xfs_vnodeops.c

来自「Linux Kernel 2.6.9 for OMAP1710」· C语言 代码 · 共 2,550 行 · 第 1/5 页

/*
 * Copyright (c) 2000-2003 Silicon Graphics, Inc.  All Rights Reserved.
 *
 * This program is free software; you can redistribute it and/or modify it
 * under the terms of version 2 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.
 *
 * Further, this software is distributed without any warranty that it is
 * free of the rightful claim of any third person regarding infringement
 * or the like.  Any license provided herein, whether implied or
 * otherwise, applies only to this software file.  Patent licenses, if
 * any, provided herein do not apply to combinations of this program with
 * other software, or any other product whatsoever.
 *
 * You should have received a copy of the GNU General Public License along
 * with this program; if not, write the Free Software Foundation, Inc., 59
 * Temple Place - Suite 330, Boston MA 02111-1307, USA.
 *
 * Contact information: Silicon Graphics, Inc., 1600 Amphitheatre Pkwy,
 * Mountain View, CA  94043, or:
 *
 * http://www.sgi.com
 *
 * For further information regarding this notice, see:
 *
 * http://oss.sgi.com/projects/GenInfo/SGIGPLNoticeExplan/
 */

#include "xfs.h"
#include "xfs_macros.h"
#include "xfs_types.h"
#include "xfs_inum.h"
#include "xfs_log.h"
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir.h"
#include "xfs_dir2.h"
#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_alloc_btree.h"
#include "xfs_bmap_btree.h"
#include "xfs_ialloc_btree.h"
#include "xfs_itable.h"
#include "xfs_btree.h"
#include "xfs_ialloc.h"
#include "xfs_alloc.h"
#include "xfs_attr_sf.h"
#include "xfs_dir_sf.h"
#include "xfs_dir2_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode_item.h"
#include "xfs_inode.h"
#include "xfs_bmap.h"
#include "xfs_da_btree.h"
#include "xfs_attr.h"
#include "xfs_rw.h"
#include "xfs_refcache.h"
#include "xfs_error.h"
#include "xfs_bit.h"
#include "xfs_rtalloc.h"
#include "xfs_quota.h"
#include "xfs_utils.h"
#include "xfs_trans_space.h"
#include "xfs_dir_leaf.h"
#include "xfs_mac.h"
#include "xfs_log_priv.h"


/*
 * The maximum pathlen is 1024 bytes. Since the minimum file system
 * blocksize is 512 bytes, we can get a max of 2 extents back from
 * bmapi.
 */
#define SYMLINK_MAPS 2

/*
 * For xfs, we check that the file isn't too big to be opened by this kernel.
 * No other open action is required for regular files.  Devices are handled
 * through the specfs file system, pipes through fifofs.  Device and
 * fifo vnodes are "wrapped" by specfs and fifofs vnodes, respectively,
 * when a new vnode is first looked up or created.
 */
STATIC int
xfs_open(
	bhv_desc_t	*bdp,
	cred_t		*credp)
{
	int		mode;
	vnode_t		*vp;
	xfs_inode_t	*ip;

	vp = BHV_TO_VNODE(bdp);
	ip = XFS_BHVTOI(bdp);

	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
		return XFS_ERROR(EIO);

	/*
	 * If it's a directory with any blocks, read-ahead block 0
	 * as we're almost certain to have the next operation be a read there.
	 */
	if (vp->v_type == VDIR && ip->i_d.di_nextents > 0) {
		mode = xfs_ilock_map_shared(ip);
		if (ip->i_d.di_nextents > 0)
			(void)xfs_da_reada_buf(NULL, ip, 0, XFS_DATA_FORK);
		xfs_iunlock(ip, mode);
	}
	return 0;
}


/*
 * xfs_getattr
 */
STATIC int
xfs_getattr(
	bhv_desc_t	*bdp,
	vattr_t		*vap,
	int		flags,
	cred_t		*credp)
{
	xfs_inode_t	*ip;
	xfs_mount_t	*mp;
	vnode_t		*vp;

	vp  = BHV_TO_VNODE(bdp);
	vn_trace_entry(vp, __FUNCTION__, (inst_t *)__return_address);

	ip = XFS_BHVTOI(bdp);
	mp = ip->i_mount;

	if (XFS_FORCED_SHUTDOWN(mp))
		return XFS_ERROR(EIO);

	if (!(flags & ATTR_LAZY))
		xfs_ilock(ip, XFS_ILOCK_SHARED);

	vap->va_size = ip->i_d.di_size;
	if (vap->va_mask == XFS_AT_SIZE)
		goto all_done;

	vap->va_nblocks =
		XFS_FSB_TO_BB(mp, ip->i_d.di_nblocks + ip->i_delayed_blks);
	vap->va_nodeid = ip->i_ino;
#if XFS_BIG_INUMS
	vap->va_nodeid += mp->m_inoadd;
#endif
	vap->va_nlink = ip->i_d.di_nlink;

	/*
	 * Quick exit for non-stat callers
	 */
	if ((vap->va_mask &
	    ~(XFS_AT_SIZE|XFS_AT_FSID|XFS_AT_NODEID|
	      XFS_AT_NLINK|XFS_AT_BLKSIZE)) == 0)
		goto all_done;

	/*
	 * Copy from in-core inode.
	 */
	vap->va_type = vp->v_type;
	vap->va_mode = ip->i_d.di_mode & MODEMASK;
	vap->va_uid = ip->i_d.di_uid;
	vap->va_gid = ip->i_d.di_gid;
	vap->va_projid = ip->i_d.di_projid;

	/*
	 * Check vnode type block/char vs. everything else.
	 * Do it with bitmask because that's faster than looking
	 * for multiple values individually.
	 */
	if (((1 << vp->v_type) & ((1<<VBLK) | (1<<VCHR))) == 0) {
		vap->va_rdev = 0;

		if (!(ip->i_d.di_flags & XFS_DIFLAG_REALTIME)) {

#if 0
			/* Large block sizes confuse various
			 * user space programs, so letting the
			 * stripe size through is not a good
			 * idea for now.
			 */
			vap->va_blocksize = mp->m_swidth ?
				/*
				 * If the underlying volume is a stripe, then
				 * return the stripe width in bytes as the
				 * recommended I/O size.
				 */
				(mp->m_swidth << mp->m_sb.sb_blocklog) :
				/*
				 * Return the largest of the preferred buffer
				 * sizes since doing small I/Os into larger
				 * buffers causes buffers to be decommissioned.
				 * The value returned is in bytes.
				 */
				(1 << (int)MAX(mp->m_readio_log,
					       mp->m_writeio_log));

#else
			vap->va_blocksize =
				/*
				 * Return the largest of the preferred buffer
				 * sizes since doing small I/Os into larger
				 * buffers causes buffers to be decommissioned.
				 * The value returned is in bytes.
				 */
				1 << (int)MAX(mp->m_readio_log,
					       mp->m_writeio_log);
#endif
		} else {

			/*
			 * If the file blocks are being allocated from a
			 * realtime partition, then return the inode's
			 * realtime extent size or the realtime volume's
			 * extent size.
			 */
			vap->va_blocksize = ip->i_d.di_extsize ?
				(ip->i_d.di_extsize << mp->m_sb.sb_blocklog) :
				(mp->m_sb.sb_rextsize << mp->m_sb.sb_blocklog);
		}
	} else {
		vap->va_rdev = ip->i_df.if_u2.if_rdev;
		vap->va_blocksize = BLKDEV_IOSIZE;
	}

	vap->va_atime.tv_sec = ip->i_d.di_atime.t_sec;
	vap->va_atime.tv_nsec = ip->i_d.di_atime.t_nsec;
	vap->va_mtime.tv_sec = ip->i_d.di_mtime.t_sec;
	vap->va_mtime.tv_nsec = ip->i_d.di_mtime.t_nsec;
	vap->va_ctime.tv_sec = ip->i_d.di_ctime.t_sec;
	vap->va_ctime.tv_nsec = ip->i_d.di_ctime.t_nsec;

	/*
	 * Exit for stat callers.  See if any of the rest of the fields
	 * to be filled in are needed.
	 */
	if ((vap->va_mask &
	     (XFS_AT_XFLAGS|XFS_AT_EXTSIZE|XFS_AT_NEXTENTS|XFS_AT_ANEXTENTS|
	      XFS_AT_GENCOUNT|XFS_AT_VCODE)) == 0)
		goto all_done;

	/*
	 * Convert di_flags to xflags.
	 */
	vap->va_xflags = xfs_dic2xflags(&ip->i_d, ARCH_NOCONVERT);

	/*
	 * Exit for inode revalidate.  See if any of the rest of
	 * the fields to be filled in are needed.
	 */
	if ((vap->va_mask &
	     (XFS_AT_EXTSIZE|XFS_AT_NEXTENTS|XFS_AT_ANEXTENTS|
	      XFS_AT_GENCOUNT|XFS_AT_VCODE)) == 0)
		goto all_done;

	vap->va_extsize = ip->i_d.di_extsize << mp->m_sb.sb_blocklog;
	vap->va_nextents =
		(ip->i_df.if_flags & XFS_IFEXTENTS) ?
			ip->i_df.if_bytes / sizeof(xfs_bmbt_rec_t) :
			ip->i_d.di_nextents;
	if (ip->i_afp)
		vap->va_anextents =
			(ip->i_afp->if_flags & XFS_IFEXTENTS) ?
				ip->i_afp->if_bytes / sizeof(xfs_bmbt_rec_t) :
				 ip->i_d.di_anextents;
	else
		vap->va_anextents = 0;
	vap->va_gen = ip->i_d.di_gen;

 all_done:
	if (!(flags & ATTR_LAZY))
		xfs_iunlock(ip, XFS_ILOCK_SHARED);
	return 0;
}


/*
 * xfs_setattr
 */
int
xfs_setattr(
	bhv_desc_t		*bdp,
	vattr_t			*vap,
	int			flags,
	cred_t			*credp)
{
	xfs_inode_t		*ip;
	xfs_trans_t		*tp;
	xfs_mount_t		*mp;
	int			mask;
	int			code;
	uint			lock_flags;
	uint			commit_flags=0;
	uid_t			uid=0, iuid=0;
	gid_t			gid=0, igid=0;
	int			timeflags = 0;
	vnode_t			*vp;
	xfs_prid_t		projid=0, iprojid=0;
	int			mandlock_before, mandlock_after;
	struct xfs_dquot	*udqp, *gdqp, *olddquot1, *olddquot2;
	int			file_owner;
	int			need_iolock = (flags & ATTR_DMI) == 0;

	vp = BHV_TO_VNODE(bdp);
	vn_trace_entry(vp, __FUNCTION__, (inst_t *)__return_address);

	if (vp->v_vfsp->vfs_flag & VFS_RDONLY)
		return XFS_ERROR(EROFS);

	/*
	 * Cannot set certain attributes.
	 */
	mask = vap->va_mask;
	if (mask & XFS_AT_NOSET) {
		return XFS_ERROR(EINVAL);
	}

	ip = XFS_BHVTOI(bdp);
	mp = ip->i_mount;

	if (XFS_FORCED_SHUTDOWN(mp))
		return XFS_ERROR(EIO);

	/*
	 * Timestamps do not need to be logged and hence do not
	 * need to be done within a transaction.
	 */
	if (mask & XFS_AT_UPDTIMES) {
		ASSERT((mask & ~XFS_AT_UPDTIMES) == 0);
		timeflags = ((mask & XFS_AT_UPDATIME) ? XFS_ICHGTIME_ACC : 0) |
			    ((mask & XFS_AT_UPDCTIME) ? XFS_ICHGTIME_CHG : 0) |
			    ((mask & XFS_AT_UPDMTIME) ? XFS_ICHGTIME_MOD : 0);
		xfs_ichgtime(ip, timeflags);
		return 0;
	}

	olddquot1 = olddquot2 = NULL;
	udqp = gdqp = NULL;

	/*
	 * If disk quotas is on, we make sure that the dquots do exist on disk,
	 * before we start any other transactions. Trying to do this later
	 * is messy. We don't care to take a readlock to look at the ids
	 * in inode here, because we can't hold it across the trans_reserve.
	 * If the IDs do change before we take the ilock, we're covered
	 * because the i_*dquot fields will get updated anyway.
	 */
	if (XFS_IS_QUOTA_ON(mp) && (mask & (XFS_AT_UID|XFS_AT_GID))) {
		uint	qflags = 0;

		if (mask & XFS_AT_UID) {
			uid = vap->va_uid;
			qflags |= XFS_QMOPT_UQUOTA;
		} else {
			uid = ip->i_d.di_uid;
		}
		if (mask & XFS_AT_GID) {
			gid = vap->va_gid;
			qflags |= XFS_QMOPT_GQUOTA;
		}  else {
			gid = ip->i_d.di_gid;
		}
		/*
		 * We take a reference when we initialize udqp and gdqp,
		 * so it is important that we never blindly double trip on
		 * the same variable. See xfs_create() for an example.
		 */
		ASSERT(udqp == NULL);
		ASSERT(gdqp == NULL);
		code = XFS_QM_DQVOPALLOC(mp, ip, uid,gid, qflags, &udqp, &gdqp);
		if (code)
			return (code);
	}

	/*
	 * For the other attributes, we acquire the inode lock and
	 * first do an error checking pass.
	 */
	tp = NULL;
	lock_flags = XFS_ILOCK_EXCL;
	if (!(mask & XFS_AT_SIZE)) {
		if ((mask != (XFS_AT_CTIME|XFS_AT_ATIME|XFS_AT_MTIME)) ||
		    (mp->m_flags & XFS_MOUNT_WSYNC)) {
			tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_NOT_SIZE);
			commit_flags = 0;
			if ((code = xfs_trans_reserve(tp, 0,
						     XFS_ICHANGE_LOG_RES(mp), 0,
						     0, 0))) {
				lock_flags = 0;
				goto error_return;
			}
		}
	} else {
		if (DM_EVENT_ENABLED (vp->v_vfsp, ip, DM_EVENT_TRUNCATE) &&
		    !(flags & ATTR_DMI)) {
			int dmflags = AT_DELAY_FLAG(flags) | DM_SEM_FLAG_WR;
			code = XFS_SEND_DATA(mp, DM_EVENT_TRUNCATE, vp,
				vap->va_size, 0, dmflags, NULL);
			if (code) {
				lock_flags = 0;
				goto error_return;
			}
		}
		if (need_iolock)
			lock_flags |= XFS_IOLOCK_EXCL;
	}

	xfs_ilock(ip, lock_flags);

	/* boolean: are we the file owner? */
	file_owner = (current_fsuid(credp) == ip->i_d.di_uid);

	/*
	 * Change various properties of a file.
	 * Only the owner or users with CAP_FOWNER
	 * capability may do these things.
	 */
	if (mask &
	    (XFS_AT_MODE|XFS_AT_XFLAGS|XFS_AT_EXTSIZE|XFS_AT_UID|
	     XFS_AT_GID|XFS_AT_PROJID)) {
		/*
		 * CAP_FOWNER overrides the following restrictions:
		 *
		 * The user ID of the calling process must be equal
		 * to the file owner ID, except in cases where the
		 * CAP_FSETID capability is applicable.
		 */
		if (!file_owner && !capable(CAP_FOWNER)) {
			code = XFS_ERROR(EPERM);
			goto error_return;
		}

		/*
		 * CAP_FSETID overrides the following restrictions:
		 *
		 * The effective user ID of the calling process shall match
		 * the file owner when setting the set-user-ID and
		 * set-group-ID bits on that file.
		 *
		 * The effective group ID or one of the supplementary group
		 * IDs of the calling process shall match the group owner of
		 * the file when setting the set-group-ID bit on that file
		 */
		if (mask & XFS_AT_MODE) {
			mode_t m = 0;

			if ((vap->va_mode & S_ISUID) && !file_owner)
				m |= S_ISUID;
			if ((vap->va_mode & S_ISGID) &&
			    !in_group_p((gid_t)ip->i_d.di_gid))
				m |= S_ISGID;
#if 0
			/* Linux allows this, Irix doesn't. */
			if ((vap->va_mode & S_ISVTX) && vp->v_type != VDIR)
				m |= S_ISVTX;
#endif
			if (m && !capable(CAP_FSETID))
				vap->va_mode &= ~m;
		}
	}

	/*
	 * Change file ownership.  Must be the owner or privileged.
	 * If the system was configured with the "restricted_chown"
	 * option, the owner is not permitted to give away the file,
	 * and can change the group id only to a group of which he
	 * or she is a member.
	 */
	if (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_PROJID)) {
		/*
		 * These IDs could have changed since we last looked at them.
		 * But, we're assured that if the ownership did change
		 * while we didn't have the inode locked, inode's dquot(s)
		 * would have changed also.
		 */
		iuid = ip->i_d.di_uid;
		iprojid = ip->i_d.di_projid;
		igid = ip->i_d.di_gid;
		gid = (mask & XFS_AT_GID) ? vap->va_gid : igid;
		uid = (mask & XFS_AT_UID) ? vap->va_uid : iuid;
		projid = (mask & XFS_AT_PROJID) ? (xfs_prid_t)vap->va_projid :
			 iprojid;

		/*
		 * CAP_CHOWN overrides the following restrictions:
		 *
		 * If _POSIX_CHOWN_RESTRICTED is defined, this capability
		 * shall override the restriction that a process cannot
		 * change the user ID of a file it owns and the restriction
		 * that the group ID supplied to the chown() function
		 * shall be equal to either the group ID or one of the
		 * supplementary group IDs of the calling process.
		 *
		 * XXX: How does restricted_chown affect projid?
		 */
		if (restricted_chown &&
		    (iuid != uid || (igid != gid &&
				     !in_group_p((gid_t)gid))) &&
		    !capable(CAP_CHOWN)) {
			code = XFS_ERROR(EPERM);
			goto error_return;
		}
		/*
		 * Do a quota reservation only if uid or gid is actually