Re: [tree-ssa] [RFC] loop versioning

[Jan Hubicka <hubicka at ucw dot cz>]

> 
> On Jan 5, 2004, at 10:35 PM, Richard Henderson wrote:
> 
> >Anyway, you should know that there are already routines to copy
> >statements and blocks.  So all of this code looks redundant.
> 
> tree_duplicate_bb () is more suitable to copy loop header. For
> other basic blocks inside loop, while duplicating a basic block
> I may not have corresponding edge available. Because it is possible
> that bb at other end of edge does not yet have corresponding
> duplicate bb.  So what I do is, duplicate all basic blocks and
> then make edge for second loop by following edges of original loop.
> Plus there can be more than one incoming edges and tree_duplicate_bb()
> does not handle it.

In addition do basic block duplication code, we may have something like
duplicate_zone that will take a set of basic blocks and duplicate the
code together with it's internal edges.  I would like to see this
implemented at the top of duplicate_bb so we get it working for both RTL
and trees.  (duplcate_bb shall be virtualized) For RTL this is already
done in Zdenek's loop unroller code so one can use this as a base.

Honza
> 
> As I understand tree_duplicate_bb() performs this operation
> 
> before
>      ---- edge e1 ----> [bb1]
> 
> after
>      --- edge e1 --+    [bb1]
>                  |
>                  +---> [bb2]
> 
> What I need to do for loop versioning is:
> 
> before
>     ---- edge e1 ---> [bb1] ---- edge e2 --->
> after
>     ---- edge e1 ---> [bb1] ---- edge e2 --->
>     ---- edge e21 ---> [bb21] ---- edge e22 --->
> 
> 
> And comment on top of tree_duplicate_bb () says that it does not
> work over SSA. At the moment, I'm trying to make loop versioning
> work over SSA.
> 
> If you think, I should update tree_duplicate_bb () and related APIs
> to satisfy these requirements then I'll follow that path.
> 
> Thank you,
> -
> Devang