super.c

ocfs1.4.1 oracle分布式文件系统

#ifdef OCFS2_ORACORE_WORKAROUNDS
	if (opts & OCFS2_MOUNT_COMPAT_OCFS)
		seq_printf(s, ",datavolume");
#endif

	if (osb->preferred_slot != OCFS2_INVALID_SLOT)
		seq_printf(s, ",preferred_slot=%d", osb->preferred_slot);

	if (osb->s_atime_quantum != OCFS2_DEFAULT_ATIME_QUANTUM)
		seq_printf(s, ",atime_quantum=%u", osb->s_atime_quantum);

	if (osb->osb_commit_interval)
		seq_printf(s, ",commit=%u",
			   (unsigned) (osb->osb_commit_interval / HZ));

	if (osb->local_alloc_size != OCFS2_DEFAULT_LOCAL_ALLOC_SIZE)
		seq_printf(s, ",localalloc=%d", osb->local_alloc_size);

	if (opts & OCFS2_MOUNT_LOCALFLOCKS)
		seq_printf(s, ",localflocks,");

	return 0;
}

static int __init ocfs2_init(void)
{
	int status;

	mlog_entry_void();

	ocfs2_print_version();

	status = init_ocfs2_uptodate_cache();
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

	status = ocfs2_initialize_mem_caches();
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

	ocfs2_wq = create_singlethread_workqueue("ocfs2_wq");
	if (!ocfs2_wq) {
		status = -ENOMEM;
		goto leave;
	}

	ocfs2_debugfs_root = debugfs_create_dir("ocfs2", NULL);
	if (!ocfs2_debugfs_root) {
		status = -EFAULT;
		mlog(ML_ERROR, "Unable to create ocfs2 debugfs root.\n");
	}

leave:
	if (status < 0) {
		ocfs2_free_mem_caches();
		exit_ocfs2_uptodate_cache();
	}

	mlog_exit(status);

	if (status >= 0) {
		return register_filesystem(&ocfs2_fs_type);
	} else
		return -1;
}

static void __exit ocfs2_exit(void)
{
	mlog_entry_void();

	if (ocfs2_wq) {
		flush_workqueue(ocfs2_wq);
		destroy_workqueue(ocfs2_wq);
	}

	debugfs_remove(ocfs2_debugfs_root);

	ocfs2_free_mem_caches();

	unregister_filesystem(&ocfs2_fs_type);

	exit_ocfs2_uptodate_cache();

	mlog_exit_void();
}

static void ocfs2_put_super(struct super_block *sb)
{
	mlog_entry("(0x%p)\n", sb);

	ocfs2_sync_blockdev(sb);
	ocfs2_dismount_volume(sb, 0);

	mlog_exit_void();
}

static inline int _ocfs2_statfs(struct dentry *dentry, struct kstatfs *buf)
{
	struct ocfs2_super *osb;
	u32 numbits, freebits;
	int status;
	struct ocfs2_dinode *bm_lock;
	struct buffer_head *bh = NULL;
	struct inode *inode = NULL;

	mlog_entry("(%p, %p)\n", dentry->d_sb, buf);

	osb = OCFS2_SB(dentry->d_sb);

	inode = ocfs2_get_system_file_inode(osb,
					    GLOBAL_BITMAP_SYSTEM_INODE,
					    OCFS2_INVALID_SLOT);
	if (!inode) {
		mlog(ML_ERROR, "failed to get bitmap inode\n");
		status = -EIO;
		goto bail;
	}

	status = ocfs2_inode_lock(inode, &bh, 0);
	if (status < 0) {
		mlog_errno(status);
		goto bail;
	}

	bm_lock = (struct ocfs2_dinode *) bh->b_data;

	numbits = le32_to_cpu(bm_lock->id1.bitmap1.i_total);
	freebits = numbits - le32_to_cpu(bm_lock->id1.bitmap1.i_used);

#ifdef OCFS2_ORACORE_WORKAROUNDS
	if (osb->s_mount_opt & OCFS2_MOUNT_COMPAT_OCFS)
		buf->f_type = OCFS_SUPER_MAGIC;
	else
#endif
		buf->f_type = OCFS2_SUPER_MAGIC;
	buf->f_bsize = dentry->d_sb->s_blocksize;
	buf->f_namelen = OCFS2_MAX_FILENAME_LEN;
	buf->f_blocks = ((sector_t) numbits) *
			(osb->s_clustersize >> osb->sb->s_blocksize_bits);
	buf->f_bfree = ((sector_t) freebits) *
		       (osb->s_clustersize >> osb->sb->s_blocksize_bits);
	buf->f_bavail = buf->f_bfree;
	buf->f_files = numbits;
	buf->f_ffree = freebits;

	brelse(bh);

	ocfs2_inode_unlock(inode, 0);
	status = 0;
bail:
	if (inode)
		iput(inode);

	mlog_exit(status);

	return status;
}

#ifdef STATFS_TAKES_SB
static int ocfs2_statfs(struct super_block *sb, struct kstatfs *buf)
{
	return _ocfs2_statfs(sb->s_root, buf);
}
#else
static int ocfs2_statfs(struct dentry *dentry, struct kstatfs *buf)
{
	return _ocfs2_statfs(dentry, buf);
}
#endif

static void ocfs2_inode_init_once(void *data,
				  struct kmem_cache *cachep,
				  unsigned long flags)
{
	struct ocfs2_inode_info *oi = data;

	oi->ip_flags = 0;
	oi->ip_open_count = 0;
	spin_lock_init(&oi->ip_lock);
	ocfs2_extent_map_init(&oi->vfs_inode);
	INIT_LIST_HEAD(&oi->ip_io_markers);
	oi->ip_created_trans = 0;
	oi->ip_last_trans = 0;
	oi->ip_dir_start_lookup = 0;

	init_rwsem(&oi->ip_alloc_sem);
	mutex_init(&oi->ip_io_mutex);

	oi->ip_blkno = 0ULL;
	oi->ip_clusters = 0;

	ocfs2_lock_res_init_once(&oi->ip_rw_lockres);
	ocfs2_lock_res_init_once(&oi->ip_inode_lockres);
	ocfs2_lock_res_init_once(&oi->ip_open_lockres);

	ocfs2_metadata_cache_init(&oi->vfs_inode);

	inode_init_once(&oi->vfs_inode);
}

static int ocfs2_initialize_mem_caches(void)
{
	ocfs2_inode_cachep = kapi_kmem_cache_create("ocfs2_inode_cache",
				       sizeof(struct ocfs2_inode_info),
				       0,
				       (SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
						SLAB_MEM_SPREAD),
				       ocfs2_inode_init_once);
	if (!ocfs2_inode_cachep)
		return -ENOMEM;

	return 0;
}

static void ocfs2_free_mem_caches(void)
{
	if (ocfs2_inode_cachep)
		kmem_cache_destroy(ocfs2_inode_cachep);

	ocfs2_inode_cachep = NULL;
}

static int ocfs2_get_sector(struct super_block *sb,
			    struct buffer_head **bh,
			    int block,
			    int sect_size)
{
	if (!sb_set_blocksize(sb, sect_size)) {
		mlog(ML_ERROR, "unable to set blocksize\n");
		return -EIO;
	}

	*bh = sb_getblk(sb, block);
	if (!*bh) {
		mlog_errno(-EIO);
		return -EIO;
	}
	lock_buffer(*bh);
	if (!buffer_dirty(*bh))
		clear_buffer_uptodate(*bh);
	unlock_buffer(*bh);
	ll_rw_block(READ, 1, bh);
	wait_on_buffer(*bh);
	return 0;
}

/* ocfs2 1.0 only allows one cluster and node identity per kernel image. */
static int ocfs2_fill_local_node_info(struct ocfs2_super *osb)
{
	int status;

	/* XXX hold a ref on the node while mounte?  easy enough, if
	 * desirable. */
	if (ocfs2_mount_local(osb))
		osb->node_num = 0;
	else
		osb->node_num = o2nm_this_node();

	if (osb->node_num == O2NM_MAX_NODES) {
		mlog(ML_ERROR, "could not find this host's node number\n");
		status = -ENOENT;
		goto bail;
	}

	mlog(0, "I am node %d\n", osb->node_num);

	status = 0;
bail:
	return status;
}

static int ocfs2_mount_volume(struct super_block *sb)
{
	int status = 0;
	int unlock_super = 0;
	struct ocfs2_super *osb = OCFS2_SB(sb);

	mlog_entry_void();

	if (ocfs2_is_hard_readonly(osb))
		goto leave;

	status = ocfs2_fill_local_node_info(osb);
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

	status = ocfs2_dlm_init(osb);
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

	status = ocfs2_super_lock(osb, 1);
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}
	unlock_super = 1;

	/* This will load up the node map and add ourselves to it. */
	status = ocfs2_find_slot(osb);
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

	/* load all node-local system inodes */
	status = ocfs2_init_local_system_inodes(osb);
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

	status = ocfs2_check_volume(osb);
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

	status = ocfs2_truncate_log_init(osb);
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

	if (ocfs2_mount_local(osb))
		goto leave;

leave:
	if (unlock_super)
		ocfs2_super_unlock(osb, 1);

	mlog_exit(status);
	return status;
}

/* we can't grab the goofy sem lock from inside wait_event, so we use
 * memory barriers to make sure that we'll see the null task before
 * being woken up */
static int ocfs2_recovery_thread_running(struct ocfs2_super *osb)
{
	mb();
	return osb->recovery_thread_task != NULL;
}

static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err)
{
	int tmp;
	struct ocfs2_super *osb = NULL;
	char nodestr[8];

	mlog_entry("(0x%p)\n", sb);

	BUG_ON(!sb);
	osb = OCFS2_SB(sb);
	BUG_ON(!osb);

	ocfs2_shutdown_local_alloc(osb);

	ocfs2_truncate_log_shutdown(osb);

	/* disable any new recovery threads and wait for any currently
	 * running ones to exit. Do this before setting the vol_state. */
	mutex_lock(&osb->recovery_lock);
	osb->disable_recovery = 1;
	mutex_unlock(&osb->recovery_lock);
	wait_event(osb->recovery_event, !ocfs2_recovery_thread_running(osb));

	/* At this point, we know that no more recovery threads can be
	 * launched, so wait for any recovery completion work to
	 * complete. */
	flush_workqueue(ocfs2_wq);

	ocfs2_journal_shutdown(osb);

	ocfs2_sync_blockdev(sb);

	/* No dlm means we've failed during mount, so skip all the
	 * steps which depended on that to complete. */
	if (osb->dlm) {
		tmp = ocfs2_super_lock(osb, 1);
		if (tmp < 0) {
			mlog_errno(tmp);
			return;
		}
	}

	if (osb->slot_num != OCFS2_INVALID_SLOT)
		ocfs2_put_slot(osb);

	if (osb->dlm)
		ocfs2_super_unlock(osb, 1);

	ocfs2_release_system_inodes(osb);

	if (osb->dlm)
		ocfs2_dlm_shutdown(osb);

	debugfs_remove(osb->osb_debug_root);

	if (!mnt_err)
		ocfs2_stop_heartbeat(osb);

	atomic_set(&osb->vol_state, VOLUME_DISMOUNTED);

	if (ocfs2_mount_local(osb))
		snprintf(nodestr, sizeof(nodestr), "local");
	else
		snprintf(nodestr, sizeof(nodestr), "%d", osb->node_num);

	printk(KERN_INFO "ocfs2: Unmounting device (%s) on (node %s)\n",
	       osb->dev_str, nodestr);

	ocfs2_delete_osb(osb);
	kfree(osb);
	sb->s_dev = 0;
	sb->s_fs_info = NULL;
}

static int ocfs2_setup_osb_uuid(struct ocfs2_super *osb, const unsigned char *uuid,
				unsigned uuid_bytes)
{
	int i, ret;
	char *ptr;

	BUG_ON(uuid_bytes != OCFS2_VOL_UUID_LEN);

	osb->uuid_str = kzalloc(OCFS2_VOL_UUID_LEN * 2 + 1, GFP_KERNEL);
	if (osb->uuid_str == NULL)
		return -ENOMEM;

	for (i = 0, ptr = osb->uuid_str; i < OCFS2_VOL_UUID_LEN; i++) {
		/* print with null */
		ret = snprintf(ptr, 3, "%02X", uuid[i]);
		if (ret != 2) /* drop super cleans up */
			return -EINVAL;
		/* then only advance past the last char */
		ptr += 2;
	}

	return 0;
}

static int ocfs2_initialize_super(struct super_block *sb,
				  struct buffer_head *bh,
				  int sector_size)
{
	int status;
	int i, cbits, bbits;
	struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data;
	struct inode *inode = NULL;
	struct ocfs2_journal *journal;
	__le32 uuid_net_key;
	struct ocfs2_super *osb;

	mlog_entry_void();

	osb = kzalloc(sizeof(struct ocfs2_super), GFP_KERNEL);
	if (!osb) {
		status = -ENOMEM;
		mlog_errno(status);
		goto bail;
	}

	sb->s_fs_info = osb;
	sb->s_op = &ocfs2_sops;
	sb->s_export_op = &ocfs2_export_ops;
	osb->osb_locking_proto = ocfs2_locking_protocol;
	sb->s_time_gran = 1;
	sb->s_flags |= MS_NOATIME;
	/* this is needed to support O_LARGEFILE */
	cbits = le32_to_cpu(di->id2.i_super.s_clustersize_bits);