Re: 9Front on VMM on Ryzen Hardware

2020-12-15 Thread Aaron Bieber


e...@disroot.org writes:

> Hello, I hope that this is the right mailing list to send this query to.
>
> First some background. It is possible to run 9front on OpenBSD using
> vmm, this is well documented and I have gotten it working before on a
> ThinkPad X220.
> Where I run into trouble is trying to install it on a T14 AMD, which
> uses an AMD processor. Essentially when you begin to run the live cd,
> the 9front kernel loads, and then immediately vmd restarts the virtual
> machine, presumably because it crashed somewhere in the boot process.
>
> Now to the question, how would I go about debugging this? I know that
> this install works on Intel, this is on OpenBSD -current.
>
> The 9front IRC told me that it was a vmm issue, as there are different
> implementations on AMD and Intel, is this true?
> If so, what debugging should I run to help the OpenBSD developers fix
> this issue?
>
> If it's a 9front issue, is there any way for me to be able to take some
> kind of memory dump so that the 9front developers can handle this?
>
> Hopefully this wasn't too off topic, I have read the relevant manual
> pages for vmm, but I couldn't work out what debugger to use, I'm not
> here to get others to debug it for me, only to work out where to start.
>
> Thank you

I have run into this as well.. There was a change in 9front some time
before the release of the amd64 ISOs that seems to have caused it.

I was able to boot the 386 ISO (9front-7408.1d345066125a.386.iso) and a
amd64 kernel built from the source contained within that ISO. There was
about a full year of development between that ISO and when I started
seeing the issue, so it's not a very useful data point :P.

cinap on #cat-v had some pointers for troubleshooting:

2020-05-29 07:19:47 cinap_lenrekso go to /sys/src/boot/pc
2020-05-29 07:19:55 cinap_lenrekin the mkfile, theres a test.iso target 
or something
2020-05-29 07:20:02 cinap_lenrekyou can adjust that
2020-05-29 07:20:07 qbitok
2020-05-29 07:20:24 cinap_lenrekbasically, you want a workflow where 
you just run a command to generate a new iso with the kernel
2020-05-29 07:20:28 cinap_lenreknothing else
2020-05-29 07:20:35 cinap_lenrekthats good enougth to troubleshoot this
2020-05-29 07:20:41 cinap_lenrekand then boot it from vmd

Sorry it isn't much help!

Cheers,
Aaron



Re: 9Front on VMM on Ryzen Hardware

2020-12-15 Thread Mike Larkin
On Tue, Dec 15, 2020 at 12:10:42PM +, e...@disroot.org wrote:
> Hello, I hope that this is the right mailing list to send this query to.
>
> First some background. It is possible to run 9front on OpenBSD using
> vmm, this is well documented and I have gotten it working before on a
> ThinkPad X220.
> Where I run into trouble is trying to install it on a T14 AMD, which
> uses an AMD processor. Essentially when you begin to run the live cd,
> the 9front kernel loads, and then immediately vmd restarts the virtual
> machine, presumably because it crashed somewhere in the boot process.
>
> Now to the question, how would I go about debugging this? I know that
> this install works on Intel, this is on OpenBSD -current.
>
> The 9front IRC told me that it was a vmm issue, as there are different
> implementations on AMD and Intel, is this true?
> If so, what debugging should I run to help the OpenBSD developers fix
> this issue?
>
> If it's a 9front issue, is there any way for me to be able to take some
> kind of memory dump so that the 9front developers can handle this?
>
> Hopefully this wasn't too off topic, I have read the relevant manual
> pages for vmm, but I couldn't work out what debugger to use, I'm not
> here to get others to debug it for me, only to work out where to start.
>
> Thank you
>

If you know how, build a kernel with VMM_DEBUG enabled and send me the
output from dmesg (it will be a lot) when the VM crashes.

VMM_DEBUG is settable at the top of vmm.c, fwiw.

-ml



9Front on VMM on Ryzen Hardware

2020-12-15 Thread edl
Hello, I hope that this is the right mailing list to send this query to.

First some background. It is possible to run 9front on OpenBSD using
vmm, this is well documented and I have gotten it working before on a
ThinkPad X220.
Where I run into trouble is trying to install it on a T14 AMD, which
uses an AMD processor. Essentially when you begin to run the live cd,
the 9front kernel loads, and then immediately vmd restarts the virtual
machine, presumably because it crashed somewhere in the boot process.

Now to the question, how would I go about debugging this? I know that
this install works on Intel, this is on OpenBSD -current.

The 9front IRC told me that it was a vmm issue, as there are different
implementations on AMD and Intel, is this true?
If so, what debugging should I run to help the OpenBSD developers fix
this issue?

If it's a 9front issue, is there any way for me to be able to take some
kind of memory dump so that the 9front developers can handle this?

Hopefully this wasn't too off topic, I have read the relevant manual
pages for vmm, but I couldn't work out what debugger to use, I'm not
here to get others to debug it for me, only to work out where to start.

Thank you



Re: 9Front on VMM on Ryzen Hardware

2020-12-15 Thread Bodie




On 15.12.2020 13:10, e...@disroot.org wrote:
Hello, I hope that this is the right mailing list to send this query 
to.


First some background. It is possible to run 9front on OpenBSD using
vmm, this is well documented and I have gotten it working before on a
ThinkPad X220.
Where I run into trouble is trying to install it on a T14 AMD, which
uses an AMD processor. Essentially when you begin to run the live cd,
the 9front kernel loads, and then immediately vmd restarts the virtual
machine, presumably because it crashed somewhere in the boot process.

Now to the question, how would I go about debugging this? I know that
this install works on Intel, this is on OpenBSD -current.

The 9front IRC told me that it was a vmm issue, as there are different
implementations on AMD and Intel, is this true?
If so, what debugging should I run to help the OpenBSD developers fix
this issue?


Described here https://www.openbsd.org/report.html

dmesg(8) is a good start and you may play with 
https://man.openbsd.org/boot_config




If it's a 9front issue, is there any way for me to be able to take some
kind of memory dump so that the 9front developers can handle this?

Hopefully this wasn't too off topic, I have read the relevant manual
pages for vmm, but I couldn't work out what debugger to use, I'm not
here to get others to debug it for me, only to work out where to start.

Thank you