RE: [plugins-ici-cloning-instrumentation] new GCC plugin developements

["Grigori Fursin" <grigori dot fursin at inria dot fr>]

Hi Joern,

Actually, I think I realized why you still see many differences in ICI 2.0 and GCC 4.5 -
my memory is not working well after being away for 2 months and I am trying to remember
what happened with ICI in summer - sorry about that.

So, if I remember correctly, when we saw the new GCC plugin branch open and that it overlaps
considerably with ICI we decided to change all our function names that had the same functionality 
as plugin branch to sync it later with plugin branch goes to the mainline (and that is happening
now in GCC 4.5) while we continued using baseline GCC 4.4.0 for all our further experiments.

In parallel Zbigniew prepared a patch that synchronized ICI with the plugin branch
and if I am correct, he removed all the duplicating functionality so the changes became
quite minimal. BUT we did not use that branch for GSOC projects, so what you see
in the gcc-4.4.0-ici2.0-adapt is the ICI with duplicate functionality of the old plugin branch
and normally the current mainline GCC 4.5. That's I think you start having conflict names.

I sent another private email to Zbigniew and you to check if I am right and if so,
I think you should look at the Zbigniew patch from June and try to update it according
to the current GCC 4.5 mainline plus add small extension from GSOC'09 about types of
the plugin parameters. After that we will synchronize GSOC'09 branches with this new
ICI and with the mainline GCC 4.5 but not versa!!!

Please, tell me if it makes sense. If someone thinks that I am still wrong (particularly
Zbigniew, Yuanjie and Liang) please respond to that so that we could decide on
our work plan ...

By the way, by accident I added 2 emails to the CC due to auto completing of emails:
announce@hipeac.net and ctuning-announce@googlegroups.com - I removed them
since they are not related to this discussion. Replying to this message is now fine
but if you reply to other messages, please remove those emails ...
I am sorry about that...

Cheers,
Grigori



> -----Original Message-----
> From: ctuning-discussions@googlegroups.com [mailto:ctuning-discussions@googlegroups.com] On
> Behalf Of Joern Rennecke
> Sent: Friday, November 06, 2009 6:14 PM
> To: Grigori Fursin
> Cc: 'Zbigniew Chamski'; 'Richard Guenther'; 'Basile STARYNKEVITCH'; 'Ian Lance Taylor'; 'GCC
> Mailing List'; 'Albert Cohen'; ctuning-discussions@googlegroups.com; 'Yuri Kashnikoff';
> Yuanjie Huang; Liang Peng; announce@hipeac.net; ctuning-announce@googlegroups.com;
> dorit@il.ibm.com; Mircea Namolaru; 'Diego Novillo'
> Subject: Re: [plugins-ici-cloning-instrumentation] new GCC plugin developements
> 
> 
> We got a name clash.
> In gcc 4.5, passes.c:register_pass is used to register a pass in a
> specific place in the pass tree.
> In gcc-4.4.0-ici, pass-manager.c:register_pass is used to register a
> pass by name.
> 
> --~--~---------~--~----~------------~-------~--~----~
> You received this message because you are subscribed to the Google Groups "ctuning-
> discussions" group.
> To post to this group, send email to ctuning-discussions@googlegroups.com
> To unsubscribe from this group, send email to ctuning-discussions+unsubscribe@googlegroups.com
> For more options, visit this group at http://groups.google.com/group/ctuning-discussions?hl=en
> -~----------~----~----~----~------~----~------~--~---