This small patch just changes around the code slightly to
make it easier to understand that the cases were handling diamond
shaped BB for both do_store_elim/do_hoist_loads.
There is no effect on code output at all since all of the checks
are the same still.

Note this depends on
https://gcc.gnu.org/pipermail/gcc-patches/2023-April/616241.html
as that allows this to be done.

OK? Bootstrapped and tested on x86_64-linux-gnu with no regressions.

gcc/ChangeLog:

        * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker):
        Change the code around slightly to move diamond
        handling for do_store_elim/do_hoist_loads out of
        the big if/else.
---
 gcc/tree-ssa-phiopt.cc | 46 +++++++++++++++++++-----------------------
 1 file changed, 21 insertions(+), 25 deletions(-)

diff --git a/gcc/tree-ssa-phiopt.cc b/gcc/tree-ssa-phiopt.cc
index 8c5c8d8c250..d886c88215b 100644
--- a/gcc/tree-ssa-phiopt.cc
+++ b/gcc/tree-ssa-phiopt.cc
@@ -175,28 +175,38 @@ tree_ssa_phiopt_worker (bool do_store_elim, bool 
do_hoist_loads, bool early_p)
          std::swap (bb1, bb2);
          std::swap (e1, e2);
        }
-      else if (do_store_elim
-              && EDGE_SUCC (bb1, 0)->dest == EDGE_SUCC (bb2, 0)->dest)
+      else if (EDGE_SUCC (bb1, 0)->dest == EDGE_SUCC (bb2, 0)->dest)
+       {
+         diamond_p = true;
+         e2 = EDGE_SUCC (bb2, 0);
+       }
+      else
+       continue;
+
+      e1 = EDGE_SUCC (bb1, 0);
+
+      /* Make sure that bb1 is just a fall through.  */
+      if (!single_succ_p (bb1)
+         || (e1->flags & EDGE_FALLTHRU) == 0)
+       continue;
+
+      if (do_store_elim && diamond_p)
        {
-         basic_block bb3 = EDGE_SUCC (bb1, 0)->dest;
+         basic_block bb3 = e1->dest;
 
-         if (!single_succ_p (bb1)
-             || (EDGE_SUCC (bb1, 0)->flags & EDGE_FALLTHRU) == 0
-             || !single_succ_p (bb2)
-             || (EDGE_SUCC (bb2, 0)->flags & EDGE_FALLTHRU) == 0
+         if (!single_succ_p (bb2)
+             || (e2->flags & EDGE_FALLTHRU) == 0
              || EDGE_COUNT (bb3->preds) != 2)
            continue;
          if (cond_if_else_store_replacement (bb1, bb2, bb3))
            cfgchanged = true;
          continue;
        }
-      else if (do_hoist_loads
-              && EDGE_SUCC (bb1, 0)->dest == EDGE_SUCC (bb2, 0)->dest)
+      else if (do_hoist_loads && diamond_p)
        {
-         basic_block bb3 = EDGE_SUCC (bb1, 0)->dest;
+         basic_block bb3 = e1->dest;
 
          if (!FLOAT_TYPE_P (TREE_TYPE (gimple_cond_lhs (cond_stmt)))
-             && single_succ_p (bb1)
              && single_succ_p (bb2)
              && single_pred_p (bb1)
              && single_pred_p (bb2)
@@ -209,20 +219,6 @@ tree_ssa_phiopt_worker (bool do_store_elim, bool 
do_hoist_loads, bool early_p)
            hoist_adjacent_loads (bb, bb1, bb2, bb3);
          continue;
        }
-      else if (EDGE_SUCC (bb1, 0)->dest == EDGE_SUCC (bb2, 0)->dest)
-       {
-         diamond_p = true;
-         e2 = EDGE_SUCC (bb2, 0);
-       }
-      else
-       continue;
-
-      e1 = EDGE_SUCC (bb1, 0);
-
-      /* Make sure that bb1 is just a fall through.  */
-      if (!single_succ_p (bb1)
-         || (e1->flags & EDGE_FALLTHRU) == 0)
-        continue;
 
       if (do_store_elim && !diamond_p)
        {
-- 
2.31.1

Reply via email to