From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751933AbZGNAgr (ORCPT ); Mon, 13 Jul 2009 20:36:47 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751372AbZGNAgq (ORCPT ); Mon, 13 Jul 2009 20:36:46 -0400 Received: from smtp-out.google.com ([216.239.33.17]:31345 "EHLO smtp-out.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751365AbZGNAgp convert rfc822-to-8bit (ORCPT ); Mon, 13 Jul 2009 20:36:45 -0400 DomainKey-Signature: a=rsa-sha1; s=beta; d=google.com; c=nofws; q=dns; h=mime-version:in-reply-to:references:date:message-id:subject:from:to: cc:content-type:content-transfer-encoding:x-system-of-record; b=w4yO+NHAf1/DEYcye+FHIOeG3Neesk4pFt0FqUGrqWiE7tfAz6wAhcH4boSyBwhLM AfCIV6YItcHV73nIg5+Zg== MIME-Version: 1.0 In-Reply-To: <1247530581-31416-2-git-send-email-vbuzov@embeddedalley.com> References: <1246998310-16764-1-git-send-email-vbuzov@embeddedalley.com> <1247530581-31416-1-git-send-email-vbuzov@embeddedalley.com> <1247530581-31416-2-git-send-email-vbuzov@embeddedalley.com> Date: Mon, 13 Jul 2009 17:36:40 -0700 Message-ID: <6599ad830907131736w4397d336xad733f274c812690@mail.gmail.com> Subject: Re: [PATCH 1/2] Resource usage threshold notification addition to res_counter (v3) From: Paul Menage To: Vladislav Buzov Cc: Linux Kernel Mailing List , Linux Containers Mailing List , Linux memory management list , Dan Malek , Andrew Morton , KAMEZAWA Hiroyuki , Balbir Singh Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8BIT X-System-Of-Record: true Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org As I mentioned in another thread, I think that associating the threshold with the res_counter rather than with each individual waiter is a mistake, since it creates global state and makes it hard to have multiple waiters on the same cgroup. Paul On Mon, Jul 13, 2009 at 5:16 PM, Vladislav Buzov wrote: > This patch updates the Resource Counter to add a configurable resource usage > threshold notification mechanism. > > Signed-off-by: Vladislav Buzov > Signed-off-by: Dan Malek > --- >  Documentation/cgroups/resource_counter.txt |   21 ++++++++- >  include/linux/res_counter.h                |   69 ++++++++++++++++++++++++++++ >  kernel/res_counter.c                       |    7 +++ >  3 files changed, 95 insertions(+), 2 deletions(-) > > diff --git a/Documentation/cgroups/resource_counter.txt b/Documentation/cgroups/resource_counter.txt > index 95b24d7..1369dff 100644 > --- a/Documentation/cgroups/resource_counter.txt > +++ b/Documentation/cgroups/resource_counter.txt > @@ -39,7 +39,20 @@ to work with it. >        The failcnt stands for "failures counter". This is the number of >        resource allocation attempts that failed. > > - c. spinlock_t lock > + e. unsigned long long threshold > + > +       The resource usage threshold to notify the resouce controller. This is > +       the minimal difference between the resource limit and current usage > +       to fire a notification. > + > + f. void (*threshold_notifier)(struct res_counter *counter) > + > +       The threshold notification callback installed by the resource > +       controller. Called when the usage reaches or exceeds the threshold. > +       Should be fast and not sleep because called when interrupts are > +       disabled. > + > + g. spinlock_t lock > >        Protects changes of the above values. > > @@ -140,6 +153,7 @@ counter fields. They are recommended to adhere to the following rules: >        usage           usage_in_ >        max_usage       max_usage_in_ >        limit           limit_in_ > +       threshold       notify_threshold_in_ >        failcnt         failcnt >        lock            no file :) > > @@ -153,9 +167,12 @@ counter fields. They are recommended to adhere to the following rules: >        usage           prohibited >        max_usage       reset to usage >        limit           set the limit > +       threshold       set the threshold >        failcnt         reset to zero > > - > + d. Notification is enabled by installing the threshold notifier callback. It > +    is up to the resouce controller to communicate the notification to user > +    space tasks. > >  5. Usage example > > diff --git a/include/linux/res_counter.h b/include/linux/res_counter.h > index 511f42f..5ec98d7 100644 > --- a/include/linux/res_counter.h > +++ b/include/linux/res_counter.h > @@ -9,6 +9,11 @@ >  * >  * Author: Pavel Emelianov >  * > + * Resouce usage threshold notification update > + * Copyright 2009 CE Linux Forum and Embedded Alley Solutions, Inc. > + * Author: Dan Malek > + * Author: Vladislav Buzov > + * >  * See Documentation/cgroups/resource_counter.txt for more >  * info about what this counter is. >  */ > @@ -35,6 +40,19 @@ struct res_counter { >         */ >        unsigned long long limit; >        /* > +        * the resource usage threshold to notify the resouce controller. This > +        * is the minimal difference between the resource limit and current > +        * usage to fire a notification. > +        */ > +       unsigned long long threshold; > +       /* > +        * the threshold notification callback installed by the resource > +        * controller. Called when the usage reaches or exceeds the threshold. > +        * Should be fast and not sleep because called when interrupts are > +        * disabled. > +        */ > +       void (*threshold_notifier)(struct res_counter *counter); > +       /* >         * the number of unsuccessful attempts to consume the resource >         */ >        unsigned long long failcnt; > @@ -87,6 +105,7 @@ enum { >        RES_MAX_USAGE, >        RES_LIMIT, >        RES_FAILCNT, > +       RES_THRESHOLD, >  }; > >  /* > @@ -132,6 +151,21 @@ static inline bool res_counter_limit_check_locked(struct res_counter *cnt) >        return false; >  } > > +static inline bool res_counter_threshold_check_locked(struct res_counter *cnt) > +{ > +       if (cnt->usage + cnt->threshold < cnt->limit) > +               return true; > + > +       return false; > +} > + > +static inline void res_counter_threshold_notify_locked(struct res_counter *cnt) > +{ > +       if (!res_counter_threshold_check_locked(cnt) && > +           cnt->threshold_notifier) > +               cnt->threshold_notifier(cnt); > +} > + >  /* >  * Helper function to detect if the cgroup is within it's limit or >  * not. It's currently called from cgroup_rss_prepare() > @@ -147,6 +181,21 @@ static inline bool res_counter_check_under_limit(struct res_counter *cnt) >        return ret; >  } > > +/* > + * Helper function to detect if the cgroup usage is under it's threshold or > + * not. > + */ > +static inline bool res_counter_check_under_threshold(struct res_counter *cnt) > +{ > +       bool ret; > +       unsigned long flags; > + > +       spin_lock_irqsave(&cnt->lock, flags); > +       ret = res_counter_threshold_check_locked(cnt); > +       spin_unlock_irqrestore(&cnt->lock, flags); > +       return ret; > +} > + >  static inline void res_counter_reset_max(struct res_counter *cnt) >  { >        unsigned long flags; > @@ -174,6 +223,26 @@ static inline int res_counter_set_limit(struct res_counter *cnt, >        spin_lock_irqsave(&cnt->lock, flags); >        if (cnt->usage <= limit) { >                cnt->limit = limit; > +               if (limit <= cnt->threshold) > +                       cnt->threshold = 0; > +               else > +                       res_counter_threshold_notify_locked(cnt); > +               ret = 0; > +       } > +       spin_unlock_irqrestore(&cnt->lock, flags); > +       return ret; > +} > + > +static inline int res_counter_set_threshold(struct res_counter *cnt, > +               unsigned long long threshold) > +{ > +       unsigned long flags; > +       int ret = -EINVAL; > + > +       spin_lock_irqsave(&cnt->lock, flags); > +       if (cnt->limit > threshold) { > +               cnt->threshold = threshold; > +               res_counter_threshold_notify_locked(cnt); >                ret = 0; >        } >        spin_unlock_irqrestore(&cnt->lock, flags); > diff --git a/kernel/res_counter.c b/kernel/res_counter.c > index e1338f0..9b36748 100644 > --- a/kernel/res_counter.c > +++ b/kernel/res_counter.c > @@ -5,6 +5,10 @@ >  * >  * Author: Pavel Emelianov >  * > + * Resouce usage threshold notification update > + * Copyright 2009 CE Linux Forum and Embedded Alley Solutions, Inc. > + * Author: Dan Malek > + * Author: Vladislav Buzov >  */ > >  #include > @@ -32,6 +36,7 @@ int res_counter_charge_locked(struct res_counter *counter, unsigned long val) >        counter->usage += val; >        if (counter->usage > counter->max_usage) >                counter->max_usage = counter->usage; > +       res_counter_threshold_notify_locked(counter); >        return 0; >  } > > @@ -101,6 +106,8 @@ res_counter_member(struct res_counter *counter, int member) >                return &counter->limit; >        case RES_FAILCNT: >                return &counter->failcnt; > +       case RES_THRESHOLD: > +               return &counter->threshold; >        }; > >        BUG(); > -- > 1.5.6.3 > >