[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <32834312.1213622802513.kamezawa.hiroyu@jp.fujitsu.com>
Date: Mon, 16 Jun 2008 22:26:42 +0900 (JST)
From: kamezawa.hiroyu@...fujitsu.com
To: balbir@...ux.vnet.ibm.com
Cc: kamezawa.hiroyu@...fujitsu.com, linux-mm@...ck.org,
LKML <linux-kernel@...r.kernel.org>, menage@...gle.com,
xemul@...nvz.org, yamamoto@...inux.co.jp,
nishimura@....nes.nec.co.jp, lizf@...fujitsu.com
Subject: Re: Re: [PATCH 1/6] res_counter: handle limit change
----- Original Message -----
>> 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().
>
Slab_counter is a candidate. But ok, if everyone doesn't like this,
I'll abandon the whole and rewrite it as v3.
And condidering your point, my high-low-watermark patch set should be
implemented within memcg and adding high/low to res_counter is too bad.
I'll change my plan. But res_counter is less useful rather than I thought of ;
)
Besides it doesn't support any feedbacks, it just restricts the access to para
meters.
BTW, I believe current res_counter's behavior to return success
at usage > limit case is very bad. I'd like to return -EBUSY.
How do you think ?
(And I also think res_counter_charge returns -ENOMEM is BUG. It should be
-EBUSY.)
Thanks,
-Kame
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists