Kernel bug while VM shuts down - SuSE Linux 10.0 AMD64 (Alsa)

[zcermak]

Hello,

I have a problem with VirtualBox running on the SuSE Linux 10.0 AMD64 host system. I am running the Windows XP as a guest OS. There are following messages loged in the syslog while the vboxdrv is loading:

Aug 31 19:56:56 home-cermak kernel: vboxdrv: module not supported by Novell, setting U taint flag.
Aug 31 19:56:56 home-cermak kernel: vboxdrv: Trying to deactivate the NMI watchdog permanently...

When I shut down the VM there are following messages in the syslog:

Aug 31 20:22:17 home-cermak kernel: ----------- [cut here ] --------- [please bite here ] ---------
Aug 31 20:22:17 home-cermak kernel: Kernel BUG at "mm/rmap.c":493
Aug 31 20:22:17 home-cermak kernel: invalid operand: 0000 [1]
Aug 31 20:22:17 home-cermak kernel: CPU 0
Aug 31 20:22:17 home-cermak kernel: Modules linked in: vboxdrv ipt_pkttype ipt_LOG ipt_limit hfsplus subfs sg cpufreq_ondemand cpufreq_userspace cpufreq_powersave powernow_k8 freq_table edd snd_pcm_oss snd_mixer_oss snd_seq button battery ac rfcomm hidp l2cap hci_usb bluetooth ip6t_REJECT usb_storage ipt_REJECT ipt_state floppy nvidia iptable_mangle ehci_hcd snd_via82xx gameport snd_ac97_codec snd_ac97_bus iptable_nat via_rhine mii snd_pcm snd_timer snd_page_alloc snd_mpu401_uart snd_rawmidi iptable_filter snd_seq_device i2c_viapro i2c_core snd soundcore uhci_hcd usbcore generic ip6table_mangle shpchp pci_hotplug ip_conntrack ip_tables ip6table_filter ip6_tables ipv6 parport_pc lp parport nls_iso8859_1 nls_cp437 vfat fat nls_utf8 ntfs dm_mod reiserfs ide_cd cdrom fan thermal processor via82cxxx sata_via libata sd_mod scsi_mod ide_disk ide_core
Aug 31 20:22:17 home-cermak kernel: Pid: 6997, comm: VirtualBox Tainted: P U 2.6.13-15.16-default
Aug 31 20:22:17 home-cermak kernel: RIP: 0010:[<ffffffff8016e409>] <ffffffff8016e409>{page_remove_rmap+41}
Aug 31 20:22:17 home-cermak kernel: RSP: 0018:ffff810017dabde0 EFLAGS: 00010286
Aug 31 20:22:17 home-cermak kernel: RAX: 00000000ffffffff RBX: 0000000000000020 RCX: ffffffff80421f80
Aug 31 20:22:17 home-cermak kernel: RDX: 0000000000000000 RSI: 0000000000000001 RDI: ffff81000101f9c0
Aug 31 20:22:17 home-cermak kernel: RBP: ffff81000101f9c0 R08: 0000000000000212 R09: ffff810017dabc30
Aug 31 20:22:17 home-cermak kernel: R10: 0000000000000001 R11: ffff810017dabd68 R12: ffff8100115af610
Aug 31 20:22:17 home-cermak kernel: R13: 00002aaaaaae2000 R14: 00002aaaaaac2000 R15: 00002aaaaaae2000
Aug 31 20:22:17 home-cermak kernel: FS: 0000000040626960(0000) GS:ffffffff8049c800(0000) knlGS:00000000567e2bb0[[BR]] Aug 31 20:22:17 home-cermak kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
Aug 31 20:22:17 home-cermak kernel: CR2: 00002aaaac2dc02c CR3: 0000000007e72000 CR4: 00000000000006e0
Aug 31 20:22:17 home-cermak kernel: Process VirtualBox (pid: 6997, threadinfo ffff810017daa000, task ffff810017da9830)
Aug 31 20:22:17 home-cermak kernel: Stack: ffffffff80166fb1 ffff810017dabed8 ffffffffffffffff 0000000000000000
Aug 31 20:22:17 home-cermak kernel: ffff810017ddeae8 ffff810007e28100 ffff810017dabee0 000000000007c000
Aug 31 20:22:17 home-cermak kernel: 00002aaaaaae2000 0000000000000000
Aug 31 20:22:17 home-cermak kernel: Call Trace:<ffffffff80166fb1>{unmap_vmas+1137} <ffffffff8016a8f4>{exit_mmap+116}
Aug 31 20:22:17 home-cermak kernel: <ffffffff801317eb>{mmput+27} <ffffffff80135a6d>{do_exit+461}
Aug 31 20:22:17 home-cermak kernel: <ffffffff801364a1>{do_group_exit+177} <ffffffff8010e95a>{system_call+126}
Aug 31 20:22:17 home-cermak kernel:
Aug 31 20:22:17 home-cermak kernel:
Aug 31 20:22:17 home-cermak kernel: Code: 0f 0b a3 88 b8 35 80 ff ff ff ff c2 ed 01 48 c7 c6 ff ff ff
Aug 31 20:22:17 home-cermak kernel: RIP <ffffffff8016e409>{page_remove_rmap+41} RSP <ffff810017dabde0>
Aug 31 20:22:17 home-cermak kernel: <1>Fixing recursive fault but reboot is needed!

I am using the VirtualBox_1.4.0_Linux_amd64.run. The same problem occured with the new VirtualBox_1.5.0_Linux_amd64.run. As a consequence of the problem reported by the above syslog messages I am not able to shut down the host Linux. It freezes while the "swapoff -a" command is executed. The VM and the guest WinXP runs OK. The only problem occurs when the WM shuts down. The same problem occurs even when I start VM with "empty" disk (without installed guest OS). I have fully patched SuSE Linux Eterprise Desktop 10.0 running on the "AMD Athlon 64 Processor 3000+" with 512MB RAM.

Zdenek Cermak

[zcermak]

I have found solution for me. I had enabled Audio for the VM and used ALSA driver. When I changed the Audio driver to the OSS sound driver or when I disabled Audio for VM the bug didn't occur (no kernel-fail messages in syslog, host Linux turns off OK). When I used the OSS driver I had problem with Skype running in host Linux: it had problem with sound device. Therefore I disabled Audio for VM and I am happy :-).

It works in the 1.5.0 even in the 1.4.0.

Zdenek Cermak

[Frank Mehnert]

Seems to be somehow related to #521

[Frank Mehnert]

Please reopen if necessary.