transaction.c

嵌入式系统设计与实例开发实验教材二源码 多线程应用程序设计 串行端口程序设计 AD接口实验 CAN总线通信实验 GPS通信实验 Linux内核移植与编译实验 IC卡读写实验 SD驱动使

			__journal_remove_journal_head(bh);
			__brelse(bh);
			if (!buffer_jbd(bh)) {
				spin_unlock(&journal_datalist_lock);
				unlock_journal(journal);
				__bforget(bh);
				return;
			}
		}
		
	} else if (jh->b_transaction) {
		J_ASSERT_JH(jh, (jh->b_transaction == 
				 journal->j_committing_transaction));
		/* However, if the buffer is still owned by a prior
		 * (committing) transaction, we can't drop it yet... */
		JBUFFER_TRACE(jh, "belongs to older transaction");
		/* ... but we CAN drop it from the new transaction if we
		 * have also modified it since the original commit. */

		if (jh->b_next_transaction) {
			J_ASSERT(jh->b_next_transaction == transaction);
			jh->b_next_transaction = NULL;
		}
	}

not_jbd:
	spin_unlock(&journal_datalist_lock);
	unlock_journal(journal);
	__brelse(bh);
	return;
}

#if 0	/* Unused */
/*
 * journal_sync_buffer: flush a potentially-journaled buffer to disk.
 *
 * Used for O_SYNC filesystem operations.  If the buffer is journaled,
 * we need to complete the O_SYNC by waiting for the transaction to
 * complete.  It is an error to call journal_sync_buffer before
 * journal_stop!
 */

void journal_sync_buffer(struct buffer_head *bh)
{
	transaction_t *transaction;
	journal_t *journal;
	long sequence;
	struct journal_head *jh;

	/* If the buffer isn't journaled, this is easy: just sync it to
	 * disk.  */
	BUFFER_TRACE(bh, "entry");

	spin_lock(&journal_datalist_lock);
	if (!buffer_jbd(bh)) {
		spin_unlock(&journal_datalist_lock);
		return;
	}
	jh = bh2jh(bh);
	if (jh->b_transaction == NULL) {
		/* If the buffer has already been journaled, then this
		 * is a noop. */
		if (jh->b_cp_transaction == NULL) {
			spin_unlock(&journal_datalist_lock);
			return;
		}
		atomic_inc(&bh->b_count);
		spin_unlock(&journal_datalist_lock);
		ll_rw_block (WRITE, 1, &bh);
		wait_on_buffer(bh);
		__brelse(bh);
		goto out;
	}
	
	/* Otherwise, just wait until the transaction is synced to disk. */
	transaction = jh->b_transaction;
	journal = transaction->t_journal;
	sequence = transaction->t_tid;
	spin_unlock(&journal_datalist_lock);

	jbd_debug(2, "requesting commit for jh %p\n", jh);
	log_start_commit (journal, transaction);
	
	while (tid_gt(sequence, journal->j_commit_sequence)) {
		wake_up(&journal->j_wait_done_commit);
		sleep_on(&journal->j_wait_done_commit);
	}
	JBUFFER_TRACE(jh, "exit");
out:
	return;
}
#endif

/*
 * All done for a particular handle.
 *
 * There is not much action needed here.  We just return any remaining
 * buffer credits to the transaction and remove the handle.  The only
 * complication is that we need to start a commit operation if the
 * filesystem is marked for synchronous update.
 *
 * journal_stop itself will not usually return an error, but it may
 * do so in unusual circumstances.  In particular, expect it to 
 * return -EIO if a journal_abort has been executed since the
 * transaction began.
 */

int journal_stop(handle_t *handle)
{
	transaction_t *transaction = handle->h_transaction;
	journal_t *journal = transaction->t_journal;
	int old_handle_count, err;
	
	if (!handle)
		return 0;

	J_ASSERT (transaction->t_updates > 0);
	J_ASSERT (journal_current_handle() == handle);
	
	if (is_handle_aborted(handle))
		err = -EIO;
	else
		err = 0;
	
	if (--handle->h_ref > 0) {
		jbd_debug(4, "h_ref %d -> %d\n", handle->h_ref + 1,
			  handle->h_ref);
		return err;
	}

	jbd_debug(4, "Handle %p going down\n", handle);

	/*
	 * Implement synchronous transaction batching.  If the handle
	 * was synchronous, don't force a commit immediately.  Let's
	 * yield and let another thread piggyback onto this transaction.
	 * Keep doing that while new threads continue to arrive.
	 * It doesn't cost much - we're about to run a commit and sleep
	 * on IO anyway.  Speeds up many-threaded, many-dir operations
	 * by 30x or more...
	 */
	if (handle->h_sync) {
		do {
			old_handle_count = transaction->t_handle_count;
			set_current_state(TASK_RUNNING);
			current->policy |= SCHED_YIELD;
			schedule();
		} while (old_handle_count != transaction->t_handle_count);
	}

	current->journal_info = NULL;
	transaction->t_outstanding_credits -= handle->h_buffer_credits;
	transaction->t_updates--;
	if (!transaction->t_updates) {
		wake_up(&journal->j_wait_updates);
		if (journal->j_barrier_count)
			wake_up(&journal->j_wait_transaction_locked);
	}

	/* 
	 * If the handle is marked SYNC, we need to set another commit
	 * going!  We also want to force a commit if the current
	 * transaction is occupying too much of the log, or if the
	 * transaction is too old now.
	 */
	if (handle->h_sync ||
			transaction->t_outstanding_credits >
				journal->j_max_transaction_buffers ||
	    		time_after_eq(jiffies, transaction->t_expires)) {
		/* Do this even for aborted journals: an abort still
		 * completes the commit thread, it just doesn't write
		 * anything to disk. */
		tid_t tid = transaction->t_tid;
		
		jbd_debug(2, "transaction too old, requesting commit for "
					"handle %p\n", handle);
		/* This is non-blocking */
		log_start_commit(journal, transaction);
		
		/*
		 * Special case: JFS_SYNC synchronous updates require us
		 * to wait for the commit to complete.  
		 */
		if (handle->h_sync && !(current->flags & PF_MEMALLOC))
			log_wait_commit(journal, tid);
	}
	kfree(handle);
	return err;
}

/*
 * For synchronous operations: force any uncommitted trasnactions
 * to disk.  May seem kludgy, but it reuses all the handle batching
 * code in a very simple manner.
 */
int journal_force_commit(journal_t *journal)
{
	handle_t *handle;
	int ret = 0;

	lock_kernel();
	handle = journal_start(journal, 1);
	if (IS_ERR(handle)) {
		ret = PTR_ERR(handle);
		goto out;
	}
	handle->h_sync = 1;
	journal_stop(handle);
out:
	unlock_kernel();
	return ret;
}

/*
 *
 * List management code snippets: various functions for manipulating the
 * transaction buffer lists.
 *
 */

/*
 * Append a buffer to a transaction list, given the transaction's list head
 * pointer.
 * journal_datalist_lock is held.
 */

static inline void 
__blist_add_buffer(struct journal_head **list, struct journal_head *jh)
{
	if (!*list) {
		jh->b_tnext = jh->b_tprev = jh;
		*list = jh;
	} else {
		/* Insert at the tail of the list to preserve order */
		struct journal_head *first = *list, *last = first->b_tprev;
		jh->b_tprev = last;
		jh->b_tnext = first;
		last->b_tnext = first->b_tprev = jh;
	}
}

/* 
 * Remove a buffer from a transaction list, given the transaction's list
 * head pointer.
 *
 * Called with journal_datalist_lock held, and the journal may not
 * be locked.
 */

static inline void
__blist_del_buffer(struct journal_head **list, struct journal_head *jh)
{
	if (*list == jh) {
		*list = jh->b_tnext;
		if (*list == jh)
			*list = 0;
	}
	jh->b_tprev->b_tnext = jh->b_tnext;
	jh->b_tnext->b_tprev = jh->b_tprev;
}

/* 
 * Remove a buffer from the appropriate transaction list.
 *
 * Note that this function can *change* the value of
 * bh->b_transaction->t_sync_datalist, t_async_datalist, t_buffers, t_forget,
 * t_iobuf_list, t_shadow_list, t_log_list or t_reserved_list.  If the caller
 * is holding onto a copy of one of thee pointers, it could go bad.
 * Generally the caller needs to re-read the pointer from the transaction_t.
 *
 * If bh->b_jlist is BJ_SyncData or BJ_AsyncData then we may have been called
 * via journal_try_to_free_buffer() or journal_clean_data_list().  In that
 * case, journal_datalist_lock will be held, and the journal may not be locked.
 */
void __journal_unfile_buffer(struct journal_head *jh)
{
	struct journal_head **list = 0;
	transaction_t * transaction;

	assert_spin_locked(&journal_datalist_lock);
	transaction = jh->b_transaction;

#ifdef __SMP__
	J_ASSERT (current->lock_depth >= 0);
#endif
	J_ASSERT_JH(jh, jh->b_jlist < BJ_Types);

	if (jh->b_jlist != BJ_None)
		J_ASSERT_JH(jh, transaction != 0);

	switch (jh->b_jlist) {
	case BJ_None:
		return;
	case BJ_SyncData:
		list = &transaction->t_sync_datalist;
		break;
	case BJ_AsyncData:
		list = &transaction->t_async_datalist;
		break;
	case BJ_Metadata:
		transaction->t_nr_buffers--;
		J_ASSERT_JH(jh, transaction->t_nr_buffers >= 0);
		list = &transaction->t_buffers;
		break;
	case BJ_Forget:
		list = &transaction->t_forget;
		break;
	case BJ_IO:
		list = &transaction->t_iobuf_list;
		break;
	case BJ_Shadow:
		list = &transaction->t_shadow_list;
		break;
	case BJ_LogCtl:
		list = &transaction->t_log_list;
		break;
	case BJ_Reserved:
		list = &transaction->t_reserved_list;
		break;
	}
	
	__blist_del_buffer(list, jh);
	jh->b_jlist = BJ_None;
	if (test_and_clear_bit(BH_JBDDirty, &jh2bh(jh)->b_state)) {
		set_bit(BH_Dirty, &jh2bh(jh)->b_state);
	}
}

void journal_unfile_buffer(struct journal_head *jh)
{
	spin_lock(&journal_datalist_lock);
	__journal_unfile_buffer(jh);
	spin_unlock(&journal_datalist_lock);
}

/*
 * Called from journal_try_to_free_buffers().  The journal is not
 * locked. lru_list_lock is not held.
 *
 * Here we see why journal_datalist_lock is global and not per-journal.
 * We cannot get back to this buffer's journal pointer without locking
 * out journal_clean_data_list() in some manner.
 *
 * One could use journal_datalist_lock to get unracy access to a
 * per-journal lock.
 *
 * Called with journal_datalist_lock held.
 *
 * Returns non-zero iff we were able to free the journal_head.
 */
static int __journal_try_to_free_buffer(struct buffer_head *bh,
					int *locked_or_dirty)
{
	struct journal_head *jh;

	assert_spin_locked(&journal_datalist_lock);

	jh = bh2jh(bh);

	if (buffer_locked(bh) || buffer_dirty(bh)) {
		*locked_or_dirty = 1;
		goto out;
	}

	if (!buffer_uptodate(bh))
		goto out;

	if (jh->b_next_transaction != 0)
		goto out;

	if (jh->b_transaction != 0 && jh->b_cp_transaction == 0) {
		if (jh->b_jlist == BJ_SyncData || jh->b_jlist==BJ_AsyncData) {
			/* A written-back ordered data buffer */
			JBUFFER_TRACE(jh, "release data");
			__journal_unfile_buffer(jh);
			jh->b_transaction = 0;
			__journal_remove_journal_head(bh);
			__brelse(bh);
		}
	}
	else if (jh->b_cp_transaction != 0 && jh->b_transaction == 0) {
		/* 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);
		}
	}
	return !buffer_jbd(bh);

out:
	return 0;
}

/*
 * journal_try_to_free_buffers().  For all the buffers on this page,
 * if they are fully written out ordered data, move them onto BUF_CLEAN
 * so try_to_free_buffers() can reap them.  Called with lru_list_lock
 * not held.  Does its own locking.
 *
 * This complicates JBD locking somewhat.  We aren't protected by the
 * BKL here.  We wish to remove the buffer from its committing or
 * running transaction's ->t_datalist via __journal_unfile_buffer.
 *
 * This may *change* the value of transaction_t->t_datalist, so anyone
 * who looks at t_datalist needs to lock against this function.
 *
 * Even worse, someone may be doing a journal_dirty_data on this
 * buffer.  So we need to lock against that.  journal_dirty_data()
 * will come out of the lock with the buffer dirty, which makes it
 * ineligible for release here.