From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1Kt11N-0007p4-Rl for qemu-devel@nongnu.org; Thu, 23 Oct 2008 10:18:41 -0400 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1Kt11N-0007oY-2C for qemu-devel@nongnu.org; Thu, 23 Oct 2008 10:18:41 -0400 Received: from [199.232.76.173] (port=40711 helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1Kt11M-0007oP-Rg for qemu-devel@nongnu.org; Thu, 23 Oct 2008 10:18:40 -0400 Received: from gecko.sbs.de ([194.138.37.40]:21210) by monty-python.gnu.org with esmtps (TLS-1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.60) (envelope-from ) id 1Kt11M-0000rK-GH for qemu-devel@nongnu.org; Thu, 23 Oct 2008 10:18:40 -0400 Message-ID: <49008787.3050202@siemens.com> Date: Thu, 23 Oct 2008 16:17:43 +0200 From: Jan Kiszka MIME-Version: 1.0 References: <1224771556-11146-1-git-send-email-glommer@redhat.com> <1224771556-11146-2-git-send-email-glommer@redhat.com> <490080A2.7040802@us.ibm.com> In-Reply-To: <490080A2.7040802@us.ibm.com> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Subject: [Qemu-devel] Re: [PATCH 01/32] use anonymous memory for kqemu. Reply-To: qemu-devel@nongnu.org List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Anthony Liguori Cc: Glauber Costa , jes@sgi.com, qemu-devel@nongnu.org, avi@qumranet.com, dmitry.baryshkov@siemens.com Anthony Liguori wrote: > The idea behind this patch is that currently kqemu uses /dev/shm for > memory allocations instead of anonymous memory (like the rest of QEMU). > Instead of introducing YA hook, we can just switch kqemu to use > anonymous memory and eliminate the special case. > > If I recall correctly, the reason for using /dev/shm was concern that > get_user_pages() didn't do the right thing for anonymous memory and the > use of /dev/shm was a hack around that. However, I'm not sure that was > ever the case. Certainly, with any sufficiently modern kernel > get_user_pages() does what one would expect. KVM uses get_user_pages() > on anonymous memory in roughly the same way kqemu uses it now. > > So I think it's safe to make the switch. Fabrice, what do you think? This hack-around, was it purely Linux-motivated? Or did/do other OSes have similar issues? Jan -- Siemens AG, Corporate Technology, CT SE 2 Corporate Competence Center Embedded Linux