[wwwdocs] IPA and LTO updates

Jan Hubicka hubicka@ucw.cz
Mon Nov 15 22:29:00 GMT 2010


> * Andi Kleen wrote on Mon, Nov 15, 2010 at 03:23:46PM CET:
> > On Mon, Nov 15, 2010 at 03:16:49PM +0100, Jan Hubicka wrote:
> > > > 
> > > > It's -flto=jobserver isn't it? (unless you renamed it recently)
> > > It is, thanks ;)
> > > BTW, it would be nice to add the autodetection of jobserver. It is something you plan to look
> > > into or should I figure out what environent vars to test?
> > 
> > I didn't do it because Ralf thought it was a bad idea with
> > some danger of deadlock? 
> > 
> > If that's resolved it's fairly easy: just check for the environment 
> > variable. I can do that if there is consensus.
> 
> Well you should be able to try it out if it works.
> Be sure to try it without '+' prepended to the rule,
> and with a recursive makefile where somewhere along
> the lines, plain 'make' rather than '$(MAKE)' is used.
> 
> I think deadlock is only likely when you do not implement
> the jobserver protocol correctly.

We still eventually call make to connect to he jobserver, so there should not
be problems with jobserver protocol as we implement none of it?

Honza
> 
> Cheers,
> Ralf



More information about the Gcc-patches mailing list