From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1DMBz0-00062x-Cq for qemu-devel@nongnu.org; Thu, 14 Apr 2005 17:34:43 -0400 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1DMByi-0005vN-Ts for qemu-devel@nongnu.org; Thu, 14 Apr 2005 17:34:27 -0400 Received: from [199.232.76.173] (helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1DMByi-0005sM-MN for qemu-devel@nongnu.org; Thu, 14 Apr 2005 17:34:24 -0400 Received: from [62.241.162.31] (helo=galaxy.systems.pipex.net) by monty-python.gnu.org with esmtp (Exim 4.34) id 1DMBqN-0008OH-4v for qemu-devel@nongnu.org; Thu, 14 Apr 2005 17:25:47 -0400 Received: from moon.base (81-179-219-236.dsl.pipex.com [81.179.219.236]) by galaxy.systems.pipex.net (Postfix) with ESMTP id 6CA8DE0000E5 for ; Thu, 14 Apr 2005 22:25:01 +0100 (BST) Received: from james by moon.base with local (Exim 4.34) id 1DMBvM-0001Nx-QI for qemu-devel@nongnu.org; Thu, 14 Apr 2005 22:30:56 +0100 Date: Thu, 14 Apr 2005 22:30:56 +0100 From: James Stone Subject: Re: [Qemu-devel] Open source qemu x86 accelerator module. Message-ID: <20050414213056.GA5312@moon.base> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable 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 Just a quick report on how this worked for me. I am not using devfs or udev, and major device 222 is no longer valid for /dev/qvm86.. I created the required device number 10.63 As expected, Win98 caused a segfault when using qvm86. Here is the dmesg output: qvm86: Created device 10.63 qvm86: Creating guest VM esp:00002f7c 00082000 fec00000 Unable to handle kernel NULL pointer dereference at virtual address 00000000 printing eip: e26144b0 *pde =3D 00000000 Oops: 0000 [#1] PREEMPT=20 Modules linked in: qvm86 realtime parport_pc lp parport ipt_REDIRECT ipt_owne pt_LOG ipt_limit ipt_state ipt_multiport iptable_mangle ip_nat_irc ip_nat_ftp table_nat ip_conntrack_irc ip_conntrack_ftp ip_conntrack emu10k1_gp snd_bt87x ner bttv video_buf i2c_algo_bit v4l2_common btcx_risc tveeprom videodev nls_i 859_1 nls_cp437 asb100 i2c_sensor i2c_nforce2 i2c_core usblp usb_storage viso sbserial sidewinder gameport joydev usbhid sg ide_cd cdrom n_hdlc ppp_synctty p_generic slhc ehci_hcd ohci_hcd snd_virmidi snd_seq_virmidi snd_seq_oss snd_ _midi snd_seq_midi_event snd_seq snd_usb_audio snd_usb_lib usbcore snd_intel8 snd_emu10k1 snd_rawmidi snd_seq_device snd_ac97_codec snd_pcm_oss snd_mixer_o snd_pcm snd_timer snd_page_alloc snd_util_mem snd_hwdep snd soundcore nvidia gart forcedeth ipt_REJECT iptable_filter ip_tables CPU: 0 EIP: 0060:[] Tainted: P VLI EFLAGS: 00210286 (2.6.11.6)=20 EIP is at qvm86_flush_shadow+0xa0/0x160 [qvm86] eax: dd984000 ebx: 00005000 ecx: e2601000 edx: 00000000 esi: 00000000 edi: 0000000c ebp: 00005000 esp: dd985d80 ds: 007b es: 007b ss: 0068 Process qemu (pid: 4060, threadinfo=3Ddd984000 task=3Dd4303ae0) Stack: e2601000 046d423c 44a1423c 00000000 00000400 0000000c dd984000 000003f 00000001 36472eb9 dd985ebc c9b00000 e2601000 e2614ce4 e2601000 36472eb 00082000 fec00000 00000000 00000000 dd985e20 c01f7d36 0036472e 000001b Call Trace: [] qvm86_exec+0x4b4/0x710 [qvm86] [] copy_from_user+0x46/0x90 [] qvm86_linux_ioctl+0xdc/0x140 [qvm86] [] do_ioctl+0x6f/0xa0 [] vfs_ioctl+0x65/0x1d0 [] sys_ioctl+0x67/0xa0 [] syscall_call+0x7/0xb Code: 8b 7c 24 14 89 dd 81 e5 00 f0 ff ff 8b 41 20 8b 34 38 c7 44 24 1c ff 03 00 b8 00 e0 ff ff 21 e0 89 44 24 18 8d b6 00 00 00 00 <8b> 1e f6 c3 01 75 17 06 00 00 00 00 83 c6 04 83 c5 04 ff 4c=20 <6>qvm86: Device closed =20 However, I noticed a significant speed up in performance in qemu performance running win98 without the module installed.. Must be something in the patch I suppose, but being somewhat clueless about these things, I have no idea what. The speed up was in the order of about 50%: Loading windows under qemu normally takes 30secs. With the qvm86 patch, it took 20secs. Also performance was noticably more snappy.. Very nice! Keep up the good work! James