namei.c

ocfs1.2.7 源码

	const int rlen = le16_to_cpu(de->rec_len);

	if (rlen < OCFS2_DIR_REC_LEN(1))
		error_msg = "rec_len is smaller than minimal";
	else if (rlen % 4 != 0)
		error_msg = "rec_len % 4 != 0";
	else if (rlen < OCFS2_DIR_REC_LEN(de->name_len))
		error_msg = "rec_len is too small for name_len";
	else if (((char *) de - bh->b_data) + rlen > dir->i_sb->s_blocksize)
		error_msg = "directory entry across blocks";

	if (error_msg != NULL)
		mlog(ML_ERROR, "bad entry in directory #%"MLFu64": %s - "
		     "offset=%lu, inode=%"MLFu64", rec_len=%d, name_len=%d\n",
		     OCFS2_I(dir)->ip_blkno, error_msg, offset,
		     le64_to_cpu(de->inode), rlen, de->name_len);
	return error_msg == NULL ? 1 : 0;
}

/* we don't always have a dentry for what we want to add, so people
 * like orphan dir can call this instead.
 *
 * If you pass me insert_bh, I'll skip the search of the other dir
 * blocks and put the record in there.
 */
static int __ocfs2_add_entry(struct ocfs2_journal_handle *handle,
			     struct inode *dir,
			     const char *name, int namelen,
			     struct inode *inode, u64 blkno,
			     struct buffer_head *parent_fe_bh,
			     struct buffer_head *insert_bh)
{
	unsigned long offset;
	unsigned short rec_len;
	struct ocfs2_dir_entry *de, *de1;
	struct super_block *sb;
	int retval, status;

	mlog_entry_void();

	sb = dir->i_sb;

	if (!namelen)
		return -EINVAL;

	rec_len = OCFS2_DIR_REC_LEN(namelen);
	offset = 0;
	de = (struct ocfs2_dir_entry *) insert_bh->b_data;
	while (1) {
		BUG_ON((char *)de >= sb->s_blocksize + insert_bh->b_data);
		/* These checks should've already been passed by the
		 * prepare function, but I guess we can leave them
		 * here anyway. */
		if (!ocfs2_check_dir_entry(dir, de, insert_bh, offset)) {
			retval = -ENOENT;
			goto bail;
		}
		if (ocfs2_match(namelen, name, de)) {
			retval = -EEXIST;
			goto bail;
		}
		if (((le64_to_cpu(de->inode) == 0) &&
		     (le16_to_cpu(de->rec_len) >= rec_len)) ||
		    (le16_to_cpu(de->rec_len) >=
		     (OCFS2_DIR_REC_LEN(de->name_len) + rec_len))) {
			dir->i_mtime = dir->i_ctime = CURRENT_TIME;
			retval = ocfs2_mark_inode_dirty(handle, dir, parent_fe_bh);
			if (retval < 0) {
				mlog_errno(retval);
				goto bail;
			}

			status = ocfs2_journal_access(handle, dir, insert_bh,
						      OCFS2_JOURNAL_ACCESS_WRITE);
			/* By now the buffer is marked for journaling */
			offset += le16_to_cpu(de->rec_len);
			if (le64_to_cpu(de->inode)) {
				de1 = (struct ocfs2_dir_entry *)((char *) de +
					OCFS2_DIR_REC_LEN(de->name_len));
				de1->rec_len =
					cpu_to_le16(le16_to_cpu(de->rec_len) -
					OCFS2_DIR_REC_LEN(de->name_len));
				de->rec_len = cpu_to_le16(OCFS2_DIR_REC_LEN(de->name_len));
				de = de1;
			}
			de->file_type = OCFS2_FT_UNKNOWN;
			if (blkno) {
				de->inode = cpu_to_le64(blkno);
				ocfs2_set_de_type(de, inode->i_mode);
			} else
				de->inode = 0;
			de->name_len = namelen;
			memcpy(de->name, name, namelen);

			dir->i_version++;
			status = ocfs2_journal_dirty(handle, insert_bh);
			retval = 0;
			goto bail;
		}
		offset += le16_to_cpu(de->rec_len);
		de = (struct ocfs2_dir_entry *) ((char *) de + le16_to_cpu(de->rec_len));
	}

	/* when you think about it, the assert above should prevent us
	 * from ever getting here. */
	retval = -ENOSPC;
bail:

	mlog_exit(retval);
	return retval;
}


/*
 * ocfs2_delete_entry deletes a directory entry by merging it with the
 * previous entry
 */
static int ocfs2_delete_entry(struct ocfs2_journal_handle *handle,
			      struct inode *dir,
			      struct ocfs2_dir_entry *de_del,
			      struct buffer_head *bh)
{
	struct ocfs2_dir_entry *de, *pde;
	int i, status = -ENOENT;

	mlog_entry("(0x%p, 0x%p, 0x%p, 0x%p)\n", handle, dir, de_del, bh);

	i = 0;
	pde = NULL;
	de = (struct ocfs2_dir_entry *) bh->b_data;
	while (i < bh->b_size) {
		if (!ocfs2_check_dir_entry(dir, de, bh, i)) {
			status = -EIO;
			mlog_errno(status);
			goto bail;
		}
		if (de == de_del)  {
			status = ocfs2_journal_access(handle, dir, bh,
						      OCFS2_JOURNAL_ACCESS_WRITE);
			if (status < 0) {
				status = -EIO;
				mlog_errno(status);
				goto bail;
			}
			if (pde)
				pde->rec_len =
					cpu_to_le16(le16_to_cpu(pde->rec_len) +
						    le16_to_cpu(de->rec_len));
			else
				de->inode = 0;
			dir->i_version++;
			status = ocfs2_journal_dirty(handle, bh);
			goto bail;
		}
		i += le16_to_cpu(de->rec_len);
		pde = de;
		de = (struct ocfs2_dir_entry *)((char *)de + le16_to_cpu(de->rec_len));
	}
bail:
	mlog_exit(status);
	return status;
}

/*
 * Returns 0 if not found, -1 on failure, and 1 on success
 */
static int inline ocfs2_search_dirblock(struct buffer_head *bh,
					struct inode *dir,
					const char *name, int namelen,
					unsigned long offset,
					struct ocfs2_dir_entry **res_dir)
{
	struct ocfs2_dir_entry *de;
	char *dlimit, *de_buf;
	int de_len;
	int ret = 0;

	mlog_entry_void();

	de_buf = bh->b_data;
	dlimit = de_buf + dir->i_sb->s_blocksize;

	while (de_buf < dlimit) {
		/* this code is executed quadratically often */
		/* do minimal checking `by hand' */

		de = (struct ocfs2_dir_entry *) de_buf;

		if (de_buf + namelen <= dlimit &&
		    ocfs2_match(namelen, name, de)) {
			/* found a match - just to be sure, do a full check */
			if (!ocfs2_check_dir_entry(dir, de, bh, offset)) {
				ret = -1;
				goto bail;
			}
			*res_dir = de;
			ret = 1;
			goto bail;
		}

		/* prevent looping on a bad block */
		de_len = le16_to_cpu(de->rec_len);
		if (de_len <= 0) {
			ret = -1;
			goto bail;
		}

		de_buf += de_len;
		offset += de_len;
	}

bail:
	mlog_exit(ret);
	return ret;
}

struct buffer_head *ocfs2_find_entry(const char *name, int namelen,
				     struct inode *dir,
				     struct ocfs2_dir_entry **res_dir)
{
	struct super_block *sb;
	struct buffer_head *bh_use[NAMEI_RA_SIZE];
	struct buffer_head *bh, *ret = NULL;
	unsigned long start, block, b;
	int ra_max = 0;		/* Number of bh's in the readahead
				   buffer, bh_use[] */
	int ra_ptr = 0;		/* Current index into readahead
				   buffer */
	int num = 0;
	int nblocks, i, err;

	mlog_entry_void();

	*res_dir = NULL;
	sb = dir->i_sb;

	nblocks = i_size_read(dir) >> sb->s_blocksize_bits;
	start = OCFS2_I(dir)->ip_dir_start_lookup;
	if (start >= nblocks)
		start = 0;
	block = start;

restart:
	do {
		/*
		 * We deal with the read-ahead logic here.
		 */
		if (ra_ptr >= ra_max) {
			/* Refill the readahead buffer */
			ra_ptr = 0;
			b = block;
			for (ra_max = 0; ra_max < NAMEI_RA_SIZE; ra_max++) {
				/*
				 * Terminate if we reach the end of the
				 * directory and must wrap, or if our
				 * search has finished at this block.
				 */
				if (b >= nblocks || (num && block == start)) {
					bh_use[ra_max] = NULL;
					break;
				}
				num++;

				/* XXX: questionable readahead stuff here */
				bh = ocfs2_bread(dir, b++, &err, 1);
				bh_use[ra_max] = bh;
#if 0		// ???
				if (bh)
					ll_rw_block(READ, 1, &bh);
#endif
			}
		}
		if ((bh = bh_use[ra_ptr++]) == NULL)
			goto next;
		wait_on_buffer(bh);
		if (!buffer_uptodate(bh)) {
			/* read error, skip block & hope for the best */
			brelse(bh);
			goto next;
		}
		i = ocfs2_search_dirblock(bh, dir, name, namelen,
					  block << sb->s_blocksize_bits,
					  res_dir);
		if (i == 1) {
			OCFS2_I(dir)->ip_dir_start_lookup = block;
			ret = bh;
			goto cleanup_and_exit;
		} else {
			brelse(bh);
			if (i < 0)
				goto cleanup_and_exit;
		}
	next:
		if (++block >= nblocks)
			block = 0;
	} while (block != start);

	/*
	 * If the directory has grown while we were searching, then
	 * search the last part of the directory before giving up.
	 */
	block = nblocks;
	nblocks = i_size_read(dir) >> sb->s_blocksize_bits;
	if (block < nblocks) {
		start = 0;
		goto restart;
	}

cleanup_and_exit:
	/* Clean up the read-ahead blocks */
	for (; ra_ptr < ra_max; ra_ptr++)
		brelse(bh_use[ra_ptr]);

	mlog_exit_ptr(ret);
	return ret;
}

static int ocfs2_blkno_stringify(u64 blkno, char *name)
{
	int status, namelen;

	mlog_entry_void();

	namelen = snprintf(name, OCFS2_ORPHAN_NAMELEN + 1, "%016"MLFx64,
			   blkno);
	if (namelen <= 0) {
		if (namelen)
			status = namelen;
		else
			status = -EINVAL;
		mlog_errno(status);
		goto bail;
	}
	if (namelen != OCFS2_ORPHAN_NAMELEN) {
		status = -EINVAL;
		mlog_errno(status);
		goto bail;
	}

	mlog(0, "built filename '%s' for orphan dir (len=%d)\n", name,
	     namelen);

	status = 0;
bail:
	mlog_exit(status);
	return status;
}

static int ocfs2_prepare_orphan_dir(struct ocfs2_super *osb,
				    struct ocfs2_journal_handle *handle,
				    struct inode *inode,
				    char *name,
				    struct buffer_head **de_bh)
{
	struct inode *orphan_dir_inode = NULL;
	struct buffer_head *orphan_dir_bh = NULL;
	int status = 0;

	status = ocfs2_blkno_stringify(OCFS2_I(inode)->ip_blkno, name);
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

	orphan_dir_inode = ocfs2_get_system_file_inode(osb,
						       ORPHAN_DIR_SYSTEM_INODE,
						       osb->slot_num);
	if (!orphan_dir_inode) {
		status = -ENOENT;
		mlog_errno(status);
		goto leave;
	}

	ocfs2_handle_add_inode(handle, orphan_dir_inode);
	status = ocfs2_meta_lock(orphan_dir_inode, handle, &orphan_dir_bh, 1);
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

	status = ocfs2_prepare_dir_for_insert(osb, orphan_dir_inode,
					      orphan_dir_bh, name,
					      OCFS2_ORPHAN_NAMELEN, de_bh);
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

leave:
	if (orphan_dir_inode)
		iput(orphan_dir_inode);

	if (orphan_dir_bh)
		brelse(orphan_dir_bh);

	mlog_exit(status);
	return status;
}

static int ocfs2_orphan_add(struct ocfs2_super *osb,
			    struct ocfs2_journal_handle *handle,
			    struct inode *inode,
			    struct ocfs2_dinode *fe,
			    char *name,
			    struct buffer_head *de_bh)
{
	struct inode *orphan_dir_inode = NULL;
	struct buffer_head *orphan_dir_bh = NULL;
	int status = 0;
	struct ocfs2_dinode *orphan_fe;

	mlog_entry("(inode->i_ino = %lu)\n", inode->i_ino);

	orphan_dir_inode = ocfs2_get_system_file_inode(osb,
						       ORPHAN_DIR_SYSTEM_INODE,
						       osb->slot_num);
	if (!orphan_dir_inode) {
		status = -ENOENT;
		mlog_errno(status);
		goto leave;
	}

	status = ocfs2_read_block(osb,
				  OCFS2_I(orphan_dir_inode)->ip_blkno,
				  &orphan_dir_bh, OCFS2_BH_CACHED,
				  orphan_dir_inode);
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

	status = ocfs2_journal_access(handle, orphan_dir_inode, orphan_dir_bh,
				      OCFS2_JOURNAL_ACCESS_WRITE);
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

	/* we're a cluster, and nlink can change on disk from
	 * underneath us... */
	orphan_fe = (struct ocfs2_dinode *) orphan_dir_bh->b_data;
	if (S_ISDIR(inode->i_mode))
		le16_add_cpu(&orphan_fe->i_links_count, 1);
	orphan_dir_inode->i_nlink = le16_to_cpu(orphan_fe->i_links_count);

	status = ocfs2_journal_dirty(handle, orphan_dir_bh);
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

	status = __ocfs2_add_entry(handle, orphan_dir_inode, name,
				   OCFS2_ORPHAN_NAMELEN, inode,
				   OCFS2_I(inode)->ip_blkno,
				   orphan_dir_bh, de_bh);
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

	le32_add_cpu(&fe->i_flags, OCFS2_ORPHANED_FL);

	/* Record which orphan dir our inode now resides
	 * in. delete_inode will use this to determine which orphan
	 * dir to lock. */
	spin_lock(&OCFS2_I(inode)->ip_lock);
	OCFS2_I(inode)->ip_orphaned_slot = osb->slot_num;
	spin_unlock(&OCFS2_I(inode)->ip_lock);

	mlog(0, "Inode %"MLFu64" orphaned in slot %d\n",
	     OCFS2_I(inode)->ip_blkno, osb->slot_num);

leave:
	if (orphan_dir_inode)
		iput(orphan_dir_inode);

	if (orphan_dir_bh)
		brelse(orphan_dir_bh);

	mlog_exit(status);
	return status;
}

/* unlike orphan_add, we expect the orphan dir to already be locked here. */
int ocfs2_orphan_del(struct ocfs2_super *osb,
		     struct ocfs2_journal_handle *handle,
		     struct inode *orphan_dir_inode,
		     struct inode *inode,
		     struct buffer_head *orphan_dir_bh)
{
	char name[OCFS2_ORPHAN_NAMELEN + 1];
	struct ocfs2_dinode *orphan_fe;
	int status = 0;
	struct buffer_head *target_de_bh = NULL;
	struct ocfs2_dir_entry *target_de = NULL;

	mlog_entry_void();

	status = ocfs2_blkno_stringify(OCFS2_I(inode)->ip_blkno, name);
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

	mlog(0, "removing '%s' from orphan dir %"MLFu64" (namelen=%d)\n",
	     name, OCFS2_I(orphan_dir_inode)->ip_blkno, OCFS2_ORPHAN_NAMELEN);

	/* find it's spot in the orphan directory */
	target_de_bh = ocfs2_find_entry(name, OCFS2_ORPHAN_NAMELEN,
					orphan_dir_inode, &target_de);
	if (!target_de_bh) {
		status = -ENOENT;
		mlog_errno(status);
		goto leave;
	}

	/* remove it from the orphan directory */
	status = ocfs2_delete_entry(handle, orphan_dir_inode, target_de,
				    target_de_bh);
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

	status = ocfs2_journal_access(handle,orphan_dir_inode,  orphan_dir_bh,
				      OCFS2_JOURNAL_ACCESS_WRITE);
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

	/* do the i_nlink dance! :) */
	orphan_fe = (struct ocfs2_dinode *) orphan_dir_bh->b_data;
	if (S_ISDIR(inode->i_mode))
		le16_add_cpu(&orphan_fe->i_links_count, -1);
	orphan_dir_inode->i_nlink = le16_to_cpu(orphan_fe->i_links_count);

	status = ocfs2_journal_dirty(handle, orphan_dir_bh);
	if (status < 0) {
		mlog_errno(status);
		goto leave;
	}

leave:
	if (target_de_bh)
		brelse(target_de_bh);

	mlog_exit(status);
	return status;
}

struct inode_operations ocfs2_dir_iops = {
	.create		= ocfs2_create,
	.lookup		= ocfs2_lookup,
	.link		= ocfs2_link,
	.unlink		= ocfs2_unlink,
	.rmdir		= ocfs2_unlink,
	.symlink	= ocfs2_symlink,
	.mkdir		= ocfs2_mkdir,
	.mknod		= ocfs2_mknod,
	.rename		= ocfs2_rename,
	.setattr	= ocfs2_setattr,
	.getattr	= ocfs2_getattr,
};