From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932085Ab1AKNGr (ORCPT ); Tue, 11 Jan 2011 08:06:47 -0500 Received: from mail-wy0-f174.google.com ([74.125.82.174]:44404 "EHLO mail-wy0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755469Ab1AKNGo convert rfc822-to-8bit (ORCPT ); Tue, 11 Jan 2011 08:06:44 -0500 DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :content-type:content-transfer-encoding; b=uDCPH7Lv9EFxi5GGT0qkw75A41O7y9+hIt6DR6PCvEf5snJ71BewyHZ6icwYPd6b9z k2q+hqJWwcOwhMCaryIpwQOtDbbXqC+ZFc3+ZxPw4wwiimDzyTHuVbO3Ij+Qx2iFK62y L8/vot8cXKGLGoa1zh1/rlzXlZ2Z5GqKMnDb0= MIME-Version: 1.0 In-Reply-To: <20110111010800.GE4879@mail.oracle.com> References: <20110111010800.GE4879@mail.oracle.com> Date: Wed, 12 Jan 2011 00:06:42 +1100 Message-ID: Subject: Re: Important for fs devs: rcu-walk merged upstream From: Nick Piggin To: Nick Piggin , linux-fsdevel , "linux-kernel@vger.kernel.org" Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8BIT Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tuesday, January 11, 2011, Joel Becker wrote: > On Sat, Jan 08, 2011 at 01:54:47PM +1100, Nick Piggin wrote: >> The rcu-walk stuff can be more tricky for your filesystem >> to take advantage of. >> >> If you supply a .d_revalidate, .permission, or .check_acl, >> then path walking is going to be slow and unscalable for >> you. > >        Do you mean "as slow and unscalable as it has always been" or > "even slower now"?  A quick look suggests the former, but I wanted to be > sure. Yeah it should be about the same. Bit more complex code so it might be a bit slower. A few code and branch and cache improvements so it might be a bit faster. Refcounts on the same dentry become a bit less scalable for using lock instead of atomic. On the other hand dropping the ref on leaf denty doesn't require global lock. It should not be urgent for 2.6.38. From mboxrd@z Thu Jan 1 00:00:00 1970 From: Nick Piggin Subject: Re: Important for fs devs: rcu-walk merged upstream Date: Wed, 12 Jan 2011 00:06:42 +1100 Message-ID: References: <20110111010800.GE4879@mail.oracle.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE To: Nick Piggin , linux-fsdevel , "linux-kernel@vger.kernel.org" Return-path: Received: from mail-wy0-f174.google.com ([74.125.82.174]:44404 "EHLO mail-wy0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755469Ab1AKNGo convert rfc822-to-8bit (ORCPT ); Tue, 11 Jan 2011 08:06:44 -0500 In-Reply-To: <20110111010800.GE4879@mail.oracle.com> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Tuesday, January 11, 2011, Joel Becker wrote: > On Sat, Jan 08, 2011 at 01:54:47PM +1100, Nick Piggin wrote: >> The rcu-walk stuff can be more tricky for your filesystem >> to take advantage of. >> >> If you supply a .d_revalidate, .permission, or .check_acl, >> then path walking is going to be slow and unscalable for >> you. > > =A0 =A0 =A0 =A0Do you mean "as slow and unscalable as it has always = been" or > "even slower now"? =A0A quick look suggests the former, but I wanted = to be > sure. Yeah it should be about the same. Bit more complex code so it might be a bit slower. A few code and branch and cache improvements so it might be a bit faster. Refcounts on the same dentry become a bit less scalable for using lock instead of atomic. On the other hand dropping the ref on leaf denty doesn't require global lock. It should not be urgent for 2.6.38. -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel= " in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html