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 49145: Another (zero_extend (const_int ...)) in combine
Date: Sat, 04 Jun 2011 21:22:00 -0000	[thread overview]
Message-ID: <87oc2dmgwn.fsf@firetop.home> (raw)
In-Reply-To: <201106012220.49874.ebotcazou@adacore.com> (Eric Botcazou's	message of "Wed, 1 Jun 2011 22:20:49 +0200")

Eric Botcazou <ebotcazou@adacore.com> writes:
> SUBREG and ZERO_EXTEND of CONST_INTs are treated somewhat specially in the 
> entire file, see for example do_SUBST.  This isn't the case for other unary 
> operators, presumably because this isn't really necessary here.  So I'm not 
> convinced that such a generalization is really a good thing in this case.

OK.  The version below just adds a special case tomake_compound_operation
instead.  As before, I've restricted the simplification to constants,
so that we don't inadvertently undo the effects of m_c_o itself.

Tested on x86_64-linux-gnu and mips-linux-gnu.  OK for trunk?

Richard


gcc/
	PR rtl-optimization/49145
	* combine.c (make_compound_operation): Handle ZERO_EXTEND specially.

gcc/testsuite/
	PR rtl-optimization/49145
	From Ryan Mansfield
	* gcc.c-torture/compile/pr49145.c: New test.

Index: gcc/combine.c
===================================================================
--- gcc/combine.c	2011-06-01 22:09:09.000000000 +0100
+++ gcc/combine.c	2011-06-01 22:09:26.000000000 +0100
@@ -7881,7 +7881,20 @@ make_compound_operation (rtx x, enum rtx
       code = GET_CODE (x);
     }
 
-  /* Now recursively process each operand of this operation.  */
+  /* Now recursively process each operand of this operation.  We need to
+     handle ZERO_EXTEND specially so that we don't lose track of the
+     inner mode.  */
+  if (GET_CODE (x) == ZERO_EXTEND)
+    {
+      new_rtx = make_compound_operation (XEXP (x, 0), next_code);
+      tem = simplify_const_unary_operation (ZERO_EXTEND, GET_MODE (x),
+					    new_rtx, GET_MODE (XEXP (x, 0)));
+      if (tem)
+	return tem;
+      SUBST (XEXP (x, 0), new_rtx);
+      return x;
+    }
+
   fmt = GET_RTX_FORMAT (code);
   for (i = 0; i < GET_RTX_LENGTH (code); i++)
     if (fmt[i] == 'e')
Index: gcc/testsuite/gcc.c-torture/compile/pr49145.c
===================================================================
--- /dev/null	2011-06-04 08:47:56.158317425 +0100
+++ gcc/testsuite/gcc.c-torture/compile/pr49145.c	2011-06-01 22:09:26.000000000 +0100
@@ -0,0 +1,30 @@
+static int
+func1 (int a, int b)
+{
+  return b ? a : a / b;
+}
+
+static unsigned char
+func2 (unsigned char a, int b)
+{
+  return b ? a : b;
+}
+
+int i;
+
+void
+func3 (const int arg)
+{
+  for (i = 0; i != 10; i = foo ())
+    {
+      if (!arg)
+	{
+	  int j;
+	  for (j = 0; j < 5; j += 1)
+	    {
+	      int *ptr;
+	      *ptr = func2 (func1 (arg, *ptr), foo (arg));
+	    }
+	}
+    }
+}

  reply	other threads:[~2011-06-04 21:22 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-05-29  9:30 Richard Sandiford
2011-06-01 20:21 ` Eric Botcazou
2011-06-04 21:22   ` Richard Sandiford [this message]
2011-06-05 22:52     ` 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=87oc2dmgwn.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).