[Bug tree-optimization/80492] [5/6/7 Regression] Wrong code when unrolling a loop with inline asm and local regs

[ubizjak at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80492

--- Comment #14 from Uroš Bizjak <ubizjak at gmail dot com> ---
(In reply to rguenther@suse.de from comment #13)
> On Tue, 25 Apr 2017, ubizjak at gmail dot com wrote:
> 
> > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80492
> > 
> > --- Comment #12 from Uroš Bizjak <ubizjak at gmail dot com> ---
> > (In reply to Richard Biener from comment #10)
> > > But I really wonder if it is ok to omit constraints from the asm marking
> > > the register variable as changed.
> > 
> > The register is not changed in the asm, please note that the testcase
> > intentionally changes the value of the register before each asm.
> 
> Yes.  See my followup comment.  Would removing the redundant set
> be ok if it were storing the same value each time?
> 
> Thus optimizing
> 
>   reg = 0;
>   __asm__ __volatile__("/* Some Code %0 */" :  : "r" reg);
>   reg = 0;
>   __asm__ __volatile__("/* Some Code %0 */" :  : "r" reg);
>   reg = 0;
>   __asm__ __volatile__("/* Some Code %0 */" :  : "r" reg);
>   reg = 0;
>   __asm__ __volatile__("/* Some Code %0 */" :  : "r" reg);
> 
> to
> 
>   reg = 0;
>   __asm__ __volatile__("/* Some Code %0 */" :  : "r" reg);
>   __asm__ __volatile__("/* Some Code %0 */" :  : "r" reg);
>   __asm__ __volatile__("/* Some Code %0 */" :  : "r" reg);
>   __asm__ __volatile__("/* Some Code %0 */" :  : "r" reg);
> 
> ?

This is OK, asm needs explicit clobber if it clobbers a register.