Kernel Oops on module (/etc/init.d/vboxdrv restart) => Fixed in SVN

[Michele Petrazzo]

Hi all. Just received an "NS_ERROR_FAILURE" error, so I restart vboxdrv and see the bug following. Before the problem I had 1 win2kserver, debian lenny and xp pro leaving on Vbox. All with 2 networks: one host (eth0) and one internal (tap0). win2kserver and debian was with amd-v enabled, but xp no

michele:~$ uname -a Linux michele 2.6.27-1-686 #1 SMP Tue Dec 16 02:23:23 UTC 2008 i686 GNU/Linux michele:~$ cat /proc/cpuinfo processor : 0 vendor_id : AuthenticAMD cpu family : 15 model : 79 model name : AMD Athlon(tm) 64 Processor 3800+ stepping : 2 cpu MHz : 1000.000 cache size : 512 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext fxsr_opt rdtscp lm 3dnowext 3dnow up pni cx16 lahf_lm svm extapic cr8_legacy bogomips : 1999.95 clflush size : 64 power management: ts fid vid ttp tm stc

michele:~$

The bug:

[504976.500614] BUG: unable to handle kernel paging request at 09d477e4 [504976.500620] IP: [<f8fedaf7>] [504976.500624] *pde = 00000000 [504976.500628] Oops: 0000 #1 SMP [504976.500631] Modules linked in: nls_utf8 cifs nls_base nvidia(P) agpgart ppdev lp rfcomm l2cap bluetooth vboxnetflt vboxdrv video output ac battery btrfs powernow_k8 cpufreq_userspace cpufreq_stats cpufreq_powersave cpufreq_ondemand freq_table cpufreq_conservative kvm_amd kvm dm_snapshot dm_mirror dm_log dm_mod ipv6 fuse zlib_inflate zlib_deflate libcrc32c firewire_sbp2 loop rtc_cmos rtc_core rtc_lib snd_pcsp k8temp snd_hda_intel snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd soundcore snd_page_alloc parport_pc parport button i2c_nforce2 i2c_core evdev joydev ext3 jbd mbcache raid1 md_mod sg sr_mod cdrom sd_mod crc_t10dif ide_disk usb_storage usbhid sata_nv hid ff_memless ata_generic libata scsi_mod dock floppy forcedeth firewire_ohci firewire_core crc_itu_t ide_pci_generic amd74xx ide_core ehci_hcd ohci_hcd usbcore thermal processor fan thermal_sys [last unloaded: vmnet] [504976.500677] [504976.500680] Pid: 22838, comm: VirtualBox Tainted: P (2.6.27-1-686 #1) [504976.500683] EIP: 0060:[<f8fedaf7>] EFLAGS: 00210212 CPU: 0 [504976.500686] EIP is at 0xf8fedaf7 [504976.500688] EAX: 09d487cc EBX: 09d477cc ECX: c0285609 EDX: 00000000 [504976.500690] ESI: 00000005 EDI: e127a290 EBP: dc20bed0 ESP: dc20be88 [504976.500692] DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068 [504976.500695] Process VirtualBox (pid: 22838, ti=dc20a000 task=c9e90620 task.ti=dc20a000) [504976.500696] Stack: 00000044 e50f5524 dc20bee4 e50f54ac 00000001 c015a8ab 00000000 00000000 [504976.500701] f5da73c0 e50f5558 e50f54ac dc20bee4 f5da73c0 dc20bfa0 f89cefb9 00007c5a [504976.500706] e127a290 f5e18010 00000028 f8cb0545 09d477cc 00000005 00000000 00000000 [504976.500711] Call Trace: [504976.500713] [<c015a8ab>] generic_file_aio_write+0x5b/0xa9 [504976.500721] [<f89cefb9>] ext3_file_write+0x19/0x83 [ext3] [504976.500732] [<f8cb0545>] supdrvIOCtl+0xc3e/0x117c [vboxdrv] [504976.500742] [<f8cad1e8>] VBoxDrvLinuxIOCtl+0xfa/0x16b [vboxdrv] [504976.500749] [<c0285609>] tcp_sendpage+0x39a/0x513 [504976.500756] [<c0285609>] tcp_sendpage+0x39a/0x513 [504976.500759] [<f8cad0ee>] VBoxDrvLinuxIOCtl+0x0/0x16b [vboxdrv] [504976.500766] [<c0285609>] tcp_sendpage+0x39a/0x513 [504976.500769] [<c01825b0>] vfs_ioctl+0x1c/0x5d [504976.500773] [<c0182830>] do_vfs_ioctl+0x23f/0x24d [504976.500776] [<c018287f>] sys_ioctl+0x41/0x5a [504976.500779] [<c01037f7>] sysenter_do_call+0x12/0x2f [504976.500782] [<c0285609>] tcp_sendpage+0x39a/0x513 [504976.500786] ======================= [504976.500787] Code: ff 8d 76 00 55 89 e5 56 53 83 ec 40 8b 45 14 8b 5d 08 8b 55 18 8b 75 0c 89 45 d0 8d 83 00 10 00 00 3d ff 1f 00 00 89 55 d4 76 0c <8b> 43 18 85 c0 74 05 83 fe 18 76 2d 8b 45 1c 89 44 24 08 8b 55 [504976.500809] EIP: [<f8fedaf7>] 0xf8fedaf7 SS:ESP 0068:dc20be88 [504976.500815] ---[ end trace 1e59cfc9ec07ba17 ]---

[Michele Petrazzo]

win2kserver log

[Frank Mehnert]

You should never do that as long as at least one VBox instance is active. I assume your kernel has CONFIG_MODULE_FORCE_LOAD set so thats why it is possible to remove a kernel module even if it is still used. Which should not be allowed. We will try to find out how to prevent this.

[Frank Mehnert]

Oh, I actually saw that this is indeed a bug in our code.

[Michele Petrazzo]

You should never do that as long as at least one VBox instance is active.

All the vbox instances were done before my restart.

P.s. Thanks for the quickly response!

[Frank Mehnert]

For some reason there was still a VBox instance running, see your log (please attach such logs next time, don't cut'n'paste!): Pid: 22838, comm: VirtualBox

[Michele Petrazzo]

lease attach such logs next time, don't cut'n'paste!

I attach the win2k_server log. Wasn't it? Which other one do you need?

[Frank Mehnert]

No, just look at the description of this ticket. You did cut'n'paste of the host kernel log. This log is fine and it is required, but please attach such log files as separate file. Anyway, I think we fixed that problem.

[Michele Petrazzo]

Ops. Sorry! :) Next time I'll find a a bug we sure that I'll attach the log!

Thanks, Michele