Re: [PATCH 1/2] memcg: res counter set limit

[KAMEZAWA Hiroyuki]

On Tue, 17 Jun 2008 09:10:38 +0530
Balbir Singh <balbir@xxxxxxxxxxxxxxxxxx> wrote:

> KAMEZAWA Hiroyuki wrote:
> > Helper function of res_counter for reducing usage in subsys(memcg).
> > 
> > Changelog xxx -> v5.
> >  - new file.
> > 
> > Background:
> >  In v3, I was asked to implement generic ones to res_counter.
> >  In v4. I was asked not to implement generic ones because memcg is only
> >  controller which can reduce usage by the kernel. Okay, maybe make sense.
> >  In this version, adds only necessary helpers to res_counter in 
> >  not-invasive manner.
> > 
> > Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
> > 
> > ---
> >  include/linux/res_counter.h |    7 +++++++
> >  kernel/res_counter.c        |   24 ++++++++++++++++++++++++
> >  2 files changed, 31 insertions(+)
> > 
> > Index: mm-2.6.26-rc5-mm3/include/linux/res_counter.h
> > ===================================================================
> > --- mm-2.6.26-rc5-mm3.orig/include/linux/res_counter.h
> > +++ mm-2.6.26-rc5-mm3/include/linux/res_counter.h
> > @@ -136,6 +136,12 @@ static inline bool res_counter_check_und
> >  	return ret;
> >  }
> > 
> > +/*
> > + * set new limit to the val. if usage > val, returns -EBUSY.
> > + * returns 0 at success.
> > + */
> > +int res_counter_set_limit(struct res_counter *cnt, unsigned long long limit);
> > +
> >  static inline void res_counter_reset_max(struct res_counter *cnt)
> >  {
> >  	unsigned long flags;
> > @@ -153,4 +159,5 @@ static inline void res_counter_reset_fai
> >  	cnt->failcnt = 0;
> >  	spin_unlock_irqrestore(&cnt->lock, flags);
> >  }
> > +
> 
> I don't understand this extra newline here
> 
Sorry. I'll remove this ;)

Thanks,
-Kame


> >  #endif
> > Index: mm-2.6.26-rc5-mm3/kernel/res_counter.c
> > ===================================================================
> > --- mm-2.6.26-rc5-mm3.orig/kernel/res_counter.c
> > +++ mm-2.6.26-rc5-mm3/kernel/res_counter.c
> > @@ -143,3 +143,27 @@ out_free:
> >  out:
> >  	return ret;
> >  }
> > +
> > +
> > +/**
> > + * res_counter_set_limit - set limit of res_counter.
> > + * @cnt: the res_counter
> > + * @limit: the new limit
> > + *
> > + * Note that res_coutner_write() allows the same kind of operation.
> > + * But this returns -EBUSY when new limit < usage. If you want strict control
> > + * of limit, please use this.
> > + */
> > +int res_counter_set_limit(struct res_counter *cnt, unsigned long long limit)
> > +{
> > +	unsigned long flags;
> > +	int ret = -EBUSY;
> > +
> > +	spin_lock_irqsave(&cnt->lock, flags);
> > +	if (cnt->usage <= limit) {
> > +		cnt->limit = limit;
> > +		ret = 0;
> > +	}
> > +	spin_unlock_irqrestore(&cnt->lock, flags);
> > +	return ret;
> > +}
> 
> Reviewed-by: Balbir Singh <balbir@xxxxxxxxxxxxxxxxxx>
> 
> -- 
> 	Warm Regards,
> 	Balbir Singh
> 	Linux Technology Center
> 	IBM, ISTL
> 

--
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/