xfs_vfsops.c

linux 内核源代码

/*
 * 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_types.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_dir2.h"
#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
#include "xfs_ialloc_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_dir2_sf.h"
#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_inode_item.h"
#include "xfs_btree.h"
#include "xfs_alloc.h"
#include "xfs_ialloc.h"
#include "xfs_quota.h"
#include "xfs_error.h"
#include "xfs_bmap.h"
#include "xfs_rw.h"
#include "xfs_refcache.h"
#include "xfs_buf_item.h"
#include "xfs_log_priv.h"
#include "xfs_dir2_trace.h"
#include "xfs_extfree_item.h"
#include "xfs_acl.h"
#include "xfs_attr.h"
#include "xfs_clnt.h"
#include "xfs_mru_cache.h"
#include "xfs_filestream.h"
#include "xfs_fsops.h"
#include "xfs_vnodeops.h"
#include "xfs_vfsops.h"


int
xfs_init(void)
{
	extern kmem_zone_t	*xfs_bmap_free_item_zone;
	extern kmem_zone_t	*xfs_btree_cur_zone;
	extern kmem_zone_t	*xfs_trans_zone;
	extern kmem_zone_t	*xfs_buf_item_zone;
	extern kmem_zone_t	*xfs_dabuf_zone;
#ifdef XFS_DABUF_DEBUG
	extern lock_t	        xfs_dabuf_global_lock;
	spinlock_init(&xfs_dabuf_global_lock, "xfsda");
#endif

	/*
	 * Initialize all of the zone allocators we use.
	 */
	xfs_bmap_free_item_zone = kmem_zone_init(sizeof(xfs_bmap_free_item_t),
						 "xfs_bmap_free_item");
	xfs_btree_cur_zone = kmem_zone_init(sizeof(xfs_btree_cur_t),
					    "xfs_btree_cur");
	xfs_trans_zone = kmem_zone_init(sizeof(xfs_trans_t), "xfs_trans");
	xfs_da_state_zone =
		kmem_zone_init(sizeof(xfs_da_state_t), "xfs_da_state");
	xfs_dabuf_zone = kmem_zone_init(sizeof(xfs_dabuf_t), "xfs_dabuf");
	xfs_ifork_zone = kmem_zone_init(sizeof(xfs_ifork_t), "xfs_ifork");
	xfs_acl_zone_init(xfs_acl_zone, "xfs_acl");
	xfs_mru_cache_init();
	xfs_filestream_init();

	/*
	 * The size of the zone allocated buf log item is the maximum
	 * size possible under XFS.  This wastes a little bit of memory,
	 * but it is much faster.
	 */
	xfs_buf_item_zone =
		kmem_zone_init((sizeof(xfs_buf_log_item_t) +
				(((XFS_MAX_BLOCKSIZE / XFS_BLI_CHUNK) /
				  NBWORD) * sizeof(int))),
			       "xfs_buf_item");
	xfs_efd_zone =
		kmem_zone_init((sizeof(xfs_efd_log_item_t) +
			       ((XFS_EFD_MAX_FAST_EXTENTS - 1) *
				 sizeof(xfs_extent_t))),
				      "xfs_efd_item");
	xfs_efi_zone =
		kmem_zone_init((sizeof(xfs_efi_log_item_t) +
			       ((XFS_EFI_MAX_FAST_EXTENTS - 1) *
				 sizeof(xfs_extent_t))),
				      "xfs_efi_item");

	/*
	 * These zones warrant special memory allocator hints
	 */
	xfs_inode_zone =
		kmem_zone_init_flags(sizeof(xfs_inode_t), "xfs_inode",
					KM_ZONE_HWALIGN | KM_ZONE_RECLAIM |
					KM_ZONE_SPREAD, NULL);
	xfs_ili_zone =
		kmem_zone_init_flags(sizeof(xfs_inode_log_item_t), "xfs_ili",
					KM_ZONE_SPREAD, NULL);
	xfs_icluster_zone =
		kmem_zone_init_flags(sizeof(xfs_icluster_t), "xfs_icluster",
					KM_ZONE_SPREAD, NULL);

	/*
	 * Allocate global trace buffers.
	 */
#ifdef XFS_ALLOC_TRACE
	xfs_alloc_trace_buf = ktrace_alloc(XFS_ALLOC_TRACE_SIZE, KM_SLEEP);
#endif
#ifdef XFS_BMAP_TRACE
	xfs_bmap_trace_buf = ktrace_alloc(XFS_BMAP_TRACE_SIZE, KM_SLEEP);
#endif
#ifdef XFS_BMBT_TRACE
	xfs_bmbt_trace_buf = ktrace_alloc(XFS_BMBT_TRACE_SIZE, KM_SLEEP);
#endif
#ifdef XFS_ATTR_TRACE
	xfs_attr_trace_buf = ktrace_alloc(XFS_ATTR_TRACE_SIZE, KM_SLEEP);
#endif
#ifdef XFS_DIR2_TRACE
	xfs_dir2_trace_buf = ktrace_alloc(XFS_DIR2_GTRACE_SIZE, KM_SLEEP);
#endif

	xfs_dir_startup();

#if (defined(DEBUG) || defined(INDUCE_IO_ERROR))
	xfs_error_test_init();
#endif /* DEBUG || INDUCE_IO_ERROR */

	xfs_init_procfs();
	xfs_sysctl_register();
	return 0;
}

void
xfs_cleanup(void)
{
	extern kmem_zone_t	*xfs_bmap_free_item_zone;
	extern kmem_zone_t	*xfs_btree_cur_zone;
	extern kmem_zone_t	*xfs_inode_zone;
	extern kmem_zone_t	*xfs_trans_zone;
	extern kmem_zone_t	*xfs_da_state_zone;
	extern kmem_zone_t	*xfs_dabuf_zone;
	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_icluster_zone;

	xfs_cleanup_procfs();
	xfs_sysctl_unregister();
	xfs_refcache_destroy();
	xfs_filestream_uninit();
	xfs_mru_cache_uninit();
	xfs_acl_zone_destroy(xfs_acl_zone);

#ifdef XFS_DIR2_TRACE
	ktrace_free(xfs_dir2_trace_buf);
#endif
#ifdef XFS_ATTR_TRACE
	ktrace_free(xfs_attr_trace_buf);
#endif
#ifdef XFS_BMBT_TRACE
	ktrace_free(xfs_bmbt_trace_buf);
#endif
#ifdef XFS_BMAP_TRACE
	ktrace_free(xfs_bmap_trace_buf);
#endif
#ifdef XFS_ALLOC_TRACE
	ktrace_free(xfs_alloc_trace_buf);
#endif

	kmem_zone_destroy(xfs_bmap_free_item_zone);
	kmem_zone_destroy(xfs_btree_cur_zone);
	kmem_zone_destroy(xfs_inode_zone);
	kmem_zone_destroy(xfs_trans_zone);
	kmem_zone_destroy(xfs_da_state_zone);
	kmem_zone_destroy(xfs_dabuf_zone);
	kmem_zone_destroy(xfs_buf_item_zone);
	kmem_zone_destroy(xfs_efd_zone);
	kmem_zone_destroy(xfs_efi_zone);
	kmem_zone_destroy(xfs_ifork_zone);
	kmem_zone_destroy(xfs_ili_zone);
	kmem_zone_destroy(xfs_icluster_zone);
}

/*
 * xfs_start_flags
 *
 * This function fills in xfs_mount_t fields based on mount args.
 * Note: the superblock has _not_ yet been read in.
 */
STATIC int
xfs_start_flags(
	struct xfs_mount_args	*ap,
	struct xfs_mount	*mp)
{
	/* Values are in BBs */
	if ((ap->flags & XFSMNT_NOALIGN) != XFSMNT_NOALIGN) {
		/*
		 * At this point the superblock has not been read
		 * in, therefore we do not know the block size.
		 * Before the mount call ends we will convert
		 * these to FSBs.
		 */
		mp->m_dalign = ap->sunit;
		mp->m_swidth = ap->swidth;
	}

	if (ap->logbufs != -1 &&
	    ap->logbufs != 0 &&
	    (ap->logbufs < XLOG_MIN_ICLOGS ||
	     ap->logbufs > XLOG_MAX_ICLOGS)) {
		cmn_err(CE_WARN,
			"XFS: invalid logbufs value: %d [not %d-%d]",
			ap->logbufs, XLOG_MIN_ICLOGS, XLOG_MAX_ICLOGS);
		return XFS_ERROR(EINVAL);
	}
	mp->m_logbufs = ap->logbufs;
	if (ap->logbufsize != -1 &&
	    ap->logbufsize !=  0 &&
	    (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_logbsize = ap->logbufsize;
	mp->m_fsname_len = strlen(ap->fsname) + 1;
	mp->m_fsname = kmem_alloc(mp->m_fsname_len, KM_SLEEP);
	strcpy(mp->m_fsname, ap->fsname);
	if (ap->rtname[0]) {
		mp->m_rtname = kmem_alloc(strlen(ap->rtname) + 1, KM_SLEEP);
		strcpy(mp->m_rtname, ap->rtname);
	}
	if (ap->logname[0]) {
		mp->m_logname = kmem_alloc(strlen(ap->logname) + 1, KM_SLEEP);
		strcpy(mp->m_logname, ap->logname);
	}

	if (ap->flags & XFSMNT_WSYNC)
		mp->m_flags |= XFS_MOUNT_WSYNC;
#if XFS_BIG_INUMS
	if (ap->flags & XFSMNT_INO64) {
		mp->m_flags |= XFS_MOUNT_INO64;
		mp->m_inoadd = XFS_INO64_OFFSET;
	}
#endif
	if (ap->flags & XFSMNT_RETERR)
		mp->m_flags |= XFS_MOUNT_RETERR;
	if (ap->flags & XFSMNT_NOALIGN)
		mp->m_flags |= XFS_MOUNT_NOALIGN;
	if (ap->flags & XFSMNT_SWALLOC)
		mp->m_flags |= XFS_MOUNT_SWALLOC;
	if (ap->flags & XFSMNT_OSYNCISOSYNC)
		mp->m_flags |= XFS_MOUNT_OSYNCISOSYNC;
	if (ap->flags & XFSMNT_32BITINODES)
		mp->m_flags |= XFS_MOUNT_32BITINODES;

	if (ap->flags & XFSMNT_IOSIZE) {
		if (ap->iosizelog > XFS_MAX_IO_LOG ||
		    ap->iosizelog < XFS_MIN_IO_LOG) {
			cmn_err(CE_WARN,
		"XFS: invalid log iosize: %d [not %d-%d]",
				ap->iosizelog, XFS_MIN_IO_LOG,
				XFS_MAX_IO_LOG);
			return XFS_ERROR(EINVAL);
		}

		mp->m_flags |= XFS_MOUNT_DFLT_IOSIZE;
		mp->m_readio_log = mp->m_writeio_log = ap->iosizelog;
	}

	if (ap->flags & XFSMNT_IDELETE)
		mp->m_flags |= XFS_MOUNT_IDELETE;
	if (ap->flags & XFSMNT_DIRSYNC)
		mp->m_flags |= XFS_MOUNT_DIRSYNC;
	if (ap->flags & XFSMNT_ATTR2)
		mp->m_flags |= XFS_MOUNT_ATTR2;

	if (ap->flags2 & XFSMNT2_COMPAT_IOSIZE)
		mp->m_flags |= XFS_MOUNT_COMPAT_IOSIZE;

	/*
	 * no recovery flag requires a read-only mount
	 */
	if (ap->flags & XFSMNT_NORECOVERY) {
		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);
		}
		mp->m_flags |= XFS_MOUNT_NORECOVERY;
	}

	if (ap->flags & XFSMNT_NOUUID)
		mp->m_flags |= XFS_MOUNT_NOUUID;
	if (ap->flags & XFSMNT_BARRIER)
		mp->m_flags |= XFS_MOUNT_BARRIER;
	else
		mp->m_flags &= ~XFS_MOUNT_BARRIER;

	if (ap->flags2 & XFSMNT2_FILESTREAMS)
		mp->m_flags |= XFS_MOUNT_FILESTREAMS;

	if (ap->flags & XFSMNT_DMAPI)
		mp->m_flags |= XFS_MOUNT_DMAPI;
	return 0;
}

/*
 * This function fills in xfs_mount_t fields based on mount args.
 * Note: the superblock _has_ now been read in.
 */
STATIC int
xfs_finish_flags(
	struct xfs_mount_args	*ap,
	struct xfs_mount	*mp)
{
	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)) {
		if ((ap->logbufsize <= 0) &&
		    (mp->m_sb.sb_logsunit > XLOG_BIG_RECORD_BSIZE)) {
			mp->m_logbsize = mp->m_sb.sb_logsunit;
		} else if (ap->logbufsize > 0 &&
			   ap->logbufsize < mp->m_sb.sb_logsunit) {
			cmn_err(CE_WARN,
	"XFS: logbuf size must be greater than or equal to log stripe size");
			return XFS_ERROR(EINVAL);
		}
	} else {
		/* Fail a mount if the logbuf is larger than 32K */
		if (ap->logbufsize > XLOG_BIG_RECORD_BSIZE) {
			cmn_err(CE_WARN,
	"XFS: logbuf size for version 1 logs must be 16K or 32K");
			return XFS_ERROR(EINVAL);
		}
	}

	if (XFS_SB_VERSION_HASATTR2(&mp->m_sb)) {
		mp->m_flags |= XFS_MOUNT_ATTR2;
	}

	/*
	 * prohibit r/w mounts of read-only filesystems
	 */
	if ((mp->m_sb.sb_flags & XFS_SBF_READONLY) && !ronly) {
		cmn_err(CE_WARN,
	"XFS: cannot mount a read-only filesystem as read-write");
		return XFS_ERROR(EROFS);
	}

	/*
	 * check for shared mount.
	 */
	if (ap->flags & XFSMNT_SHARED) {
		if (!XFS_SB_VERSION_HASSHARED(&mp->m_sb))
			return XFS_ERROR(EINVAL);

		/*
		 * For IRIX 6.5, shared mounts must have the shared
		 * version bit set, have the persistent readonly
		 * field set, must be version 0 and can only be mounted
		 * read-only.
		 */
		if (!ronly || !(mp->m_sb.sb_flags & XFS_SBF_READONLY) ||
		     (mp->m_sb.sb_shared_vn != 0))
			return XFS_ERROR(EINVAL);

		mp->m_flags |= XFS_MOUNT_SHARED;

		/*
		 * Shared XFS V0 can't deal with DMI.  Return EINVAL.
		 */
		if (mp->m_sb.sb_shared_vn == 0 && (ap->flags & XFSMNT_DMAPI))
			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;
}

/*
 * xfs_mount
 *
 * The file system configurations are:
 *	(1) device (partition) with data and internal log
 *	(2) logical volume with data and log subvolumes.
 *	(3) logical volume with data, log, and realtime subvolumes.
 *
 * We only have to handle opening the log and realtime volumes here if
 * they are present.  The data subvolume has already been opened by
 * get_sb_bdev() and is stored in vfsp->vfs_super->s_bdev.
 */
int
xfs_mount(
	struct xfs_mount	*mp,
	struct xfs_mount_args	*args,
	cred_t			*credp)
{
	struct block_device	*ddev, *logdev, *rtdev;
	int			flags = 0, error;

	ddev = mp->m_super->s_bdev;
	logdev = rtdev = NULL;

	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;

	/*
	 * Open real time and log devices - order is important.
	 */
	if (args->logname[0]) {
		error = xfs_blkdev_get(mp, args->logname, &logdev);
		if (error)
			return error;
	}
	if (args->rtname[0]) {
		error = xfs_blkdev_get(mp, args->rtname, &rtdev);
		if (error) {
			xfs_blkdev_put(logdev);
			return error;
		}

		if (rtdev == ddev || rtdev == logdev) {
			cmn_err(CE_WARN,
	"XFS: Cannot mount filesystem with identical rtdev and ddev/logdev.");
			xfs_blkdev_put(logdev);
			xfs_blkdev_put(rtdev);
			return EINVAL;
		}
	}

	/*
	 * Setup xfs_mount buffer target pointers
	 */
	error = ENOMEM;
	mp->m_ddev_targp = xfs_alloc_buftarg(ddev, 0);
	if (!mp->m_ddev_targp) {
		xfs_blkdev_put(logdev);
		xfs_blkdev_put(rtdev);
		return error;
	}
	if (rtdev) {
		mp->m_rtdev_targp = xfs_alloc_buftarg(rtdev, 1);
		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) {
		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(args, mp);
	if (error)
		goto error1;
	error = xfs_readsb(mp, flags);
	if (error)
		goto error1;
	error = xfs_finish_flags(args, mp);
	if (error)
		goto error2;

	/*
	 * Setup xfs_mount buffer target pointers based on superblock
	 */
	error = xfs_setsize_buftarg(mp->m_ddev_targp, mp->m_sb.sb_blocksize,
				    mp->m_sb.sb_sectsize);
	if (!error && logdev && logdev != ddev) {
		unsigned int	log_sector_size = BBSIZE;

		if (XFS_SB_VERSION_HASSECTOR(&mp->m_sb))
			log_sector_size = mp->m_sb.sb_logsectsize;
		error = xfs_setsize_buftarg(mp->m_logdev_targp,
					    mp->m_sb.sb_blocksize,
					    log_sector_size);