From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.33) id 1BkmeO-0000tP-Cu for qemu-devel@nongnu.org; Wed, 14 Jul 2004 12:30:32 -0400 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.33) id 1BkmeH-0000tA-CN for qemu-devel@nongnu.org; Wed, 14 Jul 2004 12:30:32 -0400 Received: from [199.232.76.173] (helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.33) id 1BkmeH-0000t7-9Y for qemu-devel@nongnu.org; Wed, 14 Jul 2004 12:30:25 -0400 Received: from [195.250.128.79] (helo=majordomo.vol.cz) by monty-python.gnu.org with esmtp (TLSv1:DES-CBC3-SHA:168) (Exim 4.34) id 1Bkmbm-0003OV-M2 for qemu-devel@nongnu.org; Wed, 14 Jul 2004 12:27:51 -0400 Received: from [83.148.9.39] (a5prg-279.dialup.vol.cz [83.148.9.39]) by majordomo.vol.cz (8.12.9p2/8.12.9) with ESMTP id i6EGRkrX040604 for ; Wed, 14 Jul 2004 18:27:47 +0200 (CEST) (envelope-from xnavara@volny.cz) Message-ID: <40F4A55B.2000203@volny.cz> Date: Wed, 14 Jul 2004 05:15:39 +0200 From: Filip Navara MIME-Version: 1.0 Subject: Re: [Qemu-devel] Win98: how to exchange data with Linux References: <200407122356.02502.eschmit@tin.it> <1089671145.12301.10.camel@aragorn> <200407130924.58879.vaise@votreservice.com> <1089727701.7843.58.camel@espiron.av7.local> <40F3EF20.2020802@kadu.net> <40F3F5B1.2040908@kadu.net> <40F41A65.8050807@volny.cz> <20040714022611.GB6436@jbrown.mylinuxbox.org> In-Reply-To: <20040714022611.GB6436@jbrown.mylinuxbox.org> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Reply-To: qemu-devel@nongnu.org List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Jim C. Brown wrote: >Technically speaking, one could write a file system driver that talks to qemu >instead of going through a storage driver. (I didn't say it would be easy, just >barely possible.) Qemu could then translate attempts (to write a file on Windows >using this driver would cause qemu to write the file on linux's fs natively for >example). It is more work than its worth, but it is possible. > > Strictly technically you would need to fake a storage device (~ 1000 lines code) since Windows can't mount filesystem without a device and then you can write a file system driver that will recoginze the fake device and act as you described. I agree, yes, doing this is possible, but it would be pretty complex to implement for no real gain. It's less flexible than any solution based on networking (since you would have to write a guest driver for every OS that you want to use the feature in) and the speed increase would be marginal if any. >A modified user-net which emulated SMB protocol to the guest but did the i/o >natively on the host would do the same thing and be easier to implement. > > Again, I agree, that's the way to go. Regards, Filip