Strange 3.2 Multi-monitor woes? PCI device does not exist?

[Stumpy]

I recently added another gfx card which seemed to go smothly enough
except it apparently didn't. When I tried to plug in monitors to the
integrated video ports AND the graphics card things started going
sideways (and I can't figure out why they would?). One example is most
times I boot no icons show up, not in the xfce menu nor context menu nor
on the qubes manager, just the text. When I tried to rotate the monitors
(via xfce) they acted funny the wallpaper displaying strangly but the
qubes manager window showing up ok (in the same monitor at the same
time). Also sometimes (some reboots) I am able to access the "display"
option to try to fiddle with getting the monitors right, other times
when try clicking display its like the display window shows up then
instantly closes (its so fast I can't actually tell if its the display
settings window or not). Lastly, so far, when I try to start up an appvm
I get the following:
Error starting VM <sys-net> PCI device 02:00.0 does not exist (domain
<sys-net>)
though that happens with all the appvms and the debian templates
and for redhat and whonix I get:
qubes manager
error starting vm 'fedora-26' maximum recursion depth exceeded

I am really really stumped, need help!

PS for what its worth, the gfx card is in a pci-16 slot, the integrated
gfx ports are hdmi and dvi, the gfx card ports are DP and DVI.

[Stumpy]

hmmm, is there any other information I should be providing to help
diagnose this issue?

[awokd]

On Wed, April 18, 2018 11:34 pm, Stumpy wrote:
> On 2018-04-18 04:18, Stumpy wrote:
>
>> I recently added another gfx card which seemed to go smothly enough
>> except it apparently didn't. When I tried to plug in monitors to the
>> integrated video ports AND the graphics card things started going
>> sideways (and I can't figure out why they would?). One example is most
>> times I boot no icons show up, not in the xfce menu nor context menu nor
>> on the qubes manager, just the text. When I tried to rotate the monitors
>> (via xfce) they acted funny the wallpaper displaying strangly
>> but the qubes manager window showing up ok (in the same monitor at the
>> same time).

Xen might not support what you're trying to do. Can you replace the
primary graphics card with one with more ports that can drive the amount
of monitors you want?

>> Lastly, so far, when I try to start up
>> an appvm I get the following: Error starting VM <sys-net> PCI device
>> 02:00.0 does not exist (domain
>> <sys-net>)
>> though that happens with all the appvms and the debian templates and for
>> redhat and whonix I get: qubes manager error starting vm 'fedora-26'
>> maximum recursion depth exceeded

Go into Qube Settings on sys-net, Devices tab. Unmap anything listed there
and add your NIC back in. You might have to repeat this if you remove the
secondary graphics card.

[Stumpy]

On 2018-04-19 13:49, awokd wrote:
> On Wed, April 18, 2018 11:34 pm, Stumpy wrote:
>> On 2018-04-18 04:18, Stumpy wrote:
>>
>>> I recently added another gfx card which seemed to go smothly enough
>>> except it apparently didn't. When I tried to plug in monitors to the
>>> integrated video ports AND the graphics card things started going
>>> sideways (and I can't figure out why they would?). One example is
>>> most
>>> times I boot no icons show up, not in the xfce menu nor context menu
>>> nor
>>> on the qubes manager, just the text. When I tried to rotate the
>>> monitors
>>> (via xfce) they acted funny the wallpaper displaying strangly
>>> but the qubes manager window showing up ok (in the same monitor at
>>> the
>>> same time).
>
> Xen might not support what you're trying to do. Can you replace the
> primary graphics card with one with more ports that can drive the
> amount
> of monitors you want?
>

doh, I hope it can. When you say might not support, are you referring to
4 monitors or 4 monitors spread out across 2 gfx ... devices? (not sure
what it would be called since one is integrated and the other a card
that was plugged in). Ah, a bit detail I left out of the post, the gfx
card model! Its a OEM AMD Radeon HD 6570 1GB DDR3 PCIe x16 DP DVI Video
Card pulled out of an HP machine.
As for replacing, most of the cards that seemed like they would be
compatible (AMD Radeon) cost a mint as soon as they went above the "3
ports" threshhold, so I would *really* like to try to get my current
setup (or the one I am aspiring to) to work before considering shelling
out more money.

Well the integrated gfx is what is primary now, I tried setting it to
switchable (whatever that is) and it didn't work, then to the grfx card
and that didnt work then to the primary and they all... kinda worked, at
least everything was getting a signal of some sort as opposed to
consistently getting 2 out of four monitors blank.

>>> Lastly, so far, when I try to start up
>>> an appvm I get the following: Error starting VM <sys-net> PCI device
>>> 02:00.0 does not exist (domain
>>> <sys-net>)
>>> though that happens with all the appvms and the debian templates and
>>> for
>>> redhat and whonix I get: qubes manager error starting vm 'fedora-26'
>>> maximum recursion depth exceeded
>
> Go into Qube Settings on sys-net, Devices tab. Unmap anything listed
> there
> and add your NIC back in. You might have to repeat this if you remove
> the
> secondary graphics card.

Alright, thanks. I will give it a try when I get back from my trip.

[awokd]

On Thu, April 19, 2018 3:35 pm, Stumpy wrote:
> On 2018-04-19 13:49, awokd wrote:

>>
>> Xen might not support what you're trying to do. Can you replace the
>> primary graphics card with one with more ports that can drive the amount
>> of monitors you want?
>>
>
> doh, I hope it can. When you say might not support, are you referring to 4
> monitors or 4 monitors spread out across 2 gfx ... devices?

I meant dual, active gfx cards. Unfortunately, it's not a very common
setup so I'm not sure what bugs you might hit...