Hello,
I'm not sure if this is related to the base system or some specific port.
I'm running FreeBSD as a desktop machine on a ThinkPad X220.
Ports/packages are managed by ports-mgmt/synth and packages are fetched from
After upgrading from FreeBSD 11.0-RELEASE-p1 amd64 to 11.1-RELEASE-p4 as described in (or as I understood) the handbook via
and doing
many applications crash sooner or later with a segmentation fault.
It seems that the duration until crash is somewhat related to resource intensity as www/firefox or mail/thunderbird segfault nearly immediately while www/dillo or mail/claws-mail may run for an hour (depending on usage).
This didn't happen to X window manager itself (I'm using x11-wm/xfce4 and x11-wm/i3) or command line programs (I tried www/links, net-im/profanity, and others), but I cannot say if they would crash eventually too if running long enough.
I enabled coredumps via
I thought it was unnecessary to recompile all ports when doing a minor upgrade.
Nevertheless I did a
Finally, I gave up and made a (dirty) rollback via
Everything is running as stable as before, now.
What else can I do to isolate my problem? I tried
Thanks in advance.
I'm not sure if this is related to the base system or some specific port.
I'm running FreeBSD as a desktop machine on a ThinkPad X220.
Ports/packages are managed by ports-mgmt/synth and packages are fetched from
url: "pkg+http://pkg.FreeBSD.org/${ABI}/latest"
.After upgrading from FreeBSD 11.0-RELEASE-p1 amd64 to 11.1-RELEASE-p4 as described in (or as I understood) the handbook via
Code:
# freebsd-update -r 11.1-RELEASE upgrade
# freebsd-update install
# reboot
# freebsd-update install
Code:
# portsnap fetch update
# synth update-system
It seems that the duration until crash is somewhat related to resource intensity as www/firefox or mail/thunderbird segfault nearly immediately while www/dillo or mail/claws-mail may run for an hour (depending on usage).
This didn't happen to X window manager itself (I'm using x11-wm/xfce4 and x11-wm/i3) or command line programs (I tried www/links, net-im/profanity, and others), but I cannot say if they would crash eventually too if running long enough.
I enabled coredumps via
sysctl kern.coredump=1
to see if I could narrow the problem down to a single library or function call. Unfortunately they weren't very helpful to me, for example:
Code:
(gdb) core thunderbird.core
Core was generated by `thunderbird'.
Program terminated with signal 11, Segmentation fault.
#0 0x000000080207b84a in ?? ()
(gdb) core claws-mail.core
Core was generated by `claws-mail'.
Program terminated with signal 11, Segmentation fault.
#0 0x000000080b36dba9 in ?? ()
Nevertheless I did a
synth force `pkg query %o`
, which took quite long but didn't help.Finally, I gave up and made a (dirty) rollback via
Code:
# mv /boot/kernel /boot/kernel.11.1
# mv /boot/kernel.old1 /boot/kernel
What else can I do to isolate my problem? I tried
ktrace
and kdump
, but I've to admit that I don't understand its output as it was full of error messages which may have nothing to do with the described issue.Thanks in advance.