Re: XP blue screen with qemu-kvm-0.11.0
Ross Boylan wrote: On Sat, 2009-10-31 at 15:21 +0300, Michael Tokarev wrote: [] Switch to alsa to get your audio working. I don't see an alsa option for kvm/qemu. I'm already running alsa, but under KDE which tends to grab the device. Which device you think KDE grabs? OSS device (/dev/dsp) can only be opened once, subsequent attempts will return EBUSY. But alsa devices can be opened multiple times. Grabbing alsa devices is not supported. For alsa usage in kvm, see -audio-help option. My package (from corpit.ru) is built with alsa being the default driver. The VM starts; I see the initial XP screen with the 4 colors; I see the background I get when I log in (it logs me in directly without prompt); and then (pretty fast) I get a blue screen. The stop code is 0x8E, and the text says to check disk space and BIOS options. What's the bios files your kvm uses? Are they by a change from some old qemu install? They appear to be from the latest install, since strace shows various bios files loading from /usr/local/kvm/share. The invoking environment was a little different from the real run, since strace vdeq apparently traced vdeq but not kvm calls. So I just ran the kvm bare. Ok. Does kvm deb from http://www.corpit.ru/debian/tls/kvm/ expose the same issue? Yes. However, as it fails it left a reverberating sound (fragment of the Windows login tone). That's semi-expected after a crash. Unlike with real hardware, when everything including the sound card gets reset by the bios right after a crash, your host continues running and your sound card continues executing the commands. I tried starting in safe mode. XP said there was new hardware: the video (-vga std). It could not find a driver on the internet(!? the device was identified as a VGA controller). Then it told me the driver had been installed (after I hit finish). I rebooted in regular mode. This time there was no sound, but the machine failed again with STOP 0x8E (as before). The video appeared to be working throughout this, showing a window that exceeded vanilla VGA resolution. -vga std creates a generic but 100% fake video card in guest. The card understands standard VESA/VGA commands and modes, but windows does not know who the vendor is, since it does not try to emulate any real device (like cirrus for example). It's an unnamed device of class VGA device. No doubt windows can't find a driver for a non-existing device, even in internet. But the thing is that the default generic video driver works with this adaptor just fine. But windows thinks is't not the most appropriate driver and marks this device in the device tree. But this all has nothing to do with your issues. Audio, VGA and straceing - it's all ok. Your main problem appears to be the broken disk image. May be it's kvm's fault, maybe windows - it's very difficult to say by now. /mjt -- To unsubscribe from this list: send the line unsubscribe kvm in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: XP blue screen with qemu-kvm-0.11.0
On Sat, 2009-10-31 at 15:21 +0300, Michael Tokarev wrote: Ross Boylan wrote: My XP VM was working OK, and then started crashing shortly after it logged me in. There were no obvious changes at the time. I built the latest qemu-kvm, but the problem persists. I am running 32 bit XP on Intel(R) Xeon(R) CPU E5420 @ 2.50GHz (8 cores total), Debian GNU/Linux mostly Lenny (amd64), but with some more recent stuff. In particular, the kernel is 2.6.30-8 and I pulled in the kernel-headers package to match before building kvm. However, libc6 and libc6-dev are at Lenny's 2.7-18 version. Libc is basically irrelevant here. What matters are the host kernel and kvm version. $ ./XP.sh ++ sudo vdeq bin/qemu-system-x86_64 -net nic,vlan=1,macaddr=52:54:a0:12:01:00 -net vde,vlan=1,sock=/var/run/vde2/tap0.ctl -boot c -vga std -hda /dev/turtle/XP01 -soundhw es1370 -localtime -m 1G -smp 2 arg ,vlan=1,sock=/var/run/vde2/tap0.ctl TUNGETIFF ioctl() failed: Invalid argument TUNSETOFFLOAD ioctl() failed: Bad address oss: Could not initialize DAC oss: Failed to open `/dev/dsp' oss: Reason: Device or resource busy oss: Could not initialize DAC oss: Failed to open `/dev/dsp' oss: Reason: Device or resource busy audio: Failed to create voice `es1370.dac2' # and more sound-related complaints Switch to alsa to get your audio working. I don't see an alsa option for kvm/qemu. I'm already running alsa, but under KDE which tends to grab the device. The VM starts; I see the initial XP screen with the 4 colors; I see the background I get when I log in (it logs me in directly without prompt); and then (pretty fast) I get a blue screen. The stop code is 0x8E, and the text says to check disk space and BIOS options. What's the bios files your kvm uses? Are they by a change from some old qemu install? They appear to be from the latest install, since strace shows various bios files loading from /usr/local/kvm/share. The invoking environment was a little different from the real run, since strace vdeq apparently traced vdeq but not kvm calls. So I just ran the kvm bare. Does kvm deb from http://www.corpit.ru/debian/tls/kvm/ expose the same issue? Yes. However, as it fails it left a reverberating sound (fragment of the Windows login tone). I tried starting in safe mode. XP said there was new hardware: the video (-vga std). It could not find a driver on the internet(!? the device was identified as a VGA controller). Then it told me the driver had been installed (after I hit finish). I rebooted in regular mode. This time there was no sound, but the machine failed again with STOP 0x8E (as before). The video appeared to be working throughout this, showing a window that exceeded vanilla VGA resolution. Ross -- To unsubscribe from this list: send the line unsubscribe kvm in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: XP blue screen with qemu-kvm-0.11.0
Ross Boylan wrote: My XP VM was working OK, and then started crashing shortly after it logged me in. There were no obvious changes at the time. I built the latest qemu-kvm, but the problem persists. I am running 32 bit XP on Intel(R) Xeon(R) CPU E5420 @ 2.50GHz (8 cores total), Debian GNU/Linux mostly Lenny (amd64), but with some more recent stuff. In particular, the kernel is 2.6.30-8 and I pulled in the kernel-headers package to match before building kvm. However, libc6 and libc6-dev are at Lenny's 2.7-18 version. Libc is basically irrelevant here. What matters are the host kernel and kvm version. $ ./XP.sh ++ sudo vdeq bin/qemu-system-x86_64 -net nic,vlan=1,macaddr=52:54:a0:12:01:00 -net vde,vlan=1,sock=/var/run/vde2/tap0.ctl -boot c -vga std -hda /dev/turtle/XP01 -soundhw es1370 -localtime -m 1G -smp 2 arg ,vlan=1,sock=/var/run/vde2/tap0.ctl TUNGETIFF ioctl() failed: Invalid argument TUNSETOFFLOAD ioctl() failed: Bad address oss: Could not initialize DAC oss: Failed to open `/dev/dsp' oss: Reason: Device or resource busy oss: Could not initialize DAC oss: Failed to open `/dev/dsp' oss: Reason: Device or resource busy audio: Failed to create voice `es1370.dac2' # and more sound-related complaints Switch to alsa to get your audio working. The VM starts; I see the initial XP screen with the 4 colors; I see the background I get when I log in (it logs me in directly without prompt); and then (pretty fast) I get a blue screen. The stop code is 0x8E, and the text says to check disk space and BIOS options. What's the bios files your kvm uses? Are they by a change from some old qemu install? Does kvm deb from http://www.corpit.ru/debian/tls/kvm/ expose the same issue? -no-kvm-irqchip and -no-kvm-pit make no difference. With -no-kvm I get stop code 0x24 and a suggestion to disable anti-virus, defragmentation, and backup software. This is the one obvious change between the Lenny kvm and the one I just built; with lenny kvm (kvm 72+dfsg-5~lenny3) running with -no-kvm simply seemed to hang forever (I think I waited at least 15 minutes). This disk turtle/XP01 is a read-write snapshot on turtle/XP00. The snapshot looks healthy, with about 50% allocated to the snapshot. The snapshot volume is 10G and the original is 50G. The VM starts fine if I point it to XP00 instead of XP01. Well, that's telling, isn't it? If you change disk image and it works, the problem should be in the disk image... P.S. What are the different files in my kvm/bin directory? There's no kvm/bin directory in the source tarball of qemu-kvm-0.11.0. /mjt -- To unsubscribe from this list: send the line unsubscribe kvm in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: XP blue screen with qemu-kvm-0.11.0
On Sat, 2009-10-31 at 15:21 +0300, Michael Tokarev wrote: Ross Boylan wrote: My XP VM was working OK, and then started crashing shortly after it logged me in. There were no obvious changes at the time. I built the latest qemu-kvm, but the problem persists. I am running 32 bit XP on Intel(R) Xeon(R) CPU E5420 @ 2.50GHz (8 cores total), Debian GNU/Linux mostly Lenny (amd64), but with some more recent stuff. In particular, the kernel is 2.6.30-8 and I pulled in the kernel-headers package to match before building kvm. However, libc6 and libc6-dev are at Lenny's 2.7-18 version. Libc is basically irrelevant here. What matters are the host kernel and kvm version. BTW, I do not have libvirt installed. $ ./XP.sh ++ sudo vdeq bin/qemu-system-x86_64 -net nic,vlan=1,macaddr=52:54:a0:12:01:00 -net vde,vlan=1,sock=/var/run/vde2/tap0.ctl -boot c -vga std -hda /dev/turtle/XP01 -soundhw es1370 -localtime -m 1G -smp 2 arg ,vlan=1,sock=/var/run/vde2/tap0.ctl TUNGETIFF ioctl() failed: Invalid argument TUNSETOFFLOAD ioctl() failed: Bad address Are the previous 2 messages significant? Just noise from vdeq? ... The VM starts; I see the initial XP screen with the 4 colors; I see the background I get when I log in (it logs me in directly without prompt); and then (pretty fast) I get a blue screen. The stop code is 0x8E, and the text says to check disk space and BIOS options. What's the bios files your kvm uses? How do I find out? Are they by a change from some old qemu install? Does kvm deb from http://www.corpit.ru/debian/tls/kvm/ expose the same issue? I'll try the next time I'm at the machine (I've also had problems with remote use). -no-kvm-irqchip and -no-kvm-pit make no difference. With -no-kvm I get stop code 0x24 and a suggestion to disable anti-virus, defragmentation, and backup software. This is the one obvious change between the Lenny kvm and the one I just built; with lenny kvm (kvm 72+dfsg-5~lenny3) running with -no-kvm simply seemed to hang forever (I think I waited at least 15 minutes). This disk turtle/XP01 is a read-write snapshot on turtle/XP00. The snapshot looks healthy, with about 50% allocated to the snapshot. The snapshot volume is 10G and the original is 50G. The VM starts fine if I point it to XP00 instead of XP01. Well, that's telling, isn't it? If you change disk image and it works, the problem should be in the disk image... Maybe. As I said, it was working, and I get different errors with --no-kvm. Is there a way I can mount the individual partitions on XP01 to get a look at them? It took a lot of time to create this, so I'm really hoping I can salvage it. P.S. What are the different files in my kvm/bin directory? There's no kvm/bin directory in the source tarball of qemu-kvm-0.11.0. I'm referring to the installation directories: /usr/local/kvm/bin$ ls qemu-img qemu-io qemu-nbd qemu-system-x86_64 Ross -- To unsubscribe from this list: send the line unsubscribe kvm in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: XP blue screen with qemu-kvm-0.11.0
Ross Boylan wrote: [] ++ sudo vdeq bin/qemu-system-x86_64 -net nic,vlan=1,macaddr=52:54:a0:12:01:00 -net vde,vlan=1,sock=/var/run/vde2/tap0.ctl -boot c -vga std -hda /dev/turtle/XP01 -soundhw es1370 -localtime -m 1G -smp 2 arg ,vlan=1,sock=/var/run/vde2/tap0.ctl TUNGETIFF ioctl() failed: Invalid argument TUNSETOFFLOAD ioctl() failed: Bad address Are the previous 2 messages significant? Just noise from vdeq? Not that I really know. I missed these in your original email. But again, I don't use vde, and since the argument for the -net command is a socket, I guess it gets messed up - kvm assumes it's a tun device, not a socket... The VM starts; I see the initial XP screen with the 4 colors; I see the background I get when I log in (it logs me in directly without prompt); and then (pretty fast) I get a blue screen. The stop code is 0x8E, and the text says to check disk space and BIOS options. What's the bios files your kvm uses? How do I find out? I usually use strace. Dunno really, it looks like there's no way to ask where kvm will look for bios files. [] The VM starts fine if I point it to XP00 instead of XP01. Well, that's telling, isn't it? If you change disk image and it works, the problem should be in the disk image... Maybe. As I said, it was working, and I get different errors with --no-kvm. With -no-kvm you're exposing less-tested code paths in kvm. Is there a way I can mount the individual partitions on XP01 to get a look at them? It took a lot of time to create this, so I'm really hoping I can salvage it. That's what qemu-nbd is for. Or qemu-img convert it to raw and use kpartx on it. P.S. What are the different files in my kvm/bin directory? There's no kvm/bin directory in the source tarball of qemu-kvm-0.11.0. I'm referring to the installation directories: /usr/local/kvm/bin$ ls qemu-img qemu-io qemu-nbd qemu-system-x86_64 There are manpages for each (except qemu-io which is a debugging tool). See also qemu documentation. /mjt -- To unsubscribe from this list: send the line unsubscribe kvm in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
XP blue screen with qemu-kvm-0.11.0
My XP VM was working OK, and then started crashing shortly after it logged me in. There were no obvious changes at the time. I built the latest qemu-kvm, but the problem persists. I am running 32 bit XP on Intel(R) Xeon(R) CPU E5420 @ 2.50GHz (8 cores total), Debian GNU/Linux mostly Lenny (amd64), but with some more recent stuff. In particular, the kernel is 2.6.30-8 and I pulled in the kernel-headers package to match before building kvm. However, libc6 and libc6-dev are at Lenny's 2.7-18 version. $ ./XP.sh ++ sudo vdeq bin/qemu-system-x86_64 -net nic,vlan=1,macaddr=52:54:a0:12:01:00 -net vde,vlan=1,sock=/var/run/vde2/tap0.ctl -boot c -vga std -hda /dev/turtle/XP01 -soundhw es1370 -localtime -m 1G -smp 2 arg ,vlan=1,sock=/var/run/vde2/tap0.ctl TUNGETIFF ioctl() failed: Invalid argument TUNSETOFFLOAD ioctl() failed: Bad address oss: Could not initialize DAC oss: Failed to open `/dev/dsp' oss: Reason: Device or resource busy oss: Could not initialize DAC oss: Failed to open `/dev/dsp' oss: Reason: Device or resource busy audio: Failed to create voice `es1370.dac2' # and more sound-related complaints The VM starts; I see the initial XP screen with the 4 colors; I see the background I get when I log in (it logs me in directly without prompt); and then (pretty fast) I get a blue screen. The stop code is 0x8E, and the text says to check disk space and BIOS options. -no-kvm-irqchip and -no-kvm-pit make no difference. With -no-kvm I get stop code 0x24 and a suggestion to disable anti-virus, defragmentation, and backup software. This is the one obvious change between the Lenny kvm and the one I just built; with lenny kvm (kvm 72+dfsg-5~lenny3) running with -no-kvm simply seemed to hang forever (I think I waited at least 15 minutes). This disk turtle/XP01 is a read-write snapshot on turtle/XP00. The snapshot looks healthy, with about 50% allocated to the snapshot. The snapshot volume is 10G and the original is 50G. The VM starts fine if I point it to XP00 instead of XP01. Any ideas or suggestions? Ross Boylan P.S. What are the different files in my kvm/bin directory? -- To unsubscribe from this list: send the line unsubscribe kvm in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html