Re: [RFC 02/18] cgroup_pids: track maximum pids

[Topi Miettinen]

On 06/13/16 21:12, Tejun Heo wrote:
> Hello,
> 
> On Mon, Jun 13, 2016 at 10:44:09PM +0300, Topi Miettinen wrote:
>> Track maximum pids in the cgroup, present it in cgroup pids.current_max.
> 
> "max" is often used for maximum limits in cgroup.  I think "watermark"
> or "high_watermark" would be a lot clearer.

OK, I have no preference.

> 
>> @@ -236,6 +246,14 @@ static void pids_free(struct task_struct *task)
>>  	pids_uncharge(pids, 1);
>>  }
>>  
>> +static void pids_fork(struct task_struct *task)
>> +{
>> +	struct pids_cgroup *pids = css_pids(task_css(task, pids_cgrp_id));
>> +
>> +	if (atomic64_read(&pids->cur_max) < atomic64_read(&pids->counter))
>> +		atomic64_set(&pids->cur_max, atomic64_read(&pids->counter));
>> +}
> 
> Wouldn't it make more sense to track high watermark from the charge
> functions instead?  I don't get why this requires a separate fork
> callback.  Also, racing atomic64_set's are racy.  The counter can end
> up with a lower number than it should be.
> 

I used fork callback as I don't want to lower the watermark in all cases
where the charge can be lowered, so I'd update the watermark only when
the fork really happens.

Is there a better way to compare and set? I don't think atomic_cmpxchg()
does what's needed,

>> @@ -300,6 +326,11 @@ static struct cftype pids_files[] = {
>>  		.read_s64 = pids_current_read,
>>  		.flags = CFTYPE_NOT_ON_ROOT,
>>  	},
>> +	{
>> +		.name = "current_max",
> 
> Please make this "high_watermark" field in pids.stats file.
> 
> Thanks.
> 

OK.

-Topi