xfs_iops.c

linux2.6.16版本

/*
 * 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_bit.h"
#include "xfs_log.h"
#include "xfs_inum.h"
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir.h"
#include "xfs_dir2.h"
#include "xfs_alloc.h"
#include "xfs_dmapi.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
#include "xfs_dir_sf.h"
#include "xfs_dir2_sf.h"
#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_bmap.h"
#include "xfs_btree.h"
#include "xfs_ialloc.h"
#include "xfs_rtalloc.h"
#include "xfs_error.h"
#include "xfs_itable.h"
#include "xfs_rw.h"
#include "xfs_acl.h"
#include "xfs_cap.h"
#include "xfs_mac.h"
#include "xfs_attr.h"
#include "xfs_buf_item.h"
#include "xfs_utils.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(
	struct vnode	*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.
 */
void
xfs_synchronize_atime(
	xfs_inode_t	*ip)
{
	vnode_t		*vp;

	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;
	}
}

/*
 * Change the requested timestamp in the given inode.
 * We don't lock across timestamp updates, and we don't log them but
 * we do record the fact that there is dirty information in core.
 *
 * NOTE -- callers MUST combine XFS_ICHGTIME_MOD or XFS_ICHGTIME_CHG
 *		with XFS_ICHGTIME_ACC to be sure that access time
 *		update will take.  Calling first with XFS_ICHGTIME_ACC
 *		and then XFS_ICHGTIME_MOD may fail to modify the access
 *		timestamp if the filesystem is mounted noacctm.
 */
void
xfs_ichgtime(
	xfs_inode_t	*ip,
	int		flags)
{
	struct inode	*inode = LINVFS_GET_IP(XFS_ITOV(ip));
	timespec_t	tv;

	nanotime(&tv);
	if (flags & XFS_ICHGTIME_MOD) {
		inode->i_mtime = tv;
		ip->i_d.di_mtime.t_sec = (__int32_t)tv.tv_sec;
		ip->i_d.di_mtime.t_nsec = (__int32_t)tv.tv_nsec;
	}
	if (flags & XFS_ICHGTIME_ACC) {
		inode->i_atime = tv;
		ip->i_d.di_atime.t_sec = (__int32_t)tv.tv_sec;
		ip->i_d.di_atime.t_nsec = (__int32_t)tv.tv_nsec;
	}
	if (flags & XFS_ICHGTIME_CHG) {
		inode->i_ctime = tv;
		ip->i_d.di_ctime.t_sec = (__int32_t)tv.tv_sec;
		ip->i_d.di_ctime.t_nsec = (__int32_t)tv.tv_nsec;
	}

	/*
	 * We update the i_update_core field _after_ changing
	 * the timestamps in order to coordinate properly with
	 * xfs_iflush() so that we don't lose timestamp updates.
	 * This keeps us from having to hold the inode lock
	 * while doing this.  We use the SYNCHRONIZE macro to
	 * ensure that the compiler does not reorder the update
	 * of i_update_core above the timestamp updates above.
	 */
	SYNCHRONIZE();
	ip->i_update_core = 1;
	if (!(inode->i_state & I_LOCK))
		mark_inode_dirty_sync(inode);
}

/*
 * Variant on the above which avoids querying the system clock
 * in situations where we know the Linux inode timestamps have
 * just been updated (and so we can update our inode cheaply).
 */
void
xfs_ichgtime_fast(
	xfs_inode_t	*ip,
	struct inode	*inode,
	int		flags)
{
	timespec_t	*tvp;

	/*
	 * Atime updates for read() & friends are handled lazily now, and
	 * explicit updates must go through xfs_ichgtime()
	 */
	ASSERT((flags & XFS_ICHGTIME_ACC) == 0);

	/*
	 * We're not supposed to change timestamps in readonly-mounted
	 * filesystems.  Throw it away if anyone asks us.
	 */
	if (unlikely(IS_RDONLY(inode)))
		return;

	if (flags & XFS_ICHGTIME_MOD) {
		tvp = &inode->i_mtime;
		ip->i_d.di_mtime.t_sec = (__int32_t)tvp->tv_sec;
		ip->i_d.di_mtime.t_nsec = (__int32_t)tvp->tv_nsec;
	}
	if (flags & XFS_ICHGTIME_CHG) {
		tvp = &inode->i_ctime;
		ip->i_d.di_ctime.t_sec = (__int32_t)tvp->tv_sec;
		ip->i_d.di_ctime.t_nsec = (__int32_t)tvp->tv_nsec;
	}

	/*
	 * We update the i_update_core field _after_ changing
	 * the timestamps in order to coordinate properly with
	 * xfs_iflush() so that we don't lose timestamp updates.
	 * This keeps us from having to hold the inode lock
	 * while doing this.  We use the SYNCHRONIZE macro to
	 * ensure that the compiler does not reorder the update
	 * of i_update_core above the timestamp updates above.
	 */
	SYNCHRONIZE();
	ip->i_update_core = 1;
	if (!(inode->i_state & I_LOCK))
		mark_inode_dirty_sync(inode);
}


/*
 * Pull the link count and size up from the xfs inode to the linux inode
 */
STATIC void
validate_fields(
	struct inode	*ip)
{
	vnode_t		*vp = LINVFS_GET_VP(ip);
	vattr_t		va;
	int		error;

	va.va_mask = XFS_AT_NLINK|XFS_AT_SIZE|XFS_AT_NBLOCKS;
	VOP_GETATTR(vp, &va, ATTR_LAZY, NULL, error);
	if (likely(!error)) {
		ip->i_nlink = va.va_nlink;
		ip->i_blocks = va.va_nblocks;

		/* we're under i_mutex so i_size can't change under us */
		if (i_size_read(ip) != va.va_size)
			i_size_write(ip, va.va_size);
	}
}

/*
 * Hook in SELinux.  This is not quite correct yet, what we really need
 * here (as we do for default ACLs) is a mechanism by which creation of
 * these attrs can be journalled at inode creation time (along with the
 * inode, of course, such that log replay can't cause these to be lost).
 */
STATIC int
linvfs_init_security(
	struct vnode	*vp,
	struct inode	*dir)
{
	struct inode	*ip = LINVFS_GET_IP(vp);
	size_t		length;
	void		*value;
	char		*name;
	int		error;

	error = security_inode_init_security(ip, dir, &name, &value, &length);
	if (error) {
		if (error == -EOPNOTSUPP)
			return 0;
		return -error;
	}

	VOP_ATTR_SET(vp, name, value, length, ATTR_SECURE, NULL, error);
	if (!error)
		VMODIFY(vp);

	kfree(name);
	kfree(value);
	return error;
}

/*
 * Determine whether a process has a valid fs_struct (kernel daemons
 * like knfsd don't have an fs_struct).
 *
 * XXX(hch):  nfsd is broken, better fix it instead.
 */
STATIC inline int
has_fs_struct(struct task_struct *task)
{
	return (task->fs != init_task.fs);
}

STATIC inline void
cleanup_inode(
	vnode_t		*dvp,
	vnode_t		*vp,
	struct dentry	*dentry,	
	int		mode)
{
	struct dentry   teardown = {};
	int             err2;

	/* Oh, the horror.
	 * If we can't add the ACL or we fail in 
	 * linvfs_init_security we must back out.
	 * ENOSPC can hit here, among other things.
	 */
	teardown.d_inode = LINVFS_GET_IP(vp);
	teardown.d_name = dentry->d_name;

	if (S_ISDIR(mode))
	  	VOP_RMDIR(dvp, &teardown, NULL, err2);
	else
		VOP_REMOVE(dvp, &teardown, NULL, err2);
	VN_RELE(vp);
}

STATIC int
linvfs_mknod(
	struct inode	*dir,
	struct dentry	*dentry,
	int		mode,
	dev_t		rdev)
{
	struct inode	*ip;
	vattr_t		va;
	vnode_t		*vp = NULL, *dvp = LINVFS_GET_VP(dir);
	xfs_acl_t	*default_acl = NULL;
	attrexists_t	test_default_acl = _ACL_DEFAULT_EXISTS;
	int		error;

	/*
	 * Irix uses Missed'em'V split, but doesn't want to see
	 * the upper 5 bits of (14bit) major.
	 */
	if (!sysv_valid_dev(rdev) || MAJOR(rdev) & ~0x1ff)
		return -EINVAL;

	if (test_default_acl && test_default_acl(dvp)) {
		if (!_ACL_ALLOC(default_acl))
			return -ENOMEM;
		if (!_ACL_GET_DEFAULT(dvp, default_acl)) {
			_ACL_FREE(default_acl);
			default_acl = NULL;
		}
	}

	if (IS_POSIXACL(dir) && !default_acl && has_fs_struct(current))
		mode &= ~current->fs->umask;

	memset(&va, 0, sizeof(va));
	va.va_mask = XFS_AT_TYPE|XFS_AT_MODE;
	va.va_mode = mode;

	switch (mode & S_IFMT) {
	case S_IFCHR: case S_IFBLK: case S_IFIFO: case S_IFSOCK:
		va.va_rdev = sysv_encode_dev(rdev);
		va.va_mask |= XFS_AT_RDEV;
		/*FALLTHROUGH*/
	case S_IFREG:
		VOP_CREATE(dvp, dentry, &va, &vp, NULL, error);
		break;
	case S_IFDIR:
		VOP_MKDIR(dvp, dentry, &va, &vp, NULL, error);
		break;
	default:
		error = EINVAL;
		break;
	}

	if (!error)
	{
		error = linvfs_init_security(vp, dir);
		if (error)
			cleanup_inode(dvp, vp, dentry, mode);
	}

	if (default_acl) {
		if (!error) {
			error = _ACL_INHERIT(vp, &va, default_acl);
			if (!error) 
				VMODIFY(vp);
			else
				cleanup_inode(dvp, vp, dentry, mode);
		}
		_ACL_FREE(default_acl);
	}

	if (!error) {
		ASSERT(vp);
		ip = LINVFS_GET_IP(vp);

		if (S_ISCHR(mode) || S_ISBLK(mode))
			ip->i_rdev = rdev;
		else if (S_ISDIR(mode))
			validate_fields(ip);
		d_instantiate(dentry, ip);
		validate_fields(dir);
	}
	return -error;
}

STATIC int
linvfs_create(
	struct inode	*dir,
	struct dentry	*dentry,
	int		mode,
	struct nameidata *nd)
{
	return linvfs_mknod(dir, dentry, mode, 0);
}

STATIC int
linvfs_mkdir(
	struct inode	*dir,
	struct dentry	*dentry,
	int		mode)
{
	return linvfs_mknod(dir, dentry, mode|S_IFDIR, 0);
}

STATIC struct dentry *
linvfs_lookup(
	struct inode	*dir,
	struct dentry	*dentry,
	struct nameidata *nd)
{
	struct vnode	*vp = LINVFS_GET_VP(dir), *cvp;
	int		error;

	if (dentry->d_name.len >= MAXNAMELEN)
		return ERR_PTR(-ENAMETOOLONG);

	VOP_LOOKUP(vp, dentry, &cvp, 0, NULL, NULL, error);
	if (error) {
		if (unlikely(error != ENOENT))
			return ERR_PTR(-error);
		d_add(dentry, NULL);
		return NULL;
	}

	return d_splice_alias(LINVFS_GET_IP(cvp), dentry);
}

STATIC int
linvfs_link(
	struct dentry	*old_dentry,
	struct inode	*dir,
	struct dentry	*dentry)
{
	struct inode	*ip;	/* inode of guy being linked to */
	vnode_t		*tdvp;	/* target directory for new name/link */
	vnode_t		*vp;	/* vp of name being linked */
	int		error;

	ip = old_dentry->d_inode;	/* inode being linked to */
	if (S_ISDIR(ip->i_mode))
		return -EPERM;