Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758447AbYCDDIX (ORCPT ); Mon, 3 Mar 2008 22:08:23 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1754551AbYCDDIO (ORCPT ); Mon, 3 Mar 2008 22:08:14 -0500 Received: from py-out-1112.google.com ([64.233.166.177]:41687 "EHLO py-out-1112.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754088AbYCDDIN (ORCPT ); Mon, 3 Mar 2008 22:08:13 -0500 DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:date:from:sender:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references:x-google-sender-auth; b=o8QKT3XmqTCWe2HDINDSGNAVXjBAsHeamvymEGgOfsJDrAj9EvhtHMgx4m6+TW95ZEVXJ9YCidujzzHVgJI8LLOvX0UdCJ6lvHwEsugsz4wR3/u/vh9dyYZ+r7kaJovIrYlfY/CMVjHf+zgz0eJsHy8m4oaONY4wVh2oa1W/1Pg= Message-ID: <2cd57c900803031908p376f86fx54aa74302e8b2b18@mail.gmail.com> Date: Tue, 4 Mar 2008 11:08:11 +0800 From: "Qi Yong" To: "Postfix developers" Subject: Re: var_mailbox_limit 32bit limit Cc: "Qi Yong" , linux-kernel@vger.kernel.org, lesanti@gmail.com In-Reply-To: <20080303180437.B33A51F3E9A@spike.porcupine.org> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline References: <20080303180437.B33A51F3E9A@spike.porcupine.org> X-Google-Sender-Auth: 636307784690d6c3 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2109 Lines: 54 On 04/03/2008, Wietse Venema wrote: > Leandro Santi: > > > On 3/1/08, Qi Yong wrote: > > > On 29/02/2008, Leandro Santi wrote: > > > > On 2/29/08, Qi Yong wrote: > > > > > var_mailbox_limit is int type. This caps mailbox_size_limit to 2147483647 > > > > > at the maximum. I'd like to make var_mailbox_limit a long integer. What do > > > > > you think of this idea? And objections? I have large mboxs here. Thanks. > > > Using long won't help, especially not when int==long like it is > on 32-bit boxes. It should work. long are 64bit on Linux x86_64 or FreeBSD amd64. And most if not all newer servers support that 'long mode'. > > To handle >2GB files, you either need a grown-up OS, or you need (Seems people here think Linux is 'toy operating' and not 'grown-up'. Actually if you have to blame, you'd blame GNU, not the kernel.) I think that GNU/Linux was supporting large file in a different way from *BSD as you think was because they were afraid making off_t 64bit would break old wrong things. And there are also other considerations. Then they introduced off64_t or loff_t. But the above does not cover all the facts. On Linux/x86_64, off_t is 64 bit too. The library has things like _FILE_OFFSET_BITS, __USE_FILE_OFFSET64 and typedef __off64_t off_t, check /usr/include/unistd.h. > to recompile Postfix in the large file model. Is Postfix has a large file model? > > However, that may or may not be sufficient. You may also have to > to recompile libraries that Postfix depends on in the large file > model. I think the pre-built libraries in an amd64 distro are already done so. So my proposal is using long for var_mailbox_limit et al in order to get rid of the 2G limit on amd64. And if that is acceptable, I'd like to try. -- Qi Yong -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/