Re: [RFC][PATCH 4/5] utsname namespaces: sysctl hack

[ebiederm@xmission.com (Eric W. Biederman)]

Cedric Le Goater <clg@fr.ibm.com> writes:

> Hello !
>
> Kirill Korotaev wrote:
>> Serge,
>> 
>>> Please look closer at the patch.
>>> I *am* doing nothing with sysctls.
>>>
>>> system_utsname no longer exists, and the way to get to that is by using
>>> init_uts_ns.name.  That's all this does.
>> Sorry for being not concrete enough.
>> I mean switch () in the code. Until we decided how to virtualize
>> sysctls/proc, I believe no dead code/hacks should be commited. IMHO.
>
> How could we improve that hack ? Removing the modification of the static
> table can easily be worked around but getting rid of the switch() statement
> is more difficult. Any idea ?

Store offsetof in data.  Not that for such a small case it really matters,
but it probably improves maintenance by a little bit.

>> FYI, I strongly object against virtualizing sysctls this way as it is
>> not flexible and is a real hack from my POV.
>
> what is the issue with flexibility ?

The only other thing I would like to see is the process argument passed
in. 

Eric