Regressions on mainline\

[Jeffrey A Law]

On Mon, 2004-06-21 at 16:42, Mark Mitchell wrote:
> Steven Bosscher wrote:
> 
> >On Tuesday 22 June 2004 00:32, Andrew Pinski wrote:
> >  
> >
> >>gcov-1.C is failing for mainly because the profiling stuff is not working
> >>right on the mainline  after tree-ssa was merged in, Jan is working on
> >>this.
> >>    
> >>
> >
> >There are also patches pending for profiling stuff, perhaps they help
> >(and perhaps not...).  Unfortunately they've been ignored four times now :-(
> >http://gcc.gnu.org/ml/gcc-patches/2004-06/msg01674.html
> >  
> >
> Diego, Jeff -- would one of you please take a look at these patches? 
> 
> We need to get these regressions hammered out of the mainline ASAP.
The change to make_forwarder_block and split_block are OK and I've
installed those changes into the mainline.

I don't know the value-prof.c code well enough to evaluate that
change.  You might ping Zdenek since I believe he wrote the code
to get an evaluation of that change.

jeff