Solved Persistent No screens found Plasma6 error with amd Picasso/raven2

[Sivan!]

I am on releng-14.1. -n267741 GENERIC amd64. Plasma6-plasma built from git 2025q3.

I have in my

~/.xinitrc

#!/bin/sh
export DISPLAY=:0.0
export TERM=xterm
exec dbus-launch --exit-with-x11 ck-launch-session startplasma-x11

Kldstat

amdgpu.ko

xorg installed

Xorg -configure

List of video drivers
amdgpu, scfb, modesetting vesa says No Devices to configure, configuration failed. EE server terminated with error (2)

sddm started, dbus running,

pkg info

drm-515-kmod-5.15.160.1401000_6 DRM drivers module libdrm-2.4.123,1 Direct rendering Manager library and heders # retyped

exec dbus-launch --exit-with-x11 ck-launch-session startplasma-x11

net local stream send space and recv space is 65536

ck-launch-session startplasma-x11

startx

Each of these commands logs me out of terminal. Also, the no screens found error is persistent. What do I do?

 

[nxjoseph]

I had entered to the thread to see if drm-xx-kmod is compiled for your kernel version but it's censored

drm-515-kmod-5.15.160.14xxxxx

14.1 release is end of life by the way.

 

[Sivan!]

I had entered to the thread to see if drm-xx-kmod is compiled for your kernel version but it's censored

drm-515-kmod-5.15.160.14xxxxx

14.1 release is end of life by the way.

1401000_6. End of life, yes. I started working on fixing xorg errors, that lead me to some understanding of the previous installation mismatches, took help from Chatgpt4 and then 5, (hugely helpful, yes, true, but looses overview at times), over 3 weeks of long endless night built git Q3 kernel which I think is releng 14.1, didn't know that it was important to buildworld also, or I builtworld and there is still a mismatch , built plasma6 and KDE from the same tree port after port with allow unsupported system make jobs unsafe flags, over chatgpt help that runs into a 1000pages+ conversation that I am unable to copy and paste. I have everything from the gitq3 2025 ports tree, most of KDE and plasma installed, but back to the xorg error after everything.

 

[SirDice]

Start with upgrading to 14.3-RELEASE.

 

[Sivan!]

Thank you SirDice

 

[Sivan!]

umame -a

14.3 release p2 freeBSD 14.3 release GENERIC

I have from the previous 14.1release p8 boot environment Plasma6 / kde/ qt6 packages built from git 2025Q3 ports tree src, though with some tweaks and fixes (when the errors were going round and round) For example, graphics ports were built with these flags:

sudo make ALLOW_UNSUPPORTED_SYSTEM=yes MAKE_JOBS_UNSAFE=yes \
  CMAKE_ARGS="-DENABLE_QT5=OFF -DENABLE_QT6=ON -DBUILD_TESTING=OFF -DBUILD_QT6_TESTS=OFF" \
  install

Or

graphics_opencv_UNSET=        CONTRIB
graphics_opencv_CONFIGURE_ARGS+= -DBUILD_opencv_sfm=OFF

The build took 6-10 or more hours a day for 2 or more weeks. Sp much time even with tweaks like these or with code edits or by skipping an entire module that was considered troublesome/non-essential (this by editing make files or with don't make flags without knowing a line of any code)

Most qt6, plama6, kde ports were built and installed, but doesn't show on [ CODE]pkg info plasma6*[/CODE] but showed up as installed by some other grep / egrep command.

Now that I am in 14.3 BE with a matching kernel, should I update qt6 +kde + plasma6 by the pkg update command? Or, would a pkg update upgrade command fix this? Should I remove anything before doing pkg update /upgrade?

Even in 14.3:

 start 
starplasma-x11 exec ck-launch-session startplasma-x11 
startx

logs me out.

dmesg -a | egrep -i 'amdgpu|drmn|ttm|linuxkpi|firmware' | tail -n 120 grep -iE 'amdgpu|drmn|firmware' /var/log/messages | tail -n 120

throws so many errors post upgrade such as :

kld amd gpu picasso gpu bin.ko depends on kernel
drmn0: could not load firmware image amdgpu/picasso_gpu_info.bin
[drm];amdgpu kernel modesetting enabled.
vgapci0: child drmn0 requested pci_enable_io (typed not verbatim)

freebsd-uodate install,
reboot, 
freebsd-update install

done already. No errors.

Is there anything else to be done before pkg update && pkg upgrade such as make clean?

Thank you.

(Still not well formatted, posting this from phone)

 

[Sivan!]

Question 1.


freebsd-verson-kru


I have 14.3 p2 (all three lines)

kldstat
shows other drivers except amdgpu even when I boot with NO kernel security, but in NO kernel security mode I can load amdgpu with the kldload command and after that startx brings up plasma, it is functional.

After reboot --with kern security, kldstat doesn't show amdgpu, kldload command is not permitted, and while all other rc settings take effect, amdgpu does not load, I have as one of the first few lines

kld_list="amdgpu"

and in boot loader.conf

amdgpu_load ="YES".

I also tried creating etc/rc.early with

kldload -n /boot/modules/amdgpu.ko 2>&1 | logger -t rc.early || true.

amdgpu does not load on boot despite the setting as above. I have installed drm 515 from port, (and it works in insecure boot) The problem is narrowed down to Amdgpu not loading; Neither with Kernel Security, nor without it; The problem is narrowed down to that of making amdgpu.ko load like the rest of the drivers on boot.

Question 1 (a).

I thought this might, just might be due to the error message that appeared on my grub screen "loaded needs to be updated". I searched for a fix for this error, there was a freeBSD forum thread, and based on the discussion in the thread I ran these commands:

grub menu shows loader needs updating, but freebsd boots fine. Though the grub menu appears, I don't find grub.d, not even usr local grub.conf


efibootmgr -v


Timeout 1 second
Boot order 0001 2 3 4
Boot order 1 uefi os HD (1,gpt, 97xxxx, 0x28,0x82000)/File(EFI/BOOT/BOOTX64.EFI)
gpt(efiboot0:/EFI/BOOT/BOOTX64.EFI /boot/efi//EFI/BOOT/BOOTX64.EFI


Gpart show


ada0 GPT
1 EFI 260 M
2 freebsd-boot 512 k

efibootmgr -v

Boot to FW : false
Timeout 1 second
Boot order 0001 2 3 4
Boot order 1 uefi os HD (1,gpt, 97xxxx, 0x28,0x82000)/File(EFI/BOOT/BOOTX64.EFI)
gpt(efiboot0:/EFI/BOOT/BOOTX64.EFI /boot/efi//EFI/BOOT/BOOTX64.EFI

[B]cp /boot/loader.efi /boot/efi/efi/freebsd/loader.efi<br>cp /boot/loader.efi /boot/efi/efi/boot/bootx64.efi[/B]

But this solution was offered to someone whose efibootmgr BOOT file was different, +Boot0000* FreeBSD HD(1,GPT,5e3e3d6d-f486-11ee-8ac6-a820661e35b1,0x28,0x82000)/File(\efi\freebsd\loader.efi)<br> gpt/efiboot0:/efi/freebsd/loader.efi /boot/efi//efi/freebsd/loader.efi

In my case the boot file is already EFI/BOOT/BOOTX64.EFI, not loader.efi as in the case of the user in the forum post to whom the cp /boot/loader.efi .... command was offered as a solution. Do I have to run these copy commands? If so, how do I MODIFY these commands?

 

[Sivan!]

Resolved all 14.3 upgrade issues and config conflicts, kldstat shows drm and amd, plasma boots to secure level.