dir.c

linux 内核源代码

/* -*- mode: c; c-basic-offset: 8; -*-
 * vim: noexpandtab sw=8 ts=8 sts=0:
 *
 * dir.c
 *
 * Creates, reads, walks and deletes directory-nodes
 *
 * Copyright (C) 2002, 2004 Oracle.  All rights reserved.
 *
 *  Portions of this code from linux/fs/ext3/dir.c
 *
 *  Copyright (C) 1992, 1993, 1994, 1995
 *  Remy Card ([e-mail])
 *  Laboratoire MASI - Institut Blaise pascal
 *  Universite Pierre et Marie Curie (Paris VI)
 *
 *   from
 *
 *   linux/fs/minix/dir.c
 *
 *   Copyright (C) 1991, 1992 Linux Torvalds
 *
 * 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; either
 * version 2 of the License, or (at your option) any later version.
 *
 * This program is distributed in the hope that it will 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 to the
 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 * Boston, MA 021110-1307, USA.
 */

#include <linux/fs.h>
#include <linux/types.h>
#include <linux/slab.h>
#include <linux/highmem.h>

#define MLOG_MASK_PREFIX ML_NAMEI
#include <cluster/masklog.h>

#include "ocfs2.h"

#include "alloc.h"
#include "dir.h"
#include "dlmglue.h"
#include "extent_map.h"
#include "file.h"
#include "inode.h"
#include "journal.h"
#include "namei.h"
#include "suballoc.h"
#include "super.h"
#include "uptodate.h"

#include "buffer_head_io.h"

#define NAMEI_RA_CHUNKS  2
#define NAMEI_RA_BLOCKS  4
#define NAMEI_RA_SIZE        (NAMEI_RA_CHUNKS * NAMEI_RA_BLOCKS)
#define NAMEI_RA_INDEX(c,b)  (((c) * NAMEI_RA_BLOCKS) + (b))

static unsigned char ocfs2_filetype_table[] = {
	DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK
};

static int ocfs2_extend_dir(struct ocfs2_super *osb,
			    struct inode *dir,
			    struct buffer_head *parent_fe_bh,
			    unsigned int blocks_wanted,
			    struct buffer_head **new_de_bh);
static int ocfs2_do_extend_dir(struct super_block *sb,
			       handle_t *handle,
			       struct inode *dir,
			       struct buffer_head *parent_fe_bh,
			       struct ocfs2_alloc_context *data_ac,
			       struct ocfs2_alloc_context *meta_ac,
			       struct buffer_head **new_bh);

/*
 * bh passed here can be an inode block or a dir data block, depending
 * on the inode inline data flag.
 */
static int ocfs2_check_dir_entry(struct inode * dir,
				 struct ocfs2_dir_entry * de,
				 struct buffer_head * bh,
				 unsigned long offset)
{
	const char *error_msg = NULL;
	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 #%llu: %s - "
		     "offset=%lu, inode=%llu, rec_len=%d, name_len=%d\n",
		     (unsigned long long)OCFS2_I(dir)->ip_blkno, error_msg,
		     offset, (unsigned long long)le64_to_cpu(de->inode), rlen,
		     de->name_len);
	return error_msg == NULL ? 1 : 0;
}

static inline int ocfs2_match(int len,
			      const char * const name,
			      struct ocfs2_dir_entry *de)
{
	if (len != de->name_len)
		return 0;
	if (!de->inode)
		return 0;
	return !memcmp(name, de->name, len);
}

/*
 * 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,
					char *first_de,
					unsigned int bytes,
					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 = first_de;
	dlimit = de_buf + bytes;

	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;
}

static struct buffer_head *ocfs2_find_entry_id(const char *name,
					       int namelen,
					       struct inode *dir,
					       struct ocfs2_dir_entry **res_dir)
{
	int ret, found;
	struct buffer_head *di_bh = NULL;
	struct ocfs2_dinode *di;
	struct ocfs2_inline_data *data;

	ret = ocfs2_read_block(OCFS2_SB(dir->i_sb), OCFS2_I(dir)->ip_blkno,
			       &di_bh, OCFS2_BH_CACHED, dir);
	if (ret) {
		mlog_errno(ret);
		goto out;
	}

	di = (struct ocfs2_dinode *)di_bh->b_data;
	data = &di->id2.i_data;

	found = ocfs2_search_dirblock(di_bh, dir, name, namelen, 0,
				      data->id_data, i_size_read(dir), res_dir);
	if (found == 1)
		return di_bh;

	brelse(di_bh);
out:
	return NULL;
}

static struct buffer_head *ocfs2_find_entry_el(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();

	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++;

				bh = ocfs2_bread(dir, b++, &err, 1);
				bh_use[ra_max] = bh;
			}
		}
		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 */
			ocfs2_error(dir->i_sb, "reading directory %llu, "
				    "offset %lu\n",
				    (unsigned long long)OCFS2_I(dir)->ip_blkno,
				    block);
			brelse(bh);
			goto next;
		}
		i = ocfs2_search_dirblock(bh, dir, name, namelen,
					  block << sb->s_blocksize_bits,
					  bh->b_data, sb->s_blocksize,
					  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;
}

/*
 * Try to find an entry of the provided name within 'dir'.
 *
 * If nothing was found, NULL is returned. Otherwise, a buffer_head
 * and pointer to the dir entry are passed back.
 *
 * Caller can NOT assume anything about the contents of the
 * buffer_head - it is passed back only so that it can be passed into
 * any one of the manipulation functions (add entry, delete entry,
 * etc). As an example, bh in the extent directory case is a data
 * block, in the inline-data case it actually points to an inode.
 */
struct buffer_head *ocfs2_find_entry(const char *name, int namelen,
				     struct inode *dir,
				     struct ocfs2_dir_entry **res_dir)
{
	*res_dir = NULL;

	if (OCFS2_I(dir)->ip_dyn_features & OCFS2_INLINE_DATA_FL)
		return ocfs2_find_entry_id(name, namelen, dir, res_dir);

	return ocfs2_find_entry_el(name, namelen, dir, res_dir);
}

/*
 * Update inode number and type of a previously found directory entry.
 */
int ocfs2_update_entry(struct inode *dir, handle_t *handle,
		       struct buffer_head *de_bh, struct ocfs2_dir_entry *de,
		       struct inode *new_entry_inode)
{
	int ret;

	/*
	 * The same code works fine for both inline-data and extent
	 * based directories, so no need to split this up.
	 */

	ret = ocfs2_journal_access(handle, dir, de_bh,
				   OCFS2_JOURNAL_ACCESS_WRITE);
	if (ret) {
		mlog_errno(ret);
		goto out;
	}

	de->inode = cpu_to_le64(OCFS2_I(new_entry_inode)->ip_blkno);
	ocfs2_set_de_type(de, new_entry_inode->i_mode);

	ocfs2_journal_dirty(handle, de_bh);

out:
	return ret;
}

static int __ocfs2_delete_entry(handle_t *handle, struct inode *dir,
				struct ocfs2_dir_entry *de_del,
				struct buffer_head *bh, char *first_de,
				unsigned int bytes)
{
	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 *) first_de;
	while (i < bytes) {
		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;
}

static inline int ocfs2_delete_entry_id(handle_t *handle,
					struct inode *dir,
					struct ocfs2_dir_entry *de_del,
					struct buffer_head *bh)
{
	int ret;
	struct buffer_head *di_bh = NULL;
	struct ocfs2_dinode *di;
	struct ocfs2_inline_data *data;

	ret = ocfs2_read_block(OCFS2_SB(dir->i_sb), OCFS2_I(dir)->ip_blkno,
			       &di_bh, OCFS2_BH_CACHED, dir);
	if (ret) {
		mlog_errno(ret);
		goto out;
	}

	di = (struct ocfs2_dinode *)di_bh->b_data;
	data = &di->id2.i_data;

	ret = __ocfs2_delete_entry(handle, dir, de_del, bh, data->id_data,
				   i_size_read(dir));

	brelse(di_bh);
out:
	return ret;
}

static inline int ocfs2_delete_entry_el(handle_t *handle,
					struct inode *dir,
					struct ocfs2_dir_entry *de_del,
					struct buffer_head *bh)
{
	return __ocfs2_delete_entry(handle, dir, de_del, bh, bh->b_data,
				    bh->b_size);
}

/*
 * ocfs2_delete_entry deletes a directory entry by merging it with the
 * previous entry
 */
int ocfs2_delete_entry(handle_t *handle,
		       struct inode *dir,
		       struct ocfs2_dir_entry *de_del,
		       struct buffer_head *bh)
{
	if (OCFS2_I(dir)->ip_dyn_features & OCFS2_INLINE_DATA_FL)
		return ocfs2_delete_entry_id(handle, dir, de_del, bh);

	return ocfs2_delete_entry_el(handle, dir, de_del, bh);
}

/*
 * Check whether 'de' has enough room to hold an entry of
 * 'new_rec_len' bytes.
 */
static inline int ocfs2_dirent_would_fit(struct ocfs2_dir_entry *de,
					 unsigned int new_rec_len)
{
	unsigned int de_really_used;

	/* Check whether this is an empty record with enough space */
	if (le64_to_cpu(de->inode) == 0 &&
	    le16_to_cpu(de->rec_len) >= new_rec_len)
		return 1;

	/*
	 * Record might have free space at the end which we can
	 * use.
	 */
	de_really_used = OCFS2_DIR_REC_LEN(de->name_len);
	if (le16_to_cpu(de->rec_len) >= (de_really_used + new_rec_len))
	    return 1;

	return 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.
 */
int __ocfs2_add_entry(handle_t *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 ocfs2_dinode *di = (struct ocfs2_dinode *)parent_fe_bh->b_data;
	struct super_block *sb = dir->i_sb;
	int retval, status;
	unsigned int size = sb->s_blocksize;
	char *data_start = insert_bh->b_data;

	mlog_entry_void();

	if (!namelen)
		return -EINVAL;

	if (OCFS2_I(dir)->ip_dyn_features & OCFS2_INLINE_DATA_FL) {
		data_start = di->id2.i_data.id_data;
		size = i_size_read(dir);

		BUG_ON(insert_bh != parent_fe_bh);
	}

	rec_len = OCFS2_DIR_REC_LEN(namelen);
	offset = 0;
	de = (struct ocfs2_dir_entry *) data_start;
	while (1) {
		BUG_ON((char *)de >= (size + data_start));

		/* 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 (ocfs2_dirent_would_fit(de, 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;
}