[Solved] Loss of output to primary monitor?

IrvineHimself · 2021-10-19 16:32:28

For a number of years I have had a progressively worsening problem with my built in laptop monitor suffering a kind of tearing. I say kind of... because, technically, the diagnostic description doesn't fit classical tearing. (As a fail safe, I have applied All the wiki suggestions regarding tearing, such as early loading of kms...., etc

The tearing starts at the bottom of the screen with colour distorted lines which become black as they rise up the screen. Finally, after a few seconds, I am left with a blank primary monitor.

Initially, it happened infrequently and was little more than a minor inconvenience. However, a couple of years ago it started to happen with gradually increasing frequency, until, earlier this year, it became a major issue with even the the log in screen, (immediately after grub,) being affected. Note: I do not use A Display Manager

By happenstance, I had noticed that my external monitor, was unaffected. So, as an experiment, I dived into the laptops Bios settings, and redirected the primary output to the second monitor and configured my various compositors, (Gnome, Sway, Wayfire and Labwc,) to redirect output back to the primary.

Believe it or not, this more or less solved the problem completely. (There was a minor recurrence after I switched to Wayfire as my daily driver, however switching to Wayfire-git seems to have restored the above as a viable workaround.

Interestingly, for switching primary output to work, the external monitor doesn't need to be powered on. It just needs to exist and be plugged into the laptop.

Another point to note is that both my Recovery Usbs: Fedora & Xubuntu, are similarly affected.

While I am not certain, I have managed to convince myself that this is a Bios problem. Which, since I can't afford to 'brick' my laptop, and, as stated, am not 100% certain this is the 'actual' source of the problem, I am reluctant to update.

I use the open source 'Nouveau drivers', but even on the rare occasions I have installed the 'Nvidia drivers', it made little noticeable difference.

Here are some Bios and Hardware details:

Main: /sys/devices/virtual/dmi/id/            

Product Vendor           --     TOSHIBA                                 
Product Name             --     SATELLITE P50-C                         
Product Family           --     Type1Family                             
Product Version          --     PSPT2E-01000WEN                          
Product sku              --     PSPT2E                                  
    
Ec Firmware Release      --     1.10                                    
Bios Vendor              --     INSYDE Corp.                            
Bios Version             --     1.50                                    
Bios Release             --     1.50                                    
Bios Date                --     01/15/2016

As it stands at the moment:

* My laptop is virtually unusable without a secondary monitor, which, as I stated above, doesn't need to be powered on.
* In general, while redirecting primary output seems to be a viable work around, I believe it is the source of some configuration problems I am having with another project.

Has anyone ever come across anything like this, and, if so, did they ever find a solution that didn't involve updating the Bios?

Irvine

Last edited by IrvineHimself (2021-10-28 14:31:54)

seth · 2021-10-20 07:14:26

"lspci" - is this a dual GPU system?
(Because then likely the internal output is driven by the intel chip and the external one by nvidia, and the artifacts simply down to fbc or psr)

IrvineHimself · 2021-10-20 10:00:27

Thanks for the reply. As you suspected it has dual Intel/Nvidia GPUs

Here is the output from 'lspci'

=$ lspci
00:00.0 Host bridge: Intel Corporation Xeon E3-1200 v5/E3-1500 v5/6th Gen Core Processor Host Bridge/DRAM Registers (rev 08)
00:02.0 VGA compatible controller: Intel Corporation Skylake GT2 [HD Graphics 520] (rev 07)
00:14.0 USB controller: Intel Corporation Sunrise Point-LP USB 3.0 xHCI Controller (rev 21)
00:16.0 Communication controller: Intel Corporation Sunrise Point-LP CSME HECI #1 (rev 21)
00:17.0 SATA controller: Intel Corporation Sunrise Point-LP SATA Controller [AHCI mode] (rev 21)
00:1c.0 PCI bridge: Intel Corporation Sunrise Point-LP PCI Express Root Port #1 (rev f1)
00:1c.4 PCI bridge: Intel Corporation Sunrise Point-LP PCI Express Root Port #5 (rev f1)
00:1c.5 PCI bridge: Intel Corporation Sunrise Point-LP PCI Express Root Port #6 (rev f1)
00:1d.0 PCI bridge: Intel Corporation Sunrise Point-LP PCI Express Root Port #9 (rev f1)
00:1f.0 ISA bridge: Intel Corporation Sunrise Point-LP LPC Controller (rev 21)
00:1f.2 Memory controller: Intel Corporation Sunrise Point-LP PMC (rev 21)
00:1f.3 Audio device: Intel Corporation Sunrise Point-LP HD Audio (rev 21)
00:1f.4 SMBus: Intel Corporation Sunrise Point-LP SMBus (rev 21)
01:00.0 3D controller: NVIDIA Corporation GM108M [GeForce 930M] (rev a2)
02:00.0 Network controller: Intel Corporation Wireless 3165 (rev 81)
03:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller (rev 0c)
04:00.0 Unassigned class [ff00]: Realtek Semiconductor Co., Ltd. RTS5229 PCI Express Card Reader (rev 01)
=$

I have just done a quick read-up on DRM power saving, and, at first glance, since I am normally not mobile, I could probably live without this feature. (Although waste heat may become an issue?)

Are there kernel parameters I can set to disable these features temporaly?

Also, among others, the relevant Wikipedia page is quite informative, Is there an Arch Wiki page covering this?

Irvine

seth · 2021-10-20 11:48:31

It is "just"

01:00.0 3D controller: NVIDIA Corporation GM108M [GeForce 930M] (rev a2)

meaning that it doesn't provide a crtc an no output can be connected there.

How exactly do you go about

configured my various compositors, (Gnome, Sway, Wayfire and Labwc,) to redirect output back to the primary.

https://wiki.archlinux.org/title/Intel_graphics
https://wiki.archlinux.org/title/Intel_ … flickering
https://wiki.archlinux.org/title/Intel_ … nable_fbc)

IrvineHimself · 2021-10-20 14:23:01

seth wrote:

It is "just"
01:00.0 3D controller: NVIDIA Corporation GM108M [GeForce 930M] (rev a2)
meaning that it doesn't provide a crtc an no output can be connected there.

I am not certain about the speciffic details of which video-buffer the bios is using. It is a setting in the advanced section of the bios boot interface which allows the user to choose which monitor to use at boot:

'Auto'
'Laptop'
'Usb-#'
'HDMI-#'

I am paraphrasing the details from memory, but I am sure you get the picture.

seth wrote:

How exactly do you go about
configured my various compositors, (Gnome, Sway, Wayfire and Labwc,) to redirect output back to the primary.

On Gnome I used the system display GUI to mirror outputs.

With Labwc I was able to mirror outputs by placing the following as the first entry in the autostart file.

wlr-randr --output HDMI-A-1 --pos 0,0 --output eDP-1 --pos 0,0  --preferred

Unfortunately, mirroring outputs with the other wlroots based compositors, (Sway and Wayfire), is problematic.

According to the Wayfire wiki, the configuration file supports:

[output:HDMI-A-1]
mode = 1920x1080@60000
position = 0,0
transform = normal
scale = 1.000000

[output:eDP-1]
mode = mirror HDMI-A-1

But, for me, it just results in a blank screen.

Additionally, with Wayfire, for me:

Attempting to use the generic auto output configuration results in seriously screwed up screen resolutions.
Applying settings using the 'configure output' GUI freezes the system necessitating a reboot with a resulting messed up output configuration.
Similarly, using wlr-randr with Wayfire to set output parameters freezes my system

The best I can manage with Wayfire that works, is placing the following in wayfire.ini:

[output:HDMI-A-1]
mode = 1920x1080@60000
position = 0,0
transform = normal
scale = 1.000000

[output:eDP-1]
mode = 1920x1080@60000 --preferred
position = 1920,0
transform = normal
scale = 1.000000

As with Wayfire, using wlr-randr with Sway, to set output parameters also freezes my system.

What works is

output HDMI-A-1  pos 0 0 res 1920x1080
output eDP-1 pos 1920 0 res 1920x1080

workspace  1 output eDP-1
workspace  2 output eDP-1
workspace  3 output eDP-1
workspace  4 output eDP-1
workspace  5 output eDP-1

workspace  6 output HDMI-A-1
workspace  7 output HDMI-A-1
workspace  8 output HDMI-A-1
workspace  9 output HDMI-A-1
workspace 10 output HDMI-A-1

For both Sway and Wayfire I can start applications in which ever monitor has focus and pass them between work spaces, or, in the case of Wayfire, between monitors using the mouse.

Thank you for the links, it will take me a while to read through them and make sure I understand what is going on
Irvine

seth · 2021-10-20 14:39:26

Ah, so technically it's not redirected - you're just cloning the output.
There's a very good chance that the multiscreen config keeps the chip in performance mode, so …

IrvineHimself · 2021-10-20 14:45:12

seth wrote:

....that the multiscreen config keeps the chip in performance mode, so …

I am sorry, I don't quite follow what you are getting at, could you expand on that a bit

Irvine

seth · 2021-10-20 14:50:26

Disabling psr and fbc (and maybe even setting max_cstate=1) sounds promising (and impactless) in the light of your workaround.

IrvineHimself · 2021-10-20 14:56:32

Thanks, I have a few real world concerns to take care of, but will try out your suggestions and get back to you as soon as possible.

Irvine

IrvineHimself · 2021-10-21 04:04:24

Okay, a quick update:

I've added the following to grub

intel_idle.max_cstate=1
i915.enable_fbc=0
i915.enable_psr=0

After resetting the bios's default monitor to auto I was able repeatedly boot and run my various compositors without any problems, (touch wood.) Additionally, since in the process of making the mods, I discovered a few problems with my polkit setup, the above fix is being seriously tested.

For the last six months or so, without my original work-around, I would, by now, be tearing my hair out and screaming profanities. So, while it has only been 3 or four hours, and realy needs to run for a few day's to be certain, I would say things are looking very positive.

I don't want to jinx things, so I will leave the issue open for a couple of days to be sure. But regardless, you have my profuse thanks. This has been a minor irritation for a long time, and for the last six months, even with my work-around, has been a major cause for concern.

Irvine

seth · 2021-10-21 06:28:13

You want to check whether one of them is more important than the others - it's probably more about psr or fbc than restricting the c-state.

IrvineHimself · 2021-10-28 14:29:07

Sorry it took so long to get back, but I was testing over 5 compositors. (I subsequently added hikari to my install list.)

With the exception of hikari, which I have not yet fully configured to suit my personal usability criteria, I ran each compositor continuously until I could authoritatively confirm whether or not tearing was still a problem. (I rebooted once a day as part of my normal update schedule.)

After applying the above kernel parameters and undoing my workaround, the results are:

Gnome: I have tested gnome for over 3 days, and there has been no recurrence of the problem.
Sway: I ran sway for almost 24 hours, before there was tearing and loss of output.
Wayfire: tearing and loss of output generally occur after about 3 or 4 hours. Though, occasionally, they can occur almost immediately
Labwc: The results were similar to wayfire, with tearing and loss of output occurring after 3 or 4 hours.
Hikari I ran it for an hour without problems. For me, it is a new compositor and I was more interested in getting a feel for it's use than running any tests on my current problem.

Summary:
With gnome, it appears to be, at worst, a minor irritation. I am not really surprised at this, gnome/mutter are fairly mature projects with a lot of contributors/testers, and, in wayland, a reputation for stability.

While I like wayfire and labwc, and would love to use either of them as a daily driver, they are most definitely in the beta stage of development. So the fact that there are problems is to be expected.

I am not a great fan of tiling compositors; they are great for project work, but for general purpose/leisure, they are a bit cumbersome. Even so, I was a bit disappointed with the results for sway.

Conclusion:
While not a total success, this is an improvement on the results I was getting from my previous workaround, not least because I can now use my laptop without a second monitor.

In general, I suspect that the root of the problem with sway, wayfire and labwc lies with wlroots still being at an early stage of it's development.

The fact that it looks like I am going to be stuck with gnome until I can afford new hardware is disappointing, but I have been looking at custom builds and other ways I can configure it to my tastes. These, particularly the latter, look like they could be very interesting projects. So, as they say, "Every little cloud has a silver lining."

Thanks again for your help
Irvine

PS
As yet, I haven't really had a chance to check which kernel parameters are actually required, but I will do so shortly.

Additionally
If any developer comes across this post, I have an unconfirmed suspicion, (or inkling,) that the problems with wlroots compositors seems to be related to rapid, erratic changes. For example, trying to accurately scroll to text when a web page has lots of images.

Last edited by IrvineHimself (2021-10-28 14:34:12)

Arch Linux

#1 2021-10-19 16:32:28

[Solved] Loss of output to primary monitor?

#2 2021-10-20 07:14:26

Re: [Solved] Loss of output to primary monitor?

#3 2021-10-20 10:00:27

Re: [Solved] Loss of output to primary monitor?

#4 2021-10-20 11:48:31

Re: [Solved] Loss of output to primary monitor?

#5 2021-10-20 14:23:01

Re: [Solved] Loss of output to primary monitor?

#6 2021-10-20 14:39:26

Re: [Solved] Loss of output to primary monitor?

#7 2021-10-20 14:45:12

Re: [Solved] Loss of output to primary monitor?

#8 2021-10-20 14:50:26

Re: [Solved] Loss of output to primary monitor?

#9 2021-10-20 14:56:32

Re: [Solved] Loss of output to primary monitor?

#10 2021-10-21 04:04:24

Re: [Solved] Loss of output to primary monitor?

#11 2021-10-21 06:28:13

Re: [Solved] Loss of output to primary monitor?

#12 2021-10-28 14:29:07

Re: [Solved] Loss of output to primary monitor?

Board footer