public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: Alan Modra <amodra@gmail.com>
To: Sebastian Huber <sebastian.huber@embedded-brains.de>
Cc: GCC Patches <gcc-patches@gcc.gnu.org>
Subject: Re: [Patch] Potential fix for PR55033
Date: Wed, 24 Oct 2012 15:17:00 -0000	[thread overview]
Message-ID: <20121024145004.GZ3263@bubble.grove.modra.org> (raw)
In-Reply-To: <5086C507.5080905@embedded-brains.de>

On Tue, Oct 23, 2012 at 06:25:43PM +0200, Sebastian Huber wrote:
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55033
> 
> This patch fixes my problem, but I am absolutely not sure if this is the
> right way.
[snip]

This is http://gcc.gnu.org/bugzilla/show_bug.cgi?id=9571 all over again.

IMHO your patch is not wrong, but a better idea is that if we've used
categorize_decl_for_section to choose a section name, then we ought to
also use categorize_decl_for_section to choose section flags.

My original fix for pr9571 was to pass the decl down to
get_named_section.
http://gcc.gnu.org/ml/gcc-patches/2004-11/msg02487.html
That hit the assert in get_named_section when building libgfortran for
ia64, and my knee-jerk patch to fix that was to simply satisfy the
assert.  After looking at all the target section_type_flags functions,
I believe the following is what I should have done way back then.
Bootstrapped and regression tested powerpc64-linux.

	* varasm.c (default_elf_select_section): Move !DECL_P check..
	(get_named_section): ..to here before calling get_section_name.
	Adjust assertion.
	(default_section_type_flags): Add DECL_P check.
	* config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
	* config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.

Index: gcc/varasm.c
===================================================================
--- gcc/varasm.c	(revision 192660)
+++ gcc/varasm.c	(working copy)
@@ -403,12 +403,16 @@ get_named_section (tree decl, const char *name, in
 {
   unsigned int flags;
 
-  gcc_assert (!decl || DECL_P (decl));
   if (name == NULL)
-    name = TREE_STRING_POINTER (DECL_SECTION_NAME (decl));
+    {
+      gcc_assert (decl && DECL_P (decl) && DECL_SECTION_NAME (decl));
+      name = TREE_STRING_POINTER (DECL_SECTION_NAME (decl));
+    }
 
   flags = targetm.section_type_flags (decl, name, reloc);
 
+  if (decl && !DECL_P (decl))
+    decl = NULL_TREE;
   return get_section (name, flags, decl);
 }
 
@@ -5943,7 +5947,7 @@ default_section_type_flags (tree decl, const char
 	flags |= SECTION_RELRO;
     }
 
-  if (decl && DECL_ONE_ONLY (decl))
+  if (decl && DECL_P (decl) && DECL_ONE_ONLY (decl))
     flags |= SECTION_LINKONCE;
 
   if (decl && TREE_CODE (decl) == VAR_DECL && DECL_THREAD_LOCAL_P (decl))
@@ -6299,8 +6303,6 @@ default_elf_select_section (tree decl, int reloc,
       gcc_unreachable ();
     }
 
-  if (!DECL_P (decl))
-    decl = NULL_TREE;
   return get_named_section (decl, sname, reloc);
 }
 
Index: gcc/config/i386/winnt.c
===================================================================
--- gcc/config/i386/winnt.c	(revision 192660)
+++ gcc/config/i386/winnt.c	(working copy)
@@ -476,7 +476,7 @@ i386_pe_section_type_flags (tree decl, const char
 	flags |= SECTION_PE_SHARED;
     }
 
-  if (decl && DECL_ONE_ONLY (decl))
+  if (decl && DECL_P (decl) && DECL_ONE_ONLY (decl))
     flags |= SECTION_LINKONCE;
 
   /* See if we already have an entry for this section.  */
Index: gcc/config/rs6000/rs6000.c
===================================================================
--- gcc/config/rs6000/rs6000.c	(revision 192660)
+++ gcc/config/rs6000/rs6000.c	(working copy)
@@ -25689,7 +25689,7 @@ rs6000_xcoff_section_type_flags (tree decl, const
   unsigned int flags = default_section_type_flags (decl, name, reloc);
 
   /* Align to at least UNIT size.  */
-  if (flags & SECTION_CODE || !decl)
+  if ((flags & SECTION_CODE) != 0 || !decl || !DECL_P (decl))
     align = MIN_UNITS_PER_WORD;
   else
     /* Increase alignment of large objects if not already stricter.  */

-- 
Alan Modra
Australia Development Lab, IBM

  reply	other threads:[~2012-10-24 14:50 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-10-23 16:38 Sebastian Huber
2012-10-24 15:17 ` Alan Modra [this message]
2012-10-25  9:10   ` Sebastian Huber
2012-10-30 14:18   ` Sebastian Huber
2013-01-29 14:40   ` Sebastian Huber
2013-01-29 17:05     ` Sebastian Huber
2012-10-26 12:35 ` Sebastian Huber
2012-10-30 14:12   ` Sebastian Huber
2012-10-31  0:36     ` Alan Modra
2012-10-31 14:02       ` Sebastian Huber

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=20121024145004.GZ3263@bubble.grove.modra.org \
    --to=amodra@gmail.com \
    --cc=gcc-patches@gcc.gnu.org \
    --cc=sebastian.huber@embedded-brains.de \
    /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).