[Ocfs2-devel] [PATCH 08/18] ocfs2: Use metadata-specific ocfs2_journal_access_*() functions.

[Joel Becker]

On Wed, Dec 10, 2008 at 10:31:50AM +0800, Tao Ma wrote:
> Joel Becker wrote:
>> +static int ocfs2_path_bh_journal_access(handle_t *handle,
>> +					struct inode *inode,
>> +					struct ocfs2_path *path,
>> +					int idx)
>> +{
>> +	ocfs2_journal_access_func access = path_root_access(path);
>> +
>> +	if (!access)
>> +		access = ocfs2_journal_access;
>> +
>> +	if (idx)
>> +		access = ocfs2_journal_access_eb;
>> +
>> +	return access(handle, inode, path->p_node[idx].bh,
>> +		      OCFS2_JOURNAL_ACCESS_WRITE);
>>  }
> In most cases, idx>0 and access isn't NULL. So the following code would be 
> more efficient I guess.
>
> 	if (idx)
> 		access = ocfs2_journal_access_eb;
> 	else if (!access)
> 		access = ocfs2_journal_access;

	I like the way mine reads better.  This sort of efficiency isn't
needed outside of tight loops.

>> @@ -1918,25 +1966,23 @@ static int ocfs2_rotate_subtree_right(struct inode 
>> *inode,
>>  	root_bh = left_path->p_node[subtree_index].bh;
>>  	BUG_ON(root_bh != right_path->p_node[subtree_index].bh);
>>  -	ret = ocfs2_journal_access(handle, inode, root_bh,
>> -				   OCFS2_JOURNAL_ACCESS_WRITE);
>> +	ret = ocfs2_path_bh_journal_access(handle, inode, right_path,
>> +					   subtree_index);
>>  	if (ret) {
>>  		mlog_errno(ret);
>>  		goto out;
>>  	}
>>   	for(i = subtree_index + 1; i < path_num_items(right_path); i++) {
>> -		ret = ocfs2_journal_access(handle, inode,
>> -					   right_path->p_node[i].bh,
>> -					   OCFS2_JOURNAL_ACCESS_WRITE);
>> +		ret = ocfs2_path_bh_journal_access(handle, inode,
>> +						   right_path, i);
>>  		if (ret) {
>>  			mlog_errno(ret);
>>  			goto out;
>>  		}
>>  -		ret = ocfs2_journal_access(handle, inode,
>> -					   left_path->p_node[i].bh,
>> -					   OCFS2_JOURNAL_ACCESS_WRITE);
>> +		ret = ocfs2_path_bh_journal_access(handle, inode,
>> +						   left_path, i);
> I guess when can use ocfs2_journal_access_eb directly since it is 
> "subtree_index+1" now. No chance of be the root. The same goes for 
> ocfs2_rotate_subtree_left, ocfs2_merge_rec_right, ocfs2_merge_rec_left.

	But then you start dereferencing the path bh list.  That's
breaks the abstraction of the path structure.  It also drops the
consistency of always using ocfs2_path_bh_journal_access() for paths.
Conversely, there is no real loss to calling
ocfs2_path_bh_journal_access(); the extra function call is
insignificant.

>> @@ -4594,8 +4633,8 @@ int ocfs2_add_clusters_in_btree(struct ocfs2_super 
>> *osb,
>>  	BUG_ON(num_bits > clusters_to_add);
>>   	/* reserve our write early -- insert_extent may update the inode */
>> -	status = ocfs2_journal_access(handle, inode, et->et_root_bh,
>> -				      OCFS2_JOURNAL_ACCESS_WRITE);
>> +	status = ocfs2_et_root_journal_access(handle, inode, et,
>> +					      OCFS2_JOURNAL_ACCESS_WRITE);
> This is really a good chance for us to modify the comments also. ;)

	You mean something like 'may update the tree root'?

Joel

-- 

Life's Little Instruction Book #24

	"Drink champagne for no reason at all."

Joel Becker
Principal Software Developer
Oracle
E-mail: joel.becker at oracle.com
Phone: (650) 506-8127