static void
__journal_try_to_free_buffer(journal_t *journal, struct buffer_head *bh)
{
	struct journal_head *jh;

	jh = bh2jh(bh);

	if (buffer_locked(bh) || buffer_dirty(bh))
		goto out;

	if (jh->b_next_transaction != NULL)
		goto out;

	spin_lock(&journal->j_list_lock);
	if (jh->b_transaction != NULL && jh->b_cp_transaction == NULL) {
		if (jh->b_jlist == BJ_SyncData || jh->b_jlist == BJ_Locked) {
			/* A written-back ordered data buffer */
			JBUFFER_TRACE(jh, "release data");
			__journal_unfile_buffer(jh);
			journal_remove_journal_head(bh);
			__brelse(bh);
		}
	} else if (jh->b_cp_transaction != NULL && jh->b_transaction == NULL) {
		/* written-back checkpointed metadata buffer */
		if (jh->b_jlist == BJ_None) {
			JBUFFER_TRACE(jh, "remove from checkpoint list");
			__journal_remove_checkpoint(jh);
			journal_remove_journal_head(bh);
			__brelse(bh);
		}
	}
	spin_unlock(&journal->j_list_lock);
out:
	return;
}

static void sync_buffers(kdev_t dev, int wait)
{
    register struct buffer_head *bh;

    for (bh = bh_chain; bh != NULL; bh = bh->b_next) {
	if (dev && bh->b_dev != dev)
	    continue;
	/*
	 *      Skip clean buffers.
	 */
	if (buffer_clean(bh))
	    continue;
	/*
	 *      Locked buffers..
	 *
	 *      Buffer is locked; skip it unless wait is requested
	 *      AND pass > 0.
	 */
	if (buffer_locked(bh) && wait)
	    continue;
	else
	    wait_on_buffer(bh);
	/*
	 *      Do the stuff
	 */
	bh->b_count++;
	ll_rw_blk(WRITE, bh);
	bh->b_count--;
    }
    return;
}

static void sprintf_buffer_head (char * buf, struct buffer_head * bh) 
{
  sprintf (buf, "dev %s, size %d, blocknr %ld, count %d, list %d, state 0x%lx, page %p, (%s, %s, %s)",
	   kdevname (bh->b_dev), bh->b_size, bh->b_blocknr, atomic_read (&(bh->b_count)), bh->b_list,
	   bh->b_state, bh->b_page,
	   buffer_uptodate (bh) ? "UPTODATE" : "!UPTODATE",
	   buffer_dirty (bh) ? "DIRTY" : "CLEAN",
	   buffer_locked (bh) ? "LOCKED" : "UNLOCKED");
}

/*
 * Try to flush one buffer from the checkpoint list to disk.
 *
 * Return 1 if something happened which requires us to abort the current
 * scan of the checkpoint list.  
 *
 * Called with j_list_lock held and drops it if 1 is returned
 * Called under jbd_lock_bh_state(jh2bh(jh)), and drops it
 */
static int __process_buffer(journal_t *journal, struct journal_head *jh,
			struct buffer_head **bhs, int *batch_count)
{
	struct buffer_head *bh = jh2bh(jh);
	int ret = 0;

	if (buffer_locked(bh)) {
		atomic_inc(&bh->b_count);
		spin_unlock(&journal->j_list_lock);
		jbd_unlock_bh_state(bh);
		wait_on_buffer(bh);
		/* the journal_head may have gone by now */
		BUFFER_TRACE(bh, "brelse");
		__brelse(bh);
		ret = 1;
	} else if (jh->b_transaction != NULL) {
		transaction_t *t = jh->b_transaction;
		tid_t tid = t->t_tid;

		spin_unlock(&journal->j_list_lock);
		jbd_unlock_bh_state(bh);
		log_start_commit(journal, tid);
		log_wait_commit(journal, tid);
		ret = 1;
	} else if (!buffer_dirty(bh)) {
		J_ASSERT_JH(jh, !buffer_jbddirty(bh));
		BUFFER_TRACE(bh, "remove from checkpoint");
		__journal_remove_checkpoint(jh);
		spin_unlock(&journal->j_list_lock);
		jbd_unlock_bh_state(bh);
		journal_remove_journal_head(bh);
		__brelse(bh);
		ret = 1;
	} else {
		/*
		 * Important: we are about to write the buffer, and
		 * possibly block, while still holding the journal lock.
		 * We cannot afford to let the transaction logic start
		 * messing around with this buffer before we write it to
		 * disk, as that would break recoverability.  
		 */
		BUFFER_TRACE(bh, "queue");
		get_bh(bh);
		J_ASSERT_BH(bh, !buffer_jwrite(bh));
		set_buffer_jwrite(bh);
		bhs[*batch_count] = bh;
		__buffer_relink_io(jh);
		jbd_unlock_bh_state(bh);
		(*batch_count)++;
		if (*batch_count == NR_BATCH) {
			spin_unlock(&journal->j_list_lock);
			__flush_batch(journal, bhs, batch_count);
			ret = 1;
		}
	}
	return ret;
}

static void lock_journal_bh_wait(struct buffer_head *bh, journal_t *journal)
{
repeat:
	wait_on_buffer(bh);
	lock_journal(journal);
	if (buffer_locked(bh)) {
		unlock_journal(journal);
		goto repeat;
	}
}

static int do_readahead(journal_t *journal, unsigned int start)
{
	int err;
	unsigned int max, nbufs, next, blocknr;
	struct buffer_head *bh;
	
	struct buffer_head * bufs[MAXBUF];
	
	/* Do up to 128K of readahead */
	max = start + (128 * 1024 / journal->j_blocksize);
	if (max > journal->j_maxlen)
		max = journal->j_maxlen;

	/* Do the readahead itself.  We'll submit MAXBUF buffer_heads at
	 * a time to the block device IO layer. */
	
	nbufs = 0;
	
	for (next = start; next < max; next++) {
		blocknr = next;
		if (journal->j_inode)
			blocknr = bmap(journal->j_inode, next);
		if (!blocknr) {
			printk (KERN_ERR "JFS: bad block at offset %u\n",
				next);
			err = -EIO;
			goto failed;
		}
		
		bh = getblk(journal->j_dev, blocknr, journal->j_blocksize);
		if (!bh) {
			err = -ENOMEM;
			goto failed;
		}

		if (!buffer_uptodate(bh) && !buffer_locked(bh)) {
			bufs[nbufs++] = bh;
			if (nbufs == MAXBUF) {
				ll_rw_block(READ, nbufs, bufs);
				brelse_array(bufs, nbufs);
				nbufs = 0;
			}
		} else
			brelse(bh);
	}

	if (nbufs)
		ll_rw_block(READ, nbufs, bufs);
	err = 0;
	
failed:	
	if (nbufs) 
		brelse_array(bufs, nbufs);
	return err;
}

void wait_on_buffer(register struct buffer_head *bh)
{
    if (buffer_locked(bh)) {

	bh->b_count++;

	wait_set(&bh->b_wait);

	goto chk_buf;
	do {
	    current->state = TASK_UNINTERRUPTIBLE;
	    schedule();
	    current->state = TASK_RUNNING;
    chk_buf:
	    ;
	} while(buffer_locked(bh));

	wait_clear(&bh->b_wait);
	bh->b_count--;
    }
}

/*
 * Try to release a checkpointed buffer from its transaction.
 * Returns 1 if we released it and 2 if we also released the
 * whole transaction.
 *
 * Requires j_list_lock
 */
static int __try_to_free_cp_buf(struct journal_head *jh)
{
	int ret = 0;
	struct buffer_head *bh = jh2bh(jh);

	if (jh->b_transaction == NULL && !buffer_locked(bh) &&
	    !buffer_dirty(bh) && !buffer_write_io_error(bh)) {
		JBUFFER_TRACE(jh, "remove from checkpoint list");
		ret = __jbd2_journal_remove_checkpoint(jh) + 1;
	}
	return ret;
}

static int __process_buffer(journal_t *journal, struct journal_head *jh,
                            int *batch_count, transaction_t *transaction)
{
    struct buffer_head *bh = jh2bh(jh);
    int ret = 0;

    if (buffer_locked(bh)) {
        get_bh(bh);
        spin_unlock(&journal->j_list_lock);
        wait_on_buffer(bh);

        BUFFER_TRACE(bh, "brelse");
        __brelse(bh);
        ret = 1;
    } else if (jh->b_transaction != NULL) {
        transaction_t *t = jh->b_transaction;
        tid_t tid = t->t_tid;

        transaction->t_chp_stats.cs_forced_to_close++;
        spin_unlock(&journal->j_list_lock);
        if (unlikely(journal->j_flags & JBD2_UNMOUNT))
            printk(KERN_ERR "JBD2: %s: "
                   "Waiting for Godot: block %llu\n",
                   journal->j_devname,
                   (unsigned long long) bh->b_blocknr);
        jbd2_log_start_commit(journal, tid);
        jbd2_log_wait_commit(journal, tid);
        ret = 1;
    } else if (!buffer_dirty(bh)) {
        ret = 1;
        if (unlikely(buffer_write_io_error(bh)))
            ret = -EIO;
        get_bh(bh);
        BUFFER_TRACE(bh, "remove from checkpoint");
        __jbd2_journal_remove_checkpoint(jh);
        spin_unlock(&journal->j_list_lock);
        __brelse(bh);
    } else {
        BUFFER_TRACE(bh, "queue");
        get_bh(bh);
        J_ASSERT_BH(bh, !buffer_jwrite(bh));
        journal->j_chkpt_bhs[*batch_count] = bh;
        __buffer_relink_io(jh);
        transaction->t_chp_stats.cs_written++;
        (*batch_count)++;
        if (*batch_count == JBD2_NR_BATCH) {
            spin_unlock(&journal->j_list_lock);
            __flush_batch(journal, batch_count);
            ret = 1;
        }
    }
    return ret;
}

static int do_readahead(journal_t *journal, unsigned int start)
{
	int err;
	unsigned int max, nbufs, next;
	unsigned long long blocknr;
	struct buffer_head *bh;

	struct buffer_head * bufs[MAXBUF];

	
	max = start + (128 * 1024 / journal->j_blocksize);
	if (max > journal->j_maxlen)
		max = journal->j_maxlen;


	nbufs = 0;

	for (next = start; next < max; next++) {
		err = jbd2_journal_bmap(journal, next, &blocknr);

		if (err) {
			printk(KERN_ERR "JBD2: bad block at offset %u\n",
				next);
			goto failed;
		}

		bh = __getblk(journal->j_dev, blocknr, journal->j_blocksize);
		if (!bh) {
			err = -ENOMEM;
			goto failed;
		}

		if (!buffer_uptodate(bh) && !buffer_locked(bh)) {
			bufs[nbufs++] = bh;
			if (nbufs == MAXBUF) {
				ll_rw_block(READ, nbufs, bufs);
				journal_brelse_array(bufs, nbufs);
				nbufs = 0;
			}
		} else
			brelse(bh);
	}

	if (nbufs)
		ll_rw_block(READ, nbufs, bufs);
	err = 0;

failed:
	if (nbufs)
		journal_brelse_array(bufs, nbufs);
	return err;
}

static void sprintf_buffer_head(char *buf, struct buffer_head *bh)
{
	char b[BDEVNAME_SIZE];

	sprintf(buf,
		"dev %s, size %zd, blocknr %llu, count %d, state 0x%lx, page %p, (%s, %s, %s)",
		bdevname(bh->b_bdev, b), bh->b_size,
		(unsigned long long)bh->b_blocknr, atomic_read(&(bh->b_count)),
		bh->b_state, bh->b_page,
		buffer_uptodate(bh) ? "UPTODATE" : "!UPTODATE",
		buffer_dirty(bh) ? "DIRTY" : "CLEAN",
		buffer_locked(bh) ? "LOCKED" : "UNLOCKED");
}

static int nilfs_mdt_read_block(struct inode *inode, unsigned long block,
				int readahead, struct buffer_head **out_bh)
{
	struct buffer_head *first_bh, *bh;
	unsigned long blkoff;
	int i, nr_ra_blocks = NILFS_MDT_MAX_RA_BLOCKS;
	int err;

	err = nilfs_mdt_submit_block(inode, block, REQ_OP_READ, 0, &first_bh);
	if (err == -EEXIST) /* internal code */
		goto out;

	if (unlikely(err))
		goto failed;

	if (readahead) {
		blkoff = block + 1;
		for (i = 0; i < nr_ra_blocks; i++, blkoff++) {
			err = nilfs_mdt_submit_block(inode, blkoff, REQ_OP_READ,
						     REQ_RAHEAD, &bh);
			if (likely(!err || err == -EEXIST))
				brelse(bh);
			else if (err != -EBUSY)
				break;
				/* abort readahead if bmap lookup failed */
			if (!buffer_locked(first_bh))
				goto out_no_wait;
		}
	}

	wait_on_buffer(first_bh);

 out_no_wait:
	err = -EIO;
	if (!buffer_uptodate(first_bh)) {
		nilfs_msg(inode->i_sb, KERN_ERR,
			  "I/O error reading meta-data file (ino=%lu, block-offset=%lu)",
			  inode->i_ino, block);
		goto failed_bh;
	}
 out:
	*out_bh = first_bh;
	return 0;

 failed_bh:
	brelse(first_bh);
 failed:
	return err;
}

int jbd2_journal_get_create_access(handle_t *handle, struct buffer_head *bh)
{
	transaction_t *transaction = handle->h_transaction;
	journal_t *journal = transaction->t_journal;
	struct journal_head *jh = jbd2_journal_add_journal_head(bh);
	int err;

	jbd_debug(5, "journal_head %p\n", jh);
	err = -EROFS;
	if (is_handle_aborted(handle))
		goto out;
	err = 0;

	JBUFFER_TRACE(jh, "entry");
	jbd_lock_bh_state(bh);
	spin_lock(&journal->j_list_lock);
	J_ASSERT_JH(jh, (jh->b_transaction == transaction ||
		jh->b_transaction == NULL ||
		(jh->b_transaction == journal->j_committing_transaction &&
			  jh->b_jlist == BJ_Forget)));

	J_ASSERT_JH(jh, jh->b_next_transaction == NULL);
	J_ASSERT_JH(jh, buffer_locked(jh2bh(jh)));

	if (jh->b_transaction == NULL) {
		clear_buffer_dirty(jh2bh(jh));
		
		jh->b_modified = 0;

		JBUFFER_TRACE(jh, "file as BJ_Reserved");
		__jbd2_journal_file_buffer(jh, transaction, BJ_Reserved);
	} else if (jh->b_transaction == journal->j_committing_transaction) {
		
		jh->b_modified = 0;

		JBUFFER_TRACE(jh, "set next transaction");
		jh->b_next_transaction = transaction;
	}
	spin_unlock(&journal->j_list_lock);
	jbd_unlock_bh_state(bh);

	JBUFFER_TRACE(jh, "cancelling revoke");
	jbd2_journal_cancel_revoke(handle, jh);
out:
	jbd2_journal_put_journal_head(jh);
	return err;
}

/*
 * Try to release a checkpointed buffer from its transaction.
 * Returns 1 if we released it and 2 if we also released the
 * whole transaction.
 *
 * Requires j_list_lock
 * Called under jbd_lock_bh_state(jh2bh(jh)), and drops it
 */
static int __try_to_free_cp_buf(struct journal_head *jh)
{
	int ret = 0;
	struct buffer_head *bh = jh2bh(jh);

	if (jh->b_jlist == BJ_None && !buffer_locked(bh) && !buffer_dirty(bh)) {
		JBUFFER_TRACE(jh, "remove from checkpoint list");
		ret = __jbd2_journal_remove_checkpoint(jh) + 1;
		jbd_unlock_bh_state(bh);
		jbd2_journal_remove_journal_head(bh);
		BUFFER_TRACE(bh, "release");
		__brelse(bh);
	} else {
		jbd_unlock_bh_state(bh);
	}
	return ret;
}

/*
 * Try to release a checkpointed buffer from its transaction.
 * Returns 1 if we released it.
 * Requires journal_datalist_lock
 */
static int __try_to_free_cp_buf(struct journal_head *jh)
{
	int ret = 0;
	struct buffer_head *bh = jh2bh(jh);

	if (jh->b_jlist == BJ_None && !buffer_locked(bh) && !buffer_dirty(bh)) {
		JBUFFER_TRACE(jh, "remove from checkpoint list");
		__journal_remove_checkpoint(jh);
		__journal_remove_journal_head(bh);
		BUFFER_TRACE(bh, "release");
		/* BUF_LOCKED -> BUF_CLEAN (fwiw) */
		refile_buffer(bh);
		__brelse(bh);
		ret = 1;
	}
	return ret;
}

/*
 * Try to flush one buffer from the checkpoint list to disk.
 *
 * Return 1 if something happened which requires us to abort the current
 * scan of the checkpoint list.  
 *
 * Called with j_list_lock held.
 * Called under jbd_lock_bh_state(jh2bh(jh)), and drops it
 */
static int __flush_buffer(journal_t *journal, struct journal_head *jh,
			struct buffer_head **bhs, int *batch_count,
			int *drop_count)
{
	struct buffer_head *bh = jh2bh(jh);
	int ret = 0;

	if (buffer_dirty(bh) && !buffer_locked(bh) && jh->b_jlist == BJ_None) {
		J_ASSERT_JH(jh, jh->b_transaction == NULL);

		/*
		 * Important: we are about to write the buffer, and
		 * possibly block, while still holding the journal lock.
		 * We cannot afford to let the transaction logic start
		 * messing around with this buffer before we write it to
		 * disk, as that would break recoverability.  
		 */
		BUFFER_TRACE(bh, "queue");
		get_bh(bh);
		J_ASSERT_BH(bh, !buffer_jwrite(bh));
		set_buffer_jwrite(bh);
		bhs[*batch_count] = bh;
		jbd_unlock_bh_state(bh);
		(*batch_count)++;
		if (*batch_count == NR_BATCH) {
			__flush_batch(journal, bhs, batch_count);
			ret = 1;
		}
	} else {
		int last_buffer = 0;
		if (jh->b_cpnext == jh) {
			/* We may be about to drop the transaction.  Tell the
			 * caller that the lists have changed.
			 */
			last_buffer = 1;
		}
		if (__try_to_free_cp_buf(jh)) {
			(*drop_count)++;
			ret = last_buffer;
		}
	}
	return ret;
}

/*
 * Clean up transaction's list of buffers submitted for io.
 * We wait for any pending IO to complete and remove any clean
 * buffers. Note that we take the buffers in the opposite ordering
 * from the one in which they were submitted for IO.
 *
 * Called with j_list_lock held.
 */
static void __wait_cp_io(journal_t *journal, transaction_t *transaction)
{
	struct journal_head *jh;
	struct buffer_head *bh;
	tid_t this_tid;
	int released = 0;

	this_tid = transaction->t_tid;
restart:
	/* Did somebody clean up the transaction in the meanwhile? */
	if (journal->j_checkpoint_transactions != transaction ||
			transaction->t_tid != this_tid)
		return;
	while (!released && transaction->t_checkpoint_io_list) {
		jh = transaction->t_checkpoint_io_list;
		bh = jh2bh(jh);
		if (!jbd_trylock_bh_state(bh)) {
			jbd_sync_bh(journal, bh);
			spin_lock(&journal->j_list_lock);
			goto restart;
		}
		if (buffer_locked(bh)) {
			atomic_inc(&bh->b_count);
			spin_unlock(&journal->j_list_lock);
			jbd_unlock_bh_state(bh);
			wait_on_buffer(bh);
			/* the journal_head may have gone by now */
			BUFFER_TRACE(bh, "brelse");
			__brelse(bh);
			spin_lock(&journal->j_list_lock);
			goto restart;
		}
		/*
		 * Now in whatever state the buffer currently is, we know that
		 * it has been written out and so we can drop it from the list
		 */
		released = __journal_remove_checkpoint(jh);
		jbd_unlock_bh_state(bh);
		journal_remove_journal_head(bh);
		__brelse(bh);
	}
}

static int nilfs_mdt_read_block(struct inode *inode, unsigned long block,
				struct buffer_head **out_bh)
{
	struct buffer_head *first_bh, *bh;
	unsigned long blkoff;
	int i, nr_ra_blocks = NILFS_MDT_MAX_RA_BLOCKS;
	int err;

	err = nilfs_mdt_submit_block(inode, block, READ, &first_bh);
	if (err == -EEXIST) /* internal code */
		goto out;

	if (unlikely(err))
		goto failed;

	blkoff = block + 1;
	for (i = 0; i < nr_ra_blocks; i++, blkoff++) {
		err = nilfs_mdt_submit_block(inode, blkoff, READA, &bh);
		if (likely(!err || err == -EEXIST))
			brelse(bh);
		else if (err != -EBUSY)
			break; /* abort readahead if bmap lookup failed */

		if (!buffer_locked(first_bh))
			goto out_no_wait;
	}

	wait_on_buffer(first_bh);

 out_no_wait:
	err = -EIO;
	if (!buffer_uptodate(first_bh))
		goto failed_bh;
 out:
	*out_bh = first_bh;
	return 0;

 failed_bh:
	brelse(first_bh);
 failed:
	return err;
}

static int __wait_cp_io(journal_t *journal, transaction_t *transaction)
{
	struct journal_head *jh;
	struct buffer_head *bh;
	tid_t this_tid;
	int released = 0;
	int ret = 0;

	this_tid = transaction->t_tid;
restart:
	/*                                                         */
	if (journal->j_checkpoint_transactions != transaction ||
			transaction->t_tid != this_tid)
		return ret;
	while (!released && transaction->t_checkpoint_io_list) {
		jh = transaction->t_checkpoint_io_list;
		bh = jh2bh(jh);
		get_bh(bh);
		if (buffer_locked(bh)) {
			spin_unlock(&journal->j_list_lock);
			wait_on_buffer(bh);
			/*                                       */
			BUFFER_TRACE(bh, "brelse");
			__brelse(bh);
			spin_lock(&journal->j_list_lock);
			goto restart;
		}
		if (unlikely(buffer_write_io_error(bh)))
			ret = -EIO;

		/*
                                                                
                                                                
   */
		released = __jbd2_journal_remove_checkpoint(jh);
		__brelse(bh);
	}

	return ret;
}

static void do_stram_request( void )
{
	unsigned long start, len;

	while( CURRENT ) {
		if (MAJOR(CURRENT->rq_dev) != MAJOR_NR)
			panic("stram: request list destroyed");
		if (CURRENT->bh) {
			if (!buffer_locked(CURRENT->bh))
				panic("stram: block not locked");
		}
		
		start = swap_start + (CURRENT->sector << 9);
		len   = CURRENT->current_nr_sectors << 9;
		if ((start + len) > swap_end) {
			printk( KERN_ERR "stram: bad access beyond end of device: "
					"block=%ld, count=%ld\n",
					CURRENT->sector,
					CURRENT->current_nr_sectors );
			end_request( 0 );
			continue;
		}

		if (CURRENT->cmd == READ) {
			memcpy( CURRENT->buffer, (char *)start, len );
#ifdef DO_PROC
			stat_swap_read += N_PAGES(len);
#endif
		}
		else {
			memcpy( (char *)start, CURRENT->buffer, len );
#ifdef DO_PROC
			stat_swap_write += N_PAGES(len);
#endif
		}
		end_request( 1 );
	}
}