From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mingming Cao Subject: Re: [Ext2-devel] Re: [RFC][PATCH 0/2]Extend ext3 filesystem limit from 8TB to 16TB Date: Mon, 10 Apr 2006 09:57:59 -0700 Message-ID: <1144688279.3964.7.camel@dyn9047017067.beaverton.ibm.com> References: <20060325223358sho@rifu.tnes.nec.co.jp> <1143485147.3970.23.camel@dyn9047017067.beaverton.ibm.com> <20060327131049.2c6a5413.akpm@osdl.org> <20060327225847.GC3756@localhost.localdomain> <1143530126.11560.6.camel@openx2.frec.bull.fr> <1143568905.3935.13.camel@dyn9047017067.beaverton.ibm.com> <1143623605.5046.11.camel@openx2.frec.bull.fr> <1143682730.4045.145.camel@dyn9047017067.beaverton.ibm.com> <20060329175446.67149f32.akpm@osdl.org> <1144660270.5816.3.camel@openx2.frec.bull.fr> Reply-To: cmm@us.ibm.com Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Andrew Morton , Takashi Sato , linux-kernel@vger.kernel.org, ext2-devel , linux-fsdevel@vger.kernel.org Return-path: Received: from e36.co.us.ibm.com ([32.97.110.154]:29065 "EHLO e36.co.us.ibm.com") by vger.kernel.org with ESMTP id S1750873AbWDJQ6G (ORCPT ); Mon, 10 Apr 2006 12:58:06 -0400 To: Laurent Vivier In-Reply-To: <1144660270.5816.3.camel@openx2.frec.bull.fr> Sender: linux-fsdevel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org On Mon, 2006-04-10 at 11:11 +0200, Laurent Vivier wrote: > Le jeu 30/03/2006 =C3=A0 03:54, Andrew Morton a =C3=A9crit : > > Mingming Cao wrote: > > > > > > The things need to be done to complete this work is the issue wit= h > > > current percpu counter, which could not handle u32 type count we= ll.=20 > >=20 > > I'm surprised there's much of a problem here. It is a 32-bit value= , so it > > should mainly be a matter of treating the return value from > > percpu_counter_read() as unsigned long. > >=20 > > However a stickier problem is when dealing with a filesystem which = has, > > say, 0xffff_ff00 blocks. Because percpu counters are approximate, = and a > > counter which really has a value of 0xffff_feee might return 0x0000= 0123.=20 > > What do we do then? > >=20 > > Of course the simple option is to nuke the percpu counters in ext3 = and use > > atomic_long_t (which is signed, so appropriate treat-it-as-unsigned= code > > would be needed). I doubt if the percpu counters in ext3 are gaini= ng us > > much. >=20 > I tried to make something in this way. > Does the attached patch look like the thing you though about ? >=20 I tried the other way -- I am trying to keep the percpu counter in use in ext2/3 as much as possible. I proposed a fix for percpu counter to deal with the possible "overflow" (i.e, a counter really has a value of 0xfff_feee and after updating one local counter it truens 0x00000123). Will send the proposed patch out for review and comments soon. Mingming - 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