[PATCH][4.3][RFT] Fix PR36519 - deal with unpartitionable but call-clobbered SFTs

[Richard Guenther]

On Wed, 18 Jun 2008, Mark Mitchell wrote:

> Richard Guenther wrote:
> 
> > So, what do people think?  Is it worth to enable building those testcases at
> > -O2
> > for this cost?  I would obviously say yes.
> 
> I don't think we have much a choice.  If we can't build benchmarks at -O2,
> we've got real problems.  So, I think we have to take the hit. :-(

Applied as r137033.

Richard.