Re: [PATCH] [RFC] pidns: don't zap processes several times

From: Andrey Wagin
Date: Tue Oct 09 2012 - 13:41:02 EST


2012/10/9 Oleg Nesterov <oleg@xxxxxxxxxx>:
> On 10/08, Andrey Wagin wrote:
>>
>> 2012/10/7 Oleg Nesterov <oleg@xxxxxxxxxx>:
>> >
>> > Perhaps we should MAX_PID_NS_LEVEL instead?
>>
>> Yes, we can.
>>
>> Could I just define MAX_PID_NS_LEVEL in a code:
>> #define MAX_PID_NS_LEVEL ((PAGE_SIZE - offsetof(struct pid, numbers))
>> / sizeof(struct upid))
>
> Or even less. But looks reasonable.
>
>> Or should it be added in a config?
>
> Personally I think that "define" is fine, we can add config/sysctl
> later if needed.

Ok, I'm going to send a patch.

>
>
> Hmm. This is off-topic, but...
>
> create_pid_namespace:
>
> unsigned int level = parent_pid_ns->level + 1;
> ns->pid_cachep = create_pid_cachep(level + 1);

Yes, it's correct, because pid->numbers[ns->level] should be valid, so
a size of an array pid->numbers should be (level + 1).

/*
....
* @nr_ids: the number of numerical ids this pid will have to carry
*/
static struct kmem_cache *create_pid_cachep(int nr_ids)

>
> is it correct? is seems that only one "+ 1" is needed?
>
> Oleg.
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/