Hello Sébastien.
I think it's quite important to have it fixed in stable since it's quite
tricky to debug and results in an error while installing or upgrading
octave. Does that qualify as a major impact? :)
Tks.
Il 05/11/2021 14:35, Sébastien Villemot ha scritto:
Control: retitle -1 segfault when running out of memory
Control: tags -1 = upstream fixed-upstream
Dear Diego,
Le vendredi 05 novembre 2021 à 06:34 +0100, Diego Zuccato a écrit :
Il 04/11/2021 12:27, Sébastien Villemot ha scritto:
I happen to have access to a machine with exactly the same CPU (and
Debian Bullseye). I fail to replicate your problem there. I can start
octave with libopenblas0-pthread and there is no segfault.
So there must be another factor at play.Uhm...
I use pam-limits and cgroups to limit the memory available to users.
I confirm that I can replicate the segfault with:
$ ulimit -d 1024000
$ octave
The crash occurs in function blas_memory_alloc(), as in your backtrace.
Thanks to your analysis, upstream has fixed the issue. I will therefore
fix it in Debian unstable (either with the next upstream release, or
possibly even earlier with a targeted patch).
Are you interested in having this bug fixed in Debian stable
(bullseye)? In theory this possibility is only open for bugs of
severity “important” or higher, while the bug is currently marked with
severity “normal”. A bug of severity important is defined as having “a
major effect on the usability of a package, without rendering it
completely unusable to everyone”. What do you think?
Best,
--
Diego Zuccato
DIFA - Dip. di Fisica e Astronomia
Servizi Informatici
Alma Mater Studiorum - Università di Bologna
V.le Berti-Pichat 6/2 - 40127 Bologna - Italy
tel.: +39 051 20 95786
