Re: [PATCH v3 3/4] ree: Main functionality to Improve ree pass for rs6000 target

[Jeff Law <jeffreyalaw@gmail.com>]

On 4/19/23 12:00, Ajit Agarwal wrote:
> Hello All:
> 
> This is patch-3 to improve ree pass for rs6000 target.
> Main functionality routines to imprve ree pass.
> 
> Bootstrapped and regtested on powerpc64-gnu-linux.
> 
> Thanks & Regards
> Ajit
> 
> 	ree: Improve ree pass for rs6000 target.
> 
> 	For rs6000 target we see redundant zero and sign
> 	extension and done to improve ree pass to eliminate
> 	such redundant zero and sign extension. Support of
> 	zero_extend/sign_extend/AND.
> 
> 	2023-04-19  Ajit Kumar Agarwal  <aagarwa1@linux.ibm.com>
> 
> gcc/ChangeLog:
> 
> 	* ree.cc (eliminate_across_bbs_p): Add checks to enable extension
> 	elimination across and within basic blocks.
> 	(def_arith_p): New function to check definition has arithmetic
> 	operation.
> 	(combine_set_extension): Modification to incorporate AND
> 	and current zero_extend and sign_extend instruction.
> 	(merge_def_and_ext): Add calls to eliminate_across_bbs_p and
> 	zero_extend sign_extend and AND instruction.
> 	(rtx_is_zext_p): New function.
> 	(reg_used_set_between_p): New function.
> 
> gcc/testsuite/ChangeLog:
> 
> 	* g++.target/powerpc/zext-elim.C: New testcase.
> 	* g++.target/powerpc/zext-elim-1.C: New testcase.
> 	* g++.target/powerpc/zext-elim-2.C: New testcase.
> 	* g++.target/powerpc/sext-elim.C: New testcase.
> ---
>   gcc/ree.cc                                    | 451 ++++++++++++++++--
>   gcc/testsuite/g++.target/powerpc/sext-elim.C  |  18 +
>   .../g++.target/powerpc/zext-elim-1.C          |  19 +
>   .../g++.target/powerpc/zext-elim-2.C          |  11 +
>   gcc/testsuite/g++.target/powerpc/zext-elim.C  |  30 ++
>   5 files changed, 482 insertions(+), 47 deletions(-)
>   create mode 100644 gcc/testsuite/g++.target/powerpc/sext-elim.C
>   create mode 100644 gcc/testsuite/g++.target/powerpc/zext-elim-1.C
>   create mode 100644 gcc/testsuite/g++.target/powerpc/zext-elim-2.C
>   create mode 100644 gcc/testsuite/g++.target/powerpc/zext-elim.C
> 
> diff --git a/gcc/ree.cc b/gcc/ree.cc
> index 413aec7c8eb..053db2e8ff3 100644
> --- a/gcc/ree.cc
> +++ b/gcc/ree.cc
> @@ -253,6 +253,71 @@ struct ext_cand
>   
>   static int max_insn_uid;
>   
> +bool
> +reg_used_set_between_p (rtx set, rtx_insn *def_insn, rtx_insn *insn
> +{
> +  if (reg_used_between_p (set, def_insn, insn)
> +      || reg_set_between_p (set, def_insn, insn))
> +    return true;
> +
> +  return false;
> +}
This seems general enough that it should go into the same file as 
reg_used_between_p and reg_set_between_p.  It needs a function comment 
as well.


> +static unsigned int
> +rtx_is_zext_p (rtx insn)
> +{
> +  if (GET_CODE (insn) == AND)
> +    {
> +      rtx set = XEXP (insn, 0);
> +      if (REG_P (set))
> +	{
> +	  if (XEXP (insn, 1) == const1_rtx)
> +	    return 1;
> +	}
> +      else
> +	return 0;
> +    }
> +
> +  return 0;
> +}
So my comment from the prior version stands.  Testing for const1_rtx is 
just wrong.  The optimization you're trying to perform (If I understand 
it correctly) works for many other constants and the set of constants 
supported will vary based on the input and output modes.

Similarly in rtx_is_zext_p.

You still have numerous formatting issues which makes reading the patch 
more difficult than it should be.  Please review the formatting 
guidelines and follow them.   In particular please review how to indent 
multi-line conditionals.





You sti
> @@ -698,6 +777,226 @@ get_sub_rtx (rtx_insn *def_insn)
>     return sub_rtx;
>   }
>   
> +/* Check if the def insn is ASHIFT and LSHIFTRT.
> +  Inputs: insn for which def has to be checked.
> +	  source operand rtx.
> +   Output: True or false if def has arithmetic
> +	   peration like ASHIFT and LSHIFTRT.  */
This still needs work.  Between the comments and code, I still don't 
know what you're really trying to do here.  I can make some guesses, but 
it's really your job to write clear comments about what you're doing so 
that a review or someone looking at the code in the future don't have to 
guess.

It looks like you want to look at all the reaching definitions of INSN 
for ORIG_SRC and if they are ASHIFT/LSHIFTRT do...  what?

Why are ASHIFT/LSHIFTRT interesting here?  Why are you looking for them?



> +
> +/* Find feasibility of extension elimination
> +   across basic blocks.
> +   Input: candiate to check the feasibility.
> +	  def_insn of candidate.
> +   Output: Returns true or false if feasible or not.  */
Function comments should read like complete sentences.  Arguments should 
be in all caps.  There are numerous examples in ree.cc you can follow.

There's no comments in this code which describe the properties of the 
CFG/blocks you are looking for (domination, post-domination, whatever). 
It's jsut a series of tests with no clear explanation of what you're 
looking for or why any particular test exists.

As far as I can tell you're looking at the predecessors of cand->insn 
and make some decisions based on what you find.  In some cases you 
return false in others you return true.  But there's zero indication of 
why you do anything.

You're checking RTL in these cases, but I suspect you really want to be 
doing some kind of CFG property checks.  But since you haven't described 
what you're looking for, I can't make suggestions for the right queries 
to make.

I stopped trying to review the function at this point.  It needs major 
work.  Let's start with the block/CFG properties you're looking for. 
We'll then have to go through each section of code and repeat the process.

In fact I might recommend pulling the code which is trying to determine 
CFG properties into its own function.  Then try to come up with a good 
name for that function and a good function comment.

THe next chunk of code you start looking at the properties of the 
candidate insn.  That seems like it ought to break out into its own 
function as well.

THe process of refactoring, choosing good names and function comments 
should make the overall intent of this code clearer.



After that's done we'll review that work and perhaps go further.

jeff