Bug#519006: mips/ld: non-dynamic relocations refer to dynamic symbol
On Mon, Aug 30, 2010 at 10:44:25PM +0200, Aurelien Jarno wrote: Hi all, I have made some progress on this bug, though it has not progress in the expected direction. In other words I got some surprises. First of all it the bug has been introduced in binutils when introducing the MIPS PLT support [1] (commits around 2008-08-08). I am still convinced it is actually a binutils bug. Secondly, if the problem is actually not present in gcc-snapshot in gcc 4.5 we have *in Debian*, it is not due to the fix linked in the bugzilla entry. It is due to the fact we default to -mplt on these compilers. It is possible to workaround the bug on gcc 4.4 using -mplt (like it is possible to workaround it by dropping the -g). Given all of that, the only forseen solution for this fix is to also default to -mplt on gcc 4.4. It also brings speed improvements, as well as some more standard binaries with regard to other architectures. On the other hand, I understand it is something quite risky so close to a release. If we go for this solution, we may want to trigger a rebuild of part of the archive as it is currently done on sparc. I have done a rebuilt of part of the archive using gcc-4.4 that defaults to -mplt. Unfortunately I don't have enough CPU power to rebuild the whole archive, so I have rebuilt all packages with priority = important as well as 436 packages selected randomly with priority optional (see list below). I have got a few build failures but, not due to the -mplt change. failures apt-rpm_0.5.15lorg3.2-6 (#533164) libavg_1.0.1-1 (#580678) libgdchart-gd2_0.11.5-6 (#564374) mixxx_1.7.2-1 (#587110) time_1.7-23 (sbuild bug) successes - 3depict_0.0.1-1 aboot_1.0~pre20040408-3 acl_2.2.49-3 adanaxisgpl_1.2.5.dfsg.1-3 afio_2.5-5 agave_0.4.7-1 agg_2.5+dfsg1-3 aide_0.15-2 alure_1.0-2 amide_0.9.2-1 apron_0.9.10-5 apt_0.8.3 apt-build_0.12.37 aptitude_0.6.3-3.1 aptsh_0.0.7+nmu1 ap-utils_1.5-2 archivemount_0.6.1-1 arpon_2.0-2 artha_1.0.1-1 at_3.1.12-1 atom4_4.1-4 attr_2.4.44-2 autoclass_3.3.6-1 autodocksuite_4.2.3-1 autounit_0.20.1-3 avida_2.0b7-4.2 awffull_3.8.2-1.3 ballz_1.0.1-2 bandwidthd_2.0.1+cvs20090917-4 barrage_1.0.2-4 base-files_5.9 base-passwd_3.5.22 bash_4.1-3 bc_1.06.95-2 bibindex_2.10-10 bind9_9.7.1.dfsg.P2-2 bison_1:2.4.1.dfsg-3 bitmeter_1.2-2 bkhive_1.1.1-1 blast_1.1-22 boa_0.94.14rc21-3.1 bognor-regis_0.5.2-2 bombardier_0.8.3+nmu1 boost1.42_1.42.0-4 bsd-mailx_8.1.2-0.20100314cvs-1 bsdmainutils_8.0.13 bygfoot_2.3.2-1 bzip2_1.0.5-5 c2hs_0.16.2-1 cachefilesd_0.9-3 cairo-5c_1.4 canorus_0.7-2 cassbeam_1.0-8 ccache_3.0.1-1 cheesetracker_0.9.15.3-4 cherokee_1.0.8-1 chise-base_0.3.0-2 claws-mail_3.7.6-2 clex_3.15-1 cln_1.3.1-2 cl-uffi_2.0.0-1 coq_8.2.pl2+dfsg-1 coreutils_8.5-1 couchdb_0.11.0-2 cpio_2.11-4 crimson_0.5.2-1 cron_3.0pl1-114 crossfire_1.50.0-1 cryptgps_0.2.1-7 cwidget_0.5.16-3 cyrus-sasl2_2.1.23.dfsg1-6 darcs_2.4.4-2 darkice_0.20.1-3 dash_0.5.5.1-7 db4.5_4.5.20-13.2 db4.6_4.6.21-16 db4.7_4.7.25-9 db4.8_4.8.30-2 dctrl-tools_2.14.1 ddns3-client_1.8-7 dealer_0.20040530-3 debian-installer_20100722 debianutils_3.4 dhcp3_3.1.3-2 diffpdf_1.1.4-1 diffutils_3.0-1 dime_0.20030921-2 docbook-to-man_1:2.0.0-28 doscan_0.3.1-3 dossizola_1.0-8.2 dovecot-antispam_1.2+20090702-1 dpkg_1.15.8.4 dvblast_1.2-1 e2fsprogs_1.41.12-2 eboard_1.1.1-4.1 eglibc_2.11.2-5 ejabberd_2.1.5-1 ekiga_3.2.7-2 elementtidy_1.0-7 elinks_0.12~pre5-2 encfs_1.6.1-1 espeak_1.43.03-2 evolution-webcal_2.28.1-1 exempi_2.1.1-1 exim4_4.72-1 exiv2_0.20-2 expect_5.44.1.15-3 fastdep_0.16-13 fast-user-switch-applet_2.24.0-4 fetchmail_6.3.17-4 file_5.04-5 filters_2.46 findutils_4.4.2-1 firebird2.1_2.1.3.18185-0.ds1-11 firestarter_1.0.3-8 fkiss_0.35-1.1 floatbg_1.0-28 focal_1.0.1-1 font-manager_0.5.6-1 freefem_3.5.8-4.2 frox_0.7.18-4 fspanel_0.7-9 fsvs_1.1.17-1 fteqcc_3343+svn3400-2 fxload_0.0.20081013-1 g2_0.72-1 g3dviewer_0.2.99.5~svn130-1 gcc-4.1_4.1.2-29 gcin_1.4.6-3 gcolor2_0.4-2 gconf-editor_2.30.0-2 gdbm_1.8.3-9 gedit_2.30.3-1 geg_1.0.2-6 gelemental_1.2.0-4 gettext_0.18.1.1-1 gigolo_0.4.1-1 gimp-plugin-registry_3.5.1-1 gip_1.7.0-1-2 gkremldk_0.9.7-2 glabels_2.2.8-1 gle_3.1.0-7 glfw_2.6-2 gloox_1.0-1 glpk_4.43-1 glui_2.36-4 gnash_0.8.8-2 gnet_2.0.8-2 gnokii_0.6.29.dfsg-1 gnomint_1.1.0-1 gnubg_0.90+20091206-4 gnupg_1.4.10-4 gnuspool_1.7 gnustep-examples_1:1.2.0-3 gnutls26_2.8.6-1 goattracker_2.70-1 gocr_0.48-1 gpe-su_0.20-1 gpe-todo_0.58-1 gpm_1.20.4-3.3 gpx2shp_0.69-3 grads_2.0.a8-1 granule_1.3.0-1 grass_6.4.0~rc6+42329-1 grep_2.6.3-3 groff_1.20.1-10 gsl_1.14+dfsg-1 gtkboard_0.11pre0-12 gtk-theme-switch_2.1.0-2 gtools_2.6.2-1 gtweakui_0.4.0-2 guayadeque_0.2.5-2 gwaei_1.4.1-2 gzip_1.3.12-9 havp_0.91-1.1 hfsplus_1.0.4-12 hildon-theme-tools_0.5.2.debian.1-2 hmisc_3.8-2-1 hostname_3.04 hspell-gui_0.2.6-4.1 html2text_1.3.2a-15 icecc-monitor_2.0~svn951472-2 ifenslave-2.6_1.1.0-17 ifrit_3.3.2-1 ifupdown_0.6.10 igaelic_0.50-7 ii_1.4-3 imagination_2.1-2 infernal_1.0.2-1 inkscape_0.47.0-2 iproute_20100519-3
Bug#519006: mips/ld: non-dynamic relocations refer to dynamic symbol
reassign 519006 binutils
tag 519006 + patch
forwarded 519006 http://sourceware.org/ml/binutils/2010-09/msg00216.html
thanks
On Tue, Sep 14, 2010 at 05:27:04PM +0200, Aurelien Jarno wrote:
I have done a rebuilt of part of the archive using gcc-4.4 that defaults
to -mplt. Unfortunately I don't have enough CPU power to rebuild the
whole archive, so I have rebuilt all packages with priority = important
as well as 436 packages selected randomly with priority optional (see
list below). I have got a few build failures but, not due to the -mplt
change.
Unfortunately we've just discovered that while this approach works for
most of the cases, it doesn't work for all packages.
On the other hand the good news is that Maciej Rozycki has written a
patch for binutils which fixes the real issue. I have converted it to
dpatch and attached to this mail. I have tested on a few packages that
currently fail because of this bug, and they now all build correctly.
--
Aurelien Jarno GPG: 1024D/F1BCDB73
aurel...@aurel32.net http://www.aurel32.net
#!/bin/sh -e
## 150_pr10144.dpatch
##
## DP: Description: http://sourceware.org/ml/binutils/2010-09/msg00216.html
if [ $# -ne 1 ]; then
echo 2 `basename $0`: script expects -patch|-unpatch as argument
exit 1
fi
[ -f debian/patches/00patch-opts ] . debian/patches/00patch-opts
patch_opts=${patch_opts:--f --no-backup-if-mismatch}
case $1 in
-patch) patch $patch_opts -p0 $0;;
-unpatch) patch $patch_opts -p0 -R $0;;
*)
echo 2 `basename $0`: script expects -patch|-unpatch as
argument
exit 1;;
esac
exit 0
2010-09-15 Maciej W. Rozycki ma...@codesourcery.com
PR ld/10144
bfd/
* elfxx-mips.c (_bfd_mips_elf_check_relocs)
[R_MIPS_32, R_MIPS_REL32, R_MIPS_64]: Ignore relocs from
SEC_DEBUGGING sections.
ld/testsuite/
* lib/ld-lib.exp (run_ld_link_tests): Handle sources from other
directories.
(run_ld_link_exec_tests): Likewise.
(run_cc_link_tests): Likewise.
* ld-elf/comm-data1.sd: New test.
* ld-elf/comm-data1.s: Source for the new test.
* ld-elf/comm-data2.sd: New test.
* ld-elf/comm-data2.rd: Likewise.
* ld-elf/comm-data2.xd: Likewise.
* ld-elf/comm-data2.s: Source for the new tests.
* ld-elf/comm-data.exp: Run the new tests.
* ld-mips-elf/comm-data.exp: Likewise.
Index: ld/testsuite/ld-elf/comm-data1.s
===
--- ld/testsuite/ld-elf/comm-data1.s(revision 0)
+++ ld/testsuite/ld-elf/comm-data1.s(revision 0)
@@ -0,0 +1,6 @@
+ .section .rodata,a,%progbits
+ .balign 8
+ .globl foo
+ .type foo,%object
+foo:
+ .skip 4, 0
Index: ld/testsuite/ld-elf/comm-data2.xd
===
--- ld/testsuite/ld-elf/comm-data2.xd (revision 0)
+++ ld/testsuite/ld-elf/comm-data2.xd (revision 0)
@@ -0,0 +1,2 @@
+Hex dump of section '\.debug_foo':
+ +0x0+ + +\.\.\.\.\.\.\.\.\.\.\.\.\.\.\.\.
Index: ld/testsuite/ld-elf/comm-data2.s
===
--- ld/testsuite/ld-elf/comm-data2.s(revision 0)
+++ ld/testsuite/ld-elf/comm-data2.s(revision 0)
@@ -0,0 +1,14 @@
+ .text
+ .globl _start
+ .globl __start
+_start:
+__start:
+ .comm foo, 4, 4
+ .section .debug_foo,,%progbits
+ .balign 16
+ .ifdef ELF64
+ .8byte foo
+ .else
+ .4byte foo
+ .endif
+ .balign 16
Index: ld/testsuite/ld-elf/comm-data.exp
===
--- ld/testsuite/ld-elf/comm-data.exp (revision 0)
+++ ld/testsuite/ld-elf/comm-data.exp (revision 0)
@@ -0,0 +1,76 @@
+# Expect script for common symbol override.
+#
+# Copyright 2010 Free Software Foundation, Inc.
+#
+# This file is part of the GNU Binutils.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, write to the Free Software
+# Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston,
+# MA 02110-1301, USA.
+#
+
+#
+# Written by Maciej W. Rozycki ma...@codesourcery.com
+#
+
+# Exclude non-ELF targets.
+if ![is_elf_format] {
+return
+}
+
+# Exclude non-Linux targets; feel free to include your
Processed: Re: Bug#519006: mips/ld: non-dynamic relocations refer to dynamic symbol
Processing commands for cont...@bugs.debian.org: reassign 519006 binutils Bug #519006 [gcc-4.4] mips/ld: non-dynamic relocations refer to dynamic symbol Bug reassigned from package 'gcc-4.4' to 'binutils'. Bug No longer marked as found in versions gcc-4.4/4.4.2-8. tag 519006 + patch Bug #519006 [binutils] mips/ld: non-dynamic relocations refer to dynamic symbol Added tag(s) patch. forwarded 519006 http://sourceware.org/ml/binutils/2010-09/msg00216.html Bug #519006 [binutils] mips/ld: non-dynamic relocations refer to dynamic symbol Changed Bug forwarded-to-address to 'http://sourceware.org/ml/binutils/2010-09/msg00216.html' from 'http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45387' thanks Stopping processing here. Please contact me if you need assistance. -- 519006: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=519006 Debian Bug Tracking System Contact ow...@bugs.debian.org with problems -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#519006: mips/ld: non-dynamic relocations refer to dynamic symbol
Aurelien Jarno aurel...@aurel32.net writes: Hi all, Hi, I have made some progress on this bug, though it has not progress in the expected direction. In other words I got some surprises. First of all it the bug has been introduced in binutils when introducing the MIPS PLT support [1] (commits around 2008-08-08). I am still convinced it is actually a binutils bug. Secondly, if the problem is actually not present in gcc-snapshot in gcc 4.5 we have *in Debian*, it is not due to the fix linked in the bugzilla entry. It is due to the fact we default to -mplt on these compilers. It is possible to workaround the bug on gcc 4.4 using -mplt (like it is possible to workaround it by dropping the -g). To sum up, there's no fix upstream for the bug but the solution/workaround for now is -mplt, right ? Given all of that, the only forseen solution for this fix is to also default to -mplt on gcc 4.4. It also brings speed improvements, as well as some more standard binaries with regard to other architectures. On the other hand, I understand it is something quite risky so close to a release. If we go for this solution, we may want If the bug is fixed by this change and as a side effect, there are improvements, there's no reason to not include it. I don't think it can be worse than the current situation. Moreover (iirc) this bug is a release blocker for mips so seeing it closed would be really nice. to trigger a rebuild of part of the archive as it is currently done on sparc. There's a bunch of packages known to be bitten by this bug but there may be others we're not aware of. imho a rebuild is mandatory. This would also give us a good up-to-date status of the mips/mipsel ports for the release. My 2 cents, Arnaud -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#519006: mips/ld: non-dynamic relocations refer to dynamic symbol
Hi all, I have made some progress on this bug, though it has not progress in the expected direction. In other words I got some surprises. First of all it the bug has been introduced in binutils when introducing the MIPS PLT support [1] (commits around 2008-08-08). I am still convinced it is actually a binutils bug. Secondly, if the problem is actually not present in gcc-snapshot in gcc 4.5 we have *in Debian*, it is not due to the fix linked in the bugzilla entry. It is due to the fact we default to -mplt on these compilers. It is possible to workaround the bug on gcc 4.4 using -mplt (like it is possible to workaround it by dropping the -g). Given all of that, the only forseen solution for this fix is to also default to -mplt on gcc 4.4. It also brings speed improvements, as well as some more standard binaries with regard to other architectures. On the other hand, I understand it is something quite risky so close to a release. If we go for this solution, we may want to trigger a rebuild of part of the archive as it is currently done on sparc. Cheers, Aurelien [1] http://readlist.com/lists/gcc.gnu.org/gcc/4/22455.html diff -u gcc-4.4-4.4.4/debian/changelog gcc-4.4-4.4.4/debian/changelog --- gcc-4.4-4.4.4/debian/changelog +++ gcc-4.4-4.4.4/debian/changelog @@ -1,3 +1,10 @@ +gcc-4.4 (4.4.4-11+mipsplt) UNRELEASED; urgency=low + + [ Aurelien Jarno ] + * Default to -mplt on mips(el). + + -- Aurelien Jarno aure...@debian.org Mon, 30 Aug 2010 15:58:29 + + gcc-4.4 (4.4.4-11) unstable; urgency=medium * Update to SVN 20100824 from the gcc-4_4-branch (r163521). diff -u gcc-4.4-4.4.4/debian/rules2 gcc-4.4-4.4.4/debian/rules2 --- gcc-4.4-4.4.4/debian/rules2 +++ gcc-4.4-4.4.4/debian/rules2 @@ -384,12 +384,14 @@ endif ifneq (,$(findstring mips-linux,$(DEB_TARGET_GNU_TYPE))) + CONFARGS += --with-mips-plt ifeq ($(biarchn32)-$(biarch64),yes-yes) CONFARGS += --enable-targets=all endif endif ifneq (,$(findstring mipsel-linux,$(DEB_TARGET_GNU_TYPE))) + CONFARGS += --with-mips-plt ifeq ($(biarchn32)-$(biarch64),yes-yes) CONFARGS += --enable-targets=all endif -- Aurelien Jarno GPG: 1024D/F1BCDB73 aurel...@aurel32.net http://www.aurel32.net -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#519006: mips/ld: non-dynamic relocations refer to dynamic symbol
reassign 519006 gcc-4.4 found 519006 4.4.2-8 thanks This bug is actually a GCC bug, present in 4.4 and fixed in 4.5. The patch that fixes the problem is available [1], but is not easily backportable without good GCC knowledge. A workaround to this bug is to not compile with -g. [1] http://gcc.gnu.org/ml/gcc-patches/2009-09/msg02070.html -- Aurelien Jarno GPG: 1024D/F1BCDB73 aurel...@aurel32.net http://www.aurel32.net -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#519006: mips/ld: non-dynamic relocations refer to dynamic symbol
Hi, while fixing #531937 I found out how to create a very minimalistic testcase. Apparently the diagnosis that ld cannot “override” libc symbols with local symbols was correct. I have created an attachment to the upstream-filed bug at: http://sources.redhat.com/bugzilla/show_bug.cgi?id=10144 I eventually fixed #531937 since the real bug was something else that only happened to be noticed because of this toolchain bug, but actually may have affected the programme negatively, so this bug can be used for testing ☺ HTH HAND, //mirabilos -- Sometimes they [people] care too much: pretty printers [and syntax highligh- ting, d.A.] mechanically produce pretty output that accentuates irrelevant detail in the program, which is as sensible as putting all the prepositions in English text in bold font. -- Rob Pike in Notes on Programming in C -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
