Bug#496028: linux-2.6: host kernel oopses when adding a cpu to a kvm virtual machine
reassign 496028 kvm thanks On Fri, Aug 22, 2008 at 11:18:20PM +0200, rtheys wrote: Hi, Quoting maximilian attems m...@stro.at: On Fri, 22 Aug 2008, Rik Theys wrote: I've noticed some kvm fixes in upstreams 2.6.26.3 release but am not sure they fix this issue. right please try out sid snapshots, see apt lines - http://wiki.debian.org/DebianKernel they carry 2.6.26.3 I've tried the kernel targetted for sid and the kernel no longer oopses when I enter the command, but kvm segfaults now. So instead of oopsing the host kernel, the virtual machine now dies. I'm also not sure the command I enter is supposed to work as I can't find any documentation about it. I guess you can reassign this bug to kvm when the new kernel enters unstable/lenny. Doing so. Cheers, Moritz -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#496028: linux-2.6: host kernel oopses when adding a cpu to a kvm virtual machine
Package: linux-2.6 Version: 2.6.26-3 Severity: normal I'm running a kvm virtual machine (with CentOS 5.2 as the guest OS, but it doesn't seem to matter) on a Lenny host (using the 2.6.26 kernel and kvm from unstable). When I try to hot-add a cpu to the virtual machine, the host kernel oopses and becomes unresponsive after a few minutes/seconds. I'm not sure this is a kernel or kvm bug. I've added the cpu to the virtual machine by logging in on the kvm monitor interface of the virtual machine and entering: cpu_set 1 online I'm not 100% sure this is the correct way to hot-add a CPU. I can't seem to find the documented way to do this using google. Also asked on #kvm but got no response. The kvm virtual machine was started with the following command: kvm -m 1024 -smp 1 -monitor tcp:localhost:,server,nowait -localtime -drive file=/test2-local.img,if=ide,boot=on -net nic,macaddr=00:16:3e:00:00:02,vlan=0, -net tap,script=/etc/kvm/kvm-ifup,vlan=0 -vnc 127.0.0.1:1 -daemonize I've noticed some kvm fixes in upstreams 2.6.26.3 release but am not sure they fix this issue. Here's the oops I get: Message from [EMAIL PROTECTED] at Fri Aug 22 09:24:16 2008 ... lo kernel: [89264.841448] Oops: 0002 [1] SMP Message from [EMAIL PROTECTED] at Fri Aug 22 09:24:16 2008 ... lo kernel: [89264.842569] Code: 48 89 df 48 83 c3 48 e8 51 86 ff ff 4c 89 e7 48 29 d8 48 89 c3 e8 8c bf eb df 48 c1 e8 0c 48 c1 eb 06 48 6b c0 38 48 8b 44 28 10 0f ab 58 40 45 85 ff 0f 85 e8 00 00 00 49 8b 04 24 48 3b 05 63 Message from [EMAIL PROTECTED] at Fri Aug 22 09:24:16 2008 ... lo kernel: [89264.842778] CR2: 0040 Aug 22 09:24:16 lo kernel: [89264.841352] BUG: unable to handle kernel NULL pointer dereference at 0040 Aug 22 09:24:16 lo kernel: [89264.841389] IP: [a03663ed] :kvm:mmu_set_spte+0x209/0x340 Aug 22 09:24:16 lo kernel: [89264.841428] PGD 7d839067 PUD 4447b067 PMD 0 Aug 22 09:24:16 lo kernel: [89264.841448] Oops: 0002 [1] SMP Aug 22 09:24:16 lo kernel: [89264.841466] CPU 0 Aug 22 09:24:16 lo kernel: [89264.841479] Modules linked in: tun kvm_intel kvm video output ac battery cpufreq_ondemand cpufreq_userspace cpufreq_powersave cpufreq_stats freq_table cpufreq_conservative ib_iser rdma_cm ib_cm iw_cm ib_sa ib_mad ib_core ib_addr iscsi_tcp libiscsi scsi_transport_iscsi bridge ipv6 loop psmouse snd_hda_intel i2c_i801 intel_agp button evdev serio_raw parport_pc parport snd_pcm snd_timer snd soundcore snd_page_alloc i2c_core iTCO_wdt dcdbas pcspkr ext3 jbd mbcache dm_mirror dm_log dm_snapshot dm_mod sg sr_mod cdrom sd_mod ata_generic usbhid hid ff_memless ata_piix libata scsi_mod dock floppy ide_pci_generic ide_core tg3 ehci_hcd uhci_hcd thermal processor fan thermal_sys Aug 22 09:24:16 lo kernel: [89264.841772] Pid: 3579, comm: kvm Not tainted 2.6.26-1-amd64 #1 Aug 22 09:24:16 lo kernel: [89264.841788] RIP: 0010:[a03663ed] [a03663ed] :kvm:mmu_set_spte+0x209/0x340 Aug 22 09:24:16 lo kernel: [89264.841827] RSP: 0018:81007add5b38 EFLAGS: 00010212 Aug 22 09:24:16 lo kernel: [89264.841843] RAX: RBX: 0001 RCX: 81007ad78080 Aug 22 09:24:16 lo kernel: [89264.841862] RDX: 0004 RSI: 0003fff2 RDI: 8100790b8030 Aug 22 09:24:16 lo kernel: [89264.841880] RBP: e200 R08: 0001 R09: 0009 Aug 22 09:24:16 lo kernel: [89264.841899] R10: 810032893410 R11: a0381fa3 R12: 8100790b8030 Aug 22 09:24:16 lo kernel: [89264.841917] R13: 81007adf00c0 R14: f001 R15: Aug 22 09:24:16 lo kernel: [89264.841937] FS: 513c1950() GS:8053b000() knlGS: Aug 22 09:24:16 lo kernel: [89264.841964] CS: 0010 DS: 002b ES: 002b CR0: 8005003b Aug 22 09:24:16 lo kernel: [89264.841980] CR2: 0040 CR3: 7d95d000 CR4: 26e0 Aug 22 09:24:16 lo kernel: [89264.841999] DR0: DR1: DR2: Aug 22 09:24:16 lo kernel: [89264.842017] DR3: DR6: 0ff0 DR7: 0400 Aug 22 09:24:16 lo kernel: [89264.842036] Process kvm (pid: 3579, threadinfo 81007add4000, task 810037878ee0) Aug 22 09:24:16 lo kernel: [89264.842063] Stack: 0001 0009 81007adf00c0 Aug 22 09:24:16 lo kernel: [89264.842099] 0007 000b 8100790b8030 a03681b0 Aug 22 09:24:16 lo kernel: [89264.842134] 0040 81007add5c54 0003fff2 Aug 22 09:24:16 lo kernel: [89264.842158] Call Trace: Aug 22 09:24:16 lo kernel: [89264.842193] [a03681b0] ? :kvm:paging64_page_fault+0x37f/0x3d1 Aug 22 09:24:16 lo kernel: [89264.842231] [a0366bc4] ? :kvm:kvm_mmu_page_fault+0x19/0x81 Aug 22 09:24:16 lo kernel: [89264.842259] [a03642be] ? :kvm:kvm_arch_vcpu_ioctl_run+0x463/0x615 Aug 22 09:24:16 lo kernel: [89264.842289]
Bug#496028: linux-2.6: host kernel oopses when adding a cpu to a kvm virtual machine
On Fri, 22 Aug 2008, Rik Theys wrote: Package: linux-2.6 Version: 2.6.26-3 Severity: normal I'm running a kvm virtual machine (with CentOS 5.2 as the guest OS, but it doesn't seem to matter) on a Lenny host (using the 2.6.26 kernel and kvm from unstable). When I try to hot-add a cpu to the virtual machine, the host kernel oopses and becomes unresponsive after a few minutes/seconds. I'm not sure this is a kernel or kvm bug. I've added the cpu to the virtual machine by logging in on the kvm monitor interface of the virtual machine and entering: cpu_set 1 online I'm not 100% sure this is the correct way to hot-add a CPU. I can't seem to find the documented way to do this using google. Also asked on #kvm but got no response. The kvm virtual machine was started with the following command: kvm -m 1024 -smp 1 -monitor tcp:localhost:,server,nowait -localtime -drive file=/test2-local.img,if=ide,boot=on -net nic,macaddr=00:16:3e:00:00:02,vlan=0, -net tap,script=/etc/kvm/kvm-ifup,vlan=0 -vnc 127.0.0.1:1 -daemonize I've noticed some kvm fixes in upstreams 2.6.26.3 release but am not sure they fix this issue. right please try out sid snapshots, see apt lines - http://wiki.debian.org/DebianKernel they carry 2.6.26.3 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#496028: linux-2.6: host kernel oopses when adding a cpu to a kvm virtual machine
Hi, Quoting maximilian attems [EMAIL PROTECTED]: On Fri, 22 Aug 2008, Rik Theys wrote: I've noticed some kvm fixes in upstreams 2.6.26.3 release but am not sure they fix this issue. right please try out sid snapshots, see apt lines - http://wiki.debian.org/DebianKernel they carry 2.6.26.3 I've tried the kernel targetted for sid and the kernel no longer oopses when I enter the command, but kvm segfaults now. So instead of oopsing the host kernel, the virtual machine now dies. I'm also not sure the command I enter is supposed to work as I can't find any documentation about it. I guess you can reassign this bug to kvm when the new kernel enters unstable/lenny. Regards, Rik Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]