This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH][LTO] "Kill" ltrans-driver(?)
On Wed, 16 Sep 2009, Diego Novillo wrote:
> On Wed, Sep 16, 2009 at 11:06, Richard Guenther <rguenther@suse.de> wrote:
>
> > Diego - any preference on the -fltrans-driver functionality? ÂI could
> > preserve the old behavior if -fltrans-driver=ltrans-driver is specified
> > and make the new behavior the default. ÂOr turn on the new behavior
> > on explicit -fltrans-driver= only (or -fno-ltrans-driver?). ÂBug 39276
> > notes existing portability issues with the ltrans-driver script.
>
> Ripping out the old behaviour is exactly what I had in mind. I don't
> want lto1 to drive the LTRANS phase, I want to give it back to xgcc to
> do the driving. Getting rid of the ltrans-driver is fine. At some
> point we will need to add the capability of launching LTRANS in
> parallel (say via distcc), but that can and should be done from the
> actual driver, not lto1.
Hm, ok. I was thinking of allowing plugins to drive the LTRANS phase
communicating with whatever distributed infrastructure is deployed.
> So, toss it.
Ok, I'll do that. I'll wait for your merge to finish though and will
be travelling the next days anyway.
Richard.