[gcc r14-9412] fwprop: Restore previous behavior for forward propagation of RTL with MEMs [PR114284]

[gcc r14-9412] fwprop: Restore previous behavior for forward propagation of RTL with MEMs [PR114284]

[Jakub Jelinek]

https://gcc.gnu.org/g:3e3e4156a5f93e6d62101594ca6660ee9ce9c10e

commit r14-9412-g3e3e4156a5f93e6d62101594ca6660ee9ce9c10e
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Sat Mar 9 13:04:26 2024 +0100

    fwprop: Restore previous behavior for forward propagation of RTL with MEMs [PR114284]
    
    Before the recent PR111267 r14-8319 fwprop changes, fwprop would never try
    to propagate what was not considered PROFITABLE, where the profitable part
    actually was partly about profitability, partly about very good reasons
    not to actually propagate and partly for cases where propagation is
    completely incorrect.
    In particular, classify_result has:
      /* Allow (subreg (mem)) -> (mem) simplifications with the following
         exceptions:
         1) Propagating (mem)s into multiple uses is not profitable.
         2) Propagating (mem)s across EBBs may not be profitable if the source EBB
            runs less frequently.
         3) Propagating (mem)s into paradoxical (subreg)s is not profitable.
         4) Creating new (mem/v)s is not correct, since DCE will not remove the old
            ones.  */
      if (single_use_p
          && single_ebb_p
          && SUBREG_P (old_rtx)
          && !paradoxical_subreg_p (old_rtx)
          && MEM_P (new_rtx)
          && !MEM_VOLATILE_P (new_rtx))
        return PROFITABLE;
    and didn't mark any other MEM_P (new_rtx) or rtxes which contain
    a MEM in its subrtxes as PROFITABLE.  Now, since r14-8319 profitable_p
    method has been renamed to likely_profitable_p and has just a minor role.
    Now, rule 4) above is something that isn't about profitability, but about
    correct behavior, if you propagate mem/v, the code is miscompiled.
    This particular case has been fixed elsewhere by Haochen in r14-9379.
    But I think even the 1) and 2) and maybe 3) are a strong don't do it,
    don't rely solely on rtx costs, increasing the number of loads of the same
    memory, even when cached, is undesirable, canceling load hoisting can
    be undesirable as well.
    
    So, the following patch restores previous behavior of src contains any MEMs,
    in that case likely_profitable_p () is taken as the old profitable_p ()
    as a requirement rather than just a hint.  For propagation of something
    which doesn't load from memory this keeps the r14-8319 behavior.
    
    2024-03-09  Jakub Jelinek  <jakub@redhat.com>
    
            PR target/114284
            * fwprop.cc (try_fwprop_subst_pattern): Don't propagate
            src containing MEMs unless prop.likely_profitable_p ().

Diff:
---
 gcc/fwprop.cc | 1 +
 1 file changed, 1 insertion(+)

diff --git a/gcc/fwprop.cc b/gcc/fwprop.cc
index cb6fd6700ca..de543923b92 100644
--- a/gcc/fwprop.cc
+++ b/gcc/fwprop.cc
@@ -451,6 +451,7 @@ try_fwprop_subst_pattern (obstack_watermark &attempt, insn_change &use_change,
 
   if (!prop.likely_profitable_p ()
       && (prop.changed_mem_p ()
+	  || contains_mem_rtx_p (src)
 	  || use_insn->is_asm ()
 	  || !single_set (use_rtl)))
     {