Closed toreanderson closed 5 years ago
BTW: 3.6.0-rc4 is not really the last release candidate of anything; 3.6 was renamed into 4.0. That's why the sequence is 3.6.0-rc1
, 3.6.0-rc2
, 3.6.0-rc3
, 3.6.0-rc4
, 4.0.0-rc5
and 4.0.0
.
All of these expect the new command line syntax:
modprobe jool_siit
jool_siit instance add --netfilter -6 <pool6>
jool_siit eamt add <eam4-1> <eam6-1>
jool_siit pool6791 add <pool6791>
Whatever the error is, it might be also present in 4.0.0. __handle_jool_message()
didn't change.
You sure the new userspace client was installed correctly? I can't reproduce it because none of the jool_siit
commands are well-formed (as far as rc4
is concerned), so the requests are shot down long before they reach the kernel.
Check jool_siit --version
, please.
3.6.0-rc4 is not really the last release candidate of anything; 3.6 was renamed into 4.0.
:man_facepalming:
You sure the new userspace client was installed correctly?
No, I'm not sure. I just used my regular install script, tried to fire it up, got that error, and reverted as I didn't have time to debug further at that point. Might be the new client didn't get installed over the old version and I just didn't notice.
I'm closing this issue as it's probably just a dumb user error. Apologies for the noise. I'll let you know If I experience it after I upgrade to 4.0 (including updating all the CLI calls).
Ok, but
It won't do if all it takes to crash the kernel is to issue a command from an outdated client.
I need to look into this more.
Bug confirmed. Fixing.
After I hit #283 when attempting a kernel upgrade, I thought I'd simply try upgrading to the latest v3.6 RC. It built fine, but when starting it up, the kernel oopsed:
I did not investigate exactly at which porint in the Jool initialisation routine the oops occurred as I did not have time to do so during the maintenance window. (I reverted to
v3.5.7
with an older LTS kernel instead.) This is what the init script does, in a nutshell: