From mboxrd@z Thu Jan  1 00:00:00 1970
From: "J. Bruce Fields" <bfields-uC3wQj2KruNg9hUCZPvPmw@public.gmane.org>
Subject: Re: [PATCH 0/3 v3] dcache: make it more scalable on large system
Date: Wed, 29 May 2013 16:43:23 -0400
Message-ID: <20130529204323.GC3243@fieldses.org>
References: <1369273048-60256-1-git-send-email-Waiman.Long@hp.com>
 <20130523094201.GA24543@dastard>
 <519E8B5F.3080905@hp.com>
 <20130527020903.GR29466@dastard>
 <51A624E2.3000301@hp.com>
 <20130529184640.GA3243@fieldses.org>
 <20130529203700.GM6123@two.firstfloor.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: Waiman Long <waiman.long-VXdhtT5mjnY@public.gmane.org>,
	Dave Chinner <david-FqsqvQoI3Ljby3iVrkZq2A@public.gmane.org>,
	Alexander Viro <viro-RmSDqhL/yNMiFSDQTTA3OLVCufUGDwFn@public.gmane.org>,
	Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>,
	Miklos Szeredi <mszeredi-AlSwsSmVLrQ@public.gmane.org>, Ian Kent <raven-PKsaG3nR2I+sTnJN9+BGXg@public.gmane.org>,
	Sage Weil <sage-4GqslpFJ+cxBDgjK7y7TUQ@public.gmane.org>, Steve French <sfrench-eUNUBHrolfbYtjvyW6yDsg@public.gmane.org>,
	Trond Myklebust <Trond.Myklebust-HgOvQuBEEgTQT0dZR+AlfA@public.gmane.org>,
	Eric Paris <eparis-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>, linux-fsdevel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, autofs-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	ceph-devel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-cifs-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	linux-nfs-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	"Chandramouleeswaran, Aswin" <aswin-VXdhtT5mjnY@public.gmane.org>,
	"Norton, Scott J" <scott.norton-VXdhtT5mjnY@public.gmane.org>
To: Andi Kleen <andi-Vw/NltI1exuRpAAqCnN02g@public.gmane.org>
Return-path: <linux-cifs-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org>
Content-Disposition: inline
In-Reply-To: <20130529203700.GM6123-1g7Xle2YJi4/4alezvVtWx2eb7JE58TQ@public.gmane.org>
Sender: linux-cifs-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
List-Id: linux-fsdevel.vger.kernel.org

On Wed, May 29, 2013 at 10:37:00PM +0200, Andi Kleen wrote:
> > As Dave said before, is the last path component sufficient?  Or how
> > about an inode number?
> 
> Neither works, the profiler needs to find the file and read it.
> 
> inode searching would be incredible expensive,

How fast does it need to be?  Actually analyzing the profile is
something you only do once after everything's over, right?

> unless the file system provided a "open-by-inode" primitive

Well, I suppose there is open_by_handle_at.

--b.

> In normal operation you only have that event when starting up the
> program and when each shared library gets mapped in.
> 
> -Andi