Re: RFA: Small tweak to ira-lives.c:single_reg_class

[Vladimir Makarov <vmakarov at redhat dot com>]

On 2014-06-05, 4:12 PM, Richard Sandiford wrote:
> It looks like there's a missing break after the 'G' and 'H' handling.
>
> Tested on x86_64-linux-gnu.  I also did an assembly comparison for
> a range of targets and Alpha and MIPS seemed to be the only ones
> affected.  OK to install?
>
> Sorry for the long write-up...

It is long but make me to know new things about Alpha and MIPS insn cases.

By the way, there are a lot of controversy about the changing cost for 
particular hard reg.  Sometimes it creates a lot of mess in allocation, 
when one pseudo involves several hard regs.

Also a lot of alternatives can gives different hard reg preferencing or 
wrong ones as we don't know what alternative will be used.  Therefore I 
think about earlier alternative selections and reg class cost 
calculations based on the selected alternatives.  But it will be a long 
project probably involving code selection too. I don't know how will it 
work until it is implemented which will not happen soon.

As I remember I added the code as I see small improvements on x86/x86-64 
on benchmarks.

As for the patch, it is ok for me.

Thanks, Richard.

> gcc/
> 	* ira-lives.c (single_reg_class): Add missing break.  Return NO_REGS
> 	for extra address and memory constraints, not just 'p' and
> 	TARGET_MEM_CONSTRAINT.  Likewise if the operand is a constant or
> 	equivalent to a constant and if it matches an extra constraint.
> 	Ignore other non-register operands.
>
> Index: gcc/ira-lives.c
> ===================================================================
> --- gcc/ira-lives.c	2014-06-04 22:15:23.527995920 +0100
> +++ gcc/ira-lives.c	2014-06-04 22:15:50.765243138 +0100
> @@ -839,7 +839,8 @@ single_reg_class (const char *constraint
>   		  && CONST_DOUBLE_OK_FOR_CONSTRAINT_P (equiv_const,
>   						       c, constraints)))
>   	    return NO_REGS;
> -	  /* ??? what about memory */
> +	  break;
> +
>   	case 'r':
>   	case 'a': case 'b': case 'c': case 'd': case 'e': case 'f':
>   	case 'h': case 'j': case 'k': case 'l':
> @@ -848,9 +849,22 @@ single_reg_class (const char *constraint
>   	case 'A': case 'B': case 'C': case 'D':
>   	case 'Q': case 'R': case 'S': case 'T': case 'U':
>   	case 'W': case 'Y': case 'Z':
> +#ifdef EXTRA_CONSTRAINT_STR
> +	  /* ??? Is rejecting memory the best thing to do?  */
> +	  if (EXTRA_MEMORY_CONSTRAINT (c, constraints)
> +	      || EXTRA_ADDRESS_CONSTRAINT (c, constraints))
> +	    return NO_REGS;
> +	  if (EXTRA_CONSTRAINT_STR (op, c, constraints)
> +	      || (equiv_const != NULL_RTX
> +		  && CONSTANT_P (equiv_const)
> +		  && EXTRA_CONSTRAINT_STR (equiv_const, c, constraints)))
> +	    return NO_REGS;
> +#endif
>   	  next_cl = (c == 'r'
>   		     ? GENERAL_REGS
>   		     : REG_CLASS_FROM_CONSTRAINT (c, constraints));
> +	  if (next_cl == NO_REGS)
> +	    break;
>   	  if (cl == NO_REGS
>   	      ? ira_class_singleton[next_cl][GET_MODE (op)] < 0
>   	      : (ira_class_singleton[cl][GET_MODE (op)]