Discussion:
OpenBSD VMM VMs Crash
(too old to reply)
Aaron Marcher
2018-04-06 16:55:05 UTC
Permalink
Raw Message
Ohai,

for me OpenBSD VMM VMs crash after some (undefined) time while logging
the following on the host:
vcpu_run_loop: vm 3 / vcpu 0 run ioctl failed: Invalid argument
Apart from that VMM works es expected.

Regards,
Aaron

--
Web: https://drkhsh.at/ or http://drkhsh5rv6pnahas.onion/
Gopher: gopher://drkhsh.at or gopher://drkhsh5rv6pnahas.onion
GPG: 0x7A65E38D55BE96FE
Fingerprint: 4688 907C 8720 3318 0D9F AFDE 7A65 E38D 55BE 96FE
Mike Larkin
2018-04-06 19:29:18 UTC
Permalink
Raw Message
On Fri, Apr 06, 2018 at 06:55:05PM +0200, Aaron Marcher wrote:
> Ohai,
>
> for me OpenBSD VMM VMs crash after some (undefined) time while logging the
> following on the host:
> vcpu_run_loop: vm 3 / vcpu 0 run ioctl failed: Invalid argument
> Apart from that VMM works es expected.
>
> Regards,
> Aaron
>
> --
> Web: https://drkhsh.at/ or http://drkhsh5rv6pnahas.onion/
> Gopher: gopher://drkhsh.at or gopher://drkhsh5rv6pnahas.onion
> GPG: 0x7A65E38D55BE96FE
> Fingerprint: 4688 907C 8720 3318 0D9F AFDE 7A65 E38D 55BE 96FE
>

could be anything. enable VMM_DEBUG and send me what you see in dmesg
after the crash.
Carlos Cardenas
2018-04-06 19:32:48 UTC
Permalink
Raw Message
On Fri, Apr 06, 2018 at 06:55:05PM +0200, Aaron Marcher wrote:
> Ohai,
>
> for me OpenBSD VMM VMs crash after some (undefined) time while logging the
> following on the host:
> vcpu_run_loop: vm 3 / vcpu 0 run ioctl failed: Invalid argument
> Apart from that VMM works es expected.
>
> Regards,
> Aaron
>
> --
> Web: https://drkhsh.at/ or http://drkhsh5rv6pnahas.onion/
> Gopher: gopher://drkhsh.at or gopher://drkhsh5rv6pnahas.onion
> GPG: 0x7A65E38D55BE96FE
> Fingerprint: 4688 907C 8720 3318 0D9F AFDE 7A65 E38D 55BE 96FE
>

Aaron,

This is not a very good bug report.

Please provide the following information:
* dmesg of host
* vm.conf
* Guest information (dmesg if OpenBSD)
* What is the guest doing when it dies?
* What is the host doing when the guest dies?

+--+
Carlos
Aaron Marcher
2018-04-07 09:46:50 UTC
Permalink
Raw Message
So i looked up logs and stuff and came to the point that my issues are
exactly this ones:
http://openbsd-archive.7691.n7.nabble.com/vmd-VMs-are-stopping-with-vcpu-0-run-ioctl-failed-Invalid-argument-td333259.html

It seems the issue was not further discussed and/or fixed?

Regards,
Aaron

--
Web: https://drkhsh.at/ or http://drkhsh5rv6pnahas.onion/
Gopher: gopher://drkhsh.at or gopher://drkhsh5rv6pnahas.onion
GPG: 0x7A65E38D55BE96FE
Fingerprint: 4688 907C 8720 3318 0D9F AFDE 7A65 E38D 55BE 96FE
Mike Larkin
2018-04-07 18:46:42 UTC
Permalink
Raw Message
On Sat, Apr 07, 2018 at 11:46:50AM +0200, Aaron Marcher wrote:
> So i looked up logs and stuff and came to the point that my issues are
> exactly this ones:
> http://openbsd-archive.7691.n7.nabble.com/vmd-VMs-are-stopping-with-vcpu-0-run-ioctl-failed-Invalid-argument-td333259.html
>
> It seems the issue was not further discussed and/or fixed?
>
> Regards,
> Aaron
>
> --
> Web: https://drkhsh.at/ or http://drkhsh5rv6pnahas.onion/
> Gopher: gopher://drkhsh.at or gopher://drkhsh5rv6pnahas.onion
> GPG: 0x7A65E38D55BE96FE
> Fingerprint: 4688 907C 8720 3318 0D9F AFDE 7A65 E38D 55BE 96FE
>

You are jumping to conclusions.

All we know is that your VMs failed to run at some point. Whether or not it
is "exactly the same" as that other post is unknown, since you didn't run with
VMM_DEBUG and give us any more info. Or even submit a proper bug report.

For what it's worth, I'm still working on that other issue.

-ml
Olivier Cherrier
2018-04-10 14:31:15 UTC
Permalink
Raw Message
On Sat, Apr 07, 2018 at 11:46:42AM -0700, ***@azathoth.net wrote:
> > So i looked up logs and stuff and came to the point that my issues are
> > exactly this ones:
> > http://openbsd-archive.7691.n7.nabble.com/vmd-VMs-are-stopping-with-vcpu-0-run-ioctl-failed-Invalid-argument-td333259.html
> >
> > It seems the issue was not further discussed and/or fixed?
> >
> > Regards,
> > Aaron
> >
> > --
> > Web: https://drkhsh.at/ or http://drkhsh5rv6pnahas.onion/
> > Gopher: gopher://drkhsh.at or gopher://drkhsh5rv6pnahas.onion
> > GPG: 0x7A65E38D55BE96FE
> > Fingerprint: 4688 907C 8720 3318 0D9F AFDE 7A65 E38D 55BE 96FE
> >
>
> You are jumping to conclusions.
>
> All we know is that your VMs failed to run at some point. Whether or not it
> is "exactly the same" as that other post is unknown, since you didn't run with
> VMM_DEBUG and give us any more info. Or even submit a proper bug report.
>
> For what it's worth, I'm still working on that other issue.

For me, the issue never went away.
I would say it is even more frequent now than before (end of 2017).


--
Olivier Cherrier
Phone: +352691754777
mailto:***@symacx.com
Mike Larkin
2018-04-10 17:31:07 UTC
Permalink
Raw Message
On Tue, Apr 10, 2018 at 02:31:15PM +0000, Olivier Cherrier wrote:
> On Sat, Apr 07, 2018 at 11:46:42AM -0700, ***@azathoth.net wrote:
> > > So i looked up logs and stuff and came to the point that my issues are
> > > exactly this ones:
> > > http://openbsd-archive.7691.n7.nabble.com/vmd-VMs-are-stopping-with-vcpu-0-run-ioctl-failed-Invalid-argument-td333259.html
> > >
> > > It seems the issue was not further discussed and/or fixed?
> > >
> > > Regards,
> > > Aaron
> > >
> > > --
> > > Web: https://drkhsh.at/ or http://drkhsh5rv6pnahas.onion/
> > > Gopher: gopher://drkhsh.at or gopher://drkhsh5rv6pnahas.onion
> > > GPG: 0x7A65E38D55BE96FE
> > > Fingerprint: 4688 907C 8720 3318 0D9F AFDE 7A65 E38D 55BE 96FE
> > >
> >
> > You are jumping to conclusions.
> >
> > All we know is that your VMs failed to run at some point. Whether or not it
> > is "exactly the same" as that other post is unknown, since you didn't run with
> > VMM_DEBUG and give us any more info. Or even submit a proper bug report.
> >
> > For what it's worth, I'm still working on that other issue.
>
> For me, the issue never went away.
> I would say it is even more frequent now than before (end of 2017).
>
>
> --
> Olivier Cherrier
> Phone: +352691754777
> mailto:***@symacx.com

Yep, I dusted off that tree and spent a few hours on it last night.

Hopefully have a working diff soon; the EPT misconfigurations are gone but I'm
running into a uvm page list accounting error now that results in panics when
the VMs are torn down. I know why it's happening but the alternative approach
is a bit more messy and I'd hoped to stay away from that. But maybe I'll have
no choice.

-ml
Mike Larkin
2018-04-12 20:50:07 UTC
Permalink
Raw Message
On Tue, Apr 10, 2018 at 02:31:15PM +0000, Olivier Cherrier wrote:
> On Sat, Apr 07, 2018 at 11:46:42AM -0700, ***@azathoth.net wrote:
> > > So i looked up logs and stuff and came to the point that my issues are
> > > exactly this ones:
> > > http://openbsd-archive.7691.n7.nabble.com/vmd-VMs-are-stopping-with-vcpu-0-run-ioctl-failed-Invalid-argument-td333259.html
> > >
> > > It seems the issue was not further discussed and/or fixed?
> > >
> > > Regards,
> > > Aaron
> > >
> > > --
> > > Web: https://drkhsh.at/ or http://drkhsh5rv6pnahas.onion/
> > > Gopher: gopher://drkhsh.at or gopher://drkhsh5rv6pnahas.onion
> > > GPG: 0x7A65E38D55BE96FE
> > > Fingerprint: 4688 907C 8720 3318 0D9F AFDE 7A65 E38D 55BE 96FE
> > >
> >
> > You are jumping to conclusions.
> >
> > All we know is that your VMs failed to run at some point. Whether or not it
> > is "exactly the same" as that other post is unknown, since you didn't run with
> > VMM_DEBUG and give us any more info. Or even submit a proper bug report.
> >
> > For what it's worth, I'm still working on that other issue.
>
> For me, the issue never went away.
> I would say it is even more frequent now than before (end of 2017).
>
>
> --
> Olivier Cherrier
> Phone: +352691754777
> mailto:***@symacx.com
>

I have a diff locally that I am testing now that might fix this. Stay tuned.

-ml
Loading...