[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 "ops_super.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);
	while (sdp->sd_glockd_num--)
		kthread_stop(sdp->sd_glockd_process[sdp->sd_glockd_num]);

	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_schedule_for_reclaim(ip->i_gl);
		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_num_glockd != GFS2_GLOCKD_DEFAULT)
		seq_printf(s, ",num_glockd=%u", args->ar_num_glockd);
	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_uninit;

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

	if (ip->i_di.di_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;

	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:
	gfs2_glock_dq(&ip->i_iopen_gh);
out_uninit:
	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"
	30	#include "mount.h"
	31	#include "ops_super.h"
b3b94faa DT	32	#include "quota.h"
	33	#include "recovery.h"
	34	#include "rgrp.h"
	35	#include "super.h"
	36	#include "sys.h"
5c676f6d	37	#include "util.h"
feaa7bba SW	38	#include "trans.h"
	39	#include "dir.h"
	40	#include "eattr.h"
	41	#include "bmap.h"
719ee344	42	#include "meta_io.h"
b3b94faa DT	43
	44	/**
	45	* gfs2_write_inode - Make sure the inode is stable on the disk
	46	* @inode: The inode
	47	* @sync: synchronous write flag
	48	*
	49	* Returns: errno
	50	*/
	51
	52	static int gfs2_write_inode(struct inode *inode, int sync)
	53	{
feaa7bba	54	struct gfs2_inode *ip = GFS2_I(inode);
719ee344 SW	55	struct gfs2_sbd *sdp = GFS2_SB(inode);
	56	struct gfs2_holder gh;
	57	struct buffer_head *bh;
	58	struct timespec atime;
	59	struct gfs2_dinode *di;
	60	int ret = 0;
	61
	62	/* Check this is a "normal" inode, etc */
	63	if (!test_bit(GIF_USER, &ip->i_flags) \|\|
	64	(current->flags & PF_MEMALLOC))
	65	return 0;
	66	ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
	67	if (ret)
	68	goto do_flush;
	69	ret = gfs2_trans_begin(sdp, RES_DINODE, 0);
	70	if (ret)
	71	goto do_unlock;
	72	ret = gfs2_meta_inode_buffer(ip, &bh);
	73	if (ret == 0) {
	74	di = (struct gfs2_dinode *)bh->b_data;
	75	atime.tv_sec = be64_to_cpu(di->di_atime);
	76	atime.tv_nsec = be32_to_cpu(di->di_atime_nsec);
	77	if (timespec_compare(&inode->i_atime, &atime) > 0) {
	78	gfs2_trans_add_bh(ip->i_gl, bh, 1);
	79	gfs2_dinode_out(ip, bh->b_data);
	80	}
	81	brelse(bh);
feaa7bba	82	}
719ee344 SW	83	gfs2_trans_end(sdp);
	84	do_unlock:
	85	gfs2_glock_dq_uninit(&gh);
	86	do_flush:
	87	if (sync != 0)
	88	gfs2_log_flush(GFS2_SB(inode), ip->i_gl);
	89	return ret;
b3b94faa DT	90	}
b3b94faa DT	91
9b8df98f SW	92	/**
	93	* gfs2_make_fs_ro - Turn a Read-Write FS into a Read-Only one
	94	* @sdp: the filesystem
	95	*
	96	* Returns: errno
	97	*/
	98
	99	static int gfs2_make_fs_ro(struct gfs2_sbd *sdp)
	100	{
	101	struct gfs2_holder t_gh;
	102	int error;
	103
	104	gfs2_quota_sync(sdp);
	105	gfs2_statfs_sync(sdp);
	106
	107	error = gfs2_glock_nq_init(sdp->sd_trans_gl, LM_ST_SHARED, GL_NOCACHE,
	108	&t_gh);
	109	if (error && !test_bit(SDF_SHUTDOWN, &sdp->sd_flags))
	110	return error;
	111
	112	gfs2_meta_syncfs(sdp);
	113	gfs2_log_shutdown(sdp);
	114
	115	clear_bit(SDF_JOURNAL_LIVE, &sdp->sd_flags);
	116
	117	if (t_gh.gh_gl)
	118	gfs2_glock_dq_uninit(&t_gh);
	119
	120	gfs2_quota_cleanup(sdp);
	121
	122	return error;
	123	}
	124
b3b94faa DT	125	/**
	126	* gfs2_put_super - Unmount the filesystem
	127	* @sb: The VFS superblock
	128	*
	129	*/
	130
	131	static void gfs2_put_super(struct super_block *sb)
	132	{
5c676f6d	133	struct gfs2_sbd *sdp = sb->s_fs_info;
b3b94faa DT	134	int error;
b3b94faa DT	135
b3b94faa DT	136	/* Unfreeze the filesystem, if we need to */
b3b94faa DT	137
f55ab26a	138	mutex_lock(&sdp->sd_freeze_lock);
b3b94faa DT	139	if (sdp->sd_freeze_count)
b3b94faa DT	140	gfs2_glock_dq_uninit(&sdp->sd_freeze_gh);
f55ab26a	141	mutex_unlock(&sdp->sd_freeze_lock);
b3b94faa	142
b3b94faa DT	143	kthread_stop(sdp->sd_quotad_process);
	144	kthread_stop(sdp->sd_logd_process);
	145	kthread_stop(sdp->sd_recoverd_process);
	146	while (sdp->sd_glockd_num--)
	147	kthread_stop(sdp->sd_glockd_process[sdp->sd_glockd_num]);
b3b94faa DT	148
	149	if (!(sb->s_flags & MS_RDONLY)) {
	150	error = gfs2_make_fs_ro(sdp);
	151	if (error)
	152	gfs2_io_error(sdp);
	153	}
b3b94faa DT	154	/* At this point, we're through modifying the disk */
	155
	156	/* Release stuff */
	157
f42faf4f SW	158	iput(sdp->sd_jindex);
	159	iput(sdp->sd_inum_inode);
	160	iput(sdp->sd_statfs_inode);
	161	iput(sdp->sd_rindex);
	162	iput(sdp->sd_quota_inode);
b3b94faa DT	163
	164	gfs2_glock_put(sdp->sd_rename_gl);
	165	gfs2_glock_put(sdp->sd_trans_gl);
	166
	167	if (!sdp->sd_args.ar_spectator) {
	168	gfs2_glock_dq_uninit(&sdp->sd_journal_gh);
	169	gfs2_glock_dq_uninit(&sdp->sd_jinode_gh);
	170	gfs2_glock_dq_uninit(&sdp->sd_ir_gh);
	171	gfs2_glock_dq_uninit(&sdp->sd_sc_gh);
b3b94faa	172	gfs2_glock_dq_uninit(&sdp->sd_qc_gh);
f42faf4f SW	173	iput(sdp->sd_ir_inode);
f42faf4f SW	174	iput(sdp->sd_sc_inode);
f42faf4f	175	iput(sdp->sd_qc_inode);
b3b94faa DT	176	}
	177
	178	gfs2_glock_dq_uninit(&sdp->sd_live_gh);
b3b94faa DT	179	gfs2_clear_rgrpd(sdp);
b3b94faa DT	180	gfs2_jindex_free(sdp);
b3b94faa	181	/* Take apart glock structures and buffer lists */
1bdad606	182	gfs2_gl_hash_clear(sdp);
b3b94faa DT	183	/* Unmount the locking protocol */
	184	gfs2_lm_unmount(sdp);
	185
	186	/* At this point, we're through participating in the lockspace */
b3b94faa	187	gfs2_sys_fs_del(sdp);
ff6af411	188	kfree(sdp);
b3b94faa DT	189	}
	190
	191	/**
4a221953 SW	192	* gfs2_write_super
4a221953 SW	193	* @sb: the superblock
b3b94faa	194	*
b3b94faa DT	195	*/
	196
	197	static void gfs2_write_super(struct super_block *sb)
	198	{
4a221953 SW	199	sb->s_dirt = 0;
	200	}
	201
	202	/**
	203	* gfs2_sync_fs - sync the filesystem
	204	* @sb: the superblock
	205	*
	206	* Flushes the log to disk.
	207	*/
9b8df98f	208
4a221953 SW	209	static int gfs2_sync_fs(struct super_block *sb, int wait)
	210	{
	211	sb->s_dirt = 0;
9171f5a9	212	if (wait && sb->s_fs_info)
b004157a	213	gfs2_log_flush(sb->s_fs_info, NULL);
4a221953	214	return 0;
b3b94faa DT	215	}
	216
	217	/**
	218	* gfs2_write_super_lockfs - prevent further writes to the filesystem
	219	* @sb: the VFS structure for the filesystem
	220	*
	221	*/
	222
	223	static void gfs2_write_super_lockfs(struct super_block *sb)
	224	{
5c676f6d	225	struct gfs2_sbd *sdp = sb->s_fs_info;
b3b94faa DT	226	int error;
b3b94faa DT	227
c3780511 DT	228	if (test_bit(SDF_SHUTDOWN, &sdp->sd_flags))
	229	return;
	230
b3b94faa DT	231	for (;;) {
	232	error = gfs2_freeze_fs(sdp);
	233	if (!error)
	234	break;
	235
	236	switch (error) {
	237	case -EBUSY:
	238	fs_err(sdp, "waiting for recovery before freeze\n");
	239	break;
	240
	241	default:
	242	fs_err(sdp, "error freezing FS: %d\n", error);
	243	break;
	244	}
	245
	246	fs_err(sdp, "retrying...\n");
	247	msleep(1000);
	248	}
	249	}
	250
	251	/**
	252	* gfs2_unlockfs - reallow writes to the filesystem
	253	* @sb: the VFS structure for the filesystem
	254	*
	255	*/
	256
	257	static void gfs2_unlockfs(struct super_block *sb)
	258	{
2bdbc5d7	259	gfs2_unfreeze_fs(sb->s_fs_info);
b3b94faa DT	260	}
	261
	262	/**
	263	* gfs2_statfs - Gather and return stats about the filesystem
	264	* @sb: The superblock
	265	* @statfsbuf: The buffer
	266	*
	267	* Returns: 0 on success or error code
	268	*/
	269
0c0834a3	270	static int gfs2_statfs(struct dentry dentry, struct kstatfs buf)
b3b94faa	271	{
0c0834a3	272	struct super_block *sb = dentry->d_inode->i_sb;
5c676f6d	273	struct gfs2_sbd *sdp = sb->s_fs_info;
bd209cc0	274	struct gfs2_statfs_change_host sc;
b3b94faa DT	275	int error;
b3b94faa DT	276
b3b94faa DT	277	if (gfs2_tune_get(sdp, gt_statfs_slow))
	278	error = gfs2_statfs_slow(sdp, &sc);
	279	else
	280	error = gfs2_statfs_i(sdp, &sc);
	281
	282	if (error)
	283	return error;
	284
b3b94faa DT	285	buf->f_type = GFS2_MAGIC;
	286	buf->f_bsize = sdp->sd_sb.sb_bsize;
	287	buf->f_blocks = sc.sc_total;
	288	buf->f_bfree = sc.sc_free;
	289	buf->f_bavail = sc.sc_free;
	290	buf->f_files = sc.sc_dinodes + sc.sc_free;
	291	buf->f_ffree = sc.sc_free;
	292	buf->f_namelen = GFS2_FNAMESIZE;
	293
	294	return 0;
	295	}
	296
	297	/**
	298	* gfs2_remount_fs - called when the FS is remounted
	299	* @sb: the filesystem
	300	* @flags: the remount flags
	301	* @data: extra data passed in (not used right now)
	302	*
	303	* Returns: errno
	304	*/
	305
	306	static int gfs2_remount_fs(struct super_block sb, int flags, char *data)
	307	{
5c676f6d	308	struct gfs2_sbd *sdp = sb->s_fs_info;
b3b94faa DT	309	int error;
b3b94faa DT	310
b3b94faa DT	311	error = gfs2_mount_args(sdp, data, 1);
	312	if (error)
	313	return error;
	314
	315	if (sdp->sd_args.ar_spectator)
	316	*flags \|= MS_RDONLY;
	317	else {
	318	if (*flags & MS_RDONLY) {
	319	if (!(sb->s_flags & MS_RDONLY))
	320	error = gfs2_make_fs_ro(sdp);
	321	} else if (!(*flags & MS_RDONLY) &&
	322	(sb->s_flags & MS_RDONLY)) {
	323	error = gfs2_make_fs_rw(sdp);
	324	}
	325	}
	326
b3b94faa DT	327	return error;
	328	}
	329
3b8249f6 SW	330	/**
	331	* gfs2_drop_inode - Drop an inode (test for remote unlink)
	332	* @inode: The inode to drop
	333	*
	334	* If we've received a callback on an iopen lock then its because a
	335	* remote node tried to deallocate the inode but failed due to this node
	336	* still having the inode open. Here we mark the link count zero
	337	* since we know that it must have reached zero if the GLF_DEMOTE flag
	338	* is set on the iopen glock. If we didn't do a disk read since the
	339	* remote node removed the final link then we might otherwise miss
	340	* this event. This check ensures that this node will deallocate the
	341	* inode's blocks, or alternatively pass the baton on to another
	342	* node for later deallocation.
	343	*/
9b8df98f	344
3b8249f6 SW	345	static void gfs2_drop_inode(struct inode *inode)
3b8249f6 SW	346	{
091806ed BP	347	struct gfs2_inode *ip = GFS2_I(inode);
	348
	349	if (test_bit(GIF_USER, &ip->i_flags) && inode->i_nlink) {
3b8249f6 SW	350	struct gfs2_glock *gl = ip->i_iopen_gh.gh_gl;
	351	if (gl && test_bit(GLF_DEMOTE, &gl->gl_flags))
	352	clear_nlink(inode);
	353	}
	354	generic_drop_inode(inode);
	355	}
	356
b3b94faa DT	357	/**
	358	* gfs2_clear_inode - Deallocate an inode when VFS is done with it
	359	* @inode: The VFS inode
	360	*
	361	*/
	362
	363	static void gfs2_clear_inode(struct inode *inode)
	364	{
091806ed BP	365	struct gfs2_inode *ip = GFS2_I(inode);
091806ed BP	366
feaa7bba SW	367	/* This tells us its a "real" inode and not one which only
	368	* serves to contain an address space (see rgrp.c, meta_io.c)
	369	* which therefore doesn't have its own glocks.
	370	*/
091806ed	371	if (test_bit(GIF_USER, &ip->i_flags)) {
feaa7bba	372	ip->i_gl->gl_object = NULL;
b3b94faa	373	gfs2_glock_schedule_for_reclaim(ip->i_gl);
feaa7bba SW	374	gfs2_glock_put(ip->i_gl);
feaa7bba SW	375	ip->i_gl = NULL;
d93cfa98 AD	376	if (ip->i_iopen_gh.gh_gl) {
d93cfa98 AD	377	ip->i_iopen_gh.gh_gl->gl_object = NULL;
feaa7bba	378	gfs2_glock_dq_uninit(&ip->i_iopen_gh);
d93cfa98	379	}
b3b94faa DT	380	}
	381	}
	382
9b8df98f SW	383	static int is_ancestor(const struct dentry d1, const struct dentry d2)
	384	{
	385	do {
	386	if (d1 == d2)
	387	return 1;
	388	d1 = d1->d_parent;
	389	} while (!IS_ROOT(d1));
	390	return 0;
	391	}
	392
b3b94faa DT	393	/**
	394	* gfs2_show_options - Show mount options for /proc/mounts
	395	* @s: seq_file structure
	396	* @mnt: vfsmount
	397	*
	398	* Returns: 0 on success or error code
	399	*/
	400
	401	static int gfs2_show_options(struct seq_file s, struct vfsmount mnt)
	402	{
5c676f6d	403	struct gfs2_sbd *sdp = mnt->mnt_sb->s_fs_info;
b3b94faa DT	404	struct gfs2_args *args = &sdp->sd_args;
b3b94faa DT	405
9b8df98f SW	406	if (is_ancestor(mnt->mnt_root, sdp->sd_master_dir))
9b8df98f SW	407	seq_printf(s, ",meta");
b3b94faa DT	408	if (args->ar_lockproto[0])
	409	seq_printf(s, ",lockproto=%s", args->ar_lockproto);
	410	if (args->ar_locktable[0])
	411	seq_printf(s, ",locktable=%s", args->ar_locktable);
	412	if (args->ar_hostdata[0])
	413	seq_printf(s, ",hostdata=%s", args->ar_hostdata);
	414	if (args->ar_spectator)
	415	seq_printf(s, ",spectator");
	416	if (args->ar_ignore_local_fs)
	417	seq_printf(s, ",ignore_local_fs");
	418	if (args->ar_localflocks)
	419	seq_printf(s, ",localflocks");
	420	if (args->ar_localcaching)
	421	seq_printf(s, ",localcaching");
	422	if (args->ar_debug)
	423	seq_printf(s, ",debug");
	424	if (args->ar_upgrade)
	425	seq_printf(s, ",upgrade");
	426	if (args->ar_num_glockd != GFS2_GLOCKD_DEFAULT)
	427	seq_printf(s, ",num_glockd=%u", args->ar_num_glockd);
	428	if (args->ar_posix_acl)
	429	seq_printf(s, ",acl");
	430	if (args->ar_quota != GFS2_QUOTA_DEFAULT) {
	431	char *state;
	432	switch (args->ar_quota) {
	433	case GFS2_QUOTA_OFF:
	434	state = "off";
	435	break;
	436	case GFS2_QUOTA_ACCOUNT:
	437	state = "account";
	438	break;
	439	case GFS2_QUOTA_ON:
	440	state = "on";
	441	break;
	442	default:
	443	state = "unknown";
	444	break;
	445	}
	446	seq_printf(s, ",quota=%s", state);
	447	}
	448	if (args->ar_suiddir)
	449	seq_printf(s, ",suiddir");
	450	if (args->ar_data != GFS2_DATA_DEFAULT) {
	451	char *state;
	452	switch (args->ar_data) {
	453	case GFS2_DATA_WRITEBACK:
	454	state = "writeback";
	455	break;
	456	case GFS2_DATA_ORDERED:
	457	state = "ordered";
	458	break;
	459	default:
	460	state = "unknown";
	461	break;
	462	}
	463	seq_printf(s, ",data=%s", state);
	464	}
	465
	466	return 0;
	467	}
	468
907b9bce	469	/*
feaa7bba SW	470	* We have to (at the moment) hold the inodes main lock to cover
	471	* the gap between unlocking the shared lock on the iopen lock and
	472	* taking the exclusive lock. I'd rather do a shared -> exclusive
	473	* conversion on the iopen lock, but we can change that later. This
	474	* is safe, just less efficient.
	475	*/
9b8df98f	476
feaa7bba SW	477	static void gfs2_delete_inode(struct inode *inode)
	478	{
	479	struct gfs2_sbd *sdp = inode->i_sb->s_fs_info;
	480	struct gfs2_inode *ip = GFS2_I(inode);
	481	struct gfs2_holder gh;
	482	int error;
	483
091806ed	484	if (!test_bit(GIF_USER, &ip->i_flags))
feaa7bba SW	485	goto out;
feaa7bba SW	486
d93cfa98	487	error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
feaa7bba SW	488	if (unlikely(error)) {
	489	gfs2_glock_dq_uninit(&ip->i_iopen_gh);
	490	goto out;
	491	}
	492
d93cfa98	493	gfs2_glock_dq_wait(&ip->i_iopen_gh);
feaa7bba SW	494	gfs2_holder_reinit(LM_ST_EXCLUSIVE, LM_FLAG_TRY_1CB \| GL_NOCACHE, &ip->i_iopen_gh);
	495	error = gfs2_glock_nq(&ip->i_iopen_gh);
	496	if (error)
	497	goto out_uninit;
	498
b60623c2	499	if (S_ISDIR(inode->i_mode) &&
feaa7bba SW	500	(ip->i_di.di_flags & GFS2_DIF_EXHASH)) {
	501	error = gfs2_dir_exhash_dealloc(ip);
	502	if (error)
	503	goto out_unlock;
	504	}
	505
	506	if (ip->i_di.di_eattr) {
	507	error = gfs2_ea_dealloc(ip);
	508	if (error)
	509	goto out_unlock;
	510	}
	511
	512	if (!gfs2_is_stuffed(ip)) {
	513	error = gfs2_file_dealloc(ip);
	514	if (error)
	515	goto out_unlock;
	516	}
	517
	518	error = gfs2_dinode_dealloc(ip);
16615be1 SW	519	if (error)
	520	goto out_unlock;
	521
	522	error = gfs2_trans_begin(sdp, 0, sdp->sd_jdesc->jd_blocks);
	523	if (error)
	524	goto out_unlock;
	525	/* Needs to be done before glock release & also in a transaction */
49686f71	526	truncate_inode_pages(&inode->i_data, 0);
16615be1	527	gfs2_trans_end(sdp);
feaa7bba SW	528
	529	out_unlock:
	530	gfs2_glock_dq(&ip->i_iopen_gh);
	531	out_uninit:
	532	gfs2_holder_uninit(&ip->i_iopen_gh);
	533	gfs2_glock_dq_uninit(&gh);
3b8249f6	534	if (error && error != GLR_TRYFAILED)
feaa7bba SW	535	fs_warn(sdp, "gfs2_delete_inode: %d\n", error);
	536	out:
	537	truncate_inode_pages(&inode->i_data, 0);
	538	clear_inode(inode);
	539	}
	540
320dd101 SW	541	static struct inode gfs2_alloc_inode(struct super_block sb)
320dd101 SW	542	{
320dd101 SW	543	struct gfs2_inode *ip;
	544
	545	ip = kmem_cache_alloc(gfs2_inode_cachep, GFP_KERNEL);
	546	if (ip) {
	547	ip->i_flags = 0;
	548	ip->i_gl = NULL;
320dd101 SW	549	}
	550	return &ip->i_inode;
	551	}
	552
	553	static void gfs2_destroy_inode(struct inode *inode)
	554	{
	555	kmem_cache_free(gfs2_inode_cachep, inode);
	556	}
	557
ee9b6d61	558	const struct super_operations gfs2_super_ops = {
4a221953 SW	559	.alloc_inode = gfs2_alloc_inode,
	560	.destroy_inode = gfs2_destroy_inode,
	561	.write_inode = gfs2_write_inode,
	562	.delete_inode = gfs2_delete_inode,
	563	.put_super = gfs2_put_super,
	564	.write_super = gfs2_write_super,
	565	.sync_fs = gfs2_sync_fs,
	566	.write_super_lockfs = gfs2_write_super_lockfs,
	567	.unlockfs = gfs2_unlockfs,
	568	.statfs = gfs2_statfs,
	569	.remount_fs = gfs2_remount_fs,
	570	.clear_inode = gfs2_clear_inode,
3b8249f6	571	.drop_inode = gfs2_drop_inode,
4a221953	572	.show_options = gfs2_show_options,
b3b94faa DT	573	};
b3b94faa DT	574