public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: Eric Botcazou <botcazou@adacore.com>
To: gcc-patches@gcc.gnu.org
Subject: [PATCH] Avoid creating useless debug temporaries
Date: Tue, 25 Apr 2023 11:33:23 +0200	[thread overview]
Message-ID: <3493207.iIbC2pHGDl@fomalhaut> (raw)

[-- Attachment #1: Type: text/plain, Size: 536 bytes --]

Hi,

insert_debug_temp_for_var_def has some strange code whereby it creates debug 
temporaries for SINGLE_RHS (RHS for gimple_assign_single_p) but not for other 
RHS in the same situation.  Removing it saves 25% of compilation time at -g -O 
for a pathological testcase I have.

Bootstrapped/regtested on x86-64/Linux, OK for the mainline?


2023-04-25  Eric Botcazou  <ebotcazou@adacore.com>

	* tree-ssa.cc (insert_debug_temp_for_var_def): Do not create
	superfluous debug temporaries for single GIMPLE assignments.

-- 
Eric Botcazou

[-- Attachment #2: p.diff --]
[-- Type: text/x-patch, Size: 1726 bytes --]

diff --git a/gcc/tree-ssa.cc b/gcc/tree-ssa.cc
index a5cad2d344e..4ca1f5f3104 100644
--- a/gcc/tree-ssa.cc
+++ b/gcc/tree-ssa.cc
@@ -412,8 +412,7 @@ insert_debug_temp_for_var_def (gimple_stmt_iterator *gsi, tree var)
     {
       /* If there's a single use of VAR, and VAR is the entire debug
 	 expression (usecount would have been incremented again
-	 otherwise), and the definition involves only constants and
-	 SSA names, then we can propagate VALUE into this single use,
+	 otherwise), then we can propagate VALUE into this single use,
 	 avoiding the temp.
 
 	 We can also avoid using a temp if VALUE can be shared and
@@ -424,11 +423,9 @@ insert_debug_temp_for_var_def (gimple_stmt_iterator *gsi, tree var)
 	 are deferred to a debug temp, although we could avoid temps
 	 at the expense of duplication of expressions.  */
 
-      if (CONSTANT_CLASS_P (value)
+      if (usecount == 1
 	  || gimple_code (def_stmt) == GIMPLE_PHI
-	  || (usecount == 1
-	      && (!gimple_assign_single_p (def_stmt)
-		  || is_gimple_min_invariant (value)))
+	  || CONSTANT_CLASS_P (value)
 	  || is_gimple_reg (value))
 	;
       else
@@ -466,11 +463,6 @@ insert_debug_temp_for_var_def (gimple_stmt_iterator *gsi, tree var)
       if (value)
 	{
 	  FOR_EACH_IMM_USE_ON_STMT (use_p, imm_iter)
-	    /* unshare_expr is not needed here.  vexpr is either a
-	       SINGLE_RHS, that can be safely shared, some other RHS
-	       that was unshared when we found it had a single debug
-	       use, or a DEBUG_EXPR_DECL, that can be safely
-	       shared.  */
 	    SET_USE (use_p, unshare_expr (value));
 	  /* If we didn't replace uses with a debug decl fold the
 	     resulting expression.  Otherwise we end up with invalid IL.  */

             reply	other threads:[~2023-04-25  9:34 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-04-25  9:33 Eric Botcazou [this message]
2023-04-25 15:10 ` Richard Biener
2023-04-25 15:20   ` Jakub Jelinek
2023-04-25 15:35     ` Eric Botcazou
2023-04-25 15:37       ` Jakub Jelinek
2023-04-26  9:31   ` Eric Botcazou
2023-04-26  9:48     ` 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=3493207.iIbC2pHGDl@fomalhaut \
    --to=botcazou@adacore.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).