This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: pretty-ipa merge 4: "process" cgraph node flag
> On Sat, Mar 28, 2009 at 12:15, Jan Hubicka <hubicka@ucw.cz> wrote:
>
> > OK, that sounds fine for me too. ?I still need to add profile update
> > code to new clonning infrastructure that I will probably write while
> > being offline at beggining of week.
>
> Sure. The merge that I'm more interested in is the removal of
> master_clone. LTO uses it to avoid writing the same body more than
> once, but when I ported the pretty-ipa patch, I ran into new failures
> (likely because I messed up the logic that deals with clones).
This is going to change again with clone infrastructure, where you have
clone_of pointer that is more or less same as pointer to master clone,
so you only need to save bodies with empty clone_of.
On current mainline you can probably go with cgraph_node (node->decl) ==
node
this will match only on first from the inlined node clones list.
Honza
>
>
> Diego.