This is the mail archive of the
libc-alpha@sources.redhat.com
mailing list for the glibc project.
Inconsistency detected by ld.so: dynamic-link.h: 161: elf_get_dynamic_info:Assertion `info[15] == ((void *)0)' failed!
- From: "Jeremy C. Reed" <reed at reedmedia dot net>
- To: libc-alpha at sources dot redhat dot com
- Date: Thu, 25 Sep 2003 18:31:04 -0700 (PDT)
- Subject: Inconsistency detected by ld.so: dynamic-link.h: 161: elf_get_dynamic_info:Assertion `info[15] == ((void *)0)' failed!
(If there is a better forum for assistance, please let me know. Thank
you.)
I have built glibc-2.3.2 several times over the past 25+ days on different
hardware, under different 2.2 and 2.4 Linux kernels, and using different
binutils. When building or unsuccessfully using glibc-2.3.2, I continue to
get the following error:
Inconsistency detected by ld.so: dynamic-link.h: 161:
elf_get_dynamic_info: Assertion `info[15] == ((void *)0)' failed!
This last time couple builds it happened during the sunrpc build part when
configured with --enable-add-ons=linuxthreads. I also tried with and
without --prefix=/usr, --disable-profile, --with-gd=no, --without-tls and
--without-__thread.
CPP='gcc -E -x c-header' /home/reed/glibc-build/elf/ld-linux.so.2 --library-path
/home/reed/glibc-build:/home/reed/glibc-build/math:/home/reed/glibc-build/elf:/home/reed/glibc-build/dlfcn:/home/reed/glibc-build/nss:/home/reed/glibc-build/nis:/home/reed/glibc-build/rt:/home/reed/glibc-build/resolv:/home/reed/glibc-build/crypt:/home/reed/glibc-build/linuxthreads
/home/reed/glibc-build/sunrpc/rpcgen -Y ../scripts -c
rpcsvc/bootparam_prot.x -o /home/reed/glibc-build/sunrpc/xbootparam_prot.T
Inconsistency detected by ld.so: dynamic-link.h: 161:
elf_get_dynamic_info: Assertion `info[15] == ((void *)0)' failed!
gmake[2]: *** [/home/reed/glibc-build/sunrpc/xbootparam_prot.stmp] Error 127
gmake[2]: Leaving directory `/home/reed/glibc-2.3.2/sunrpc'
I don't know if it is related, but there are many warnings like:
... inlining failed in call to ...
I am using a stdio-common/sscanf.c patch.
I was using --host=i386-pc-linux so I had to touch
stdio-common/errlist-compat.c (due to "errlist.c count 126 vs Versions
sys_errlist@GLIBC_2.0 count 123" error) before building. I didn't need to
touch stdio-common/errlist-compat.c when not using --host.
I used GCC 3.3.1. I don't know if it is related, but gcc is built with
DRIVER_DEFINES = \
-DLINK_LIBGCC_SPEC="\"%D $(RPATH_FLAG)$(prefix)/lib\"" \
and
TARGET_LIBGCC2_CFLAGS += -fPIC
And I have tried with binutils 2.13.2.1 and 2.14.0.0.
I also included the glibc-linuxthreads add-on.
It was configured with --enable-add-ons=linuxthreads. I have used
different --prefix and also tried with and without --disable-profile,
--with-gd=no, --without-tls and --without-__thread.
Sometimes, I get it to finish building. I used install_root to choose
where to install in a temporary location. And then my install logging had
many lines like:
wa_BE.ISO-8859-15@euro...Inconsistency detected by ld.so: dynamic-link.h:
161: elf_get_dynamic_info: Assertion `info[15] == ((void *)0)' failed!
Also running commands, like:
# vi
Inconsistency detected by ld.so: dynamic-link.h: 161:
elf_get_dynamic_info: Assertion `info[15] == ((void *)0)' failed!
And using the new ld-2.3.2.so always gives that error (and is unusable).
$ /home/reed/glibc-build/elf/ld-linux.so.2
Inconsistency detected by ld.so: dynamic-link.h: 161:
elf_get_dynamic_info: Assertion `info[15] == ((void *)0)' failed!
$
I searched the bug-glibc archives and looked at the FAQ but didn't see any
reference to this. I also saw some postings via google, but didn't find
(or understand) any replies.
dynamic-link.h has:
#ifdef RTLD_BOOTSTRAP
/* The dynamic linker should have none of these set. */
assert (info[DT_RUNPATH] == NULL);
assert (info[DT_RPATH] == NULL);
#else
glibc-2.3.2/elf/elf.h has:
#define DT_RPATH 15 /* Library search path (deprecated) */
Why is it "deprecated"? Is that part of my problem?
glibc-2.3.2/elf/rtld.c defines RTLD_BOOTSTRAP. It says "This #define
produces dynamic linking inline functions for bootstrap relocation instead
of general-purpose relocation."
Do I need this bootstrap relocation?
Can anyone tell me what this error means?
And what should I do to fix or workaround this?
If I overlooked some documentation about this, please point me to it.
If a developer wants to log on to one of my build machines, please let me
know.
I can provide logs, configure output, etc.
(By the way, http://www-gnats.gnu.org:8080/cgi-bin/wwwgnats.pl appears to
be down or unavailable for me and the bug-glibc list is not useful.
Hopefully, libc-alpha is correct list for this. If there is a more
appropriate forum, please let me know.)
Thank you,
Jeremy C. Reed
http://www.isp-faq.com/