On Thu, Feb 22, 2001 at 08:33:22PM +0530, amit shanbhag wrote:
> amit shanbhag wrote:
> Hi
>
> ill go step wise...
> 1.) I have registered a protcol using index ett_A
> 2.) I use it to create an item
> item=proto_tree_add_item(tree,proto_A,NullTVB,offset,END_OF_FRAME,FALSE);
> 3.) I then register a subtree in the main tree;
> A_tree=proto_item_add_subtree(item,ett_A);
>
> later on in another C file i do the following
>
> /*************************************************************************************
> *
>
> ************************************************************************************/
> 4.) ti = proto_tree_add_item(A_tree,NullTVB,offset,0,"B Data" );
> 5.) B_tree = proto_item_add_subtree(ti, ett_B;
>
> in the main C file i have acccess only to ett_B
> now i want to add some items to the B_treein the main file, But i only have ett_B
> with me, How do i get B_tree provided that i have with me A_tree and ett_B
>
> tree
> |
> --A_tree with index ett_a
> |
> --B_tree with index ett_b
>
Are you sure you have your modules defined correctly. Why is the A module
adding information to the B tree? Shouldn't all information about the B tree
be handled in the B module?
There is no way to retrieve a particular branch from the proto_tree
based on the ett_* value.
I guess you could take advantage of the fact that protocol dissection
is single threaded and set a global variable in your B module. Once processing
goes back to the A module, it could use that global variable to retrieve
the B_tree value. But do you really want to do that at the expense of
a non-clean design?
--gilbert
