Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932722AbWCPUff (ORCPT ); Thu, 16 Mar 2006 15:35:35 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S932724AbWCPUff (ORCPT ); Thu, 16 Mar 2006 15:35:35 -0500 Received: from smtp.osdl.org ([65.172.181.4]:21418 "EHLO smtp.osdl.org") by vger.kernel.org with ESMTP id S932722AbWCPUfe (ORCPT ); Thu, 16 Mar 2006 15:35:34 -0500 Date: Thu, 16 Mar 2006 12:35:22 -0800 (PST) From: Linus Torvalds To: Hugh Dickins cc: "Bryan O'Sullivan" , Andrew Morton , rdreier@cisco.com, hch@infradead.org, linux-kernel@vger.kernel.org Subject: Re: [PATCH 10 of 20] ipath - support for userspace apps using core driver In-Reply-To: Message-ID: References: <71644dd19420ddb07a75.1141922823@localhost.localdomain> <1141948516.10693.55.camel@serpentine.pathscale.com> <1141949262.10693.69.camel@serpentine.pathscale.com> <20060309163740.0b589ea4.akpm@osdl.org> <1142470579.6994.78.camel@localhost.localdomain> <1142475069.6994.114.camel@localhost.localdomain> <1142477579.6994.124.camel@localhost.localdomain> <20060315192813.71a5d31a.akpm@osdl.org> <1142485103.25297.13.camel@camp4.serpentine.com> <20060315213813.747b5967.akpm@osdl.org> <1142523201.25297.56.camel@camp4.serpentine.com> <1142538765.10950.16.camel@serpentine.pathscale.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1257 Lines: 30 On Thu, 16 Mar 2006, Hugh Dickins wrote: > On Thu, 16 Mar 2006, Bryan O'Sullivan wrote: > > > > OK. Would it be correct to say that this is what we should do, then? > > > > * On 2.6.15 and later kernels, use __GFP_COMP at allocation time, > > and get_page in ->nopage. This is what we're doing as of this > > morning, and it works. > > * For backports to 2.6.14 and earlier, avoid __GFP_COMP, mark each > > page with SetPageReserved at allocation time, and do nothing > > special in ->nopage. Do we need to ClearPageReserved before > > freeing? > > Yes, I believe that's exactly right - so long as you do ClearPageReserved > from each of its constituent 0-order-pages before freeing the >0-order > page, in the <= 2.6.14 case. The alternative is to always allocate the pages one by one ("order-0"), and do get_page() when you return them in the ->nopage handler. That will work with any kernel, so it has the simplicity thing going for it. Linus - 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/