transaction.c

linux 内核源代码

		 *
		 * So, if we have a checkpoint on the buffer, we should
		 * now refile the buffer on our BJ_Forget list so that
		 * we know to remove the checkpoint after we commit.
		 */

		if (jh->b_cp_transaction) {
			__journal_temp_unlink_buffer(jh);
			__journal_file_buffer(jh, transaction, BJ_Forget);
		} else {
			__journal_unfile_buffer(jh);
			journal_remove_journal_head(bh);
			__brelse(bh);
			if (!buffer_jbd(bh)) {
				spin_unlock(&journal->j_list_lock);
				jbd_unlock_bh_state(bh);
				__bforget(bh);
				goto drop;
			}
		}
	} 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;
			drop_reserve = 1;
		}
	}

not_jbd:
	spin_unlock(&journal->j_list_lock);
	jbd_unlock_bh_state(bh);
	__brelse(bh);
drop:
	if (drop_reserve) {
		/* no need to reserve log space for this block -bzzz */
		handle->h_buffer_credits++;
	}
	return err;
}

/**
 * int journal_stop() - complete a transaction
 * @handle: tranaction to complete.
 *
 * 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;
	pid_t pid;

	J_ASSERT(journal_current_handle() == handle);

	if (is_handle_aborted(handle))
		err = -EIO;
	else {
		J_ASSERT(transaction->t_updates > 0);
		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...
	 *
	 * But don't do this if this process was the most recent one to
	 * perform a synchronous write.  We do this to detect the case where a
	 * single process is doing a stream of sync writes.  No point in waiting
	 * for joiners in that case.
	 */
	pid = current->pid;
	if (handle->h_sync && journal->j_last_sync_writer != pid) {
		journal->j_last_sync_writer = pid;
		do {
			old_handle_count = transaction->t_handle_count;
			schedule_timeout_uninterruptible(1);
		} while (old_handle_count != transaction->t_handle_count);
	}

	current->journal_info = NULL;
	spin_lock(&journal->j_state_lock);
	spin_lock(&transaction->t_handle_lock);
	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;

		spin_unlock(&transaction->t_handle_lock);
		jbd_debug(2, "transaction too old, requesting commit for "
					"handle %p\n", handle);
		/* This is non-blocking */
		__log_start_commit(journal, transaction->t_tid);
		spin_unlock(&journal->j_state_lock);

		/*
		 * Special case: JFS_SYNC synchronous updates require us
		 * to wait for the commit to complete.
		 */
		if (handle->h_sync && !(current->flags & PF_MEMALLOC))
			err = log_wait_commit(journal, tid);
	} else {
		spin_unlock(&transaction->t_handle_lock);
		spin_unlock(&journal->j_state_lock);
	}

	lock_release(&handle->h_lockdep_map, 1, _THIS_IP_);

	jbd_free_handle(handle);
	return err;
}

/**int journal_force_commit() - force any uncommitted transactions
 * @journal: journal to force
 *
 * For synchronous operations: force any uncommitted transactions
 * 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;

	handle = journal_start(journal, 1);
	if (IS_ERR(handle)) {
		ret = PTR_ERR(handle);
	} else {
		handle->h_sync = 1;
		ret = journal_stop(handle);
	}
	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.
 *
 * j_list_lock is held.
 *
 * jbd_lock_bh_state(jh2bh(jh)) 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 j_list_lock held, and the journal may not be locked.
 *
 * jbd_lock_bh_state(jh2bh(jh)) is held.
 */

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 = NULL;
	}
	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_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.
 *
 * Called under j_list_lock.  The journal may not be locked.
 */
static void __journal_temp_unlink_buffer(struct journal_head *jh)
{
	struct journal_head **list = NULL;
	transaction_t *transaction;
	struct buffer_head *bh = jh2bh(jh);

	J_ASSERT_JH(jh, jbd_is_locked_bh_state(bh));
	transaction = jh->b_transaction;
	if (transaction)
		assert_spin_locked(&transaction->t_journal->j_list_lock);

	J_ASSERT_JH(jh, jh->b_jlist < BJ_Types);
	if (jh->b_jlist != BJ_None)
		J_ASSERT_JH(jh, transaction != NULL);

	switch (jh->b_jlist) {
	case BJ_None:
		return;
	case BJ_SyncData:
		list = &transaction->t_sync_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;
	case BJ_Locked:
		list = &transaction->t_locked_list;
		break;
	}

	__blist_del_buffer(list, jh);
	jh->b_jlist = BJ_None;
	if (test_clear_buffer_jbddirty(bh))
		mark_buffer_dirty(bh);	/* Expose it to the VM */
}

void __journal_unfile_buffer(struct journal_head *jh)
{
	__journal_temp_unlink_buffer(jh);
	jh->b_transaction = NULL;
}

void journal_unfile_buffer(journal_t *journal, struct journal_head *jh)
{
	jbd_lock_bh_state(jh2bh(jh));
	spin_lock(&journal->j_list_lock);
	__journal_unfile_buffer(jh);
	spin_unlock(&journal->j_list_lock);
	jbd_unlock_bh_state(jh2bh(jh));
}

/*
 * Called from journal_try_to_free_buffers().
 *
 * Called under jbd_lock_bh_state(bh)
 */
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;
}


/**
 * int journal_try_to_free_buffers() - try to free page buffers.
 * @journal: journal for operation
 * @page: to try and free
 * @unused_gfp_mask: unused
 *
 *
 * 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.
 *
 * This function returns non-zero if we wish try_to_free_buffers()
 * to be called. We do this if the page is releasable by try_to_free_buffers().
 * We also do it if the page has locked or dirty buffers and the caller wants
 * us to perform sync or async writeout.
 *
 * 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.
 *
 * Who else is affected by this?  hmm...  Really the only contender
 * is do_get_write_access() - it could be looking at the buffer while
 * journal_try_to_free_buffer() is changing its state.  But that
 * cannot happen because we never reallocate freed data as metadata
 * while the data is part of a transaction.  Yes?
 */
int journal_try_to_free_buffers(journal_t *journal,
				struct page *page, gfp_t unused_gfp_mask)
{
	struct buffer_head *head;
	struct buffer_head *bh;
	int ret = 0;

	J_ASSERT(PageLocked(page));

	head = page_buffers(page);
	bh = head;
	do {
		struct journal_head *jh;

		/*
		 * We take our own ref against the journal_head here to avoid
		 * having to add tons of locking around each instance of
		 * journal_remove_journal_head() and journal_put_journal_head().
		 */
		jh = journal_grab_journal_head(bh);
		if (!jh)
			continue;

		jbd_lock_bh_state(bh);
		__journal_try_to_free_buffer(journal, bh);
		journal_put_journal_head(jh);
		jbd_unlock_bh_state(bh);
		if (buffer_jbd(bh))
			goto busy;
	} while ((bh = bh->b_this_page) != head);
	ret = try_to_free_buffers(page);
busy:
	return ret;