public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: Richard Sandiford <rdsandiford@googlemail.com>
To: Eric Botcazou <ebotcazou@adacore.com>
Cc: gcc-patches@gcc.gnu.org
Subject: Re: PR 48826: NOTE_INSN_CALL_ARG_LOCATION vs. define_split
Date: Sun, 29 May 2011 01:07:00 -0000	[thread overview]
Message-ID: <87fwnyhh4o.fsf@firetop.home> (raw)
In-Reply-To: <87aaeds3jh.fsf@firetop.home> (Richard Sandiford's message of	"Mon, 23 May 2011 18:57:38 +0100")

Richard Sandiford <rdsandiford@googlemail.com> writes:
> gcc/
> 	PR rtl-optimization/48826
> 	* emit-rtl.c (try_split): When splitting a call that is followed
> 	by a NOTE_INSN_CALL_ARG_LOCATION, move the note after the new call.

It turns out that this isn't enough.  In the attached testcase,
we have a:

    (var_location pc (nil))

note between the call and the CALL_ARG_LOCATION.  A quick look
at the var-tracking flow suggests that this is legitimate,
so we need to look at every note before the next real insn,
rather than just looking at the first.

I assume the ARM, SH and S390 reorg code would also need to cope
with this case.

If this usage isn't legitimate, and the CALL_ARG_LOCATION is supposed
to come directly after the call, then it would be great if someone more
familiar with the var-tracking code could have a look at it.

Otherwise, patch bootstrapped & regression tested on x86_64-linux-gnu.
Also tested on mips-linux-gnu.  OK to install?

Richard


gcc/
	* emit-rtl.c (try_split): Use a loop to search for
	NOTE_INSN_CALL_ARG_LOCATIONs.

gcc/testsuite/
	From Ryan Mansfield
	* gcc.dg/pr48826.c: New test.

Index: gcc/emit-rtl.c
===================================================================
--- gcc/emit-rtl.c	2011-05-28 17:08:48.000000000 +0100
+++ gcc/emit-rtl.c	2011-05-28 18:23:28.000000000 +0100
@@ -3494,16 +3494,15 @@ try_split (rtx pat, rtx trial, int last)
 	       we must move any following NOTE_INSN_CALL_ARG_LOCATION note
 	       so that it comes immediately after the new call.  */
 	    if (NEXT_INSN (insn))
-	      {
-		next = NEXT_INSN (trial);
-		if (next
-		    && NOTE_P (next)
-		    && NOTE_KIND (next) == NOTE_INSN_CALL_ARG_LOCATION)
+	      for (next = NEXT_INSN (trial);
+		   next && NOTE_P (next);
+		   next = NEXT_INSN (next))
+		if (NOTE_KIND (next) == NOTE_INSN_CALL_ARG_LOCATION)
 		  {
 		    remove_insn (next);
 		    add_insn_after (next, insn, NULL);
+		    break;
 		  }
-	      }
 	  }
     }
 
Index: gcc/testsuite/gcc.dg/pr48826.c
===================================================================
--- /dev/null	2011-05-28 09:03:52.070295797 +0100
+++ gcc/testsuite/gcc.dg/pr48826.c	2011-05-28 18:12:00.000000000 +0100
@@ -0,0 +1,10 @@
+/* { dg-options "-O -g -w" } */
+
+void bar (int *);
+
+void
+foo ()
+{
+  int *const pc = __builtin_return_address (0);
+  bar (pc);
+}

  parent reply	other threads:[~2011-05-28 17:25 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-05-23  2:07 Richard Sandiford
2011-05-23  9:30 ` Eric Botcazou
2011-05-23 20:16   ` Richard Sandiford
2011-05-23 22:46     ` Eric Botcazou
2011-05-29  1:07     ` Richard Sandiford [this message]
2011-05-29 14:31       ` Eric Botcazou

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=87fwnyhh4o.fsf@firetop.home \
    --to=rdsandiford@googlemail.com \
    --cc=ebotcazou@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).