Blank Screen on KDE5 after upgrading to FreeBSD 14.0-p3

Hi All!!!

I just got onto 14.0-p3. (Hooray!) However, I have a problem with KDE5 not starting on an NVIDIA 1070 ti. The rest of the machine is:

32 GB G. Skill
Asus ROG Strix 670X E-E
AMD 7700X

Everything used to work fine, but the first time in a while my upgrade did not go smooth.

I'm finally there with one major problem: blank screen

Yes, I have a blank screen even though everything booted up fine including X??? (according to the Xorg log file)

All screens are non-responsive.
 

Attachments

  • monitor.conf
    182 bytes · Views: 51
  • keyboard-zap.conf
    133 bytes · Views: 59
  • nvidia.conf
    88 bytes · Views: 44
  • screen.conf
    668 bytes · Views: 48
  • screen-resolution.conf
    136 bytes · Views: 43
Graham, not sure what that's going to do but render a new /etc/X11/xorg.conf.

It 'looks' fine.
 

Attachments

  • xorg.conf
    1.2 KB · Views: 53
Hi, also, is there a simple way to get the latest package version so I know which one I'm using? I just performed the freebsd-update and pkg upgrade.
 
Thanks.

quarterly rules out the blank screen symptom that I encountered yesterday with devel/dbus 1.14.10_2,1 in latest (fixed for me with 1.14.10_3,1).

GeForce GTX 1070 Ti is listed as supported at FreeBSD Display Driver – x64 | 535.146.02 | FreeBSD x64 | NVIDIA.

<https://www.freshports.org/x11/nvidia-driver/#packages> 535.146.02 is not yet packaged for FreeBSD:14:amd64, I don't imagine that slightly inferior 535.104.05_1 is a problem.

With the blankness, is there (also) a movable 'X' pointer on screen?

Hint: the pointer may be invisible until after you key Control-Alt-F2 then Alt-F9.

Does installation of graphics/nvidia-drm-515-kmod improve the situation?
 
echo $HOME to confirm the path to your home directory, then you might remove the contents of your .cache/ directory before attempting log in to Plasma (X11).

First, though, try nvidia-drm-515-kmod.
 
Good day!

Well, ctrl-alt-F2 shows me a typical curses login screen (at least that's normal again). Ctrl-Alt-F9 gives me a non-blinking cursor in the upper left corner of the screen.

I've removed everything within the .cache directory before I started this exercise.

I added the nvidia-drm-515-kmod in the /boot/loader.conf file with an appended '_load="YES"'

I guess a more direct question is what should be in the loader.conf and /etc/rc.conf for getting this all to work (I'd like to confirm and help those in the future who have this issue)

Thank you for you help n this!

Paul
 
It seldom is a good idea to load modules in loader.conf (imho).
A problem with the module may render your system unbootable.
So if there is no explicit reason for loading via loader.conf I’d always prefer rc.conf.
 
Noted and advice taken. :)

Presently, this is /boot/loader.conf:

kern.geom.label.disk_ident.enable="0"
kern.geom.label.gptid.enable="0"

#Required modules for root file system booting and on zfs
zfs_load="YES"
vfs.zfs.arc_min="1024M"
vfs.zfs.arc_max="6144M"
vfs.zfs.prefetch_disable="1"
vfs.zfs.min_auto_ashift=12
vfs.zfs.txg.timeout="5"
vm.kmem_size_max="8192M":wq::

kern.vty=vt
usbhid_load="YES"
hw.usb.usbhid.enable=1

This is /etc/rc.conf:


bsdstats_enable="YES"
cupsd_enable="YES"
dbus_enable="YES"
ddclient_enable="YES"
devd_enable="YES"
dhcpd_enable="YES"
dhcpd_flags="-q" # command option(s)
dhcpd_conf="/usr/local/etc/dhcpd.conf" # configuration file
dhcpd_ifaces="em1" # ethernet interface(s)
dhcpd_withumask="022" # file creation mask
dhcpd_chuser_enable="YES" # runs w/o privileges?
dhcpd_withuser="dhcpd" # user name to run as
dhcpd_withgroup="dhcpd" # group name to run as
dhcpd_rootdir="/var/db/dhcpd" # directory to run in
fail2ban_enable="YES"
gateway_enable="YES"
ipv6_activate_all_interfaces="YES"
log_in_vain="YES"
mailman_enable="YES"
ntpd_enable="YES"
ntpdate_enable="YES"
ntpdate_pool="pool.ntp.org"
pf_enable="YES"
pflog_enable="YES"
pflog_logfile="/var/log/pflog"
pflog_flags=""
saver="logo"
sendmail_enable="NO"
sendmail_submit_enable="NO"
sendmail_outbound_enable="NO"
sendmail_msp_queue_enable="NO"
postfix_enable="NO"
sddm_enable="NO"
tmpmfs="YES"
tmpsize="1g"
tmpmfs_flags="-m 0 -o async,noatime -S -p 1777"
syslogd_enable="YES"
syslogd_flags=""
sshd_enable="YES"
usbd_enable="YES"
postgresql_enable="YES"
plexmediaserver_enable="YES"
plexmediaserver_support_path="/usr/local/plexdata" # Plex data: logs, media metadata, settings, etc
plexmediaserver_tmp="/var/tmp/plex" # configure tmp directory used for the transcoding process if desired
plexmediaserver_maxplugins="6" # Maximum number of background plugin procs. May have to raise in rare cases.
zfs_enable="YES"

devfs_system_ruleset="system"

#kld_list="/boot/modules/nvidia /boot/modules/nvidia-modeset"
kld_list="/boot/modules/nvidia /boot/modules/nvidia-drm-515-kmod"

sndiod_enable="YES"
virtual_oss_dsp="-T /dev/sndstat -S -i 8 -C 18 -c 18 -r 48000 -b 32 -s 512 -f /dev/dsp0 -c 2 -d dsp -c 18 -d vdsp -t vdsp.ctl"

ipmi_load="YES"
mps_load="YES"
nullfs_load="YES"
tmpfs_load="YES"

That's minus all the addresses and things at the top.
 
Can you try to set exactly to
Code:
kld_list="nvidia-drm nvidia-modeset"
in that order and reboot? (That's what I have and it works under 14.0 for me)
If it still does not work please share the output of kldstat | grep nvidia
 
Whoops... dont know why that was in the buffer... sorry....

Anyhow, things are progressing!!!! I can now startx or xinit as a user or root. Yay!!!!

The last problem I'm having is baloo is crashing.

kldstat | grep nvidia

8 2 0xffffffff82582000 3d83578 nvidia.ko
19 1 0xffffffff86c3d000 11e630 nvidia-modeset.ko

I have enabled sddm in my /etc/rc.conf and the login looks great. (I made sure I was on X and not wayland for now)

However, I login and I get apps up but it's on a black background and no toolbars or menus show and there's a failure window for 'baloo' (I believe I had to turn this off before to get everything working, I can do that again, but if there's a way to make it work...??? I'm up for it.)

Brick by brick!

Paul
 
I‘d try without xorg.conf.
The preferred method is via xorg.conf.d:
See step 7 and 8 here:
 
P.S.: baloo is a resource monster imho and of no use (at least to me). Why‘d I want to index all files in a KDE database?
 
rm /etc/X11/xorg.conf

balooctl purge
balooctl disable

ctrl-alt-F9 back to the login screen.
Login.....
IT'S MAGIC!!!!
Yes, everything is back. It looks great.

Thank you both, El Grande and Graham.


Just as an aside, looking through my boot up at this time....

.
.
.
Firmware Error (ACPI): Could not resolve symbol [\134_SB.PCI0.GPP7.UP00.DP40.UP00.DP68], AE_NOT_FOUND (20221020/dswload2-315)
ACPI Error: AE_NOT_FOUND, During name lookup/catalog (20221020/psobject-372)
.
.
.
KLD nvidia-drm.ko: depends on linuxkpi_gplv2 - not available or version mismatch
linker_load_file: /boot/modules/nvidia-drm.ko - unsupported file type
nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms 535.104.05 Sat Aug 19 00:40:44 UTC 2023
.
.
.
End

I can't wait to get everything back onto my FreeBSD desktop/server. I missed it so much (Sobs gently) I can't wait to try sumiyazu (sp) and get gaming on here and move my AMD 6700 XT onto this box. (Yay, I'll be an expert on getting both AMD and NVIDIA running on FreeBSD 14.0)
 
So.... I got everything working on KDE5. It was a joy to see it all working again.

I then decided to move my AMD 6700 XT over to it for serious gfx power.

The only thing I had to do was change the /etc/rc.conf entries from nvidia to amdgpu and reboot. MAGIC!!!

I'm going to have to open some new threads as I have non GFX issues going on but this can be closed.

Thank you all for the concise and understandable assistance. FreeBSD rules and the rest drools!!!
 
Back
Top