset, int ioflags);
+int xfs_bmap(struct xfs_inode *ip, xfs_off_t offset, ssize_t count,
+		int flags, struct xfs_iomap *iomapp, int *niomaps);
+void xfs_tosspages(struct xfs_inode *inode, xfs_off_t first,
+		xfs_off_t last, int fiopt);
+int xfs_flushinval_pages(struct xfs_inode *ip, xfs_off_t first,
+		xfs_off_t last, int fiopt);
+int xfs_flush_pages(struct xfs_inode *ip, xfs_off_t first,
+		xfs_off_t last, uint64_t flags, int fiopt);
+
+#endif /* _XFS_VNODEOPS_H */