[Ocfs2-devel] [PATCH 04/15] Make extend allocation generic.

[Joel Becker]

On Fri, Aug 15, 2008 at 11:43:06PM +0800, Tao Ma wrote:
>>>>>>> +int ocfs2_add_clusters_in_btree(struct ocfs2_super *osb,
>>>>>>> +				struct inode *inode,
>>>>>>> +				u32 *logical_offset,
>>>>>>> +				u32 clusters_to_add,
>>>>>>> +				int mark_unwritten,
>>>>>>> +				struct buffer_head *root_bh,
>>>>>>> +				struct ocfs2_extent_list *root_el,
>>>>>>> +				handle_t *handle,
>>>>>>> +				struct ocfs2_alloc_context *data_ac,
>>>>>>> +				struct ocfs2_alloc_context *meta_ac,
>>>>>>> +				enum ocfs2_alloc_restarted *reason_ret,
>>>>>>> +				enum ocfs2_extent_tree_type type)

> But as for the original case which you ask me not to pass root_el, it  
> isn't suitable since root_el need less stack length and more readable  
> than we initiate a ocfs2_extent_tree here.

	That's a 12 argument function there.  That's ugly.  This isn't
about actual stack usage (you're right, we don't gain stack in my
scheme).  It's about readable functions.
	In fact, struct ocfs2_alloc_context exists so that Mark could
pass one _ac down to a function rather than 5 arguments.
	Why aren't you passing a struct ocfs2_extent_tree to this
function anyway?  You could do it in the caller, pass it down, and then
this function takes one *et rather than *root_bh+*root_el+type.  Just a
thought.
	Mark knows this code better than I, so if he jumps in and says
"I find this 12 argument function more readable than what you propose",
I will happily consider myself outvoted :-)

Joel

-- 

"If at first you don't succeed, cover all traces that you tried."
                                                        -Unknown

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