2015-11-12 19:41 GMT+03:00 Stanislav Kinsburskiy <skinsbur...@odin.com>: > > > 12.11.2015 17:11, Andrey Ryabinin пишет: > >> >> On 11/12/2015 07:08 PM, Stanislav Kinsburskiy wrote: >>> >>> >>> 12.11.2015 16:47, Andrey Ryabinin пишет: >>>> >>>> On 11/12/2015 05:59 PM, Stanislav Kinsburskiy wrote: >>>>> >>>>> 12.11.2015 15:53, Andrey Wagin пишет: >>>>>> >>>>>> 2015-11-12 17:46 GMT+03:00 Stanislav Kinsburskiу >>>>>> <skinsbur...@odin.com>: >>>>>>> >>>>>>> 12 нояб. 2015 г. 15:14 пользователь Andrey Ryabinin >>>>>>> <aryabi...@virtuozzo.com> написал: >>>>>>>> >>>>>>>> CRIU sends SIGKILL to container's init process as a part of >>>>>>>> cleanup process if restoring failed. >>>>>>>> CRIU does this from a different ve, which is currently not allowed >>>>>>>> without any apparent reason. >>>>>>> >>>>>>> The reason looks very clear to me: improve namespaces isolation. >>>>>>> It espesially applies to killing child reaper of another ve. >>>>>>> You throwed away this check, and now it's possible to kill one >>>>>>> container from another one. >>>>>>> Or I'm missing somethig? >>>>>> >>>>>> Each container has its own pidns, so you can't kill anyone who isn't >>>>>> in this pidns. >>>>> >>>>> So how CRIU sends kill signal from one ve to another then? >>>>> >>>> AFAIK, CRIU creates it's own ve namespace, but it still operates in root >>>> pid namespace. >>> >>> Hmm, ok. >>> Then nothing against this patch. >>> The only thing I'm curios: for how long we have this patch? Pid >>> namespaces are used in OpenVZ for at least last 6 years (probably more). >>> When this checks appeared? Maybe there was another reason, which is just >>> not obvious so far? >> >> I suspect that it was just blindly ported from 2.6: >> >> commit fd3207d650434ac82f2c897cadd5607e67f2c274 >> Author: Kirill Tkhai <ktk...@parallels.com> >> Date: Fri Oct 10 19:35:02 2014 +0400 >> >> ve: Ignore signals from wrong ve >> Port sig_ve_ignored(). >> This is a part of 74-diff-ve-mix-combined. >> https://jira.sw.ru/browse/PSBM-17903 >> Signed-off-by: Kirill Tkhai <ktk...@parallels.com> > > > That's for sure. > My question was about origins of this patch. > Thanks to Vasiliy: > > RCS file: > /cvs/Virtuozzo/kernel-patches/2.6.18-rhel5/diff-ve-init-signals-20070514,v > Working file: diff-ve-init-signals-20070514 > head: 1.1 > branch: > locks: strict > access list: > symbolic names: > keyword substitution: o > total revisions: 1; selected revisions: 1 > description: > ---------------------------- > revision 1.1 > date: 2007/05/18 13:24:17; author: dev; state: Exp; > Patch from Denis Lunev <d...@openvz.org> > [VE] VE init signal delivery reworked to be similar to host > Prevent VE init from receiving unexpected signals sent from VE > including fatal ones. Signals sent from VE0 are still allowed, > e.g. for fast VE stop. > Fix for sys_reboot called from VE to force VE death > (SIGKILL is sent in the context of VE). > > http://bugzilla.openvz.org/show_bug.cgi?id=533 > > Are you sure, that you are not braking the logic, this patch introduced in > past?
[root@fc22-vm ~]# unshare --fork -p [root@fc22-vm ~]# kill -9 1 [root@fc22-vm ~]# kill -9 1 [root@fc22-vm ~]# kill -9 1 [root@fc22-vm ~]# kill -9 1 [root@fc22-vm ~]# kill -USR1 1 [root@fc22-vm ~]# kill -USR1 1 [root@fc22-vm ~]# > _______________________________________________ Devel mailing list Devel@openvz.org https://lists.openvz.org/mailman/listinfo/devel
