From mboxrd@z Thu Jan  1 00:00:00 1970
From: Cyrill Gorcunov <gorcunov@openvz.org>
Subject: Re: [patch 3/7] fs, notify: Add file handle entry into
 inotify_inode_mark
Date: Tue, 13 Nov 2012 12:29:36 +0400
Message-ID: <20121113082936.GG6511@moon>
References: <20121112101440.665694060@openvz.org>
 <20121112101845.839702715@openvz.org>
 <20121112165540.2ec39f50.akpm@linux-foundation.org>
 <20121113072057.GC6511@moon>
 <20121112234001.009b2ae3.akpm@linux-foundation.org>
 <20121113080032.GF6511@moon>
 <alpine.DEB.2.00.1211130016240.18379@chino.kir.corp.google.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: Andrew Morton <akpm@linux-foundation.org>,
	linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org,
	Al Viro <viro@zeniv.linux.org.uk>,
	Alexey Dobriyan <adobriyan@gmail.com>,
	Pavel Emelyanov <xemul@parallels.com>,
	James Bottomley <jbottomley@parallels.com>,
	Matthew Helsley <matt.helsley@gmail.com>,
	aneesh.kumar@linux.vnet.ibm.com, bfields@fieldses.org
To: David Rientjes <rientjes@google.com>
Return-path: <linux-kernel-owner@vger.kernel.org>
Content-Disposition: inline
In-Reply-To: <alpine.DEB.2.00.1211130016240.18379@chino.kir.corp.google.com>
Sender: linux-kernel-owner@vger.kernel.org
List-Id: linux-fsdevel.vger.kernel.org

On Tue, Nov 13, 2012 at 12:19:25AM -0800, David Rientjes wrote:
> On Tue, 13 Nov 2012, Cyrill Gorcunov wrote:
> 
> > > The question is, how many `struct inotify_inode_mark's are instantiated
> > > system-wide?  Could be millions?
> > 
> > Well, hard to tell, to be fair. On my testing machine only apache has been
> > using inotify system as far as I remember, but for sure nothing except memory
> > limit the number of inotify. But I think if one running machine with millions
> > of inotify it's rather powerful machine with enough memory.
> > 
> 
> Seems easy to determine if you boot with slub_nomerge on the command line 
> and then read /sys/kernel/slab/inotify_inode_mark/objects.  On my system, 
> that happens to be 210, but I'm sure you could come up with a realistic 
> synthetic workload to make it much higher.

Which would give about 26K of additional memory if c/r get used here. Not a
big number i guess?