df: Don't set bbs dirty because of debug insn moves [PR104459]

Hi!

As mentioned in the PR, we get -fcompare-debug failure, which is caused by
cfg_layout_merge_blocks successfully merging two bbs where both bbs
contained just CODE_LABEL, NOTE_INSN_BASIC_BLOCK and in the -g case both
some debug insns at the end.  cfg_layout_merge_blocks calls
update_bb_for_insn_chain which for the post-label insns in the second block
(except for BARRIERs) calls df_insn_change_bb.  This function changes
the bb of the insns and for notes just punts, but for other insns calls
df_set_bb_dirty.  Now the problem is that because there were only debug
insns and notes in the second block, df_set_bb_dirty is called on both
only in the -g case and not with -g0.  df_set_bb_dirty these days
sets both the BB_MODIFIED flag and marks the bb as dirty, and the former
is what 6 spots in cfgcleanup.cc use in code-generation decisions,
in this case
          may_thread |= (target->flags & BB_MODIFIED) != 0;
in particular.  So, with -g may_thread is true while with -g0 it is not
and we diverge from that point onwards.
I've thought about introducing df_set_bb_dirty_nondebug that wouldn't
set BB_MODIFIED but would mark the bb dirty, but then I went through
history and found changes like:
https://gcc.gnu.org/legacy-ml/gcc-patches/2010-10/msg00059.html
so I've also tried just not calling df_set_bb_dirty for debug insns
at all and it passed x86_64-linux and i686-linux
--enable-checking=yes,rtl,extra,df bootstraps/regtests, so perhaps
that works too.
Now that I look at it again, if we don't need those from %d to %d messages
for debug insns in the dump files, another way to fix it would be just to
change the very first line in the hunk from
  if (!INSN_P (insn))
to
  if (!DEBUG_INSN_P (insn))
Though, df_set_bb_dirty_nondebug which will do everything but
set bb->flags |= BB_MODIFIED is yet another option I can test.
Perhaps even that PR42889 was solely about those 6 decisions in cfgcleanup
(at that point it used df_get_bb_dirty) and not about actually the
recomputation of some of the problems causing different code generations.

2022-02-10  Jakub Jelinek  <jakub@redhat.com>

	PR rtl-optimization/104459
	* df-scan.cc (df_insn_change_bb): Don't call df_set_bb_dirty when
	moving DEBUG_INSNs between bbs.

	* gcc.dg/pr104459.c: New test.


	Jakub

On Feb 10, 2022, Jakub Jelinek <jakub@redhat.com> wrote:

> 	PR rtl-optimization/104459
> 	* df-scan.cc (df_insn_change_bb): Don't call df_set_bb_dirty when
> 	moving DEBUG_INSNs between bbs.

Thanks, that looks quite reasonable to me.  I suppose if we can
reconsider a variant that distinguishes debug insns if we find this to
break something.

On Thu, 10 Feb 2022, Jakub Jelinek wrote:

> Hi!
> 
> As mentioned in the PR, we get -fcompare-debug failure, which is caused by
> cfg_layout_merge_blocks successfully merging two bbs where both bbs
> contained just CODE_LABEL, NOTE_INSN_BASIC_BLOCK and in the -g case both
> some debug insns at the end.  cfg_layout_merge_blocks calls
> update_bb_for_insn_chain which for the post-label insns in the second block
> (except for BARRIERs) calls df_insn_change_bb.  This function changes
> the bb of the insns and for notes just punts, but for other insns calls
> df_set_bb_dirty.  Now the problem is that because there were only debug
> insns and notes in the second block, df_set_bb_dirty is called on both
> only in the -g case and not with -g0.  df_set_bb_dirty these days
> sets both the BB_MODIFIED flag and marks the bb as dirty, and the former
> is what 6 spots in cfgcleanup.cc use in code-generation decisions,
> in this case
>           may_thread |= (target->flags & BB_MODIFIED) != 0;
> in particular.  So, with -g may_thread is true while with -g0 it is not
> and we diverge from that point onwards.
> I've thought about introducing df_set_bb_dirty_nondebug that wouldn't
> set BB_MODIFIED but would mark the bb dirty, but then I went through
> history and found changes like:
> https://gcc.gnu.org/legacy-ml/gcc-patches/2010-10/msg00059.html
> so I've also tried just not calling df_set_bb_dirty for debug insns
> at all and it passed x86_64-linux and i686-linux
> --enable-checking=yes,rtl,extra,df bootstraps/regtests, so perhaps
> that works too.
> Now that I look at it again, if we don't need those from %d to %d messages
> for debug insns in the dump files, another way to fix it would be just to
> change the very first line in the hunk from
>   if (!INSN_P (insn))
> to
>   if (!DEBUG_INSN_P (insn))
> Though, df_set_bb_dirty_nondebug which will do everything but
> set bb->flags |= BB_MODIFIED is yet another option I can test.
> Perhaps even that PR42889 was solely about those 6 decisions in cfgcleanup
> (at that point it used df_get_bb_dirty) and not about actually the
> recomputation of some of the problems causing different code generations.

OK.

Thanks,
Richard.

> 2022-02-10  Jakub Jelinek  <jakub@redhat.com>
> 
> 	PR rtl-optimization/104459
> 	* df-scan.cc (df_insn_change_bb): Don't call df_set_bb_dirty when
> 	moving DEBUG_INSNs between bbs.
> 
> 	* gcc.dg/pr104459.c: New test.
> 
> --- gcc/df-scan.cc.jj	2022-01-18 00:18:02.720744815 +0100
> +++ gcc/df-scan.cc	2022-02-10 11:10:54.039135547 +0100
> @@ -1769,13 +1769,15 @@ df_insn_change_bb (rtx_insn *insn, basic
>    if (!INSN_P (insn))
>      return;
>  
> -  df_set_bb_dirty (new_bb);
> +  if (!DEBUG_INSN_P (insn))
> +    df_set_bb_dirty (new_bb);
>    if (old_bb)
>      {
>        if (dump_file)
>  	fprintf (dump_file, "  from %d to %d\n",
>  		 old_bb->index, new_bb->index);
> -      df_set_bb_dirty (old_bb);
> +      if (!DEBUG_INSN_P (insn))
> +	df_set_bb_dirty (old_bb);
>      }
>    else
>      if (dump_file)
> --- gcc/testsuite/gcc.dg/pr104459.c.jj	2022-02-10 11:09:38.397181836 +0100
> +++ gcc/testsuite/gcc.dg/pr104459.c	2022-02-10 11:09:16.049490953 +0100
> @@ -0,0 +1,38 @@
> +/* PR rtl-optimization/104459 */
> +/* { dg-do compile } */
> +/* { dg-options "-O2 -funswitch-loops -fno-tree-dce -fcompare-debug -w" } */
> +
> +void
> +foo (int x, int y)
> +{
> +  unsigned int a;
> +
> +  for (;;)
> +    {
> +      short int *p = (short int *) &x;
> +      unsigned int q = 0;
> +
> +      a /= 2;
> +      if (a)
> +	{
> +	  q -= y;
> +	  while (q)
> +	    ;
> +	}
> +
> +      if (x)
> +	{
> +	  for (q = 0; q != 1; q += 2)
> +	    {
> +	      unsigned int n;
> +
> +	      n = *p ? 0 : q;
> +	      y += n < 1;
> +
> +	      n = a || *p;
> +	      if (n % x == 0)
> +		y /= x;
> +	    }
> +	}
> +    }
> +}
> 
> 	Jakub
> 
>