PIDs Controller Limit

PIDs Controller Limit

[Aleksa Sarai]

Does it make sense for the PIDs controller to allow a user to set a
limit of 0? Since we don't cancel attaches, a limit of 0 doesn't
affect anything (nothing stops attaches, and you need to have a
process in the PIDs cgroup in order for fork()s to be affected by the
limit). So I think that attempting to set pid.limit to 0 should return
an -EINVAL.

-- 
Aleksa Sarai (cyphar)
www.cyphar.com

Re: PIDs Controller Limit

[Tejun Heo]

Hello, Aleksa.

On Thu, Sep 24, 2015 at 09:42:38AM +1000, Aleksa Sarai wrote:
> Does it make sense for the PIDs controller to allow a user to set a
> limit of 0? Since we don't cancel attaches, a limit of 0 doesn't
> affect anything (nothing stops attaches, and you need to have a
> process in the PIDs cgroup in order for fork()s to be affected by the
> limit). So I think that attempting to set pid.limit to 0 should return
> an -EINVAL.

I don't know.  Why does it matter?

Thanks.

-- 
tejun

Re: PIDs Controller Limit

[Aleksa Sarai]

> On Thu, Sep 24, 2015 at 09:42:38AM +1000, Aleksa Sarai wrote:
>> Does it make sense for the PIDs controller to allow a user to set a
>> limit of 0? Since we don't cancel attaches, a limit of 0 doesn't
>> affect anything (nothing stops attaches, and you need to have a
>> process in the PIDs cgroup in order for fork()s to be affected by the
>> limit). So I think that attempting to set pid.limit to 0 should return
>> an -EINVAL.
>
> I don't know.  Why does it matter?

Well, it might be confusing that a limit of `0` is not different from
a limit of `1`. Especially since someone might think that a limit of
`0` means "no processes AT ALL", which is wrong. Although, I guess
they should've just RTFM'd in that case.

-- 
Aleksa Sarai (cyphar)
www.cyphar.com

Re: PIDs Controller Limit

[Tejun Heo]

Hello,

On Sat, Sep 26, 2015 at 09:11:02AM +1000, Aleksa Sarai wrote:
> Well, it might be confusing that a limit of `0` is not different from
> a limit of `1`. Especially since someone might think that a limit of
> `0` means "no processes AT ALL", which is wrong. Although, I guess
> they should've just RTFM'd in that case.

I really don't think it matters.

Thanks.

-- 
tejun

Re: PIDs Controller Limit

[Nikolay Borisov]

On 09/26/2015 02:11 AM, Aleksa Sarai wrote:
>> On Thu, Sep 24, 2015 at 09:42:38AM +1000, Aleksa Sarai wrote:
>>> Does it make sense for the PIDs controller to allow a user to set a
>>> limit of 0? Since we don't cancel attaches, a limit of 0 doesn't
>>> affect anything (nothing stops attaches, and you need to have a
>>> process in the PIDs cgroup in order for fork()s to be affected by the
>>> limit). So I think that attempting to set pid.limit to 0 should return
>>> an -EINVAL.
>>
>> I don't know.  Why does it matter?
> 
> Well, it might be confusing that a limit of `0` is not different from
> a limit of `1`. Especially since someone might think that a limit of
> `0` means "no processes AT ALL", which is wrong. Although, I guess
> they should've just RTFM'd in that case.

I personally would have parsed a value of 0 as "unlimited"