Control: tag -1 moreinfo On Mon, Mar 09, 2020 at 10:10:29AM +0100, Félix Sipma wrote: > Package: apt > Version: 1.8.4 > Severity: important > > Running "apt upgrade" I got: > > Segmentation fault (core dumped) > > /var/log/messages reports: > > Mar 9 10:03:48 laptop kernel: [ 3262.779582] apt[44617]: segfault at 0 > ip 00007f024bc54e6f sp 00007ffcc6b28d40 error 4 in > libapt-pkg.so.5.0.2[7f024bbe4000+148000] > Mar 9 10:03:48 laptop kernel: [ 3262.779588] Code: 08 01 48 8d 35 1c b6 > 0d 00 4c 89 fa 48 8b 38 e8 a7 04 f9 ff 41 89 c4 84 c0 74 50 45 84 ed 74 21 48 > 8b 05 7c 7d 12 00 48 8b 38 <48> 8b 07 ff 10 41 89 c4 84 c0 0f 84 e3 00 00 00 > 48 8b 43 08 c6 00 > > Attached is the error output of "strace apt upgrade".
What do you think an strace is going to tell us? Please Install gdb and debug symbols for apt, libstdc++6, libc6, and produce a backtrace. So libapt-pkg5.0-dbgsym, libstdc++6-dbgsym, libc6-dbgsym at a minimum. You might want to install systemd-coredump so that cores are always dumped and stored and you can then analyze them with coredumpctl gdb and get a backtrace there to send in bug report. You probably want to configure gdb to log to a file and get a full backtrace, then attach that file, e.g. $ gdb --args apt upgrade set logging overwrite on set logging file /tmp/apt-crash.log set logging on set pagination 0 run backtrace full quit $ ... attach /tmp/apt-crash.log to bug report ... The same thing without the run if you used coredumpctl gdb, because, um, you are analysing a prior crash rather than making it crash again :) -- debian developer - deb.li/jak | jak-linux.org - free software dev ubuntu core developer i speak de, en
signature.asc
Description: PGP signature