From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261536AbVGRTTf (ORCPT ); Mon, 18 Jul 2005 15:19:35 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261556AbVGRTTd (ORCPT ); Mon, 18 Jul 2005 15:19:33 -0400 Received: from wproxy.gmail.com ([64.233.184.196]:27840 "EHLO wproxy.gmail.com") by vger.kernel.org with ESMTP id S261545AbVGRTTc convert rfc822-to-8bit (ORCPT ); Mon, 18 Jul 2005 15:19:32 -0400 DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:reply-to:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=eb04fKPQBYtGYXmaTqpgN9hIMwKNARUUv10gVLHJJ15oehhYxvXifxYbvxmynEjQbZYYz9kkbYju5djP0jyfQlGGHYpjq6qo0/ucKrjd/igz+mHXoU2ZV6NZKIJ6h524oH9lcnH9fxRN/E7OHoCeJakmIP3ihZKbtnMhhvk3A/Y= Message-ID: <1e62d137050718121848b5080f@mail.gmail.com> Date: Tue, 19 Jul 2005 00:18:52 +0500 From: Fawad Lateef Reply-To: Fawad Lateef To: vamsi krishna Subject: Re: Avoiding BIGMEM support programatically. Cc: linux-kernel@vger.kernel.org In-Reply-To: <3faf0568050718075677c13e8@mail.gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7BIT Content-Disposition: inline References: <3faf0568050718075677c13e8@mail.gmail.com> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On 7/18/05, vamsi krishna wrote: > Hello All, > > I have a program working fine on a 2.6.xx-smp kernel, and the program > crashes on the same version kernel with bigmem i.e (2.6.xxx-bigmem). > What is your program ??? what it is doing ??? Can u explain ?? or send some code portion ?? b/c the BIGMEM kernel and smp/normal kenel has only a difference of HIGHMEM64G which allows system/kernel on x86 to use physical memory upto 64GB ..... and enabling this creates little-bit overhead on the kernel, but I don't think it will effect the working of most of the kernel modules ...... > I also found that for a same executable on bigmem kernel the virtual > address's of '&_start' and '&_etext', seem to vary in every new run. > I don't know abt this, so can't say any thing .... > Is there any way I can avoid the kernel's bigmem virtual address > mapping programatically? and still run the program on a bigmem kernel? > I think this can be done through specifying GFP_ATOMIC flag in the memory allocation function, so that it will use ZONE_NORMAL of the kernel ....... (if i m wrong, then plzz do correct me) -- Fawad Lateef