Re: RFC: bug in code hoisting

[Dale Johannesen <dalej at apple dot com>]

On Mar 11, 2005, at 12:14 PM, Jeffrey A Law wrote:
> >  The patch
> > I posted earlier makes it work as (I think) was intended; I've
> > qualified this
> > on Apple's branch (no problems), and will go ahead and qualify it on
> > mainline
> > unless there's objection.
> I didnt' look at the patch, sorry.
>
> The reason I think we've got a fundamental problem is if we do
> any walking backwards through any assignments to pseudos, then
> we run the risk of hitting the problem you've described.
>
> Add to that my worry that the walk backwards is a correctness
> issue for S_R_C machines, then we're in some ways stuck unless
> we want to take drastic measures.

I don't think it's that bad, what you want to do is stop walking 
backwards
when you find anything that isn't a store into a hard reg.  That's not 
what
my patch does, but it's easy enough.  Or are you concerned there will
be stores into psuedos intermingled with the stores into hard regs?
That doesn't appear to happen, on ppc at least, in a few experiments.