[PATCH V16] VECT: Add decrement IV iteration loop control by variable amount support

[Richard Biener]

On Thu, 25 May 2023, Richard Sandiford wrote:

> This looks good to me.  Just a couple of very minor cosmetic things:
> 
> juzhe.zhong@rivai.ai writes:
> > @@ -753,17 +846,35 @@ vect_set_loop_condition_partial_vectors (class loop *loop,
> >  	      continue;
> >  	  }
> >  
> > -	/* See whether zero-based IV would ever generate all-false masks
> > -	   or zero length before wrapping around.  */
> > -	bool might_wrap_p = vect_rgroup_iv_might_wrap_p (loop_vinfo, rgc);
> > -
> > -	/* Set up all controls for this group.  */
> > -	test_ctrl = vect_set_loop_controls_directly (loop, loop_vinfo,
> > -						     &preheader_seq,
> > -						     &header_seq,
> > -						     loop_cond_gsi, rgc,
> > -						     niters, niters_skip,
> > -						     might_wrap_p);
> > +	if (!LOOP_VINFO_USING_DECREMENTING_IV_P (loop_vinfo) || !iv_rgc
> > +	    || (iv_rgc->max_nscalars_per_iter * iv_rgc->factor
> > +		!= rgc->max_nscalars_per_iter * rgc->factor))
> 
> Coding style is to put each subcondition on a separate line when the
> whole condition doesn't fit on a single line.  So:
> 
> 	if (!LOOP_VINFO_USING_DECREMENTING_IV_P (loop_vinfo)
> 	    || !iv_rgc
> 	    || (iv_rgc->max_nscalars_per_iter * iv_rgc->factor
> 		!= rgc->max_nscalars_per_iter * rgc->factor))
> 
> > @@ -2725,6 +2726,17 @@ start_over:
> >        && !vect_verify_loop_lens (loop_vinfo))
> >      LOOP_VINFO_CAN_USE_PARTIAL_VECTORS_P (loop_vinfo) = false;
> >  
> > +  /* If we're vectorizing an loop that uses length "controls" and
> 
> s/an loop/a loop/    (Sorry for not noticing earlier.)
> 
> OK for trunk from my POV with those changes; no need to repost unless
> your policies require it.  Please give Richi a chance to comment too
> though.

LGTM as well.

Thanks,
Richard.