valgrind-users

[Valgrind-users] Assertion Failed

[Seth L. <se...@br...>]

Hello,

I'm attempting to run valgrind 1.9.4 on redhat 8.1 beta, using gcc 
3.2.1.  I'm trying to run a program under it, but I get a Assertion 
failed message.  The full text of the output is included below.  I 
checked the mailing list archives, and the nano-faq, but didn't see this 
mentioned.  Does any of this ring a bell?  Is there something I can do?

Thanks very much,
Seth

[root@disheveled .libs]# valgrind -v ./state testmode
shell-init: could not get current directory: getcwd: cannot access 
parent direct
ories: No such file or directory
sh_makepath: could not get current directory: getcwd: cannot access 
parent direc
tories: No such file or directory
==15339== Memcheck, a.k.a. Valgrind, a memory error detector for x86-linux.
==15339== Copyright (C) 2002, and GNU GPL'd, by Julian Seward.
==15339== Using valgrind-1.9.4, a program instrumentation system for 
x86-linux.
==15339== Copyright (C) 2000-2002, and GNU GPL'd, by Julian Seward.
==15339== Startup, with flags:
==15339==    --suppressions=/usr/local/lib/valgrind/default.supp
==15339==    -v
==15339== Reading suppressions file: /usr/local/lib/valgrind/default.supp
==15339== Estimated CPU clock rate is 865 MHz
==15339==

valgrind: vg_ldt.c:167 (vgPlain_do_useseg): Assertion `(seg_selector & 
7) == 7'
failed.

sched status:
Thread 1: status = Runnable, associated_mx = 0x0, associated_cv = 0x0
==15339== Reading syms from /home/seth/code/G4/src/.libs/state
==15339== Reading syms from /lib/ld-2.3.1.so
==15339== Reading syms from /usr/local/lib/valgrind/vgskin_memcheck.so
==15339== Reading syms from /usr/local/lib/valgrind/valgrind.so
==15339== Reading syms from 
/home/seth/code/G4/external/dist/lib/libcomms.so.0.0
.0
==15339== Reading syms from 
/home/seth/code/G4/external/dist/lib/libcurl.so.2.0.
2
==15339== Reading syms from 
/home/seth/code/G4/external/dist/lib/libssl.so.0.9.7
==15339== Reading syms from /lib/libdl-2.3.1.so
==15339==    object doesn't have any debug info
==15339== Reading syms from 
/home/seth/code/G4/external/dist/lib/libcore.so.0.0.
0
==15339== Reading syms from 
/home/seth/code/G4/external/dist/lib/libbrivodb.so.0
.0.0
==15339== Reading syms from 
/home/seth/code/G4/external/dist/lib/librdf.so.0.0.0
==15339== Reading syms from 
/home/seth/code/G4/external/dist/lib/libraptor.so.0.
0.0
==15339== Reading syms from 
/home/seth/code/G4/external/dist/lib/libxml2.so.2.5.
3
==15339== Reading syms from 
/home/seth/code/G4/external/dist/lib/libz.so.1.1.4
==15339== Reading syms from 
/home/seth/code/G4/external/dist/lib/libcrypto.so.0.
9.7
==15339== Reading syms from 
/home/seth/code/G4/external/dist/lib/libinterface.so
.0.0.0
==15339== Reading syms from /usr/local/lib/valgrind/libpthread.so
==15339== Reading syms from /usr/lib/libstdc++.so.5.0.2
==15339== Reading syms from /lib/tls/libm-2.3.1.so
==15339== Reading syms from /lib/libgcc_s-3.2.1-20021208.so.1
==15339== Reading syms from /lib/tls/libc-2.3.1.so
==15339==    at 0x42029BA1: __new_exitfn (in /lib/tls/libc-2.3.1.so)
==15339==    by 0x42029B58: __cxa_atexit_internal (in 
/lib/tls/libc-2.3.1.so)
==15339==    by 0x404767E7: 
__static_initialization_and_destruction_0(int, int)
(/usr/src/build/177318-i386/BUILD/gcc-3.2.1-20021208/obj-i386-redhat-linux/i386-
redhat-linux/libstdc++-v3/include/iostream:63)
==15339==    by 0x40476839: 
_GLOBAL__I__ZThn8_NSt14basic_iostreamIwSt11char_trai
tsIwEED0Ev.._.._.._.._libstdc___v3_src_io_inst.ccJGVYgb 
(/usr/src/build/177318-i
386/BUILD/gcc-3.2.1-20021208/obj-i386-redhat-linux/i386-redhat-linux/libstdc++-v
3/include/bits/locale_facets.h:107)

Please report this bug to: js...@ac...

[root@disheveled .libs]# gcc -v
Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/3.2.1/specs
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man 
--infodir=/usr/share/info --enable-shared --enable-threads=posix 
--disable-checking --with-system-zlib --enable-__cxa_atexit 
--host=i386-redhat-linux
Thread model: posix
gcc version 3.2.1 20021207 (Red Hat Linux 8.0 3.2.1-2)

Re: [Valgrind-users] Assertion Failed

[Tom H. <th...@cy...>]

In message <3E7...@br...>
        Seth Ladd <se...@br...> wrote:

> I'm attempting to run valgrind 1.9.4 on redhat 8.1 beta, using gcc
> 3.2.1.  I'm trying to run a program under it, but I get a Assertion
> failed message.  The full text of the output is included below.  I
> checked the mailing list archives, and the nano-faq, but didn't see
> this mentioned.  Does any of this ring a bell?  Is there something I
> can do?

RedHat 8.1 uses glibc 2.3 with the NPTL thread package instead of
the linuxthreads thread package and valgrind is therefore highly
unlikely to work with threaded programs under RedHat 8.1 until some
work is done to make it handle the changes to threading.

The fact that this assertion is talking about segment selectors
suggests that thread local storage is involved, and hence this is
probably a threading related problem.

Before anybody worries about the update to glibc 2.3.2 that was
released for RedHat 8.0 the other day, that version was built with
linuxthreads so there is no problem there.

Tom

-- 
Tom Hughes (th...@cy...)
Software Engineer, Cyberscience Corporation
http://www.cyberscience.com/

Re: [Valgrind-users] Assertion Failed

[Julian S. <js...@ac...>]

> > valgrind: vg_ldt.c:167 (vgPlain_do_useseg): Assertion `(seg_selector & 
> > 7) == 7'

> The fact that this assertion is talking about segment selectors
> suggests that thread local storage is involved, and hence this is
> probably a threading related problem.

The lowest 3 bits of the segment selector is a tag which indicates some
properties of the segment (iirc).  111b means a LDT selector, and that's
what the above assertion tests.   I think you've probably hit a 
GDT selector (011b iirc).

But anyway, as Tom says, R H 8.1 contains NPTL, which V can't emulate
well enough to be useful, and we don't have an ETA on fixing it either.
Best advice is to avoid 8.1 if you want to use V.  At least for the
moment.

J

Re: [Valgrind-users] Assertion Failed

[Tom H. <th...@cy...>]

In message <200...@ac...>
          Julian Seward <js...@ac...> wrote:

> But anyway, as Tom says, R H 8.1 contains NPTL, which V can't emulate
> well enough to be useful, and we don't have an ETA on fixing it either.
> Best advice is to avoid 8.1 if you want to use V.  At least for the
> moment.

I wrote my earlier message based on your comment in the documentation
that NPTL will need extra work, but I was thinking afterwards that as
valgrind replaces libpthread.so with it's own version that NPTL won't
actually get a chance to be involved will it?

Presumably there must be some other problem though, perhaps with the
use of thread local storage by the main C library or something?

Tom

-- 
Tom Hughes (th...@cy...)
Software Engineer, Cyberscience Corporation
http://www.cyberscience.com/

Re: [Valgrind-users] Assertion Failed

[Tom H. <th...@cy...>]

In message <200...@ac...>
          Julian Seward <js...@ac...> wrote:

> 
> > > valgrind: vg_ldt.c:167 (vgPlain_do_useseg): Assertion `(seg_selector &
> > > 7) == 7'
> 
> > The fact that this assertion is talking about segment selectors
> > suggests that thread local storage is involved, and hence this is
> > probably a threading related problem.
> 
> The lowest 3 bits of the segment selector is a tag which indicates some
> properties of the segment (iirc).  111b means a LDT selector, and that's
> what the above assertion tests.   I think you've probably hit a
> GDT selector (011b iirc).

I suspect the same - the NPTL documentation does mention that it
uses the GDT to handle thread local storage instead of the LDT that
is used by the old threading system.

There is also a new system call, tls, to manage the GDT so valgrind
is probably going to need to intercept that and emulate the GDT in a
similar way to that the LDT emulation is does at the moment.

Tom

-- 
Tom Hughes (th...@cy...)
Software Engineer, Cyberscience Corporation
http://www.cyberscience.com/