All of lore.kernel.org
 help / color / mirror / Atom feed
From: David Daney <david.daney@cavium.com>
To: binutils <binutils@sourceware.org>
Cc: linux-mips <linux-mips@linux-mips.org>,
	Manuel Lauss <manuel.lauss@googlemail.com>,
	Debian MIPS <debian-mips@lists.debian.org>
Subject: [Patch]: Fix ld pr11138 FAILures on mips*.
Date: Mon, 05 Dec 2011 16:49:35 -0800	[thread overview]
Message-ID: <4EDD669F.30207@cavium.com> (raw)

[-- Attachment #1: Type: text/plain, Size: 1489 bytes --]

The pr11138 testcase links an executable with a version script.  On 
mips64-linux the presence of a version script was causing the 
MIPS_RLD_MAP dynamic tag to be populated with a NULL value.  When such 
an executable was run ld.so would try to dereference this and receive 
SIGSEGV, thus killing the process.

The root cause of this is that the mips linker synthesizes a special 
symbol "__RLD_MAP", and then sets MIPS_RLD_MAP to point to it.  When a 
version script is present, this symbol gets versioned along with all the 
rest, and when it is time to take its address, the symbol can no longer 
be found as it has had version information appended to its name.

Since "__RLD_MAP" is really part of the ABI, we want to exclude it from 
symbol versioning.  To this end, I introduced a new symbol flag 
'no_sym_version' to tag this type of symbol.  When the "__RLD_MAP" 
symbol is created, we set this flag.

In _bfd_elf_link_assign_sym_version, we then skip all symbols that have 
'no_sym_version' set, and everything now works.

This problem has also been reported in the wild when linking the firefox 
executable.

Tested on mips64-linux-gnu and x86_64-linux-gnu

Ok to commit?

2011-12-05  David Daney  <david.daney@cavium.com>

	* elf-bfd.h (elf_link_hash_entry): Add no_sym_version field.
	* elflink.c (_bfd_elf_link_assign_sym_version): Don't assign a
	version if no_sym_version is set.
	* elfxx-mips.c (_bfd_mips_elf_create_dynamic_sections): Set
	no_sym_version for "__RLD_MAP".

[-- Attachment #2: dd-2.patch --]
[-- Type: text/plain, Size: 1784 bytes --]

Index: bfd/elf-bfd.h
===================================================================
RCS file: /cvs/src/src/bfd/elf-bfd.h,v
retrieving revision 1.329
diff -u -p -r1.329 elf-bfd.h
--- bfd/elf-bfd.h	17 Aug 2011 00:39:38 -0000	1.329
+++ bfd/elf-bfd.h	5 Dec 2011 20:15:49 -0000
@@ -198,6 +198,8 @@ struct elf_link_hash_entry
   unsigned int pointer_equality_needed : 1;
   /* Symbol is a unique global symbol.  */
   unsigned int unique_global : 1;
+  /* Symbol should not be versioned.  It is part of the ABI */
+  unsigned int no_sym_version : 1;
 
   /* String table index in .dynstr if this is a dynamic symbol.  */
   unsigned long dynstr_index;
Index: bfd/elflink.c
===================================================================
RCS file: /cvs/src/src/bfd/elflink.c,v
retrieving revision 1.430
diff -u -p -r1.430 elflink.c
--- bfd/elflink.c	15 Nov 2011 11:33:57 -0000	1.430
+++ bfd/elflink.c	5 Dec 2011 20:15:50 -0000
@@ -1946,6 +1946,9 @@ _bfd_elf_link_assign_sym_version (struct
   if (!h->def_regular)
     return TRUE;
 
+  if (h->no_sym_version)
+    return TRUE;
+
   bed = get_elf_backend_data (info->output_bfd);
   p = strchr (h->root.root.string, ELF_VER_CHR);
   if (p != NULL && h->verinfo.vertree == NULL)
Index: bfd/elfxx-mips.c
===================================================================
RCS file: /cvs/src/src/bfd/elfxx-mips.c,v
retrieving revision 1.296
diff -u -p -r1.296 elfxx-mips.c
--- bfd/elfxx-mips.c	29 Nov 2011 20:28:54 -0000	1.296
+++ bfd/elfxx-mips.c	5 Dec 2011 20:15:50 -0000
@@ -7260,6 +7260,7 @@ _bfd_mips_elf_create_dynamic_sections (b
 	  h = (struct elf_link_hash_entry *) bh;
 	  h->non_elf = 0;
 	  h->def_regular = 1;
+	  h->no_sym_version = 1;
 	  h->type = STT_OBJECT;
 
 	  if (! bfd_elf_link_record_dynamic_symbol (info, h))

             reply	other threads:[~2011-12-06  0:49 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-12-06  0:49 David Daney [this message]
2011-12-06  5:40 ` [Patch]: Fix ld pr11138 FAILures on mips* Alan Modra
2011-12-06 20:20   ` David Daney
2011-12-06 23:43     ` Alan Modra
2011-12-06 21:16   ` Richard Sandiford
2011-12-06 22:13     ` David Daney
2011-12-08 20:34       ` Richard Sandiford
2011-12-08 20:53         ` David Daney
2011-12-07  0:07     ` Alan Modra

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=4EDD669F.30207@cavium.com \
    --to=david.daney@cavium.com \
    --cc=binutils@sourceware.org \
    --cc=debian-mips@lists.debian.org \
    --cc=linux-mips@linux-mips.org \
    --cc=manuel.lauss@googlemail.com \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.