From owner-freebsd-hackers Thu Dec 6 22: 2: 7 2001 Delivered-To: freebsd-hackers@freebsd.org Received: from mail.viasoft.com.cn (unknown [61.153.1.177]) by hub.freebsd.org (Postfix) with ESMTP id 85D5F37B416 for ; Thu, 6 Dec 2001 22:02:02 -0800 (PST) Received: from viasoft.com.cn (davidwnt.viasoft.com.cn [192.168.1.239]) by mail.viasoft.com.cn (8.9.3/8.9.3) with ESMTP id OAA16215; Fri, 7 Dec 2001 14:09:17 +0800 Message-ID: <3C105A26.9090005@viasoft.com.cn> Date: Fri, 07 Dec 2001 13:56:54 +0800 From: David Xu User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:0.9.4) Gecko/20011019 Netscape6/6.2 X-Accept-Language: en-us MIME-Version: 1.0 To: Alfred Perlstein Cc: Nicolas Souchu , freebsd-hackers@FreeBSD.ORG Subject: Re: kld VM pager References: <20011206143022.H13566@cedar.alcove-fr> <3C10410E.9040805@viasoft.com.cn> <20011206231758.J92148@elvis.mu.org> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG you mean we can use OBJT_PHYS? do we fully support it? althougth we have phys_pager, I suspect it is not enough, because along with fault handling, graphics driver will need other operations, for examples, unmap other pages and setup hardware registers, think about a frame buffer driver need to map 32K window into 1024K address space, when fault ocurrs in second 32K window, it should unmap first 32K address and activate second 32K address and setup graphics card registers to let later access to its internal second 32K memory bank. how can you handle such issue after vm_fault? Regards, -- David Xu Alfred Perlstein wrote: >* David Xu [011206 22:15] wrote: > >>FreeBSD does not have fault hook available, all faults are processed in >>vm_fault. >>I know Linux supports that idea, you can insert a fault hook to >>monitor some >>address range where fault occurs, and then graphics frame buffer can be >>supported. >> > >I'm sure one could add a callback in vm_fault without much issue. >However this isn't the point of the hook. One can insert the >vm object into the vm map, when a fault occurs the vm_fault code >will call the pager handler to service it. > >The only good part for having a fault hook in linux is most likely >for debug purposes. > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message