From: Avi Kivity Subject: Re: Re: [PATCH 1 of 7] x86: add _PAGE_IOMAP pte flag for IO mappings Date: Wed, 10 Sep 2008 12:55:25 +0300 Message-ID: <48C7998D.1030408@qumranet.com> References: <4d93ba619c716c45cca0.1220826073@localhost> <48C67AE0.5060503@qumranet.com> <48C68C6B.7030803@goop.org> <48C68EA3.4050508@qumranet.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <48C68EA3.4050508@qumranet.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Jeremy Fitzhardinge Cc: Ingo Molnar , Xen Devel , Andi Kleen , linux-kernel@vger.kernel.org, "H. Peter Anvin" List-Id: xen-devel@lists.xenproject.org Avi Kivity wrote: >> >> I'm not sure; I still don't really understand how _PAGE_SPECIAL gets >> used, other than being user-mode mapping only. But in principle, >> _PAGE_IOMAP could be set on both kernel and user mappings (if you direct >> map a device into a process address space), so I think they would >> conflict then? >> >> > > It's a "don't refcount me" flag, which is not sematically the same as > I/O, but may be close enough. Actually it's more of a "no struct page" flag, which implies no refcounting. And not having a struct page should correspond well to a pte not requiring pfn->mfn conversion and being an I/O page. -- error compiling committee.c: too many arguments to function