From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1HFZb5-00024X-Tm for qemu-devel@nongnu.org; Fri, 09 Feb 2007 12:31:43 -0500 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1HFZb4-00023K-Ho for qemu-devel@nongnu.org; Fri, 09 Feb 2007 12:31:43 -0500 Received: from [199.232.76.173] (helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1HFZb4-00023H-CX for qemu-devel@nongnu.org; Fri, 09 Feb 2007 12:31:42 -0500 Received: from brick.kernel.dk ([62.242.22.158] helo=kernel.dk) by monty-python.gnu.org with esmtps (TLS-1.0:DHE_RSA_AES_256_CBC_SHA:32) (Exim 4.52) id 1HFZb3-0001Ny-Vo for qemu-devel@nongnu.org; Fri, 09 Feb 2007 12:31:42 -0500 Received: from carl.home.kernel.dk (carl.home.kernel.dk [192.168.0.22]) by kernel.dk (Postfix) with ESMTP id 558E7257189 for ; Fri, 9 Feb 2007 18:31:35 +0100 (CET) Date: Fri, 9 Feb 2007 18:32:04 +0100 From: Jens Axboe Subject: Re: [Qemu-devel] Xenoppix (KNOPPIX5.1.1 + Xen3.0.4 + QEMU/KVM + HTTP-FUSE) is released Message-ID: <20070209173204.GC3873@kernel.dk> References: <20070209.194958.123996617.k.suzaki@aist.go.jp> <45CC5063.60504@bellard.org> <20070209.202151.55858230.k.suzaki@aist.go.jp> <45CC8195.1030404@bellard.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <45CC8195.1030404@bellard.org> 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 On Fri, Feb 09 2007, Fabrice Bellard wrote: > Kuniyasu Suzaki wrote: > > [...] > > >>2) For best performances with kqemu, it is better to use Linux 2.4 as > > >>guest OS (I know this is far from acceptable, but it can help some > > >>people to get better performance !). > > > >Please tell me the reason why Linux 2.4 shows better performance, if > >you have time. Does the difference of clock tick cause the problem? > > The Linux 2.6 kernel has more code running with the interrupts disabled > than the 2.4 one. kqemu interprets the x86 code when the interrupts are > disabled, so it is very slow in this case. > > A potential solution I am investigating is to use the new > paravirtualization API of the kernel versions >= 2.6.20. That'd be a great idea! -- Jens Axboe