From: Andrew Pinski <apinski@marvell.com>
To: <gcc-patches@gcc.gnu.org>
Cc: Andrew Pinski <apinski@marvell.com>
Subject: [PATCH 5/6] PHIOPT: Ignore predicates for match-and-simplify phi-opt
Date: Sat, 22 Apr 2023 15:09:20 -0700 [thread overview]
Message-ID: <20230422220921.452264-6-apinski@marvell.com> (raw)
In-Reply-To: <20230422220921.452264-1-apinski@marvell.com>
This fixes a missed optimization where early phi-opt would
not work when there was predicates. The easiest fix is
to change empty_bb_or_one_feeding_into_p to ignore those
statements while checking for only feeding statement.
Note phi-opt-23.c and phi-opt-24.c still fail as we don't handle
diamond form in match_and_simplify phiopt yet.
OK? Bootstrapped and tested on x86_64-linux-gnu with no regressions.
gcc/ChangeLog:
* tree-ssa-phiopt.cc (empty_bb_or_one_feeding_into_p):
Instead of calling last_and_only_stmt, look for the last statement
manually.
gcc/testsuite/ChangeLog:
* gcc.dg/tree-ssa/ssa-ifcombine-13.c: Add -fno-ssa-phiopt.
---
.../gcc.dg/tree-ssa/ssa-ifcombine-13.c | 4 +++-
gcc/tree-ssa-phiopt.cc | 22 +++++++++++++++++--
2 files changed, 23 insertions(+), 3 deletions(-)
diff --git a/gcc/testsuite/gcc.dg/tree-ssa/ssa-ifcombine-13.c b/gcc/testsuite/gcc.dg/tree-ssa/ssa-ifcombine-13.c
index 425eb3d6481..7976544c79b 100644
--- a/gcc/testsuite/gcc.dg/tree-ssa/ssa-ifcombine-13.c
+++ b/gcc/testsuite/gcc.dg/tree-ssa/ssa-ifcombine-13.c
@@ -1,5 +1,7 @@
/* { dg-do compile } */
-/* { dg-options "-O1 -fdump-tree-optimized-details-blocks --param logical-op-non-short-circuit=1" } */
+/* Disable phi-opt as it is no longer confused by predicate which had allowed ifcombine to work in the past.
+ Note this testcase is about ifcombine rather than phi-opt. */
+/* { dg-options "-O1 -fdump-tree-optimized-details-blocks --param logical-op-non-short-circuit=1 -fno-ssa-phiopt" } */
_Bool f1(_Bool a, _Bool b)
{
diff --git a/gcc/tree-ssa-phiopt.cc b/gcc/tree-ssa-phiopt.cc
index cb4d8788b56..ffd6a4e6f35 100644
--- a/gcc/tree-ssa-phiopt.cc
+++ b/gcc/tree-ssa-phiopt.cc
@@ -962,9 +962,27 @@ empty_bb_or_one_feeding_into_p (basic_block bb,
if (!gimple_seq_empty_p (phi_nodes (bb)))
return false;
- stmt_to_move = last_and_only_stmt (bb);
+ gimple_stmt_iterator gsi;
+
+ gsi = gsi_start_nondebug_after_labels_bb (bb);
+ while (!gsi_end_p (gsi))
+ {
+ gimple *s = gsi_stmt (gsi);
+ gsi_next_nondebug (&gsi);
+ /* Skip over Predict and nop statements. */
+ if (gimple_code (s) == GIMPLE_PREDICT
+ || gimple_code (s) == GIMPLE_NOP)
+ continue;
+ /* If there is more one statement return false. */
+ if (stmt_to_move)
+ return false;
+ stmt_to_move = s;
+ }
+
+ /* The only statement here was a Predict or a nop statement
+ so return true. */
if (!stmt_to_move)
- return false;
+ return true;
if (gimple_vuse (stmt_to_move))
return false;
--
2.39.1
next prev parent reply other threads:[~2023-04-22 22:09 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-04-22 22:09 [PATCH 0/6] Improve PHIOPT match and simplify for diamond shaped bbs Andrew Pinski
2023-04-22 22:09 ` [PATCH 1/6] PHIOPT: Move check on diamond bb to tree_ssa_phiopt_worker from minmax_replacement Andrew Pinski
2023-04-24 12:01 ` Richard Biener
2023-04-22 22:09 ` [PATCH 2/6] PHIOPT: Cleanup tree_ssa_phiopt_worker code Andrew Pinski
2023-04-24 12:06 ` Richard Biener
2023-04-22 22:09 ` [PATCH 3/6] PHIOPT: Allow other diamond uses when do_hoist_loads is true Andrew Pinski
2023-04-24 12:06 ` Richard Biener
2023-04-22 22:09 ` [PATCH 4/6] PHIOPT: Factor out some code from match_simplify_replacement Andrew Pinski
2023-04-24 12:02 ` Richard Biener
2023-04-22 22:09 ` Andrew Pinski [this message]
2023-04-24 12:05 ` [PATCH 5/6] PHIOPT: Ignore predicates for match-and-simplify phi-opt Richard Biener
2023-04-22 22:09 ` [PATCH 6/6] PHIOPT: Add support for diamond shaped bb to match_simplify_replacement Andrew Pinski
2023-04-24 12:07 ` Richard Biener
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20230422220921.452264-6-apinski@marvell.com \
--to=apinski@marvell.com \
--cc=gcc-patches@gcc.gnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).