Re: [PATCH] diagnose unsupported uses of hardware register variables (PR 88000)

[Segher Boessenkool <segher@kernel.crashing.org>]

On Wed, Nov 14, 2018 at 01:27:26PM +0100, Jakub Jelinek wrote:
> On Wed, Nov 14, 2018 at 06:22:51AM -0600, Segher Boessenkool wrote:
> > Btw, if you just add
> > 
> > void *
> > retsp (void)
> > {
> >   register void *sp __asm ("sp");
> >   asm ("" : "+g" (sp));  // <-- this line
> >   return sp;
> > }
> > 
> > everything works fine.
> 
> Even in what you are proposing, i.e. handle the var as any other var
> in SSA form and only copy into the hard register right before asm and out of
> it after it?

Yes, *only* in that: with current trunk sp lives in the "sp" hard register
at the "return sp", which cannot work reliably (what value is returned?
It is unspecified).

> Because 
> {
>   void *sp;
>   asm ("" : "+g" (sp));
>   return sp;
> }
> would store into the register default definition of the SSA_NAME (the var
> has no initializer).

I'm more concerned about what it looks like in RTL, but sure :-)  What
*should* it do before RTL?  Not much at all I think, just keep track that
this var is a register asm and that's that?


Segher