kvm kernel backtrace 2.6.27.3 using kvm 78 userspace

kvm kernel backtrace 2.6.27.3 using kvm 78 userspace

[Kasper Sandberg]

Hello.

I just noticed that my dmesg had been flooded with backtraces, appearing
to be of KVM origin. I realize that my kernel is tainted, and slightly
outdated, however, i just wanted to drop the log here in case it might
be useful, but feel free to completely ignore it :)

i am not subscribed to the kvm list, so please cc me in case you need
further info..

my entire dmesg was flooded with backtraces looking practically
identical to this one, so i will only paste a few of them, however,
should you want more(though from a quick look they appear identical), i
can put online the complete log

kvm_handle_exit: unexpected, valid vectoring info and exit reason is 0x9
------------[ cut here ]------------
WARNING: at arch/x86/kvm/x86.c:185 kvm_queue_exception_e+0x63/0x70
[kvm]()
Modules linked in: tun ext3 jbd mbcache usb_storage cryptomgr cbc
aes_x86_64 aes_generic dm_crypt dm_mod crypto_blkcipher sha256_generic
crypto_algapi coretemp it87 hwmon_vid hwmon snd_seq kvm_intel kvm fuse
nvidia(P) snd_usb_audio snd_hda_intel snd_usb_lib snd_rawmidi
snd_seq_device snd_pcm snd_timer snd_hwdep psmouse thermal ehci_hcd rtc
snd ohci1394 ieee1394 soundcore i2c_core processor snd_page_alloc r8169
button
Pid: 20168, comm: kvm Tainted: P        W 2.6.27.3 #1

Call Trace:
 [<ffffffff80235c84>] warn_on_slowpath+0x64/0xb0
 [<ffffffffa0899c01>] paging64_gva_to_gpa+0x21/0x50 [kvm]
 [<ffffffffa088cc09>] gfn_to_hva+0x9/0x70 [kvm]
 [<ffffffffa088ccd1>] kvm_read_guest_page+0x61/0x70 [kvm]
 [<ffffffffa088cd28>] kvm_read_guest+0x48/0x80 [kvm]
 [<ffffffffa08931ae>] load_guest_segment_descriptor+0x6e/0xb0 [kvm]
 [<ffffffffa0893133>] kvm_queue_exception_e+0x63/0x70 [kvm]
 [<ffffffffa08941de>] kvm_task_switch+0x2ee/0x9f0 [kvm]
 [<ffffffffa088ed4c>] ioapic_service+0x1ec/0x340 [kvm]
 [<ffffffff804b2af7>] __down_read+0x17/0xa6
 [<ffffffffa0893566>] kvm_arch_vcpu_ioctl_run+0x266/0x6f0 [kvm]
 [<ffffffffa088c502>] kvm_vcpu_ioctl+0x2b2/0x480 [kvm]
 [<ffffffff8022ee76>] try_to_wake_up+0xf6/0x1b0
 [<ffffffff8022ba4a>] __wake_up_common+0x5a/0x90
 [<ffffffff8022bea3>] __wake_up+0x43/0x70
 [<ffffffff802556ff>] wake_futex+0x1f/0x30
 [<ffffffff80257961>] do_futex+0x881/0xa70
 [<ffffffffa08b2ca0>] handle_io+0x0/0x60 [kvm_intel]
 [<ffffffff802a0f46>] vfs_ioctl+0x36/0xb0
 [<ffffffff802a0f3f>] vfs_ioctl+0x2f/0xb0
 [<ffffffff802a124b>] do_vfs_ioctl+0x28b/0x2f0
 [<ffffffff802a1351>] sys_ioctl+0xa1/0xb0
 [<ffffffff8020bacb>] system_call_fastpath+0x16/0x1b

---[ end trace dc61957677c4cb77 ]---
kvm_handle_exit: unexpected, valid vectoring info and exit reason is 0x9
------------[ cut here ]------------
WARNING: at arch/x86/kvm/x86.c:185 kvm_queue_exception_e+0x63/0x70
[kvm]()
Modules linked in: tun ext3 jbd mbcache usb_storage cryptomgr cbc
aes_x86_64 aes_generic dm_crypt dm_mod crypto_blkcipher sha256_generic
crypto_algapi coretemp it87 hwmon_vid hwmon snd_seq kvm_intel kvm fuse
nvidia(P) snd_usb_audio snd_hda_intel snd_usb_lib snd_rawmidi
snd_seq_device snd_pcm snd_timer snd_hwdep psmouse thermal ehci_hcd rtc
snd ohci1394 ieee1394 soundcore i2c_core processor snd_page_alloc r8169
button
Pid: 20168, comm: kvm Tainted: P        W 2.6.27.3 #1

Call Trace:
 [<ffffffff80235c84>] warn_on_slowpath+0x64/0xb0
 [<ffffffffa0899c01>] paging64_gva_to_gpa+0x21/0x50 [kvm]
 [<ffffffffa088cc09>] gfn_to_hva+0x9/0x70 [kvm]
 [<ffffffffa088ccd1>] kvm_read_guest_page+0x61/0x70 [kvm]
 [<ffffffffa088cd28>] kvm_read_guest+0x48/0x80 [kvm]
 [<ffffffffa08931ae>] load_guest_segment_descriptor+0x6e/0xb0 [kvm]
 [<ffffffffa0893133>] kvm_queue_exception_e+0x63/0x70 [kvm]
 [<ffffffffa08941de>] kvm_task_switch+0x2ee/0x9f0 [kvm]
 [<ffffffffa088ed4c>] ioapic_service+0x1ec/0x340 [kvm]
 [<ffffffff804b2af7>] __down_read+0x17/0xa6
 [<ffffffffa0893566>] kvm_arch_vcpu_ioctl_run+0x266/0x6f0 [kvm]
 [<ffffffffa088c502>] kvm_vcpu_ioctl+0x2b2/0x480 [kvm]
 [<ffffffff8022ee76>] try_to_wake_up+0xf6/0x1b0
 [<ffffffff8022ba4a>] __wake_up_common+0x5a/0x90
 [<ffffffff8022bea3>] __wake_up+0x43/0x70
 [<ffffffff802556ff>] wake_futex+0x1f/0x30
 [<ffffffff80257961>] do_futex+0x881/0xa70
 [<ffffffffa08b2ca0>] handle_io+0x0/0x60 [kvm_intel]
 [<ffffffff802a0f46>] vfs_ioctl+0x36/0xb0
 [<ffffffff802a0f3f>] vfs_ioctl+0x2f/0xb0
 [<ffffffff802a124b>] do_vfs_ioctl+0x28b/0x2f0
 [<ffffffff802a1351>] sys_ioctl+0xa1/0xb0
 [<ffffffff8020bacb>] system_call_fastpath+0x16/0x1b

---[ end trace dc61957677c4cb77 ]---
kvm_handle_exit: unexpected, valid vectoring info and exit reason is 0x9
------------[ cut here ]------------
WARNING: at arch/x86/kvm/x86.c:185 kvm_queue_exception_e+0x63/0x70
[kvm]()
Modules linked in: tun ext3 jbd mbcache usb_storage cryptomgr cbc
aes_x86_64 aes_generic dm_crypt dm_mod crypto_blkcipher sha256_generic
crypto_algapi coretemp it87 hwmon_vid hwmon snd_seq kvm_intel kvm fuse
nvidia(P) snd_usb_audio snd_hda_intel snd_usb_lib snd_rawmidi
snd_seq_device snd_pcm snd_timer snd_hwdep psmouse thermal ehci_hcd rtc
snd ohci1394 ieee1394 soundcore i2c_core processor snd_page_alloc r8169
button
Pid: 20168, comm: kvm Tainted: P        W 2.6.27.3 #1

Call Trace:
 [<ffffffff80235c84>] warn_on_slowpath+0x64/0xb0
 [<ffffffffa0899c01>] paging64_gva_to_gpa+0x21/0x50 [kvm]
 [<ffffffffa088cc09>] gfn_to_hva+0x9/0x70 [kvm]
 [<ffffffffa088ccd1>] kvm_read_guest_page+0x61/0x70 [kvm]
 [<ffffffffa088cd28>] kvm_read_guest+0x48/0x80 [kvm]
 [<ffffffffa08931ae>] load_guest_segment_descriptor+0x6e/0xb0 [kvm]
 [<ffffffffa0893133>] kvm_queue_exception_e+0x63/0x70 [kvm]
 [<ffffffffa08941de>] kvm_task_switch+0x2ee/0x9f0 [kvm]
 [<ffffffffa088ed4c>] ioapic_service+0x1ec/0x340 [kvm]
 [<ffffffff804b2af7>] __down_read+0x17/0xa6
 [<ffffffffa0893566>] kvm_arch_vcpu_ioctl_run+0x266/0x6f0 [kvm]
 [<ffffffffa088c502>] kvm_vcpu_ioctl+0x2b2/0x480 [kvm]
 [<ffffffff8022ee76>] try_to_wake_up+0xf6/0x1b0
 [<ffffffff8022ba4a>] __wake_up_common+0x5a/0x90
 [<ffffffff8022bea3>] __wake_up+0x43/0x70
 [<ffffffff802556ff>] wake_futex+0x1f/0x30
 [<ffffffff80257961>] do_futex+0x881/0xa70
 [<ffffffffa08b2ca0>] handle_io+0x0/0x60 [kvm_intel]
 [<ffffffff802a0f46>] vfs_ioctl+0x36/0xb0
 [<ffffffff802a0f3f>] vfs_ioctl+0x2f/0xb0
 [<ffffffff802a124b>] do_vfs_ioctl+0x28b/0x2f0
 [<ffffffff802a1351>] sys_ioctl+0xa1/0xb0
 [<ffffffff8020bacb>] system_call_fastpath+0x16/0x1b

---[ end trace dc61957677c4cb77 ]---


mvh.
Kasper Sandberg

Re: kvm kernel backtrace 2.6.27.3 using kvm 78 userspace

[Avi Kivity]

Kasper Sandberg wrote:
> Hello.
>
> I just noticed that my dmesg had been flooded with backtraces, appearing
> to be of KVM origin. I realize that my kernel is tainted, and slightly
> outdated, however, i just wanted to drop the log here in case it might
> be useful, but feel free to completely ignore it :)
>
> i am not subscribed to the kvm list, so please cc me in case you need
> further info..
>
> my entire dmesg was flooded with backtraces looking practically
> identical to this one, so i will only paste a few of them, however,
> should you want more(though from a quick look they appear identical), i
> can put online the complete log
>   

What guest are you running?

-- 
I have a truly marvellous patch that fixes the bug which this
signature is too narrow to contain.

Re: kvm kernel backtrace 2.6.27.3 using kvm 78 userspace

[Kasper Sandberg]

On Sun, 2008-11-23 at 10:49 +0200, Avi Kivity wrote:
> Kasper Sandberg wrote:
> > Hello.
> >
> > I just noticed that my dmesg had been flooded with backtraces, appearing
> > to be of KVM origin. I realize that my kernel is tainted, and slightly
> > outdated, however, i just wanted to drop the log here in case it might
> > be useful, but feel free to completely ignore it :)
> >
> > i am not subscribed to the kvm list, so please cc me in case you need
> > further info..
> >
> > my entire dmesg was flooded with backtraces looking practically
> > identical to this one, so i will only paste a few of them, however,
> > should you want more(though from a quick look they appear identical), i
> > can put online the complete log
> >   
> 
> What guest are you running?

Yeah sorry, i was gonna include that in the mail, i apparently forgot. 

I have been running debian, both 32 and 64bit inside it, .18, .24
and .26 kernels, and ubuntu 8.10 .27 kernels, ATLEAST 32bit, i do not
remember if i have run 64bit ubuntu 8.10

I can however give some information that _MIGHT_ (i have no clue myself)
be relevant

on one of the debian guests i had xrandr crash the X server while trying
to change resolution, i did this a few times. Using -vga std.

also, on the ubuntu 32bit guest, i ran qemu with debian stable, where X
crashed during start inside the qemu on the ubuntu guest. This happened
only once.

im quite certain my ram and that is fine, no errors when doing memory
testing etc.


but as said, the kernel is tainted, i dont know if that affects the
results or not, i just wanted to inform in case this can help someone.
That being said, the box has NOT been rebooted since, and i can confirm
that kvm works excellently, i have spawned several VMs and used a tad,
since i wrote the mail.

> 

Re: kvm kernel backtrace 2.6.27.3 using kvm 78 userspace

[Avi Kivity]

Kasper Sandberg wrote:
>> What guest are you running?
>>     
>
> Yeah sorry, i was gonna include that in the mail, i apparently forgot. 
>
> I have been running debian, both 32 and 64bit inside it, .18, .24
> and .26 kernels, and ubuntu 8.10 .27 kernels, ATLEAST 32bit, i do not
> remember if i have run 64bit ubuntu 8.10
>
> I can however give some information that _MIGHT_ (i have no clue myself)
> be relevant
>
> on one of the debian guests i had xrandr crash the X server while trying
> to change resolution, i did this a few times. Using -vga std.
>
> also, on the ubuntu 32bit guest, i ran qemu with debian stable, where X
> crashed during start inside the qemu on the ubuntu guest. This happened
> only once.
>
> im quite certain my ram and that is fine, no errors when doing memory
> testing etc.
>
>
> but as said, the kernel is tainted, i dont know if that affects the
> results or not, i just wanted to inform in case this can help someone.
> That being said, the box has NOT been rebooted since, and i can confirm
> that kvm works excellently, i have spawned several VMs and used a tad,
> since i wrote the mail

Well, if you can find a recipe that reliably reproduces the problems, we 
want to hear of it.

-- 
error compiling committee.c: too many arguments to function

Re: kvm kernel backtrace 2.6.27.3 using kvm 78 userspace

[Kasper Sandberg]

On Sun, 2008-11-23 at 16:27 +0200, Avi Kivity wrote:
> Kasper Sandberg wrote:
> >> What guest are you running?
> >>     
> >
> > Yeah sorry, i was gonna include that in the mail, i apparently forgot. 
> >
> > I have been running debian, both 32 and 64bit inside it, .18, .24
> > and .26 kernels, and ubuntu 8.10 .27 kernels, ATLEAST 32bit, i do not
> > remember if i have run 64bit ubuntu 8.10
> >
> > I can however give some information that _MIGHT_ (i have no clue myself)
> > be relevant
> >
> > on one of the debian guests i had xrandr crash the X server while trying
> > to change resolution, i did this a few times. Using -vga std.
> >
> > also, on the ubuntu 32bit guest, i ran qemu with debian stable, where X
> > crashed during start inside the qemu on the ubuntu guest. This happened
> > only once.
> >
> > im quite certain my ram and that is fine, no errors when doing memory
> > testing etc.
> >
> >
> > but as said, the kernel is tainted, i dont know if that affects the
> > results or not, i just wanted to inform in case this can help someone.
> > That being said, the box has NOT been rebooted since, and i can confirm
> > that kvm works excellently, i have spawned several VMs and used a tad,
> > since i wrote the mail
> 
> Well, if you can find a recipe that reliably reproduces the problems, we 
> want to hear of it.
Within the coming week i will try to find the time to update to latest
kernel/kvm, and then do all the things i remember doing in the VMs, to
try and reproduce

>