From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263805AbUBRHpL (ORCPT ); Wed, 18 Feb 2004 02:45:11 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263793AbUBRHpL (ORCPT ); Wed, 18 Feb 2004 02:45:11 -0500 Received: from gate.crashing.org ([63.228.1.57]:60068 "EHLO gate.crashing.org") by vger.kernel.org with ESMTP id S263805AbUBRHpE (ORCPT ); Wed, 18 Feb 2004 02:45:04 -0500 Subject: Re: ppc suspend rewrite using inline. From: Benjamin Herrenschmidt To: Hugang Cc: Linux Kernel list , Pavel Machek In-Reply-To: <20040218153548.2690e905@localhost> References: <20040218153548.2690e905@localhost> Content-Type: text/plain Message-Id: <1077090098.10665.121.camel@gaston> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Wed, 18 Feb 2004 18:41:39 +1100 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Wed, 2004-02-18 at 18:35, Hugang wrote: > Hi Ben: > > I has finished rewriten the suspend code by inline, The swsusp2 works > fine without any problem, But pmdisk can't works, The problem that's, > when resume the current is not same as suspend. I don't understand why you write. Looking at your code, I don't see the point. By mixing C and assembly the way you do, and using those horrible inline asm constructs, you are basically making the code worse, less readable, and prone for bugs. I don't think you should even try going into that direction, it makes no sense. > I think the code can reuse by sleep. Same > I think turn off mmu copyback page is important, but I has looked x86 > resume code, it do not. Same > Please take a look. Please try to be clear. > thanks.