xfs_vnodeops.c

linux 内核源代码

		     vap->va_extsize) ) {
			code = XFS_ERROR(EINVAL);	/* EFBIG? */
			goto error_return;
		}

		/*
		 * Can't change realtime flag if any extents are allocated.
		 */
		if ((ip->i_d.di_nextents || ip->i_delayed_blks) &&
		    (mask & XFS_AT_XFLAGS) &&
		    (ip->i_d.di_flags & XFS_DIFLAG_REALTIME) !=
		    (vap->va_xflags & XFS_XFLAG_REALTIME)) {
			code = XFS_ERROR(EINVAL);	/* EFBIG? */
			goto error_return;
		}
		/*
		 * Extent size must be a multiple of the appropriate block
		 * size, if set at all.
		 */
		if ((mask & XFS_AT_EXTSIZE) && vap->va_extsize != 0) {
			xfs_extlen_t	size;

			if ((ip->i_d.di_flags & XFS_DIFLAG_REALTIME) ||
			    ((mask & XFS_AT_XFLAGS) &&
			    (vap->va_xflags & XFS_XFLAG_REALTIME))) {
				size = mp->m_sb.sb_rextsize <<
				       mp->m_sb.sb_blocklog;
			} else {
				size = mp->m_sb.sb_blocksize;
			}
			if (vap->va_extsize % size) {
				code = XFS_ERROR(EINVAL);
				goto error_return;
			}
		}
		/*
		 * If realtime flag is set then must have realtime data.
		 */
		if ((mask & XFS_AT_XFLAGS) &&
		    (vap->va_xflags & XFS_XFLAG_REALTIME)) {
			if ((mp->m_sb.sb_rblocks == 0) ||
			    (mp->m_sb.sb_rextsize == 0) ||
			    (ip->i_d.di_extsize % mp->m_sb.sb_rextsize)) {
				code = XFS_ERROR(EINVAL);
				goto error_return;
			}
		}

		/*
		 * Can't modify an immutable/append-only file unless
		 * we have appropriate permission.
		 */
		if ((mask & XFS_AT_XFLAGS) &&
		    (ip->i_d.di_flags &
				(XFS_DIFLAG_IMMUTABLE|XFS_DIFLAG_APPEND) ||
		     (vap->va_xflags &
				(XFS_XFLAG_IMMUTABLE | XFS_XFLAG_APPEND))) &&
		    !capable(CAP_LINUX_IMMUTABLE)) {
			code = XFS_ERROR(EPERM);
			goto error_return;
		}
	}

	/*
	 * Now we can make the changes.  Before we join the inode
	 * to the transaction, if XFS_AT_SIZE is set then take care of
	 * the part of the truncation that must be done without the
	 * inode lock.  This needs to be done before joining the inode
	 * to the transaction, because the inode cannot be unlocked
	 * once it is a part of the transaction.
	 */
	if (mask & XFS_AT_SIZE) {
		code = 0;
		if ((vap->va_size > ip->i_size) &&
		    (flags & ATTR_NOSIZETOK) == 0) {
			code = xfs_igrow_start(ip, vap->va_size, credp);
		}
		xfs_iunlock(ip, XFS_ILOCK_EXCL);

		/*
		 * We are going to log the inode size change in this
		 * transaction so any previous writes that are beyond the on
		 * disk EOF and the new EOF that have not been written out need
		 * to be written here. If we do not write the data out, we
		 * expose ourselves to the null files problem.
		 *
		 * Only flush from the on disk size to the smaller of the in
		 * memory file size or the new size as that's the range we
		 * really care about here and prevents waiting for other data
		 * not within the range we care about here.
		 */
		if (!code &&
		    (ip->i_size != ip->i_d.di_size) &&
		    (vap->va_size > ip->i_d.di_size)) {
			code = xfs_flush_pages(ip,
					ip->i_d.di_size, vap->va_size,
					XFS_B_ASYNC, FI_NONE);
		}

		/* wait for all I/O to complete */
		vn_iowait(ip);

		if (!code)
			code = xfs_itruncate_data(ip, vap->va_size);
		if (code) {
			ASSERT(tp == NULL);
			lock_flags &= ~XFS_ILOCK_EXCL;
			ASSERT(lock_flags == XFS_IOLOCK_EXCL);
			goto error_return;
		}
		tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_SIZE);
		if ((code = xfs_trans_reserve(tp, 0,
					     XFS_ITRUNCATE_LOG_RES(mp), 0,
					     XFS_TRANS_PERM_LOG_RES,
					     XFS_ITRUNCATE_LOG_COUNT))) {
			xfs_trans_cancel(tp, 0);
			if (need_iolock)
				xfs_iunlock(ip, XFS_IOLOCK_EXCL);
			return code;
		}
		commit_flags = XFS_TRANS_RELEASE_LOG_RES;
		xfs_ilock(ip, XFS_ILOCK_EXCL);
	}

	if (tp) {
		xfs_trans_ijoin(tp, ip, lock_flags);
		xfs_trans_ihold(tp, ip);
	}

	/* determine whether mandatory locking mode changes */
	mandlock_before = MANDLOCK(vp, ip->i_d.di_mode);

	/*
	 * Truncate file.  Must have write permission and not be a directory.
	 */
	if (mask & XFS_AT_SIZE) {
		if (vap->va_size > ip->i_size) {
			xfs_igrow_finish(tp, ip, vap->va_size,
			    !(flags & ATTR_DMI));
		} else if ((vap->va_size <= ip->i_size) ||
			   ((vap->va_size == 0) && ip->i_d.di_nextents)) {
			/*
			 * signal a sync transaction unless
			 * we're truncating an already unlinked
			 * file on a wsync filesystem
			 */
			code = xfs_itruncate_finish(&tp, ip,
					    (xfs_fsize_t)vap->va_size,
					    XFS_DATA_FORK,
					    ((ip->i_d.di_nlink != 0 ||
					      !(mp->m_flags & XFS_MOUNT_WSYNC))
					     ? 1 : 0));
			if (code)
				goto abort_return;
			/*
			 * Truncated "down", so we're removing references
			 * to old data here - if we now delay flushing for
			 * a long time, we expose ourselves unduly to the
			 * notorious NULL files problem.  So, we mark this
			 * vnode and flush it when the file is closed, and
			 * do not wait the usual (long) time for writeout.
			 */
			xfs_iflags_set(ip, XFS_ITRUNCATED);
		}
		/*
		 * Have to do this even if the file's size doesn't change.
		 */
		timeflags |= XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG;
	}

	/*
	 * Change file access modes.
	 */
	if (mask & XFS_AT_MODE) {
		ip->i_d.di_mode &= S_IFMT;
		ip->i_d.di_mode |= vap->va_mode & ~S_IFMT;

		xfs_trans_log_inode (tp, ip, XFS_ILOG_CORE);
		timeflags |= XFS_ICHGTIME_CHG;
	}

	/*
	 * 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)) {
		/*
		 * CAP_FSETID overrides the following restrictions:
		 *
		 * The set-user-ID and set-group-ID bits of a file will be
		 * cleared upon successful return from chown()
		 */
		if ((ip->i_d.di_mode & (S_ISUID|S_ISGID)) &&
		    !capable(CAP_FSETID)) {
			ip->i_d.di_mode &= ~(S_ISUID|S_ISGID);
		}

		/*
		 * Change the ownerships and register quota modifications
		 * in the transaction.
		 */
		if (iuid != uid) {
			if (XFS_IS_UQUOTA_ON(mp)) {
				ASSERT(mask & XFS_AT_UID);
				ASSERT(udqp);
				olddquot1 = XFS_QM_DQVOPCHOWN(mp, tp, ip,
							&ip->i_udquot, udqp);
			}
			ip->i_d.di_uid = uid;
		}
		if (igid != gid) {
			if (XFS_IS_GQUOTA_ON(mp)) {
				ASSERT(!XFS_IS_PQUOTA_ON(mp));
				ASSERT(mask & XFS_AT_GID);
				ASSERT(gdqp);
				olddquot2 = XFS_QM_DQVOPCHOWN(mp, tp, ip,
							&ip->i_gdquot, gdqp);
			}
			ip->i_d.di_gid = gid;
		}
		if (iprojid != projid) {
			if (XFS_IS_PQUOTA_ON(mp)) {
				ASSERT(!XFS_IS_GQUOTA_ON(mp));
				ASSERT(mask & XFS_AT_PROJID);
				ASSERT(gdqp);
				olddquot2 = XFS_QM_DQVOPCHOWN(mp, tp, ip,
							&ip->i_gdquot, gdqp);
			}
			ip->i_d.di_projid = projid;
			/*
			 * We may have to rev the inode as well as
			 * the superblock version number since projids didn't
			 * exist before DINODE_VERSION_2 and SB_VERSION_NLINK.
			 */
			if (ip->i_d.di_version == XFS_DINODE_VERSION_1)
				xfs_bump_ino_vers2(tp, ip);
		}

		xfs_trans_log_inode (tp, ip, XFS_ILOG_CORE);
		timeflags |= XFS_ICHGTIME_CHG;
	}


	/*
	 * Change file access or modified times.
	 */
	if (mask & (XFS_AT_ATIME|XFS_AT_MTIME)) {
		if (mask & XFS_AT_ATIME) {
			ip->i_d.di_atime.t_sec = vap->va_atime.tv_sec;
			ip->i_d.di_atime.t_nsec = vap->va_atime.tv_nsec;
			ip->i_update_core = 1;
			timeflags &= ~XFS_ICHGTIME_ACC;
		}
		if (mask & XFS_AT_MTIME) {
			ip->i_d.di_mtime.t_sec = vap->va_mtime.tv_sec;
			ip->i_d.di_mtime.t_nsec = vap->va_mtime.tv_nsec;
			timeflags &= ~XFS_ICHGTIME_MOD;
			timeflags |= XFS_ICHGTIME_CHG;
		}
		if (tp && (flags & ATTR_UTIME))
			xfs_trans_log_inode (tp, ip, XFS_ILOG_CORE);
	}

	/*
	 * Change XFS-added attributes.
	 */
	if (mask & (XFS_AT_EXTSIZE|XFS_AT_XFLAGS)) {
		if (mask & XFS_AT_EXTSIZE) {
			/*
			 * Converting bytes to fs blocks.
			 */
			ip->i_d.di_extsize = vap->va_extsize >>
				mp->m_sb.sb_blocklog;
		}
		if (mask & XFS_AT_XFLAGS) {
			uint	di_flags;

			/* can't set PREALLOC this way, just preserve it */
			di_flags = (ip->i_d.di_flags & XFS_DIFLAG_PREALLOC);
			if (vap->va_xflags & XFS_XFLAG_IMMUTABLE)
				di_flags |= XFS_DIFLAG_IMMUTABLE;
			if (vap->va_xflags & XFS_XFLAG_APPEND)
				di_flags |= XFS_DIFLAG_APPEND;
			if (vap->va_xflags & XFS_XFLAG_SYNC)
				di_flags |= XFS_DIFLAG_SYNC;
			if (vap->va_xflags & XFS_XFLAG_NOATIME)
				di_flags |= XFS_DIFLAG_NOATIME;
			if (vap->va_xflags & XFS_XFLAG_NODUMP)
				di_flags |= XFS_DIFLAG_NODUMP;
			if (vap->va_xflags & XFS_XFLAG_PROJINHERIT)
				di_flags |= XFS_DIFLAG_PROJINHERIT;
			if (vap->va_xflags & XFS_XFLAG_NODEFRAG)
				di_flags |= XFS_DIFLAG_NODEFRAG;
			if (vap->va_xflags & XFS_XFLAG_FILESTREAM)
				di_flags |= XFS_DIFLAG_FILESTREAM;
			if ((ip->i_d.di_mode & S_IFMT) == S_IFDIR) {
				if (vap->va_xflags & XFS_XFLAG_RTINHERIT)
					di_flags |= XFS_DIFLAG_RTINHERIT;
				if (vap->va_xflags & XFS_XFLAG_NOSYMLINKS)
					di_flags |= XFS_DIFLAG_NOSYMLINKS;
				if (vap->va_xflags & XFS_XFLAG_EXTSZINHERIT)
					di_flags |= XFS_DIFLAG_EXTSZINHERIT;
			} else if ((ip->i_d.di_mode & S_IFMT) == S_IFREG) {
				if (vap->va_xflags & XFS_XFLAG_REALTIME) {
					di_flags |= XFS_DIFLAG_REALTIME;
					ip->i_iocore.io_flags |= XFS_IOCORE_RT;
				} else {
					ip->i_iocore.io_flags &= ~XFS_IOCORE_RT;
				}
				if (vap->va_xflags & XFS_XFLAG_EXTSIZE)
					di_flags |= XFS_DIFLAG_EXTSIZE;
			}
			ip->i_d.di_flags = di_flags;
		}
		xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
		timeflags |= XFS_ICHGTIME_CHG;
	}

	/*
	 * Change file inode change time only if XFS_AT_CTIME set
	 * AND we have been called by a DMI function.
	 */

	if ( (flags & ATTR_DMI) && (mask & XFS_AT_CTIME) ) {
		ip->i_d.di_ctime.t_sec = vap->va_ctime.tv_sec;
		ip->i_d.di_ctime.t_nsec = vap->va_ctime.tv_nsec;
		ip->i_update_core = 1;
		timeflags &= ~XFS_ICHGTIME_CHG;
	}

	/*
	 * Send out timestamp changes that need to be set to the
	 * current time.  Not done when called by a DMI function.
	 */
	if (timeflags && !(flags & ATTR_DMI))
		xfs_ichgtime(ip, timeflags);

	XFS_STATS_INC(xs_ig_attrchg);

	/*
	 * If this is a synchronous mount, make sure that the
	 * transaction goes to disk before returning to the user.
	 * This is slightly sub-optimal in that truncates require
	 * two sync transactions instead of one for wsync filesystems.
	 * One for the truncate and one for the timestamps since we
	 * don't want to change the timestamps unless we're sure the
	 * truncate worked.  Truncates are less than 1% of the laddis
	 * mix so this probably isn't worth the trouble to optimize.
	 */
	code = 0;
	if (tp) {
		if (mp->m_flags & XFS_MOUNT_WSYNC)
			xfs_trans_set_sync(tp);

		code = xfs_trans_commit(tp, commit_flags);
	}

	/*
	 * If the (regular) file's mandatory locking mode changed, then
	 * notify the vnode.  We do this under the inode lock to prevent
	 * racing calls to vop_vnode_change.
	 */
	mandlock_after = MANDLOCK(vp, ip->i_d.di_mode);

	xfs_iunlock(ip, lock_flags);

	/*
	 * Release any dquot(s) the inode had kept before chown.
	 */
	XFS_QM_DQRELE(mp, olddquot1);
	XFS_QM_DQRELE(mp, olddquot2);
	XFS_QM_DQRELE(mp, udqp);
	XFS_QM_DQRELE(mp, gdqp);

	if (code) {
		return code;
	}

	if (DM_EVENT_ENABLED(ip, DM_EVENT_ATTRIBUTE) &&
	    !(flags & ATTR_DMI)) {
		(void) XFS_SEND_NAMESP(mp, DM_EVENT_ATTRIBUTE, vp, DM_RIGHT_NULL,
					NULL, DM_RIGHT_NULL, NULL, NULL,
					0, 0, AT_DELAY_FLAG(flags));
	}
	return 0;

 abort_return:
	commit_flags |= XFS_TRANS_ABORT;
	/* FALLTHROUGH */
 error_return:
	XFS_QM_DQRELE(mp, udqp);
	XFS_QM_DQRELE(mp, gdqp);
	if (tp) {
		xfs_trans_cancel(tp, commit_flags);
	}
	if (lock_flags != 0) {
		xfs_iunlock(ip, lock_flags);
	}
	return code;
}


/*
 * xfs_access
 * Null conversion from vnode mode bits to inode mode bits, as in efs.
 */
int
xfs_access(
	xfs_inode_t	*ip,
	int		mode,
	cred_t		*credp)
{
	int		error;

	vn_trace_entry(ip, __FUNCTION__, (inst_t *)__return_address);

	xfs_ilock(ip, XFS_ILOCK_SHARED);
	error = xfs_iaccess(ip, mode, credp);
	xfs_iunlock(ip, XFS_ILOCK_SHARED);
	return error;
}


/*
 * 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

STATIC int
xfs_readlink_bmap(
	xfs_inode_t	*ip,
	char		*link)
{
	xfs_mount_t	*mp = ip->i_mount;
	int		pathlen = ip->i_d.di_size;
	int             nmaps = SYMLINK_MAPS;
	xfs_bmbt_irec_t mval[SYMLINK_MAPS];
	xfs_daddr_t	d;
	int		byte_cnt;
	int		n;
	xfs_buf_t	*bp;
	int		error = 0;

	error = xfs_bmapi(NULL, ip, 0, XFS_B_TO_FSB(mp, pathlen), 0, NULL, 0,
			mval, &nmaps, NULL, NULL);
	if (error)
		goto out;

	for (n = 0; n < nmaps; n++) {
		d = XFS_FSB_TO_DADDR(mp, mval[n].br_startblock);
		byte_cnt = XFS_FSB_TO_B(mp, mval[n].br_blockcount);

		bp = xfs_buf_read(mp->m_ddev_targp, d, BTOBB(byte_cnt), 0);
		error = XFS_BUF_GETERROR(bp);
		if (error) {
			xfs_ioerror_alert("xfs_readlink",
				  ip->i_mount, bp, XFS_BUF_ADDR(bp));
			xfs_buf_relse(bp);
			goto out;
		}
		if (pathlen < byte_cnt)
			byte_cnt = pathlen;
		pathlen -= byte_cnt;

		memcpy(link, XFS_BUF_PTR(bp), byte_cnt);
		xfs_buf_relse(bp);
	}

	link[ip->i_d.di_size] = '\0';
	error = 0;

 out:
	return error;
}

int
xfs_readlink(
	xfs_inode_t     *ip,
	char		*link)
{
	xfs_mount_t	*mp = ip->i_mount;
	int		pathlen;
	int		error = 0;

	vn_trace_entry(ip, __FUNCTION__, (inst_t *)__return_address);

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

	xfs_ilock(ip, XFS_ILOCK_SHARED);

	ASSERT((ip->i_d.di_mode & S_IFMT) == S_IFLNK);
	ASSERT(ip->i_d.di_size <= MAXPATHLEN);

	pathlen = ip->i_d.di_size;