Le forum de XCAS

Xcas: un logiciel libre de calcul formel
Nous sommes actuellement le Mer Nov 22, 2017 8:59 am

Heures au format UTC




Publier un nouveau sujet Répondre au sujet  [ 16 messages ]  Aller à la page 1, 2  Suivant
Auteur Message
MessagePublié: Mar Sep 12, 2017 8:12 pm 
Hors-ligne

Inscrit le: Mar Sep 12, 2017 6:40 pm
Messages: 8
Hi,

all of xcas' tests fail on big endian architectures on Debian. You can see the build logs here:
https://buildd.debian.org/status/packag ... sg1-2%2Bb1

The big endian architectures are mips, s390x, hppa, m68k, powerpc, ppc64 and sparc64. Do you have an idea what's wrong?

The Debian bug report is here:
https://bugs.debian.org/867172

Thanks,
Tobias


Haut
 Profil  
 
MessagePublié: Jeu Sep 14, 2017 6:10 am 
Hors-ligne

Inscrit le: Mar Déc 20, 2005 4:02 pm
Messages: 4104
I don't see any information in the log except FAIL ...
What failed? Is it a segfault of the executable (can we get a stack trace?) or a result that is not the same as the expected result?


Haut
 Profil  
 
MessagePublié: Jeu Sep 14, 2017 8:43 am 
Hors-ligne

Inscrit le: Mar Sep 12, 2017 6:40 pm
Messages: 8
Here are the test logs from s390x: https://people.debian.org/~thansen/make ... 90x.tar.xz

After chk_fhan16 it started to hang. I waited at least half an hour and aborted. Hope this helps.


Haut
 Profil  
 
MessagePublié: Jeu Sep 14, 2017 9:36 am 
Hors-ligne

Inscrit le: Mar Déc 20, 2005 4:02 pm
Messages: 4104
It's full wrong, I have no idea why. I'm afraid I can not fix anything unless I have access to a shell with emacs and gdb on a similar arch.


Haut
 Profil  
 
MessagePublié: Jeu Sep 14, 2017 10:16 am 
Hors-ligne

Inscrit le: Mar Sep 12, 2017 6:40 pm
Messages: 8
Debian has machines for this purpose, but they don't like to give access to people who are not involved with Debian...

Has Giac been tested on a big endian architecture at some point in the past? If it is expected to work, maybe somebody who sees the bug (here or in Debian) will fix it eventually.


Haut
 Profil  
 
MessagePublié: Jeu Sep 14, 2017 10:49 am 
Hors-ligne

Inscrit le: Mar Déc 20, 2005 4:02 pm
Messages: 4104
Is it related to http://xcas.e.ujf-grenoble.fr/XCAS/viewtopic.php?f=4&t=1785 ?


Haut
 Profil  
 
MessagePublié: Jeu Sep 14, 2017 10:57 am 
Hors-ligne

Inscrit le: Mar Sep 12, 2017 6:40 pm
Messages: 8
The change from there is applied and it fixed the problem on arm64, but not on the big endian architectures.


Haut
 Profil  
 
MessagePublié: Jeu Sep 14, 2017 11:36 am 
Hors-ligne

Inscrit le: Mar Déc 20, 2005 4:02 pm
Messages: 4104
Is it possible to test with non default compile flags?
For example -DDOUBLEVAL would disable some optimizations in the gen type that are probably not valid here.


Haut
 Profil  
 
MessagePublié: Jeu Sep 14, 2017 12:52 pm 
Hors-ligne

Inscrit le: Mar Sep 12, 2017 6:40 pm
Messages: 8
That helped: Now there's a crash in icas with backtrace, 5 fails, 18 passes, then it hangs. See make-check.log:

https://people.debian.org/~thansen/giac ... 90x.tar.xz


Haut
 Profil  
 
MessagePublié: Jeu Sep 14, 2017 6:18 pm 
Hors-ligne

Inscrit le: Mar Déc 20, 2005 4:02 pm
Messages: 4104
I compiled with -DDOUBLEVAL (and SMARTPTR64 disabled) to compare. I get the same diffs for chk_cas chk_fhan0 and chk_fhan2 (since DOUBLEVAL enables standard double with 53 bits of mantissa instead of truncated double with 48 bits of mantissa, roundoff errors are expected).
We have therefore only 2 failures remaining, both with segfaults. Unfortunately the trace is not useful.
I would recommend to recompile with -g, and trace the segfault with gdb and valgrind on ../src/icas testnormalize or ../src/icas testlimit


Haut
 Profil  
 
MessagePublié: Jeu Sep 14, 2017 7:24 pm 
Hors-ligne

Inscrit le: Mar Sep 12, 2017 6:40 pm
Messages: 8
I attached gdb backtraces.


Pièces jointes:
gdb_backtraces.zip [9.23 Kio]
Téléchargé 10 fois
Haut
 Profil  
 
MessagePublié: Ven Sep 15, 2017 6:13 am 
Hors-ligne

Inscrit le: Mar Déc 20, 2005 4:02 pm
Messages: 4104
It does not help much. Maybe valgrind would give more information, but if I'm not able to debug myself and move in frames to print variables and set breakpoints, I won't be able to fix something.


Haut
 Profil  
 
MessagePublié: Ven Sep 15, 2017 10:24 am 
Hors-ligne

Inscrit le: Mar Sep 12, 2017 6:40 pm
Messages: 8
Ok, thanks anyway. I enabled the -DDOUBLEVAL flag in the Debian package for big endian architectures, so if someone looks at this in the future they are one step further.


Haut
 Profil  
 
MessagePublié: Ven Sep 15, 2017 10:40 am 
Hors-ligne

Inscrit le: Mar Déc 20, 2005 4:02 pm
Messages: 4104
If you have some good pointers on how to install something like qemu for one of this arch, maybe I could try to debug myself.


Haut
 Profil  
 
MessagePublié: Ven Sep 15, 2017 10:51 am 
Hors-ligne

Inscrit le: Mar Sep 12, 2017 6:40 pm
Messages: 8
I have never used qemu. Didn't you mention in another thread that you debugged arm64 with qemu? Shouldn't it work the same way with mips or s390x?


Haut
 Profil  
 
Afficher les messages publiés depuis:  Trier par  
Publier un nouveau sujet Répondre au sujet  [ 16 messages ]  Aller à la page 1, 2  Suivant

Heures au format UTC


Qui est en ligne ?

Utilisateurs parcourant actuellement ce forum : Aucun utilisateur inscrit et 4 invités


Vous ne pouvez pas publier de nouveaux sujets dans ce forum
Vous ne pouvez pas répondre aux sujets dans ce forum
Vous ne pouvez pas éditer vos messages dans ce forum
Vous ne pouvez pas supprimer vos messages dans ce forum
Vous ne pouvez pas insérer de pièces jointes dans ce forum

Rechercher pour:
Sauter vers:  
Powered by phpBB® Forum Software © phpBB Group
Traduction réalisée par Maël Soucaze © 2009 phpBB.fr