Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1761108AbZF3XVk (ORCPT ); Tue, 30 Jun 2009 19:21:40 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1758462AbZF3XVa (ORCPT ); Tue, 30 Jun 2009 19:21:30 -0400 Received: from terminus.zytor.com ([198.137.202.10]:38612 "EHLO terminus.zytor.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1760975AbZF3XVY (ORCPT ); Tue, 30 Jun 2009 19:21:24 -0400 Message-ID: <4A4A9D8E.5080500@zytor.com> Date: Tue, 30 Jun 2009 16:19:42 -0700 From: "H. Peter Anvin" User-Agent: Thunderbird 2.0.0.21 (X11/20090320) MIME-Version: 1.0 To: Yinghai Lu CC: Linus Torvalds , Ingo Molnar , Mikael Pettersson , Matthew Wilcox , Grant Grundler , Linux Kernel Mailing List , linux-pci@vger.kernel.org Subject: Re: [BUG 2.6.31-rc1] HIGHMEM64G causes hang in PCI init on 32-bit x86 References: <200906261559.n5QFxJH8027336@pilspetsen.it.uu.se> <20090629022911.GC20297@lackof.org> <4A484A8A.9020704@zytor.com> <19016.41349.636663.515540@pilspetsen.it.uu.se> <20090629112155.GJ5480@parisc-linux.org> <19016.44061.600652.676183@pilspetsen.it.uu.se> <4A490804.3040609@zytor.com> <4A494478.7020304@kernel.org> <4A494E3C.70304@kernel.org> <4A495C0D.2020807@zytor.com> <4A4966EF.6010809@kernel.org> <4A496D4B.3040608@kernel.org> <19017.53428.834539.389495@pilspetsen.it.uu.se> <4A4A25B1.5010102@zytor.com> <4A4A6888.30001@kernel.org> <4A4A6B1C.7030405@zytor.com> <4A4A701F.3050700@kernel.org> <4A4A81C9.9070008@zytor.com> <4A4A88A2.7010509@kernel.org> <4A4A921F.7080100@kernel.org> <4A4A9C2F.80808@kernel.org> In-Reply-To: <4A4A9C2F.80808@kernel.org> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 931 Lines: 30 Yinghai Lu wrote: > Linus Torvalds wrote: >> On Tue, 30 Jun 2009, Linus Torvalds wrote: >>> I don't see any single-use trick then, and so it needs the whole statement >>> expression mess. >> Hmm. Does (((x)-1) | mask)+1) work? >> >> I haven't thought it fully through, but that _should_ take care of the >> "already aligned" case, no? > > yes. that is right. > > then how about > roundup(x,y) > round_up(x,y) > > roundup doesn't need y is 2^n > but round_up does need y is 2^n, and only for x86 > We should definitely move whatever to global. However, I do think it is valuable to have something that can avoid divides even if the argument is not necessarily a constant. -hpa -- 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/