Subject: /proc/sys/user/max_time_namespaces?

Dmitry, Andrei

[Andrei, I mistyped your openvz address in an email of a few minutes
ago ("[PATCH] ns: Fix time_for_children symlink"). You were CCed
on your gmail address though.]

In the namespaces(7) manual page there is documentation for
these files:

$ ls -1 /proc/sys/user/*namespaces
/proc/sys/user/max_cgroup_namespaces
/proc/sys/user/max_ipc_namespaces
/proc/sys/user/max_mnt_namespaces
/proc/sys/user/max_net_namespaces
/proc/sys/user/max_pid_namespaces
/proc/sys/user/max_user_namespaces
/proc/sys/user/max_uts_namespaces

These files allow one to limit on the number of namespaces
of each type.

However, there is no /proc/sys/user/max_time_namespaces file.
Was this deliberate, or an oversight?

Thanks,

Michael

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


2020-04-06 17:28:19

by Dmitry Safonov

[permalink] [raw]
Subject: Re: /proc/sys/user/max_time_namespaces?

Hi Michael,

On 4/3/20 1:21 PM, Michael Kerrisk (man-pages) wrote:
> Dmitry, Andrei
>
> [Andrei, I mistyped your openvz address in an email of a few minutes
> ago ("[PATCH] ns: Fix time_for_children symlink"). You were CCed
> on your gmail address though.]
>
> In the namespaces(7) manual page there is documentation for
> these files:
>
> $ ls -1 /proc/sys/user/*namespaces
> /proc/sys/user/max_cgroup_namespaces
> /proc/sys/user/max_ipc_namespaces
> /proc/sys/user/max_mnt_namespaces
> /proc/sys/user/max_net_namespaces
> /proc/sys/user/max_pid_namespaces
> /proc/sys/user/max_user_namespaces
> /proc/sys/user/max_uts_namespaces
>
> These files allow one to limit on the number of namespaces
> of each type.
>
> However, there is no /proc/sys/user/max_time_namespaces file.
> Was this deliberate, or an oversight?

Thanks a lot for noticing this!
Yes, it's an oversight and examining the code I can see that it's an
actual user-visible regression: max_inotify_instances controls time
namespaces limit.

I've just sent a patch to fix the issue, Cc'ing @stable:
https://lore.kernel.org/linux-api/[email protected]/

Thanks again,
Dmitry