[linux-2.6-block.git] / fs / gfs2 / ops_super.c

/*
 * Copyright (C) Sistina Software, Inc.  1997-2003 All rights reserved.
 * Copyright (C) 2004-2008 Red Hat, Inc.  All rights reserved.
 *
 * This copyrighted material is made available to anyone wishing to use,
 * modify, copy, or redistribute it subject to the terms and conditions
 * of the GNU General Public License version 2.
 */

#include <linux/sched.h>
#include <linux/slab.h>
#include <linux/spinlock.h>
#include <linux/completion.h>
#include <linux/buffer_head.h>
#include <linux/statfs.h>
#include <linux/seq_file.h>
#include <linux/mount.h>
#include <linux/kthread.h>
#include <linux/delay.h>
#include <linux/gfs2_ondisk.h>
#include <linux/crc32.h>
#include <linux/lm_interface.h>
#include <linux/time.h>

#include "gfs2.h"
#include "incore.h"
#include "glock.h"
#include "inode.h"
#include "log.h"
#include "mount.h"
#include "quota.h"
#include "recovery.h"
#include "rgrp.h"
#include "super.h"
#include "sys.h"
#include "util.h"
#include "trans.h"
#include "dir.h"
#include "eattr.h"
#include "bmap.h"
#include "meta_io.h"

/**
 * gfs2_write_inode - Make sure the inode is stable on the disk
 * @inode: The inode
 * @sync: synchronous write flag
 *
 * Returns: errno
 */

static int gfs2_write_inode(struct inode *inode, int sync)
{
	struct gfs2_inode *ip = GFS2_I(inode);
	struct gfs2_sbd *sdp = GFS2_SB(inode);
	struct gfs2_holder gh;
	struct buffer_head *bh;
	struct timespec atime;
	struct gfs2_dinode *di;
	int ret = 0;

	/* Check this is a "normal" inode, etc */
	if (!test_bit(GIF_USER, &ip->i_flags) ||
	    (current->flags & PF_MEMALLOC))
		return 0;
	ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
	if (ret)
		goto do_flush;
	ret = gfs2_trans_begin(sdp, RES_DINODE, 0);
	if (ret)
		goto do_unlock;
	ret = gfs2_meta_inode_buffer(ip, &bh);
	if (ret == 0) {
		di = (struct gfs2_dinode *)bh->b_data;
		atime.tv_sec = be64_to_cpu(di->di_atime);
		atime.tv_nsec = be32_to_cpu(di->di_atime_nsec);
		if (timespec_compare(&inode->i_atime, &atime) > 0) {
			gfs2_trans_add_bh(ip->i_gl, bh, 1);
			gfs2_dinode_out(ip, bh->b_data);
		}
		brelse(bh);
	}
	gfs2_trans_end(sdp);
do_unlock:
	gfs2_glock_dq_uninit(&gh);
do_flush:
	if (sync != 0)
		gfs2_log_flush(GFS2_SB(inode), ip->i_gl);
	return ret;
}

/**
 * gfs2_make_fs_ro - Turn a Read-Write FS into a Read-Only one
 * @sdp: the filesystem
 *
 * Returns: errno
 */

static int gfs2_make_fs_ro(struct gfs2_sbd *sdp)
{
	struct gfs2_holder t_gh;
	int error;

	gfs2_quota_sync(sdp);
	gfs2_statfs_sync(sdp);

	error = gfs2_glock_nq_init(sdp->sd_trans_gl, LM_ST_SHARED, GL_NOCACHE,
				   &t_gh);
	if (error && !test_bit(SDF_SHUTDOWN, &sdp->sd_flags))
		return error;

	gfs2_meta_syncfs(sdp);
	gfs2_log_shutdown(sdp);

	clear_bit(SDF_JOURNAL_LIVE, &sdp->sd_flags);

	if (t_gh.gh_gl)
		gfs2_glock_dq_uninit(&t_gh);

	gfs2_quota_cleanup(sdp);

	return error;
}

/**
 * gfs2_put_super - Unmount the filesystem
 * @sb: The VFS superblock
 *
 */

static void gfs2_put_super(struct super_block *sb)
{
	struct gfs2_sbd *sdp = sb->s_fs_info;
	int error;

	/*  Unfreeze the filesystem, if we need to  */

	mutex_lock(&sdp->sd_freeze_lock);
	if (sdp->sd_freeze_count)
		gfs2_glock_dq_uninit(&sdp->sd_freeze_gh);
	mutex_unlock(&sdp->sd_freeze_lock);

	kthread_stop(sdp->sd_quotad_process);
	kthread_stop(sdp->sd_logd_process);
	kthread_stop(sdp->sd_recoverd_process);

	if (!(sb->s_flags & MS_RDONLY)) {
		error = gfs2_make_fs_ro(sdp);
		if (error)
			gfs2_io_error(sdp);
	}
	/*  At this point, we're through modifying the disk  */

	/*  Release stuff  */

	iput(sdp->sd_jindex);
	iput(sdp->sd_inum_inode);
	iput(sdp->sd_statfs_inode);
	iput(sdp->sd_rindex);
	iput(sdp->sd_quota_inode);

	gfs2_glock_put(sdp->sd_rename_gl);
	gfs2_glock_put(sdp->sd_trans_gl);

	if (!sdp->sd_args.ar_spectator) {
		gfs2_glock_dq_uninit(&sdp->sd_journal_gh);
		gfs2_glock_dq_uninit(&sdp->sd_jinode_gh);
		gfs2_glock_dq_uninit(&sdp->sd_ir_gh);
		gfs2_glock_dq_uninit(&sdp->sd_sc_gh);
		gfs2_glock_dq_uninit(&sdp->sd_qc_gh);
		iput(sdp->sd_ir_inode);
		iput(sdp->sd_sc_inode);
		iput(sdp->sd_qc_inode);
	}

	gfs2_glock_dq_uninit(&sdp->sd_live_gh);
	gfs2_clear_rgrpd(sdp);
	gfs2_jindex_free(sdp);
	/*  Take apart glock structures and buffer lists  */
	gfs2_gl_hash_clear(sdp);
	/*  Unmount the locking protocol  */
	gfs2_lm_unmount(sdp);

	/*  At this point, we're through participating in the lockspace  */
	gfs2_sys_fs_del(sdp);
	kfree(sdp);
}

/**
 * gfs2_write_super
 * @sb: the superblock
 *
 */

static void gfs2_write_super(struct super_block *sb)
{
	sb->s_dirt = 0;
}

/**
 * gfs2_sync_fs - sync the filesystem
 * @sb: the superblock
 *
 * Flushes the log to disk.
 */

static int gfs2_sync_fs(struct super_block *sb, int wait)
{
	sb->s_dirt = 0;
	if (wait && sb->s_fs_info)
		gfs2_log_flush(sb->s_fs_info, NULL);
	return 0;
}

/**
 * gfs2_write_super_lockfs - prevent further writes to the filesystem
 * @sb: the VFS structure for the filesystem
 *
 */

static void gfs2_write_super_lockfs(struct super_block *sb)
{
	struct gfs2_sbd *sdp = sb->s_fs_info;
	int error;

	if (test_bit(SDF_SHUTDOWN, &sdp->sd_flags))
		return;

	for (;;) {
		error = gfs2_freeze_fs(sdp);
		if (!error)
			break;

		switch (error) {
		case -EBUSY:
			fs_err(sdp, "waiting for recovery before freeze\n");
			break;

		default:
			fs_err(sdp, "error freezing FS: %d\n", error);
			break;
		}

		fs_err(sdp, "retrying...\n");
		msleep(1000);
	}
}

/**
 * gfs2_unlockfs - reallow writes to the filesystem
 * @sb: the VFS structure for the filesystem
 *
 */

static void gfs2_unlockfs(struct super_block *sb)
{
	gfs2_unfreeze_fs(sb->s_fs_info);
}

/**
 * gfs2_statfs - Gather and return stats about the filesystem
 * @sb: The superblock
 * @statfsbuf: The buffer
 *
 * Returns: 0 on success or error code
 */

static int gfs2_statfs(struct dentry *dentry, struct kstatfs *buf)
{
	struct super_block *sb = dentry->d_inode->i_sb;
	struct gfs2_sbd *sdp = sb->s_fs_info;
	struct gfs2_statfs_change_host sc;
	int error;

	if (gfs2_tune_get(sdp, gt_statfs_slow))
		error = gfs2_statfs_slow(sdp, &sc);
	else
		error = gfs2_statfs_i(sdp, &sc);

	if (error)
		return error;

	buf->f_type = GFS2_MAGIC;
	buf->f_bsize = sdp->sd_sb.sb_bsize;
	buf->f_blocks = sc.sc_total;
	buf->f_bfree = sc.sc_free;
	buf->f_bavail = sc.sc_free;
	buf->f_files = sc.sc_dinodes + sc.sc_free;
	buf->f_ffree = sc.sc_free;
	buf->f_namelen = GFS2_FNAMESIZE;

	return 0;
}

/**
 * gfs2_remount_fs - called when the FS is remounted
 * @sb:  the filesystem
 * @flags:  the remount flags
 * @data:  extra data passed in (not used right now)
 *
 * Returns: errno
 */

static int gfs2_remount_fs(struct super_block *sb, int *flags, char *data)
{
	struct gfs2_sbd *sdp = sb->s_fs_info;
	int error;

	error = gfs2_mount_args(sdp, data, 1);
	if (error)
		return error;

	if (sdp->sd_args.ar_spectator)
		*flags |= MS_RDONLY;
	else {
		if (*flags & MS_RDONLY) {
			if (!(sb->s_flags & MS_RDONLY))
				error = gfs2_make_fs_ro(sdp);
		} else if (!(*flags & MS_RDONLY) &&
			   (sb->s_flags & MS_RDONLY)) {
			error = gfs2_make_fs_rw(sdp);
		}
	}

	return error;
}

/**
 * gfs2_drop_inode - Drop an inode (test for remote unlink)
 * @inode: The inode to drop
 *
 * If we've received a callback on an iopen lock then its because a
 * remote node tried to deallocate the inode but failed due to this node
 * still having the inode open. Here we mark the link count zero
 * since we know that it must have reached zero if the GLF_DEMOTE flag
 * is set on the iopen glock. If we didn't do a disk read since the
 * remote node removed the final link then we might otherwise miss
 * this event. This check ensures that this node will deallocate the
 * inode's blocks, or alternatively pass the baton on to another
 * node for later deallocation.
 */

static void gfs2_drop_inode(struct inode *inode)
{
	struct gfs2_inode *ip = GFS2_I(inode);

	if (test_bit(GIF_USER, &ip->i_flags) && inode->i_nlink) {
		struct gfs2_glock *gl = ip->i_iopen_gh.gh_gl;
		if (gl && test_bit(GLF_DEMOTE, &gl->gl_flags))
			clear_nlink(inode);
	}
	generic_drop_inode(inode);
}

/**
 * gfs2_clear_inode - Deallocate an inode when VFS is done with it
 * @inode: The VFS inode
 *
 */

static void gfs2_clear_inode(struct inode *inode)
{
	struct gfs2_inode *ip = GFS2_I(inode);

	/* This tells us its a "real" inode and not one which only
	 * serves to contain an address space (see rgrp.c, meta_io.c)
	 * which therefore doesn't have its own glocks.
	 */
	if (test_bit(GIF_USER, &ip->i_flags)) {
		ip->i_gl->gl_object = NULL;
		gfs2_glock_put(ip->i_gl);
		ip->i_gl = NULL;
		if (ip->i_iopen_gh.gh_gl) {
			ip->i_iopen_gh.gh_gl->gl_object = NULL;
			gfs2_glock_dq_uninit(&ip->i_iopen_gh);
		}
	}
}

static int is_ancestor(const struct dentry *d1, const struct dentry *d2)
{
	do {
		if (d1 == d2)
			return 1;
		d1 = d1->d_parent;
	} while (!IS_ROOT(d1));
	return 0;
}

/**
 * gfs2_show_options - Show mount options for /proc/mounts
 * @s: seq_file structure
 * @mnt: vfsmount
 *
 * Returns: 0 on success or error code
 */

static int gfs2_show_options(struct seq_file *s, struct vfsmount *mnt)
{
	struct gfs2_sbd *sdp = mnt->mnt_sb->s_fs_info;
	struct gfs2_args *args = &sdp->sd_args;

	if (is_ancestor(mnt->mnt_root, sdp->sd_master_dir))
		seq_printf(s, ",meta");
	if (args->ar_lockproto[0])
		seq_printf(s, ",lockproto=%s", args->ar_lockproto);
	if (args->ar_locktable[0])
		seq_printf(s, ",locktable=%s", args->ar_locktable);
	if (args->ar_hostdata[0])
		seq_printf(s, ",hostdata=%s", args->ar_hostdata);
	if (args->ar_spectator)
		seq_printf(s, ",spectator");
	if (args->ar_ignore_local_fs)
		seq_printf(s, ",ignore_local_fs");
	if (args->ar_localflocks)
		seq_printf(s, ",localflocks");
	if (args->ar_localcaching)
		seq_printf(s, ",localcaching");
	if (args->ar_debug)
		seq_printf(s, ",debug");
	if (args->ar_upgrade)
		seq_printf(s, ",upgrade");
	if (args->ar_posix_acl)
		seq_printf(s, ",acl");
	if (args->ar_quota != GFS2_QUOTA_DEFAULT) {
		char *state;
		switch (args->ar_quota) {
		case GFS2_QUOTA_OFF:
			state = "off";
			break;
		case GFS2_QUOTA_ACCOUNT:
			state = "account";
			break;
		case GFS2_QUOTA_ON:
			state = "on";
			break;
		default:
			state = "unknown";
			break;
		}
		seq_printf(s, ",quota=%s", state);
	}
	if (args->ar_suiddir)
		seq_printf(s, ",suiddir");
	if (args->ar_data != GFS2_DATA_DEFAULT) {
		char *state;
		switch (args->ar_data) {
		case GFS2_DATA_WRITEBACK:
			state = "writeback";
			break;
		case GFS2_DATA_ORDERED:
			state = "ordered";
			break;
		default:
			state = "unknown";
			break;
		}
		seq_printf(s, ",data=%s", state);
	}

	return 0;
}

/*
 * We have to (at the moment) hold the inodes main lock to cover
 * the gap between unlocking the shared lock on the iopen lock and
 * taking the exclusive lock. I'd rather do a shared -> exclusive
 * conversion on the iopen lock, but we can change that later. This
 * is safe, just less efficient.
 */

static void gfs2_delete_inode(struct inode *inode)
{
	struct gfs2_sbd *sdp = inode->i_sb->s_fs_info;
	struct gfs2_inode *ip = GFS2_I(inode);
	struct gfs2_holder gh;
	int error;

	if (!test_bit(GIF_USER, &ip->i_flags))
		goto out;

	error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
	if (unlikely(error)) {
		gfs2_glock_dq_uninit(&ip->i_iopen_gh);
		goto out;
	}

	gfs2_glock_dq_wait(&ip->i_iopen_gh);
	gfs2_holder_reinit(LM_ST_EXCLUSIVE, LM_FLAG_TRY_1CB | GL_NOCACHE, &ip->i_iopen_gh);
	error = gfs2_glock_nq(&ip->i_iopen_gh);
	if (error)
		goto out_truncate;

	if (S_ISDIR(inode->i_mode) &&
	    (ip->i_diskflags & GFS2_DIF_EXHASH)) {
		error = gfs2_dir_exhash_dealloc(ip);
		if (error)
			goto out_unlock;
	}

	if (ip->i_eattr) {
		error = gfs2_ea_dealloc(ip);
		if (error)
			goto out_unlock;
	}

	if (!gfs2_is_stuffed(ip)) {
		error = gfs2_file_dealloc(ip);
		if (error)
			goto out_unlock;
	}

	error = gfs2_dinode_dealloc(ip);
	if (error)
		goto out_unlock;

out_truncate:
	error = gfs2_trans_begin(sdp, 0, sdp->sd_jdesc->jd_blocks);
	if (error)
		goto out_unlock;
	/* Needs to be done before glock release & also in a transaction */
	truncate_inode_pages(&inode->i_data, 0);
	gfs2_trans_end(sdp);

out_unlock:
	if (test_bit(HIF_HOLDER, &ip->i_iopen_gh.gh_iflags))
		gfs2_glock_dq(&ip->i_iopen_gh);
	gfs2_holder_uninit(&ip->i_iopen_gh);
	gfs2_glock_dq_uninit(&gh);
	if (error && error != GLR_TRYFAILED)
		fs_warn(sdp, "gfs2_delete_inode: %d\n", error);
out:
	truncate_inode_pages(&inode->i_data, 0);
	clear_inode(inode);
}

static struct inode *gfs2_alloc_inode(struct super_block *sb)
{
	struct gfs2_inode *ip;

	ip = kmem_cache_alloc(gfs2_inode_cachep, GFP_KERNEL);
	if (ip) {
		ip->i_flags = 0;
		ip->i_gl = NULL;
	}
	return &ip->i_inode;
}

static void gfs2_destroy_inode(struct inode *inode)
{
	kmem_cache_free(gfs2_inode_cachep, inode);
}

const struct super_operations gfs2_super_ops = {
	.alloc_inode		= gfs2_alloc_inode,
	.destroy_inode		= gfs2_destroy_inode,
	.write_inode		= gfs2_write_inode,
	.delete_inode		= gfs2_delete_inode,
	.put_super		= gfs2_put_super,
	.write_super		= gfs2_write_super,
	.sync_fs		= gfs2_sync_fs,
	.write_super_lockfs 	= gfs2_write_super_lockfs,
	.unlockfs		= gfs2_unlockfs,
	.statfs			= gfs2_statfs,
	.remount_fs		= gfs2_remount_fs,
	.clear_inode		= gfs2_clear_inode,
	.drop_inode		= gfs2_drop_inode,
	.show_options		= gfs2_show_options,
};
Commit	Line	Data
b3b94faa DT	1	/*
b3b94faa DT	2	* Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved.
091806ed	3	* Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved.
b3b94faa DT	4	*
	5	* This copyrighted material is made available to anyone wishing to use,
	6	* modify, copy, or redistribute it subject to the terms and conditions
e9fc2aa0	7	* of the GNU General Public License version 2.
b3b94faa DT	8	*/
	9
	10	#include <linux/sched.h>
	11	#include <linux/slab.h>
	12	#include <linux/spinlock.h>
	13	#include <linux/completion.h>
	14	#include <linux/buffer_head.h>
b3b94faa DT	15	#include <linux/statfs.h>
	16	#include <linux/seq_file.h>
	17	#include <linux/mount.h>
	18	#include <linux/kthread.h>
	19	#include <linux/delay.h>
5c676f6d	20	#include <linux/gfs2_ondisk.h>
feaa7bba	21	#include <linux/crc32.h>
7d308590	22	#include <linux/lm_interface.h>
719ee344	23	#include <linux/time.h>
b3b94faa DT	24
b3b94faa DT	25	#include "gfs2.h"
5c676f6d	26	#include "incore.h"
b3b94faa DT	27	#include "glock.h"
b3b94faa DT	28	#include "inode.h"
b3b94faa DT	29	#include "log.h"
b3b94faa DT	30	#include "mount.h"
b3b94faa DT	31	#include "quota.h"
	32	#include "recovery.h"
	33	#include "rgrp.h"
	34	#include "super.h"
	35	#include "sys.h"
5c676f6d	36	#include "util.h"
feaa7bba SW	37	#include "trans.h"
	38	#include "dir.h"
	39	#include "eattr.h"
	40	#include "bmap.h"
719ee344	41	#include "meta_io.h"
b3b94faa DT	42
	43	/**
	44	* gfs2_write_inode - Make sure the inode is stable on the disk
	45	* @inode: The inode
	46	* @sync: synchronous write flag
	47	*
	48	* Returns: errno
	49	*/
	50
	51	static int gfs2_write_inode(struct inode *inode, int sync)
	52	{
feaa7bba	53	struct gfs2_inode *ip = GFS2_I(inode);
719ee344 SW	54	struct gfs2_sbd *sdp = GFS2_SB(inode);
	55	struct gfs2_holder gh;
	56	struct buffer_head *bh;
	57	struct timespec atime;
	58	struct gfs2_dinode *di;
	59	int ret = 0;
	60
	61	/* Check this is a "normal" inode, etc */
	62	if (!test_bit(GIF_USER, &ip->i_flags) \|\|
	63	(current->flags & PF_MEMALLOC))
	64	return 0;
	65	ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
	66	if (ret)
	67	goto do_flush;
	68	ret = gfs2_trans_begin(sdp, RES_DINODE, 0);
	69	if (ret)
	70	goto do_unlock;
	71	ret = gfs2_meta_inode_buffer(ip, &bh);
	72	if (ret == 0) {
	73	di = (struct gfs2_dinode *)bh->b_data;
	74	atime.tv_sec = be64_to_cpu(di->di_atime);
	75	atime.tv_nsec = be32_to_cpu(di->di_atime_nsec);
	76	if (timespec_compare(&inode->i_atime, &atime) > 0) {
	77	gfs2_trans_add_bh(ip->i_gl, bh, 1);
	78	gfs2_dinode_out(ip, bh->b_data);
	79	}
	80	brelse(bh);
feaa7bba	81	}
719ee344 SW	82	gfs2_trans_end(sdp);
	83	do_unlock:
	84	gfs2_glock_dq_uninit(&gh);
	85	do_flush:
	86	if (sync != 0)
	87	gfs2_log_flush(GFS2_SB(inode), ip->i_gl);
	88	return ret;
b3b94faa DT	89	}
b3b94faa DT	90
9b8df98f SW	91	/**
	92	* gfs2_make_fs_ro - Turn a Read-Write FS into a Read-Only one
	93	* @sdp: the filesystem
	94	*
	95	* Returns: errno
	96	*/
	97
	98	static int gfs2_make_fs_ro(struct gfs2_sbd *sdp)
	99	{
	100	struct gfs2_holder t_gh;
	101	int error;
	102
	103	gfs2_quota_sync(sdp);
	104	gfs2_statfs_sync(sdp);
	105
	106	error = gfs2_glock_nq_init(sdp->sd_trans_gl, LM_ST_SHARED, GL_NOCACHE,
	107	&t_gh);
	108	if (error && !test_bit(SDF_SHUTDOWN, &sdp->sd_flags))
	109	return error;
	110
	111	gfs2_meta_syncfs(sdp);
	112	gfs2_log_shutdown(sdp);
	113
	114	clear_bit(SDF_JOURNAL_LIVE, &sdp->sd_flags);
	115
	116	if (t_gh.gh_gl)
	117	gfs2_glock_dq_uninit(&t_gh);
	118
	119	gfs2_quota_cleanup(sdp);
	120
	121	return error;
	122	}
	123
b3b94faa DT	124	/**
	125	* gfs2_put_super - Unmount the filesystem
	126	* @sb: The VFS superblock
	127	*
	128	*/
	129
	130	static void gfs2_put_super(struct super_block *sb)
	131	{
5c676f6d	132	struct gfs2_sbd *sdp = sb->s_fs_info;
b3b94faa DT	133	int error;
b3b94faa DT	134
b3b94faa DT	135	/* Unfreeze the filesystem, if we need to */
b3b94faa DT	136
f55ab26a	137	mutex_lock(&sdp->sd_freeze_lock);
b3b94faa DT	138	if (sdp->sd_freeze_count)
b3b94faa DT	139	gfs2_glock_dq_uninit(&sdp->sd_freeze_gh);
f55ab26a	140	mutex_unlock(&sdp->sd_freeze_lock);
b3b94faa	141
b3b94faa DT	142	kthread_stop(sdp->sd_quotad_process);
	143	kthread_stop(sdp->sd_logd_process);
	144	kthread_stop(sdp->sd_recoverd_process);
b3b94faa DT	145
	146	if (!(sb->s_flags & MS_RDONLY)) {
	147	error = gfs2_make_fs_ro(sdp);
	148	if (error)
	149	gfs2_io_error(sdp);
	150	}
b3b94faa DT	151	/* At this point, we're through modifying the disk */
	152
	153	/* Release stuff */
	154
f42faf4f SW	155	iput(sdp->sd_jindex);
	156	iput(sdp->sd_inum_inode);
	157	iput(sdp->sd_statfs_inode);
	158	iput(sdp->sd_rindex);
	159	iput(sdp->sd_quota_inode);
b3b94faa DT	160
	161	gfs2_glock_put(sdp->sd_rename_gl);
	162	gfs2_glock_put(sdp->sd_trans_gl);
	163
	164	if (!sdp->sd_args.ar_spectator) {
	165	gfs2_glock_dq_uninit(&sdp->sd_journal_gh);
	166	gfs2_glock_dq_uninit(&sdp->sd_jinode_gh);
	167	gfs2_glock_dq_uninit(&sdp->sd_ir_gh);
	168	gfs2_glock_dq_uninit(&sdp->sd_sc_gh);
b3b94faa	169	gfs2_glock_dq_uninit(&sdp->sd_qc_gh);
f42faf4f SW	170	iput(sdp->sd_ir_inode);
f42faf4f SW	171	iput(sdp->sd_sc_inode);
f42faf4f	172	iput(sdp->sd_qc_inode);
b3b94faa DT	173	}
	174
	175	gfs2_glock_dq_uninit(&sdp->sd_live_gh);
b3b94faa DT	176	gfs2_clear_rgrpd(sdp);
b3b94faa DT	177	gfs2_jindex_free(sdp);
b3b94faa	178	/* Take apart glock structures and buffer lists */
1bdad606	179	gfs2_gl_hash_clear(sdp);
b3b94faa DT	180	/* Unmount the locking protocol */
	181	gfs2_lm_unmount(sdp);
	182
	183	/* At this point, we're through participating in the lockspace */
b3b94faa	184	gfs2_sys_fs_del(sdp);
ff6af411	185	kfree(sdp);
b3b94faa DT	186	}
	187
	188	/**
4a221953 SW	189	* gfs2_write_super
4a221953 SW	190	* @sb: the superblock
b3b94faa	191	*
b3b94faa DT	192	*/
	193
	194	static void gfs2_write_super(struct super_block *sb)
	195	{
4a221953 SW	196	sb->s_dirt = 0;
	197	}
	198
	199	/**
	200	* gfs2_sync_fs - sync the filesystem
	201	* @sb: the superblock
	202	*
	203	* Flushes the log to disk.
	204	*/
9b8df98f	205
4a221953 SW	206	static int gfs2_sync_fs(struct super_block *sb, int wait)
	207	{
	208	sb->s_dirt = 0;
9171f5a9	209	if (wait && sb->s_fs_info)
b004157a	210	gfs2_log_flush(sb->s_fs_info, NULL);
4a221953	211	return 0;
b3b94faa DT	212	}
	213
	214	/**
	215	* gfs2_write_super_lockfs - prevent further writes to the filesystem
	216	* @sb: the VFS structure for the filesystem
	217	*
	218	*/
	219
	220	static void gfs2_write_super_lockfs(struct super_block *sb)
	221	{
5c676f6d	222	struct gfs2_sbd *sdp = sb->s_fs_info;
b3b94faa DT	223	int error;
b3b94faa DT	224
c3780511 DT	225	if (test_bit(SDF_SHUTDOWN, &sdp->sd_flags))
	226	return;
	227
b3b94faa DT	228	for (;;) {
	229	error = gfs2_freeze_fs(sdp);
	230	if (!error)
	231	break;
	232
	233	switch (error) {
	234	case -EBUSY:
	235	fs_err(sdp, "waiting for recovery before freeze\n");
	236	break;
	237
	238	default:
	239	fs_err(sdp, "error freezing FS: %d\n", error);
	240	break;
	241	}
	242
	243	fs_err(sdp, "retrying...\n");
	244	msleep(1000);
	245	}
	246	}
	247
	248	/**
	249	* gfs2_unlockfs - reallow writes to the filesystem
	250	* @sb: the VFS structure for the filesystem
	251	*
	252	*/
	253
	254	static void gfs2_unlockfs(struct super_block *sb)
	255	{
2bdbc5d7	256	gfs2_unfreeze_fs(sb->s_fs_info);
b3b94faa DT	257	}
	258
	259	/**
	260	* gfs2_statfs - Gather and return stats about the filesystem
	261	* @sb: The superblock
	262	* @statfsbuf: The buffer
	263	*
	264	* Returns: 0 on success or error code
	265	*/
	266
0c0834a3	267	static int gfs2_statfs(struct dentry dentry, struct kstatfs buf)
b3b94faa	268	{
0c0834a3	269	struct super_block *sb = dentry->d_inode->i_sb;
5c676f6d	270	struct gfs2_sbd *sdp = sb->s_fs_info;
bd209cc0	271	struct gfs2_statfs_change_host sc;
b3b94faa DT	272	int error;
b3b94faa DT	273
b3b94faa DT	274	if (gfs2_tune_get(sdp, gt_statfs_slow))
	275	error = gfs2_statfs_slow(sdp, &sc);
	276	else
	277	error = gfs2_statfs_i(sdp, &sc);
	278
	279	if (error)
	280	return error;
	281
b3b94faa DT	282	buf->f_type = GFS2_MAGIC;
	283	buf->f_bsize = sdp->sd_sb.sb_bsize;
	284	buf->f_blocks = sc.sc_total;
	285	buf->f_bfree = sc.sc_free;
	286	buf->f_bavail = sc.sc_free;
	287	buf->f_files = sc.sc_dinodes + sc.sc_free;
	288	buf->f_ffree = sc.sc_free;
	289	buf->f_namelen = GFS2_FNAMESIZE;
	290
	291	return 0;
	292	}
	293
	294	/**
	295	* gfs2_remount_fs - called when the FS is remounted
	296	* @sb: the filesystem
	297	* @flags: the remount flags
	298	* @data: extra data passed in (not used right now)
	299	*
	300	* Returns: errno
	301	*/
	302
	303	static int gfs2_remount_fs(struct super_block sb, int flags, char *data)
	304	{
5c676f6d	305	struct gfs2_sbd *sdp = sb->s_fs_info;
b3b94faa DT	306	int error;
b3b94faa DT	307
b3b94faa DT	308	error = gfs2_mount_args(sdp, data, 1);
	309	if (error)
	310	return error;
	311
	312	if (sdp->sd_args.ar_spectator)
	313	*flags \|= MS_RDONLY;
	314	else {
	315	if (*flags & MS_RDONLY) {
	316	if (!(sb->s_flags & MS_RDONLY))
	317	error = gfs2_make_fs_ro(sdp);
	318	} else if (!(*flags & MS_RDONLY) &&
	319	(sb->s_flags & MS_RDONLY)) {
	320	error = gfs2_make_fs_rw(sdp);
	321	}
	322	}
	323
b3b94faa DT	324	return error;
	325	}
	326
3b8249f6 SW	327	/**
	328	* gfs2_drop_inode - Drop an inode (test for remote unlink)
	329	* @inode: The inode to drop
	330	*
	331	* If we've received a callback on an iopen lock then its because a
	332	* remote node tried to deallocate the inode but failed due to this node
	333	* still having the inode open. Here we mark the link count zero
	334	* since we know that it must have reached zero if the GLF_DEMOTE flag
	335	* is set on the iopen glock. If we didn't do a disk read since the
	336	* remote node removed the final link then we might otherwise miss
	337	* this event. This check ensures that this node will deallocate the
	338	* inode's blocks, or alternatively pass the baton on to another
	339	* node for later deallocation.
	340	*/
9b8df98f	341
3b8249f6 SW	342	static void gfs2_drop_inode(struct inode *inode)
3b8249f6 SW	343	{
091806ed BP	344	struct gfs2_inode *ip = GFS2_I(inode);
	345
	346	if (test_bit(GIF_USER, &ip->i_flags) && inode->i_nlink) {
3b8249f6 SW	347	struct gfs2_glock *gl = ip->i_iopen_gh.gh_gl;
	348	if (gl && test_bit(GLF_DEMOTE, &gl->gl_flags))
	349	clear_nlink(inode);
	350	}
	351	generic_drop_inode(inode);
	352	}
	353
b3b94faa DT	354	/**
	355	* gfs2_clear_inode - Deallocate an inode when VFS is done with it
	356	* @inode: The VFS inode
	357	*
	358	*/
	359
	360	static void gfs2_clear_inode(struct inode *inode)
	361	{
091806ed BP	362	struct gfs2_inode *ip = GFS2_I(inode);
091806ed BP	363
feaa7bba SW	364	/* This tells us its a "real" inode and not one which only
	365	* serves to contain an address space (see rgrp.c, meta_io.c)
	366	* which therefore doesn't have its own glocks.
	367	*/
091806ed	368	if (test_bit(GIF_USER, &ip->i_flags)) {
feaa7bba	369	ip->i_gl->gl_object = NULL;
feaa7bba SW	370	gfs2_glock_put(ip->i_gl);
feaa7bba SW	371	ip->i_gl = NULL;
d93cfa98 AD	372	if (ip->i_iopen_gh.gh_gl) {
d93cfa98 AD	373	ip->i_iopen_gh.gh_gl->gl_object = NULL;
feaa7bba	374	gfs2_glock_dq_uninit(&ip->i_iopen_gh);
d93cfa98	375	}
b3b94faa DT	376	}
	377	}
	378
9b8df98f SW	379	static int is_ancestor(const struct dentry d1, const struct dentry d2)
	380	{
	381	do {
	382	if (d1 == d2)
	383	return 1;
	384	d1 = d1->d_parent;
	385	} while (!IS_ROOT(d1));
	386	return 0;
	387	}
	388
b3b94faa DT	389	/**
	390	* gfs2_show_options - Show mount options for /proc/mounts
	391	* @s: seq_file structure
	392	* @mnt: vfsmount
	393	*
	394	* Returns: 0 on success or error code
	395	*/
	396
	397	static int gfs2_show_options(struct seq_file s, struct vfsmount mnt)
	398	{
5c676f6d	399	struct gfs2_sbd *sdp = mnt->mnt_sb->s_fs_info;
b3b94faa DT	400	struct gfs2_args *args = &sdp->sd_args;
b3b94faa DT	401
9b8df98f SW	402	if (is_ancestor(mnt->mnt_root, sdp->sd_master_dir))
9b8df98f SW	403	seq_printf(s, ",meta");
b3b94faa DT	404	if (args->ar_lockproto[0])
	405	seq_printf(s, ",lockproto=%s", args->ar_lockproto);
	406	if (args->ar_locktable[0])
	407	seq_printf(s, ",locktable=%s", args->ar_locktable);
	408	if (args->ar_hostdata[0])
	409	seq_printf(s, ",hostdata=%s", args->ar_hostdata);
	410	if (args->ar_spectator)
	411	seq_printf(s, ",spectator");
	412	if (args->ar_ignore_local_fs)
	413	seq_printf(s, ",ignore_local_fs");
	414	if (args->ar_localflocks)
	415	seq_printf(s, ",localflocks");
	416	if (args->ar_localcaching)
	417	seq_printf(s, ",localcaching");
	418	if (args->ar_debug)
	419	seq_printf(s, ",debug");
	420	if (args->ar_upgrade)
	421	seq_printf(s, ",upgrade");
b3b94faa DT	422	if (args->ar_posix_acl)
	423	seq_printf(s, ",acl");
	424	if (args->ar_quota != GFS2_QUOTA_DEFAULT) {
	425	char *state;
	426	switch (args->ar_quota) {
	427	case GFS2_QUOTA_OFF:
	428	state = "off";
	429	break;
	430	case GFS2_QUOTA_ACCOUNT:
	431	state = "account";
	432	break;
	433	case GFS2_QUOTA_ON:
	434	state = "on";
	435	break;
	436	default:
	437	state = "unknown";
	438	break;
	439	}
	440	seq_printf(s, ",quota=%s", state);
	441	}
	442	if (args->ar_suiddir)
	443	seq_printf(s, ",suiddir");
	444	if (args->ar_data != GFS2_DATA_DEFAULT) {
	445	char *state;
	446	switch (args->ar_data) {
	447	case GFS2_DATA_WRITEBACK:
	448	state = "writeback";
	449	break;
	450	case GFS2_DATA_ORDERED:
	451	state = "ordered";
	452	break;
	453	default:
	454	state = "unknown";
	455	break;
	456	}
	457	seq_printf(s, ",data=%s", state);
	458	}
	459
	460	return 0;
	461	}
	462
907b9bce	463	/*
feaa7bba SW	464	* We have to (at the moment) hold the inodes main lock to cover
	465	* the gap between unlocking the shared lock on the iopen lock and
	466	* taking the exclusive lock. I'd rather do a shared -> exclusive
	467	* conversion on the iopen lock, but we can change that later. This
	468	* is safe, just less efficient.
	469	*/
9b8df98f	470
feaa7bba SW	471	static void gfs2_delete_inode(struct inode *inode)
	472	{
	473	struct gfs2_sbd *sdp = inode->i_sb->s_fs_info;
	474	struct gfs2_inode *ip = GFS2_I(inode);
	475	struct gfs2_holder gh;
	476	int error;
	477
091806ed	478	if (!test_bit(GIF_USER, &ip->i_flags))
feaa7bba SW	479	goto out;
feaa7bba SW	480
d93cfa98	481	error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
feaa7bba SW	482	if (unlikely(error)) {
	483	gfs2_glock_dq_uninit(&ip->i_iopen_gh);
	484	goto out;
	485	}
	486
d93cfa98	487	gfs2_glock_dq_wait(&ip->i_iopen_gh);
feaa7bba SW	488	gfs2_holder_reinit(LM_ST_EXCLUSIVE, LM_FLAG_TRY_1CB \| GL_NOCACHE, &ip->i_iopen_gh);
	489	error = gfs2_glock_nq(&ip->i_iopen_gh);
	490	if (error)
1bb7322f	491	goto out_truncate;
feaa7bba	492
b60623c2	493	if (S_ISDIR(inode->i_mode) &&
383f01fb	494	(ip->i_diskflags & GFS2_DIF_EXHASH)) {
feaa7bba SW	495	error = gfs2_dir_exhash_dealloc(ip);
	496	if (error)
	497	goto out_unlock;
	498	}
	499
3767ac21	500	if (ip->i_eattr) {
feaa7bba SW	501	error = gfs2_ea_dealloc(ip);
	502	if (error)
	503	goto out_unlock;
	504	}
	505
	506	if (!gfs2_is_stuffed(ip)) {
	507	error = gfs2_file_dealloc(ip);
	508	if (error)
	509	goto out_unlock;
	510	}
	511
	512	error = gfs2_dinode_dealloc(ip);
16615be1 SW	513	if (error)
	514	goto out_unlock;
	515
1bb7322f	516	out_truncate:
16615be1 SW	517	error = gfs2_trans_begin(sdp, 0, sdp->sd_jdesc->jd_blocks);
	518	if (error)
	519	goto out_unlock;
	520	/* Needs to be done before glock release & also in a transaction */
49686f71	521	truncate_inode_pages(&inode->i_data, 0);
16615be1	522	gfs2_trans_end(sdp);
feaa7bba SW	523
feaa7bba SW	524	out_unlock:
1bb7322f SW	525	if (test_bit(HIF_HOLDER, &ip->i_iopen_gh.gh_iflags))
1bb7322f SW	526	gfs2_glock_dq(&ip->i_iopen_gh);
feaa7bba SW	527	gfs2_holder_uninit(&ip->i_iopen_gh);
feaa7bba SW	528	gfs2_glock_dq_uninit(&gh);
3b8249f6	529	if (error && error != GLR_TRYFAILED)
feaa7bba SW	530	fs_warn(sdp, "gfs2_delete_inode: %d\n", error);
	531	out:
	532	truncate_inode_pages(&inode->i_data, 0);
	533	clear_inode(inode);
	534	}
	535
320dd101 SW	536	static struct inode gfs2_alloc_inode(struct super_block sb)
320dd101 SW	537	{
320dd101 SW	538	struct gfs2_inode *ip;
	539
	540	ip = kmem_cache_alloc(gfs2_inode_cachep, GFP_KERNEL);
	541	if (ip) {
	542	ip->i_flags = 0;
	543	ip->i_gl = NULL;
320dd101 SW	544	}
	545	return &ip->i_inode;
	546	}
	547
	548	static void gfs2_destroy_inode(struct inode *inode)
	549	{
	550	kmem_cache_free(gfs2_inode_cachep, inode);
	551	}
	552
ee9b6d61	553	const struct super_operations gfs2_super_ops = {
4a221953 SW	554	.alloc_inode = gfs2_alloc_inode,
	555	.destroy_inode = gfs2_destroy_inode,
	556	.write_inode = gfs2_write_inode,
	557	.delete_inode = gfs2_delete_inode,
	558	.put_super = gfs2_put_super,
	559	.write_super = gfs2_write_super,
	560	.sync_fs = gfs2_sync_fs,
	561	.write_super_lockfs = gfs2_write_super_lockfs,
	562	.unlockfs = gfs2_unlockfs,
	563	.statfs = gfs2_statfs,
	564	.remount_fs = gfs2_remount_fs,
	565	.clear_inode = gfs2_clear_inode,
3b8249f6	566	.drop_inode = gfs2_drop_inode,
4a221953	567	.show_options = gfs2_show_options,
b3b94faa DT	568	};
b3b94faa DT	569