From: Nikanth Karthikesan <knikanth@suse.de>
Organization: suse.de
To: Andrew Morton <akpm@linux-foundation.org>
Subject: Re: [PATCH] Make /proc/sys/vm/drop_caches write only
Date: Mon, 17 Nov 2008 11:57:57 +0530
User-Agent: KMail/1.9.51 (KDE/4.0.4; ; )
Cc: linux-kernel@vger.kernel.org
References: <200811131754.11863.knikanth@suse.de> <20081114150705.c98e5004.akpm@linux-foundation.org>
In-Reply-To: <20081114150705.c98e5004.akpm@linux-foundation.org>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200811171157.57438.knikanth@suse.de>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 2672
Lines: 84

On Saturday 15 November 2008 04:37:05 Andrew Morton wrote:
> On Thu, 13 Nov 2008 17:54:11 +0530
>
> Nikanth Karthikesan <knikanth@suse.de> wrote:
> >  /proc/sys/vm/drop_caches is used only to trigger dropping caches. It
> > does not disable the cache. The ability to read the value written to this
> > file long back gives wrong impression.
> >
> > Signed-off-by: Nikanth Karthikesan <knikanth@suse.de>
> >
> > ---
> >
> > diff --git a/fs/drop_caches.c b/fs/drop_caches.c
> > index 3e5637f..9892e1b 100644
> > --- a/fs/drop_caches.c
> > +++ b/fs/drop_caches.c
> > @@ -66,11 +66,9 @@ int drop_caches_sysctl_handler(ctl_table *table, int
> > write, struct file *file, void __user *buffer, size_t *length, loff_t
> > *ppos) {
> >  	proc_dointvec_minmax(table, write, file, buffer, length, ppos);
> > -	if (write) {
> > -		if (sysctl_drop_caches & 1)
> > -			drop_pagecache();
> > -		if (sysctl_drop_caches & 2)
> > -			drop_slab();
> > -	}
> > +	if (sysctl_drop_caches & 1)
> > +		drop_pagecache();
> > +	if (sysctl_drop_caches & 2)
> > +		drop_slab();
> >  	return 0;
> >  }
> > diff --git a/kernel/sysctl.c b/kernel/sysctl.c
> > index 9d048fa..dff3c13 100644
> > --- a/kernel/sysctl.c
> > +++ b/kernel/sysctl.c
> > @@ -1015,7 +1015,7 @@ static struct ctl_table vm_table[] = {
> >  		.procname	= "drop_caches",
> >  		.data		= &sysctl_drop_caches,
> >  		.maxlen		= sizeof(int),
> > -		.mode		= 0644,
> > +		.mode		= 0200,
> >  		.proc_handler	= drop_caches_sysctl_handler,
> >  		.strategy	= &sysctl_intvec,
> >  	},
>
> True.  But it's a backward-incompatible change, and there's a (small)
> chance that someone's tools will break if we do this.
>
> We have to get these things right first time :(

hmm.. add this to the list of things for the big version change ;)

May be zeroing out the value would have slimmer chance of breaking things?! Or 
would it cause a different confusion?

Thanks
Nikanth Karthikesan

/proc/sys/vm/drop_caches is used only to trigger dropping caches. It
does not disable the cache. Unset it after clearing the cache.

Signed-off-by: Nikanth Karthikesan <knikanth@suse.de>

---

diff --git a/fs/drop_caches.c b/fs/drop_caches.c
index 3e5637f..1a03d6e 100644
--- a/fs/drop_caches.c
+++ b/fs/drop_caches.c
@@ -71,6 +71,7 @@ int drop_caches_sysctl_handler(ctl_table *table, int write,
 			drop_pagecache();
 		if (sysctl_drop_caches & 2)
 			drop_slab();
+		sysctl_drop_caches = 0;
 	}
 	return 0;
 }

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/