Re: [PATCH 1/6] res_counter: handle limit change

[Balbir Singh <balbir@linux.vnet.ibm.com>]

kamezawa.hiroyu@jp.fujitsu.com wrote:
> ----- Original Message -----
>> KAMEZAWA Hiroyuki wrote:
>>> Add a support to shrink_usage_at_limit_change feature to res_counter.
>>> memcg will use this to drop pages.
>>>
>>> Change log: xxx -> v4 (new file.)
>>>  - cut out the limit-change part from hierarchy patch set.
>>>  - add "retry_count" arguments to shrink_usage(). This allows that we don't
>>>    have to set the default retry loop count.
>>>  - res_counter_check_under_val() is added to support subsystem.
>>>  - res_counter_init() is res_counter_init_ops(cnt, NULL)
>>>
>>> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
>>>
>> Does shrink_usage() really belong to res_counters? Could a task limiter, a
>> CPU/IO bandwidth controller use this callback? Resource Counters were designe
> d
>> to be generic and work across controllers. Isn't the memory controller a bett
> er
>> place for such ops.
>>
> Definitely No. I think counters which cannot be shrink should return -EBUSY
> by shrink_usage() when it cannot do it. 

Wouldn't that be all counters except for the memory controller RSS counter? I
can't see anyone besides the memory controller supporting shrink_usage().

-- 
	Warm Regards,
	Balbir Singh
	Linux Technology Center
	IBM, ISTL