Re: [Valgrind-users] AMD64/Linux Valgrind working pretty well

[Nicholas N. <nj...@cs...>]

Paul,

Are you still seeing the problems you reported below?  Thanks.

N


On Sat, 2 Jul 2005, Paul Pluzhnikov wrote:

> Today's SVN checkout fails to build on
> Red Hat Enterprise Linux AS release 3 (Taroon Update 2) x86_64 with
> gcc (GCC) 3.2.3 20030502 (Red Hat Linux 3.2.3-39), binutils-2.14.90.0.4-35
>
> $ gcc -pie  -m64 ... m_syswrap/libsyswrap.a
> /home/paul/valgrind-3.0-svn/vex/libvex.a -ldl
> /usr/bin/ld: /home/paul/valgrind-3.0-svn/vex/libvex.a(irdefs.o):
> relocation R_X86_64_32S can not be used when making a shared object;
> recompile with -fPIC
> /home/paul/valgrind-3.0-svn/vex/libvex.a: could not read symbols: Bad value
> collect2: ld returned 1 exit status
>
> Removing '-pie' solves that, but the resulting valgrind crashes at startup.
> Adding '-fPIC' to CFLAGS in vex/Makefile also solves this problem and produces a
> working VG.
>
> Every program on this system produces several of these:
> ==15027== Warning: zero-sized CIE/FDE but not at section end in DWARF2
> CFI reading
>
> Finally, Insure++-instrumented binaries trigger:
> valgrind: syswrap-amd64-linux.c:718
> (vgSysWrap_amd64_linux_sys_arch_prctl_before): Assertion 'ARG1 ==
> VKI_ARCH_SET_FS' failed.
>
> On FC2-i686 the same test triggers:
> valgrind: syswrap-main.c:812 (vgPlain_post_syscall): Assertion
> 'eq_SyscallStatus( &sci->status, &test_status )' failed.
>
> The same binary runs fine on FC2-i686 under VG-2.4. This may be a case
> of 2 memory debugers fighting each other.
>
> Regards,
>