Re: cgroups(7): documenting cgroups v2 delegation

From: Michael Kerrisk (man-pages)
Date: Tue Jan 09 2018 - 17:48:23 EST


Hello Tejun,

On 01/09/2018 10:07 PM, Tejun Heo wrote:
> Hello,
>
> On Tue, Jan 09, 2018 at 12:27:58AM +0100, Michael Kerrisk (man-pages) wrote:
>>>> /dlgt_grp/cgroup.subtree_control
>>>> Making this file owned by the delegatee is optional.
>>>> Doing so means that that the delegatee can enable conâ
>>>> trollers (that are present in /dlgt_grp/cgroup.conâ
>>>> trollers) in order to further redistribute resources at
>>>> lower levels in the subtree. As an alternative to
>>>> changing the ownership of this file, the delegater might
>>>> instead add selected controllers to this file.
>>>
>>> I'm not sure how useful it is to describe this to be optional. In the
>>> same sense, cgroup.procs would be optional too as the delegatee can
>>> take control from its first children. Users of course can choose to
>>> do mix and match as they see fit but outside of the defined model,
>>> there can be surprises - e.g. nsdelegate or some future delegation
>>> aware feature can behave differently. I think it'd be better to keep
>>> it simple - either a subtree is delegated or not.
>>
>> So, I changed that piece to:
>>
>> /dlgt_grp/cgroup.subtree_control
>> Changing the ownership of this file means that that the
>> delegatee can enable controllers (that are present in
>> /dlgt_grp/cgroup.controllers) in order to further redisâ
>> tribute resources at lower levels in the subtree. (As an
>> alternative to changing the ownership of this file, the
>> delegater might instead add selected controllers to this
>> file.)
>>
>> Okay?
>
> I haven't thought much about not giving out cgroup.subtree_control
> when delegating. It's probably okay but there can be surprises -
> e.g. systemd or other agent failing to init resource control because
> it failed to write to subtree_control at root.
>
> Also, given that the recommended way to delegate is now chown all
> files in the /sys/kernel/cgroup/delegate file it's a bit weird to
> single out subtree_control.

I guess the point here is that it really makes no sense not to
change ownership of cgroup.procs. On the other hand, if
ownership of cgroup.subtree_control is not handed to the
delegatee, then the delegater can choose which controllers the
delegatee will be allowed to exercise, by writing only those
controllers into cgroup.subtree_control. I'm not sure if
there's a use case though.

> That said, not necessarily an objection. I'm just not sure about it.

Okay. For the moment, I'll leave that text as is.

I realized that there was also one more file that needs to be included
in the list of files that must be delegated:

/dlgt_grp/cgroup.threads
Changing the ownership of this file is necessary if a
threaded subtree is being delegated (see the description of
"thread mode", below).

Can you please confirm that it's only necessary to delegate this file
if we are delegating a threaded subtree?

>>> Roman recently added /sys/kernel/cgroup/delegate and
>>> /sys/kernel/cgroup/features. The former contains newline separated
>>> list of cgroup files which should be chowned on delegation (in
>>> addition to the directory itself) and the latter contains optional
>>> features (currently only nsdelegate).
>>
>> Oh -- and this reminds that I've been meaning to ask you for a while
>> now: could you please (please please please) CC all cgroup interface
>> changes to linux-api@xxxxxxxxxxxxxxx (and prod others to do so
>> please). There have been many of these changes in recent times
>> (addition of new v2 controllers, thread mode, nsdelegate, the
>> changes that Roman made that you refer to above), and they really
>> all should have been CCed to linux-api@. It's often the only (easy)
>> way that I have to discover changes that should be documented in
>> the manual pages. And there are many other groups that are also
>> interested in tracking kernel-user-space interface changes; see
>> https://www.kernel.org/doc/man-pages/linux-api-ml.html
>
> Sorry about having not added them before. Will try to. Please feel
> free to scream at me if I forget.

Okay. (Except I may not not e when you forget ;-).)

>>> Also, if nsdelegate is enabled, both the source and destination
>>> cgroups must be visible (cgroup namespace-wise) to the writer.
>>
>> I added the following:
>>
>> * If the cgroup v2 filesystem was mounted with the nsdelegate
>> option, the writer must be able to see the source and destinaâ
>> tion cgroup from its cgroup namespace.
>>
>> Okay?
>
> Yeah, thanks.

Thanks for checking the text, Tejun!

Cheers,

Michael

--
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/