|
From: <sa...@lu...> - 2003-10-16 22:18:23
|
Hi, What does this error mean? Memcheck: the `impossible' happened: add_MAC_chunk: shadow area is accessible Basic block ctr is approximately 111850000 Is it clear indication that something is wrong in my application or is it i= ndication that I=92m doing something that is not fully supported by Valgrin= d tool? My application seems to be stable and I have not noticed any leakag= e using other tools. The application is using commercial in memory database= , which is using shared memory to store the data. When I comment out the pa= rts that are using the IMDB from the application I don=92t get the error.= =20 Has anyone else run into this kind of problem? I can submit more details if= some one wants to study the issue. As a temporary solution I simply commented out "paranoid" check generating = the error message (lines 146 - 148 from memcheck/mac_malloc_wrappers.c file= ), at least I can now get leak summary when application exits. BR, Sami ............................................................ Maksuton s=E4hk=F6posti aina k=E4yt=F6ss=E4 http://luukku.com Kuukausimaksuton MTV3 Internet-liittym=E4 www.mtv3.fi/liittyma |
While running some SSE optimized code (not the same program as in last
message) with valgrind 2.1.0, I get the following error:
frame size: 160
0: SSE2a1_MRdQQ 0xF:0xC6:0x55:0x55(t34)
Memcheck: the `impossible' happened:
memcheck_instrument: unhandled case
Basic block ctr is approximately 0
=3D=3D16868=3D=3D at 0x401593A8: vgPlain_core_panic (vg_mylibc.c:1121)
=3D=3D16868=3D=3D by 0x401593A7: panic (vg_mylibc.c:1117)
=3D=3D16868=3D=3D by 0x401593F0: vgPlain_skin_panic (vg_mylibc.c:1127)
=3D=3D16868=3D=3D by 0x400128E3: memcheck_instrument (mc_translate.c:118=
7)
sched status:
Thread 1: status =3D Runnable, associated_mx =3D 0x0, associated_cv =3D 0x0
=3D=3D16868=3D=3D at 0x80532CA: fir_mem2_10 (in
/home/jm/dsp/speex/libspeex/testenc)
=3D=3D16868=3D=3D by 0x804AA26: nb_encode (in
/home/jm/dsp/speex/libspeex/testenc)
=3D=3D16868=3D=3D by 0x80495A4: speex_encode (in
/home/jm/dsp/speex/libspeex/testenc)
=3D=3D16868=3D=3D by 0x8048AE0: main (in /home/jm/dsp/speex/libspeex/tes=
tenc)
Setup is:
Valgrind 2.1.0
Fedora Core 1
Kernel 2.6.2-mm1 highmem (4G)
Pentium-M 1.6 GHz
Program uses only SSE1 intrinsics and compiled with -O3 -march=3Dpentium3
-msse
Jean-Marc
--=20
Jean-Marc Valin, M.Sc.A., ing. jr.
LABORIUS (http://www.gel.usherb.ca/laborius)
Universit=E9 de Sherbrooke, Qu=E9bec, Canada
|
|
From: Tom H. <th...@cy...> - 2004-02-12 08:24:08
|
In message <107...@id...>
Jean-Marc Valin <Jean-Marc.Valin@USherbrooke.ca> wrote:
> While running some SSE optimized code (not the same program as in last
> message) with valgrind 2.1.0, I get the following error:
>
> frame size: 160
> 0: SSE2a1_MRdQQ 0xF:0xC6:0x55:0x55(t34)
>
> Memcheck: the `impossible' happened:
> memcheck_instrument: unhandled case
I believe the CVS head code should handle this - the fix was in my
patch that Nick committed yesterday.
Tom
--
Tom Hughes (th...@cy...)
Software Engineer, Cyberscience Corporation
http://www.cyberscience.com/
|