Re: [OpenIndiana-discuss] A KVM switch broke my server?! Lightdm is panic'ing the kernel

2021-09-03 Thread Carl Brewer
I tried using nvidia-xconfig to generate an xorg.conf file in /etc/X11 
and then hand-adding the speeds that my monitor supports.


The monitor is an AOC 27V2Q, which, according to

supports 30-88kHz horizontal and 50-75 Hz vertical so my xorg.conf file, 
now shows (I hand-added the vendor name, model and frequencies)  :



# nvidia-xconfig: X configuration file generated by nvidia-xconfig
# nvidia-xconfig:  version 390.144  (dvs-builder@U64-P08-16)  Wed Jun  2 
22:56:17 UTC 2021


Section "ServerLayout"
    Identifier "Layout0"
    Screen  0  "Screen0"
    InputDevice    "Keyboard0" "CoreKeyboard"
    InputDevice    "Mouse0" "CorePointer"
EndSection

Section "Files"
    FontPath    "/usr/X11R6/lib/X11/fonts/misc/:unscaled"
    FontPath    "/usr/X11R6/lib/X11/fonts/100dpi/:unscaled"
    FontPath    "/usr/X11R6/lib/X11/fonts/75dpi/:unscaled"
    FontPath    "/usr/X11R6/lib/X11/fonts/misc/"
    FontPath    "/usr/X11R6/lib/X11/fonts/100dpi/"
    FontPath    "/usr/X11R6/lib/X11/fonts/75dpi/"
EndSection

Section "InputDevice"
    # generated from default
    Identifier "Mouse0"
    Driver "mouse"
    Option "Protocol" "auto"
    Option "Device" "/dev/mouse"
    Option "Emulate3Buttons" "no"
    Option "ZAxisMapping" "4 5"
EndSection

Section "InputDevice"
    # generated from default
    Identifier "Keyboard0"
    Driver "keyboard"
EndSection

Section "Monitor"
    Identifier "Monitor0"
    VendorName "AOC"
    ModelName  "27V2Q"
    HorizSync   30.0 - 88.0
    VertRefresh 50.0 - 75.0
    Option "DPMS"
EndSection

Section "Device"
    Identifier "Device0"
    Driver "nvidia"
    VendorName "NVIDIA Corporation"
EndSection

Section "Screen"
    Identifier "Screen0"
    Device "Device0"
    Monitor    "Monitor0"
    DefaultDepth    24
    SubSection "Display"
    Depth   24
    EndSubSection
EndSection

No difference, when I run startx or lightdm, the screen blanks, a white 
cursor appears on the top left of the screen, it then reboots.


What else can I try?

Tangentially, I don't really need a console, but I want a remote X11 
display - I can run vncserver and get at VirtualBox's GUI from there, 
but it's defaulting to TWM, which was nice in 1989 ... Is there some way 
to get lightdm to work just with vncserver/Xvnc and not try to run on 
the console - as a workaround, I can live without the console.






OpenPGP_signature
Description: OpenPGP digital signature
___
openindiana-discuss mailing list
openindiana-discuss@openindiana.org
https://openindiana.org/mailman/listinfo/openindiana-discuss


Re: [OpenIndiana-discuss] A KVM switch broke my server?! Lightdm is panic'ing the kernel

2021-09-03 Thread Carl Brewer

Still beating my head against this, I checked the NVidia drivers

According to the machine :

dmesg | grep -i nvi
Sep  4 11:35:11 skaro pseudo: [ID 129642 kern.info] pseudo-device: nvidia255
Sep  4 11:35:11 skaro genunix: [ID 936769 kern.info] nvidia255 is 
/pseudo/nvidia@255
Sep  4 11:35:11 skaro nvidia_modeset: [ID 107833 kern.notice] Loading 
NVIDIA Kernel Mode Setting Driver for UNIX platforms 390.144  Wed Jun  2 
22:57:33 UTC 2021

Sep  4 11:35:17 skaro nvidia_modeset: [ID 107833 kern.notice] Unloading

So it's got 390.144, and physically, it has a Geforce GT 1030 in it, 
which matches :


https://www.nvidia.com/Download/driverResults.aspx/177153/en-us

So it should be the right driver for that card.

the Xorg.0 logfile :

carl@skaro:/var/log$ grep -i nvid *
grep: can't open "authlog"
grep: can't open "lightdm"
Xorg.0.log:[   277.419] (II) Module glx: vendor="NVIDIA Corporation"
Xorg.0.log:[   277.419] (II) NVIDIA GLX Module  390.144  Wed Jun 2 
22:58:50 UTC 2021

Xorg.0.log:[   277.439] (II) LoadModule: "nvidia"
Xorg.0.log:[   277.441] (II) Loading 
/usr/X11/lib/modules/drivers/amd64/nvidia_drv.so

Xorg.0.log:[   277.445] (II) Module nvidia: vendor="NVIDIA Corporation"
Xorg.0.log:[   277.445] (II) NVIDIA dlloader X Driver  390.144 Wed Jun  
2 23:00:36 UTC 2021
Xorg.0.log:[   277.445] (II) NVIDIA Unified Driver for all Supported 
NVIDIA GPUs


Nothing weird about the driver :

ls -la /usr/X11/lib/modules/drivers/amd64/nvidia_drv.so
-r-xr-xr-x   1 root bin  7966304 Aug 13 18:01 
/usr/X11/lib/modules/drivers/amd64/nvidia_drv.so


I am *stumped*. I have run out of things to check and eliminate.

Carl




OpenPGP_signature
Description: OpenPGP digital signature
___
openindiana-discuss mailing list
openindiana-discuss@openindiana.org
https://openindiana.org/mailman/listinfo/openindiana-discuss


Re: [OpenIndiana-discuss] A KVM switch broke my server?! Lightdm is panic'ing the kernel

2021-09-03 Thread Carl Brewer

On 4/09/2021 12:36 am, s...@pandora.be wrote:


# scanpci | grep UHD
  Intel Corporation CometLake-S GT2 [UHD Graphics 630]

So in my case the 'scanpci' tool from OpenIndiana detects a UHD Graphics 630 in 
addition to the NVIDIA adapter.

Did you check 'scanpci'  and grep Intel ?



Yes, but not relevant, no UHD on this, see :

 scanpci | grep Intel
 Intel Corporation Comet Lake-S 6c Host Bridge/DRAM Controller
 Intel Corporation 6th-10th Gen Core Processor PCIe Controller (x16)
 Intel Corporation Comet Lake PCH-V USB Controller
 Intel Corporation Comet Lake PCH-V Thermal Subsystem
 Intel Corporation Device unknown
 Intel Corporation 400 Series Chipset Family SATA AHCI Controller
 Intel Corporation Device unknown
 Intel Corporation Device unknown
 Intel Corporation Device unknown
 Intel Corporation Device unknown
 Intel Corporation Memory controller
 Intel Corporation Device unknown
 Intel Corporation Comet Lake PCH-V SMBus Host Controller
 Intel Corporation Ethernet Connection (12) I219-V
root@skaro:/var/crash/skaro# scanpci | grep NV
 NVIDIA Corporation GP108 [GeForce GT 1030]
 NVIDIA Corporation GP108 High Definition Audio Controller


I am running OpenIndiana 2021.04 on a Dell Precision and also ran 'startx' to 
configure X.

# prtdiag -v | head
System Configuration: Dell Inc. Precision 3640 Tower
BIOS Configuration: Dell Inc. 1.2.3 08/06/2020

 Processor Sockets 

Version  Location Tag
 --
Intel(R) Core(TM) i3-10100 CPU @ 3.60GHz U3E1



prtdiag -v | head
System Configuration: Gigabyte Technology Co., Ltd. B460MAORUSPRO
BIOS Configuration: American Megatrends Inc. F4 08/28/2020

 Processor Sockets 

Version  Location Tag
 --
Intel(R) Core(TM) i5-10400F CPU @ 2.90GHz U3E1






Basically what I did after initial text install was run startx and check that 
it was unable to use the 'intel' driver for that card.

Anyway the NVIDIA card in my Dell Precision is working fine.

I configured /etc/X11/xorg.conf using /usr/bin tools :

/usr/bin/nvidia-SunOS-bug-report.sh
/usr/bin/nvidia-xconfig
/usr/bin/nvidia-settings


Basically when running /usr/bin/nvidia-xconfig I save the xorg.conf file from 
there.


I've just tried that, will see what happens, I just ran the 
nvidia-xconfig and saw that it generated an xorg.conf file in /etc/X11



Thanks!

Carl




OpenPGP_signature
Description: OpenPGP digital signature
___
openindiana-discuss mailing list
openindiana-discuss@openindiana.org
https://openindiana.org/mailman/listinfo/openindiana-discuss


Re: [OpenIndiana-discuss] A KVM switch broke my server?! Lightdm is panic'ing the kernel

2021-09-03 Thread Carl Brewer

This is the panic message :


root@skaro:/var/log# fmdump -Vp -u 017daaba-0d44-c582-d73d-b554bcf017a5
TIME UUID SUNW-MSG-ID
Sep 04 2021 09:50:42.514056000 017daaba-0d44-c582-d73d-b554bcf017a5 
SUNOS-8000-KL


  TIME CLASS ENA
  Sep 04 09:50:41.8497 ireport.os.sunos.panic.dump_pending_on_device 
0x


nvlist version: 0
    version = 0x0
    class = list.suspect
    uuid = 017daaba-0d44-c582-d73d-b554bcf017a5
    code = SUNOS-8000-KL
    diag-time = 1630713042 443560
    de = fmd:///module/software-diagnosis
    fault-list-sz = 0x1
    fault-list = (array of embedded nvlists)
    (start fault-list[0])
    nvlist version: 0
    version = 0x0
    class = defect.sunos.kernel.panic
    certainty = 0x64
    asru = 
sw:///:path=/var/crash/skaro/.017daaba-0d44-c582-d73d-b554bcf017a5
    resource = 
sw:///:path=/var/crash/skaro/.017daaba-0d44-c582-d73d-b554bcf017a5

    savecore-succcess = 0
    os-instance-uuid = 017daaba-0d44-c582-d73d-b554bcf017a5
    panicstr = hat_devload: loading a mapping to free page 
fe0001007830
    panicstack = unix:hat_devload+1ba () | 
gfx_private:gfxp_map_kernel_space+b5 () | nvidia:_nv027867rm+58 () | 
9c400 () | nvidia:_nv002650rm+0 () | nvidia:_nv002231rm+0 () |

    crashtime = 1630713007
    panic-time = September  4, 2021 at 09:50:07 AM AEST AEST
    (end fault-list[0])

    fault-status = 0x1
    severity = Major
    __ttl = 0x1
    __tod = 0x6132b4d2 0x1ea3df40


Suggests that it's something in the Nvidia driver? This stuff is way 
beyond my debugging skill level.


dmesg reports

Sep  4 09:50:41 skaro genunix: [ID 936769 kern.info] winlock0 is 
/pseudo/winlock@0
Sep  4 09:50:41 skaro devfsadmd[684]: [ID 511948 daemon.error] di_init 
failed for /pci@0,0/pci1458,5007@14/input: No such device or address
Sep  4 09:50:41 skaro svc.startd[9]: [ID 652011 daemon.warning] 
svc:/application/virtualbox/zoneaccess:default: Method 
"/opt/VirtualBox/VBoxZoneAccess" failed with exit status 127.
Sep  4 09:50:41 skaro savecore: [ID 570001 auth.error] reboot after 
panic: hat_devload: loading a mapping to free page fe0001007830
Sep  4 09:50:41 skaro savecore: [ID 620374 auth.error] Panic crashdump 
pending on dump device but dumpadm -n in effect; run savecore(1M) 
manually to extract. Image UUID 017daaba-0d44-c582-d73d-b554bcf017a5.
Sep  4 09:50:42 skaro svc.startd[9]: [ID 652011 daemon.warning] 
svc:/application/virtualbox/zoneaccess:default: Method 
"/opt/VirtualBox/VBoxZoneAccess" failed with exit status 127.

Sep  4 09:50:42 skaro last message repeated 1 time
Sep  4 09:50:42 skaro svc.startd[9]: [ID 748625 daemon.error] 
application/virtualbox/zoneaccess:default failed: transitioned to 
maintenance (see 'svcs -xv' for details)
Sep  4 09:50:42 skaro svc.startd[9]: [ID 748625 daemon.error] 
network/fail2ban:default failed repeatedly: transitioned to maintenance 
(see 'svcs -xv' for details)
Sep  4 09:50:42 skaro fmd: [ID 377184 daemon.error] SUNW-MSG-ID: 
SUNOS-8000-KL, TYPE: Defect, VER: 1, SEVERITY: Major

Sep  4 09:50:42 skaro EVENT-TIME: Sat Sep  4 09:50:42 AEST 2021
Sep  4 09:50:42 skaro PLATFORM: B460MAORUSPRO, CSN: Default-string, 
HOSTNAME: skaro

Sep  4 09:50:42 skaro SOURCE: software-diagnosis, REV: 0.1
Sep  4 09:50:42 skaro EVENT-ID: 017daaba-0d44-c582-d73d-b554bcf017a5
Sep  4 09:50:42 skaro DESC: The system has rebooted after a kernel 
panic.  Refer to http://illumos.org/msg/SUNOS-8000-KL for more information.
Sep  4 09:50:42 skaro AUTO-RESPONSE: The failed system image was dumped 
to the dump device.  If savecore is enabled (see dumpadm(1M)) a copy of 
the dump will be written to the savecore directory .
Sep  4 09:50:42 skaro IMPACT: There may be some performance impact while 
the panic is copied to the savecore directory.  Disk space usage by 
panics can be substantial.
Sep  4 09:50:42 skaro REC-ACTION: If savecore is not enabled then please 
take steps to preserve the crash image.
Sep  4 09:50:42 skaro Use 'fmdump -Vp -u 
017daaba-0d44-c582-d73d-b554bcf017a5' to view more panic detail. Please 
refer to the knowledge article for additional information.
Sep  4 09:50:43 skaro mac: [ID 435574 kern.info] NOTICE: e1000g0 link 
up, 1000 Mbps, full duplex

Sep  4 09:50:47 skaro nvidia_modeset: [ID 107833 kern.notice] Unloading

I don't think the Virtualbox stuff is relevant.

I have debugging info in /var/crash/skaro after running savecore :

root@skaro:/var/cron# mkdir -p /var/crash/skaro
root@skaro:/var/cron# savecore
savecore: System dump time: Sat Sep  4 09:50:07 2021

savecore: Saving compressed system crash dump in /var/crash/skaro/vmdump.0
savecore: Decompress the crash dump with
'savecore -vf /var/crash/skaro/vmdump.0'
root@skaro:/var/cron# cd /var/crash/skaro/

Re: [OpenIndiana-discuss] A KVM switch broke my server?! Lightdm is panic'ing the kernel

2021-09-03 Thread Carl Brewer

On 4/09/2021 12:36 am, s...@pandora.be wrote:

You wrote that you are running a Intel i5-10400F CPU.

That is a 6 core CPU according and the specs do not list an integrated Graphics 
card so I think it is an Intel CPU *without* integrated graphics :



That's correct.  That's why I'm using Nvidia cards in it.



https://www.intel.com/content/www/us/en/products/sku/199278/intel-core-i510400f-processor-12m-cache-up-to-4-30-ghz/specifications.html

The following is based on my experience with an Intel i3 with integrated 
graphics so that's different from your case, but may be of use anyway :

Have you checked for error messages in /var/log/Xorg.0.log ?

Are you using a specific /etc/X11/xorg.conf configuration file or do you use 
the probe and autodetection ?



Totally vanilla setup.  Whatever that happens to be. I just booted off 
the USB image and it worked, or, it did work ... until yesterday's plug 
into the KVM switch anyway.


I'll check the Xorg log file.

thank you

Carl




OpenPGP_signature
Description: OpenPGP digital signature
___
openindiana-discuss mailing list
openindiana-discuss@openindiana.org
https://openindiana.org/mailman/listinfo/openindiana-discuss


Re: [OpenIndiana-discuss] A KVM switch broke my server?! Lightdm is panic'ing the kernel

2021-09-03 Thread s...@pandora.be


You wrote that you are running a Intel i5-10400F CPU.

That is a 6 core CPU according and the specs do not list an integrated Graphics 
card so I think it is an Intel CPU *without* integrated graphics :

https://www.intel.com/content/www/us/en/products/sku/199278/intel-core-i510400f-processor-12m-cache-up-to-4-30-ghz/specifications.html

The following is based on my experience with an Intel i3 with integrated 
graphics so that's different from your case, but may be of use anyway :

Have you checked for error messages in /var/log/Xorg.0.log ?

Are you using a specific /etc/X11/xorg.conf configuration file or do you use 
the probe and autodetection ?

See man Xorg :

   Xorg has a mechanism for automatically generating a built-in
   configuration at run-time when no xorg.conf file or xorg.conf.d files
   are present.

 I had problems (and still have in some sense) with

# scanpci | grep UHD
 Intel Corporation CometLake-S GT2 [UHD Graphics 630]

So in my case the 'scanpci' tool from OpenIndiana detects a UHD Graphics 630 in 
addition to the NVIDIA adapter.

Did you check 'scanpci'  and grep Intel ?

I reported on the OpenIndiana Hardware Compatibility guide that I was (or still 
am) unable to get the Intel UHD630 card working with X on OpenIndiana.

http://docs.openindiana.org/community-hcl/components/

I added an entry there "Intel CometLake-S GT2 [UHD Graphics 630]"  that is NOT 
working (as far as I know).

I am running OpenIndiana 2021.04 on a Dell Precision and also ran 'startx' to 
configure X.

# prtdiag -v | head
System Configuration: Dell Inc. Precision 3640 Tower
BIOS Configuration: Dell Inc. 1.2.3 08/06/2020

 Processor Sockets 

Version  Location Tag
 --
Intel(R) Core(TM) i3-10100 CPU @ 3.60GHz U3E1


Basically what I did after initial text install was run startx and check that 
it was unable to use the 'intel' driver for that card.

Anyway the NVIDIA card in my Dell Precision is working fine.

I configured /etc/X11/xorg.conf using /usr/bin tools :

/usr/bin/nvidia-SunOS-bug-report.sh
/usr/bin/nvidia-xconfig
/usr/bin/nvidia-settings


Basically when running /usr/bin/nvidia-xconfig I save the xorg.conf file from 
there.

By the way, if there is way to get the Intel UHD630 card working I would be 
interested to learn about it,
but in my case the way I work with X is explicitely asking NOT to use the intel 
card in the xorg.conf ...


Regards,
David Stes


- Op 3 sep 2021 om 14:06 schreef Carl Brewer c...@bl.echidna.id.au:

> On 3/09/2021 9:35 pm, Tony Brian Albers wrote:
>> Have you tried just booting with lightdm disabled, logging in through
>> the cli and running startx?
> 
> It did a pretty good kernel panic!
> 
> This is weird.
> 
> 
> 
> 
> ___
> openindiana-discuss mailing list
> openindiana-discuss@openindiana.org
> https://openindiana.org/mailman/listinfo/openindiana-discuss

___
openindiana-discuss mailing list
openindiana-discuss@openindiana.org
https://openindiana.org/mailman/listinfo/openindiana-discuss


Re: [OpenIndiana-discuss] A KVM switch broke my server?! Lightdm is panic'ing the kernel

2021-09-03 Thread Carl Brewer

On 3/09/2021 9:35 pm, Tony Brian Albers wrote:

Have you tried just booting with lightdm disabled, logging in through
the cli and running startx?


It did a pretty good kernel panic!

This is weird.





OpenPGP_signature
Description: OpenPGP digital signature
___
openindiana-discuss mailing list
openindiana-discuss@openindiana.org
https://openindiana.org/mailman/listinfo/openindiana-discuss


Re: [OpenIndiana-discuss] A KVM switch broke my server?! Lightdm is panic'ing the kernel

2021-09-03 Thread Tony Brian Albers
Have you tried just booting with lightdm disabled, logging in through 
the cli and running startx?

Not sure if it would work, but it's worth a shot.

/tony

Carl Brewer wrote:
> 
> I've almost finished setting up my new i5/NVIDIA 1030 server. latest pkg 
> image-update today, my WD red pros arrived at last so I could build my 
> ZFS mirror for the data store, it's all coming together.  I've turned 
> off atime on rpool and my other zpools. Rebooted, everything seems fine.
> 
> Everything works on my dining room table, rebooted, it's solid. So it's 
> time to move it into the rack in my office.
> 
> I have a HDMI and USB KVM switch I plug it into. Not really paying much 
> attention, I didn't watch the console when it booted, I don't think I 
> even had it plugged in to a monitor at the time.
> 
> Long story short, it keeps kernel panic'ing if lightdm is enabled. I can 
> boot without lightdm and it's fine. I'd like the GUI though.  The ASCII 
> console is pretty close to useless, it's slow as mud.  Inherited that 
> from SPARC I guess ... In the OId Days, when I had a Sun 3/60(!) as my 
> desktop, the ASCII console was faster. heh.  IPX, Sparc 2, no quicker. I 
> never cared, I ran X11 on them, meh ...
> 
> Things I have tried, after getting rid of the KVM switch from the 
> situation.
> 
> Connected directly to the original (previously working) monitor with the 
> originally working HDMI cable.
> 
> Replacing the 1030 with a NVIDIA 1660 GFX card I had lying around.
> 
> different HDMI cable
> 
> different monitor
> 
> HDMI/SVGA adaptor cable instead of KVM switch
> 
> No luck. I've even tried booting off the (working, or it used to!) USB 
> drive Aurelian set up for me. It won't boot lightdm up. Kernel panic. I 
> don't get it. It worked this afternoon!  Reverted back to last week's OI 
> boot environment (which worked!), nope ...
> 
> I'm stumped, short of resetting the BIOS, I can't think of anything else 
> to try to revert it back to how it was 4 hours ago when it worked on the 
> table! It's back there now, and I can log in remotely.
> 
> If anyone has any suggestions for how to debug it? Or how to just get at 
> it using a VNC client - all the "set up VNC for OI" instructions I've 
> found are ancient and don't mention lightdm. If I can't have a GUI 
> console, that's ok, as long as I can get at it using VNC.
> 
> 
> ___
> openindiana-discuss mailing list
> openindiana-discuss@openindiana.org
> https://openindiana.org/mailman/listinfo/openindiana-discuss
> 


-- 
Tony Albers - Systems Architect - Data Department, Royal Danish Library, 
Victor Albecks Vej 1, 8000 Aarhus C, Denmark
Tel: +45 2566 2383 - CVR/SE: 2898 8842 - EAN: 5798000792142
___
openindiana-discuss mailing list
openindiana-discuss@openindiana.org
https://openindiana.org/mailman/listinfo/openindiana-discuss