From owner-freebsd-current@FreeBSD.ORG Wed Apr 4 17:18:00 2007 Return-Path: X-Original-To: current@freebsd.org Delivered-To: freebsd-current@FreeBSD.ORG Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id EBD8D16A408; Wed, 4 Apr 2007 17:18:00 +0000 (UTC) (envelope-from bright@elvis.mu.org) Received: from elvis.mu.org (elvis.mu.org [192.203.228.196]) by mx1.freebsd.org (Postfix) with ESMTP id D7A7413C46A; Wed, 4 Apr 2007 17:18:00 +0000 (UTC) (envelope-from bright@elvis.mu.org) Received: by elvis.mu.org (Postfix, from userid 1192) id 4FA901A4DA7; Wed, 4 Apr 2007 10:18:00 -0700 (PDT) Date: Wed, 4 Apr 2007 10:18:00 -0700 From: Alfred Perlstein To: Howard Su Message-ID: <20070404171800.GW61362@elvis.mu.org> References: <20070404101222.GU61362@elvis.mu.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.4.2.2i Cc: current@freebsd.org Subject: Re: [Review] Remove procfs dependency of truss X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 04 Apr 2007 17:18:01 -0000 * Howard Su [070404 04:35] wrote: > On 4/4/07, Alfred Perlstein wrote: > >* Howard Su [070404 01:20] wrote: > >> Following the suggestion in idea page, I proposed the attached patch. > >> I didn't change any kernel part because I think PTRACE(2) is > >> functional although man page didn't document it. > >> > >> I tested the patch under i386 and amd64 box. The help on testing and > >> code review will be appreciated. > > > >wow, well done! any draw backs to using ptrace over procfs? > I didn't see. > > > >have you tested performance? > Not yet. Base on the number of kernel syscall, new implementaion keep > in a same level. However ptrace calls has a short code path compare to > generic read syscall. I suppose there will be some improvement. > Anyway, I will try to get perf data. Thank you very much for the work, perhaps if the performance is slower we can make it a runtime option? Regardless, very well done, it's nice not to have this depend on procfs any longer! -- - Alfred Perlstein