libcall changes for the scheduler

[Mark Mitchell]

>> Mark's call on this one.  Though personally, given the amount of
>> headaches associated with this patch, I'd prefer to see it "prove"
>> itself in the mainline sources longer before backporting it to the GCC
>> 3.1 branch.
>
> What is the status on this one?

This is pretty thorny.  That's a lot of changing at this late date, but
we do have a regression to fix.

The USE/CLOBBER counts against issue rate *sounds* like it's an
optimization problem, not a correctness problem.  Is that correct?  If
so, maybe we could leave out those bits -- but they also sound the
least risky.

> powerpc-linux-gnu I recompiled stuff (including kernel, glibc, XFree86
> and dozens of other packages without) all over the place in the meantime
> without problems.

And, I assume, ran those packages?

-- 
Mark Mitchell                mark@codesourcery.com
CodeSourcery, LLC            http://www.codesourcery.com