Howdy,
On my new rig, I've got everything installed. I mostly been on a
console which has worked without issue. Now I've started using the GUI,
KDE, and I'm having issues. I wanted to run a command to generate a xorg.conf file and it generate all the needed info regarding hardware
and such since the GUI wasn't working right without a config file. The
only one I found is the nvidia one. It is minimal at best. Anyway,
when I try to start display-manager, with or without a config file, I
get the sddm login. I login and if I just let it sit there, after a
minute or so, the monitor goes black. It is still powered up but
nothing on the screen at all. I've moved the mouse and pressed buttons
on the keyboard to make sure it isn't powering off but nothing. Also,
the resolution is pretty low too. It should run in 1080P easily. The
card supports 4K I think. The monitor tho has ran 1080P on my main rig before, for years I might add.
I used tail -f to watch a few error logs. I watched sddm, messages and Xorg.0.log. The only thing that got added when the monitor went black
was this:
Jun 21 13:27:41 Gentoo-1 kernel: nvidia-modeset: WARNING: GPU:0: Unable
to read EDID for display device DP-3
Jun 21 13:29:01 Gentoo-1 kernel: nvidia-modeset: WARNING: GPU:0: Unable
to read EDID for display device LG Electronics W2253 (DP-3)
Jun 21 13:29:02 Gentoo-1 kernel: nvidia-modeset: WARNING: GPU:0: Unable
to read EDID for display device DP-3
Jun 21 13:29:03 Gentoo-1 kernel: nvidia-modeset: WARNING: GPU:0: Unable
to read EDID for display device LG Electronics W2253 (DP-3)
If I let it sit for a good while, it comes back on, sort of. The plasma panel thing on the bottom, where the app menu, clock and all is, doesn't
come back tho. Also, the background is just black, no picture like
usual. The only thing that shows up is gkrellm. I'm pretty sure plasma died. I thought maybe it was heat, the fan isn't running on the video
card or something. Nvidia shows the card between 25 and 30C. The fan
nvidia says is at 34%. If that is correct, then the fan is running and
heat is not a issue. Out comes the flashlight and a mirror. Yep, fan spinning. According to IR temp thingy, nothing even gets into the 90F
area. I think if it was heat, I'd see something getting hot with the IR
temp thing.
There is two versions of Nvidia driver for this card in the tree. I've
tried both, no change at all. Screen goes black and after a while comes
back but most of the desktop has crashed.
By the way, I ran the tail command over ssh. Sometimes when the monitor
goes black, it doesn't come back. I can use ssh to reboot and repeat tho.
Did I miss something during the install? Does the error above cause
this problem? If so, how do I fix it? If you need info, just let me
know the command to run. I monitored all I could think of. The one
above is all I saw that showed a problem exists. Given this rig is
still in testing, I can reboot or anything else as needed.
Thanks.
Dale
:-) :-)
On Friday, 21 June 2024 20:02:22 BST Dale wrote:
Howdy,
On my new rig, I've got everything installed. I mostly been on a
console which has worked without issue. Now I've started using the GUI, KDE, and I'm having issues. I wanted to run a command to generate a xorg.conf file and it generate all the needed info regarding hardware
and such since the GUI wasn't working right without a config file. The only one I found is the nvidia one. It is minimal at best. Anyway,
when I try to start display-manager, with or without a config file, I
get the sddm login. I login and if I just let it sit there, after a
minute or so, the monitor goes black. It is still powered up but
nothing on the screen at all. I've moved the mouse and pressed buttons
on the keyboard to make sure it isn't powering off but nothing. Also,
the resolution is pretty low too. It should run in 1080P easily. The
card supports 4K I think. The monitor tho has ran 1080P on my main rig before, for years I might add.
I used tail -f to watch a few error logs. I watched sddm, messages and Xorg.0.log. The only thing that got added when the monitor went black
was this:
Jun 21 13:27:41 Gentoo-1 kernel: nvidia-modeset: WARNING: GPU:0: Unable
to read EDID for display device DP-3
Jun 21 13:29:01 Gentoo-1 kernel: nvidia-modeset: WARNING: GPU:0: Unable
to read EDID for display device LG Electronics W2253 (DP-3)
Jun 21 13:29:02 Gentoo-1 kernel: nvidia-modeset: WARNING: GPU:0: Unable
to read EDID for display device DP-3
Jun 21 13:29:03 Gentoo-1 kernel: nvidia-modeset: WARNING: GPU:0: Unable
to read EDID for display device LG Electronics W2253 (DP-3)
There are reports in the interwebs about LG monitors having buggy EDID tables:
https://gist.github.com/kj800x/be3001c07c49fdb36970633b0bc6defb
What is connected at DP-1 and is the problem resolved if you change the port you connect your monitor on the card?
Do you get an EDID table shown in your Xorg.0.log, or with xrandr --verbose?
Does your ~/.local/share/sddm/xorg-session.log reveal anything more?
If I let it sit for a good while, it comes back on, sort of. The plasma panel thing on the bottom, where the app menu, clock and all is, doesn't come back tho. Also, the background is just black, no picture like
usual. The only thing that shows up is gkrellm. I'm pretty sure plasma died. I thought maybe it was heat, the fan isn't running on the video
card or something. Nvidia shows the card between 25 and 30C. The fan nvidia says is at 34%. If that is correct, then the fan is running and heat is not a issue. Out comes the flashlight and a mirror. Yep, fan spinning. According to IR temp thingy, nothing even gets into the 90F area. I think if it was heat, I'd see something getting hot with the IR temp thing.
There is two versions of Nvidia driver for this card in the tree. I've tried both, no change at all. Screen goes black and after a while comes back but most of the desktop has crashed.
By the way, I ran the tail command over ssh. Sometimes when the monitor goes black, it doesn't come back. I can use ssh to reboot and repeat tho.
Did I miss something during the install? Does the error above cause
this problem? If so, how do I fix it? If you need info, just let me
know the command to run. I monitored all I could think of. The one
above is all I saw that showed a problem exists. Given this rig is
still in testing, I can reboot or anything else as needed.
Thanks.
Dale
:-) :-)
If this is a buggy EDID, or monitor chip, you should be able to extract the EDID table and store it as a firmware blob for the video card to load.
Dale wrote:
Michael wrote:
root@Gentoo-1 ~ # xrandr --verbose
Can't open display
root@Gentoo-1 ~ #
That's after I started display-manager but this time, it did nothing.
The screen stayed on a console. I did move the cable to another port
before booting up.
This is the other info except nothing xorg there, just wayland. See
below for more on that.
root@Gentoo-1 ~ # cat /home/dale/.local/share/sddm/wayland-session.log
dbus-daemon[2766]: [session uid=1000 pid=2766 pidfd=5] Successfully
activated service 'org.freedesktop.portal.Documents'
fuse: device not found, try 'modprobe fuse' first
error: fuse init failed: Can't mount path /run/user/1000/doc
No backend specified, automatically choosing drm[snip ...]
kwin_wayland_drm: No suitable DRM devices have been found
The Wayland connection broke. Did the Wayland compositor die?
The Wayland connection broke. Did the Wayland compositor die?
The Wayland connection broke. Did the Wayland compositor die?
qt.qpa.wayland: Creating a fake screen in order for Qt not to crash dbus-daemon[2766]: [session uid=1000 pid=2766 pidfd=5] Activated service 'org.freedesktop.impl.portal.desktop.kde' failed: Process org.freedesktop.impl.portal.desktop.kde exited with status 1
(/usr/libexec/xdg-desktop-portal:2787): xdg-desktop-portal-WARNING **: 03:20:14.512: Failed to create settings proxy: Error calling StartServiceByName for org.freedesktop.impl.portal.desktop.kde: Process org.freedesktop.impl.portal.desktop.kde exited with status 1
(/usr/libexec/xdg-desktop-portal:2787): xdg-desktop-portal-WARNING **: 03:20:14.512: No skeleton to export
kdeinit5: Communication error with launcher. Exiting!
dbus-daemon[2766]: [session uid=1000 pid=2766 pidfd=5] Activating
service name='org.freedesktop.impl.portal.desktop.kde' requested by
':1.5' (uid=1000 pid=2787 comm="/usr/libexec/xdg-desktop-portal" label="kernel")
Error: could not determine $DISPLAY.
Error: Can not contact kdeinit5!
org.kde.startup: "kdeinit5_shutdown" () exited with code 255 startplasma-wayland: Shutting down...
startplasmacompositor: Shutting down...
startplasmacompositor: Done.
root@Gentoo-1 ~ #
Since it complains about DRM. Here's this.
root@Gentoo-1 ~ # lsmod
Module Size Used by
nvidia_drm 102400 0
nvidia_modeset 1544192 1 nvidia_drm
nvidia 59744256 1 nvidia_modeset
root@Gentoo-1 ~ #
What's that about fuse???
To add to this, this is a new video card. It's a P1000, Nvidia of
course. Could it be a bad card that works to a point but then dies? I
got a older PCIe V2 card I'm going to try. That was the card that was
in during the early stages. The P1000 was the one that got hung up by
the post office and arrived days late. I'm going to test the other
slower card shortly, got to remember to downgrade nvidia-drivers to
tho. ;-)
On the above wayland info, I recall some having issues with wayland. I
tried to disable wayland by putting -wayland in make.conf and
rebuilding. Since it wanted to rebuild a lot, I removed everything GUI related from the world file and then ran --depclean to clean house.
Then I emerge everything I removed from the world file again, while I
napped. The output above could be old but may still provide a clue. If
I could figure out how to go back to KDE V5 instead of the new V6, I
would, to eliminate a wayland issue if nothing else. I removed
everything from /etc/portage/package.keyword file but I think KDE V6 is stable now. The odd part, right now, even sddm isn't coming up.
This is the other info except nothing xorg there, just wayland. SeeI thought you said you were having a problem starting a X session, not starting a Wayland session. What desktop session are you selecting in your SDDM?
below for more on that.
root@Gentoo-1 ~ # cat/home/dale/.local/share/sddm/wayland-session.log
Michael wrote:
On Saturday, 22 June 2024 19:13:42 BST Dale wrote:
Dale wrote:
Michael wrote:root@Gentoo-1 ~ # xrandr --verbose
Can't open display
root@Gentoo-1 ~ #
That's after I started display-manager but this time, it did nothing.
The screen stayed on a console. I did move the cable to another port
before booting up.
The xerver is not running.
How do I start it? I did a /etc/init.d/ <tab twice> and I see display-manager and xdm. I been using display-manager. I have sddm set
for display-manager. Like this.
root@Gentoo-1 ~ # cat /etc/conf.d/xdm
# We always try and start X on a static VT. The various DMs normally default # to using VT7. If you wish to use the xdm init script, then you should ensure
# that the VT checked is the same VT your DM wants to use. We do this
check to
# ensure that you haven't accidentally configured something to run on the VT # in your /etc/inittab file so that you don't get a dead keyboard.
CHECKVT=7
# What display manager do you use ? [ xdm | gdm | sddm | gpe | lightdm
| entrance ]
# NOTE: If this is set in /etc/rc.conf, that setting will override this one. DISPLAYMANAGER="sddm"
This is the other info except nothing xorg there, just wayland. See
below for more on that.
root@Gentoo-1 ~ # cat /home/dale/.local/share/sddm/wayland-session.log
I thought you said you were having a problem starting a X session, not starting a Wayland session. What desktop session are you selecting in
your
SDDM?
[snip ...]
That made me think. After a few attempts, I got the sddm login screen
to come up. I checked and it was logging me into a KDE wayland session,
as a default.
I changed it to x11.
The first time, it logged in but
nothing plasma showed up. No panel at the bottom and a black screen for
the background. The only thing there was gkrellm. I leave it there so
I can tell if it is working, to some degree at least. Otherwise, it is
just a black screen.
kde-plasma/plasma-meta-6.1.0 is still in testing - I sync'ed earlier
today:
~ $ eshowkw kde-plasma/plasma-meta
Keywords for kde-plasma/plasma-meta:
| | u |
|
| a a p s l r a | n |
| m r h p p i o i s l m m | e u s | r
| d a m p p c a x a o s 3 p 6 i | a s l | e
| 6 r 6 p p 6 r 8 6 n c 9 h 8 p | p e o | p
| 4 m 4 a c 4 c 6 4 g v 0 a k s | i d t | o
--------------+-------------------------------+-------+------- [I]5.27.11-r1 | + ~ + o o ~ o + o ~ ~ o o o o | 8 o 5 | gentoo --------------+-------------------------------+-------+-------
6.1.0 | ~ o o o o o o o o o o o o o o | 8 o 6 | gentoo
I see some qt stuff that is version 6. KDE has gotten pretty
confusing. It depends on qt so much that KDE depends heavily on
versions matching up somehow.
Is there a guide about installing the GUI parts, including KDE? I found
a Xorg guide. I'm good all the way down to startx. At that point,
nothing. I installed xterm just so startx should show some windows on
the screen. Nothing. Monitor goes to sleep.
Top posting for consistency.
I booted the Gentoo GUI media. I opened a window just in case it
rebooted or something while I took a little nap. The resolution is
1080P which is what I expected the monitor to run at. When I got back
up a few minutes ago, the same window was there. It ran for hours
without the monitor powering off. I think this eliminates hardware.
I also found something new. I can go in the BIOS menu and boot a USB
stick directly from that by just clicking on it. O_O WOW!!! I said
the new BIOS had improved by a lot. Can that thing wash dishes too???
LOL
Still open to ideas if anyone has any. If not, I may safe key files and start over, and fix my partition layout too.
Michael wrote:
On Sunday, 23 June 2024 08:53:01 BST Dale wrote:
Top posting for consistency.
I booted the Gentoo GUI media. I opened a window just in case it
rebooted or something while I took a little nap. The resolution is
1080P which is what I expected the monitor to run at. When I got back
up a few minutes ago, the same window was there. It ran for hours
without the monitor powering off. I think this eliminates hardware.
Yes! :-)
I was worried that $100 video card I bought, most I ever spent on a
video card, was for nothing. I'd hate for it to be defective somehow.
I also found something new. I can go in the BIOS menu and boot a USB
stick directly from that by just clicking on it. O_O WOW!!! I said
the new BIOS had improved by a lot. Can that thing wash dishes too???
LOL
It is the same as the legacy BIOS in this respect, only the GUI is different. With the old BIOS you would press F2 or Delete at POST and
then go into the boot menu to select 'Removable Media' or some such.
Well, on this one, I thought I was going to select it for a temporary
boot after exiting the BIOS, one time thing, but when I clicked it, the
menu screen for Knoppix popped up. I didn't exit the BIOS or anything.
Just click and off it went to boot what I clicked. o_O
Still open to ideas if anyone has any. If not, I may safe key files and >> start over, and fix my partition layout too.
Diff the settings used by the LiveUSB and the settings you have configured on your installation. Starting over with your installation, only to
follow the same path and without *knowing* what you need to change, will not necessarily resolve your issue.
That's my thinking. The only benefit to reinstalling is correcting the partition boo boo.
I also might learn something, maybe. Of course,
there is that blind squirrel tho. I think I'll get the config file info
from the Gentoo media tho.
Since my last message, it's been sitting on a Knoppix screen. It's
still sitting there just like I left it for my second nap. So, hardware
is certainly in the clear. A config issue is the problem, which means I
was right, I missed something. Somewhere.
Now to go find it. ;?
Michael wrote:
On Sunday, 23 June 2024 13:19:18 BST Dale wrote:
That's my thinking. The only benefit to reinstalling is correcting the
partition boo boo.
What in particular are you referring to? I thought you created an ESP, / and /home partitions, if I recall.
I was thinking the EFI partition, ef02, and /boot were the same
partition. I only created one that was huge. I made the ef02 partition
8GBs I think when 8MBs would have been more than enough from what I've
read.
I kinda like /boot on its own partition. If /boot gets corrupted
somehow, I can get the kernel and config again from /usr/src/linux.
Building the init thingy again is trivial. So is reinstalling grub. If
I were to lose root somehow, I'd have to reinstall but I got the kernel
and its config file. It ain't much but it's something. As it is, /boot
is on the same partition as root. If root goes bad, all is lost, except
for any backup copies I might have. If I redo the install, I'd have a
EFI partition and a separate /boot partition as well. The EFI would be
like 8MBs or so and the /boot partition would be ext2 and 8GBs or so.
Plenty of room for expansion.
I saved messages, sddm log and the Xorg log. Honestly tho, they will be different because the live DVD uses the kernel drivers, nouveau, where I
use Nvidia. Still, I got them anyway.
Several years ago, I took the CPU cooler off my main rig. I dunked the
fin part into some heavy duty cleaner that cleans off dirt and dust.
They wasn't to bad really since I blow them clean with compressed air
from my compressor pretty regular. Still, it looked new when I was
done. I used alcohol and a toothbrush to clean the old paste off of the base. The original paste was what came with the CPU I think. I'm
pretty sure I put Arctic Silver on the second time. I can't remember
the exact amount but it did cool better after a few updates and some compiling. I don't know if it was the fin cleaning or the thermal paste
or both tho. I think Arctic Silver is still considered a good brand and
very good product. I'm pretty sure it never dries out. Even some
cheaper generic brands are quite good.
I'm going to compare some data between the Gentoo live DVD and my
install. If I don't see something obvious, I'm going to fix my
partition boo boo with a fresh start. While at it. What is the best
way to wipe the partition data from a m.2 stick? They not spinning rust
so don't want to try to dd or use shred on the whole thing. Doesn't
gdisk have a wipe partition option? Curious what you think is the best
way to do that. Don't want to shorten the life of my m.2 stick.
Now to ponder what comes next.
Dale
:-) :-)
Now to ponder what comes next.
William Kenworthy wrote:
...
Now to ponder what comes next.
Dale
:-) :-)
Hi Dale, did I see in one of your early emails you created an
xorg.conf for nvidia? Have you followed the gentoo Xorg guide where
it says to try first without that file? I doubt the knoppix etc use a
conf file and so must depend on the auto detection.
Based on my own experience I can say using an xorg.conf (though it was
a radeon card) is an easy way to shoot yourself in the foot!
BillK
At first I did not have a xorg.conf file. When it didn't work right, I
tried creating one to see if it would work. It didn't work with or
without one.
Michael wrote:
On Sunday, 23 June 2024 23:37:15 BST Dale wrote:
Michael wrote:
On Sunday, 23 June 2024 13:19:18 BST Dale wrote:
I kinda like /boot on its own partition. If /boot gets corrupted
somehow, I can get the kernel and config again from /usr/src/linux.
Building the init thingy again is trivial. So is reinstalling grub. If >> I were to lose root somehow, I'd have to reinstall but I got the kernel
and its config file. It ain't much but it's something. As it is, /boot >> is on the same partition as root. If root goes bad, all is lost, except >> for any backup copies I might have. If I redo the install, I'd have a
EFI partition and a separate /boot partition as well. The EFI would be
like 8MBs or so and the /boot partition would be ext2 and 8GBs or so.
Plenty of room for expansion.
I see. I don't think you need to redo the install. All you need to do
is:
1. Back up the ESP contents, just in case.
2. Shrink the ESP partition, down to a reasonable size. 500M or 1G would be more than enough.
3. Create a new partition, say ~7G in the space your just freed up, of
type
8300.
4. Check if the content of the ESP fs is intact (it should be, but we're talking about FAT here) and if not reformat as FAT32 and copy over the files from the ESP backup.
5. Format the new /boot partition and copy over the files from your
current / boot tdirectory to the new /boot partition you created.
6. Adjust your fstab and reboot.
NOTE: The GPT partition numbering order will be messed up, but this does not alter their functionality. If it annoys you, then use gdisk to re-order them.Oh, I'd mess that up quick. o_O
[snip ...]I'm going to compare some data between the Gentoo live DVD and my
install. If I don't see something obvious, I'm going to fix my
partition boo boo with a fresh start. While at it. What is the best
way to wipe the partition data from a m.2 stick? They not spinning rust >> so don't want to try to dd or use shred on the whole thing. Doesn't
gdisk have a wipe partition option? Curious what you think is the best
way to do that. Don't want to shorten the life of my m.2 stick.
In this case do not reinstall. Most of it, if not all, would be unnecessarily deleting and rewriting the same data.
gdisk can destroy all the GPT data structures on a disk. Press x, then z. However, I suggest you don't this. Use Gparted to shrink your ESP and add a new partition for /boot as I explained above. The focus on sorting out your graphic card.
Now to ponder what comes next.
Dale
:-) :-)
If a person is trying to copy another install and runs into a failure in
a package to compile, skip ahead and deal with the locale section first
then come back.
Michael wrote:
You should at least try to launch a X11 session using a console and see what is printed out on the CLI after you exit (or if it crashes).
~ $ exec dbus-launch --exit-with-session startplasma-x11
I tried every combination I could think of. It seems the more I tried,
the worse the failure got.
Anyway, I've started a fresh install. I fixed my partition boo boo. I didn't need much of anything except to just get a fresh start so I used
the zap feature of gdisk. After that, I used cgdisk to set up
everything else, making sure the efi partition was ef00 and had the old
fat type partition on it. Or was it vfat? Whatever the docs said. I
copy and paste some of it except for the partition info.
I did notice one thing, I copied some USE flags over from my current
rig. I found some that might not should have been copied over before.
I wonder if that could have caused a problem or two. I realized I chose
the merged /usr stage3 but had split-usr USE flag in make.conf. That
was one I remembered. There was a couple others as well.
Dale wrote:
If a person is trying to copy another install and runs into a failure in
a package to compile, skip ahead and deal with the locale section first
then come back. This failure is between syncing the tree and during the Optional: Updating the @world set section. I moved the USE line over
from my old system with some changes so I expected quite a few updates.
I ran into libdrm failing, no matter what USE flags I try. It failed
with this:
ninja: build stopped: subcommand failed.[snip ...]
* ERROR: x11-libs/libdrm-2.4.120::gentoo failed (compile phase):
* ninja -v -j16 -l10 failed
* Working directory: '/var/tmp/portage/x11-libs/libdrm-2.4.120/work/libdrm-2.4.120-abi_x86_64.amd 64' * S: '/var/tmp/portage/x11-libs/libdrm-2.4.120/work/libdrm-2.4.120' /var/tmp/portage/x11-libs/libdrm-2.4.120/temp/environment: line 87:
warning: setlocale: LC_MESSAGES: cannot change locale (en_US.UTF8): No such file or directory
/var/tmp/portage/x11-libs/libdrm-2.4.120/temp/environment: line 90:
warning: setlocale: LC_NUMERIC: cannot change locale (en_US.UTF8): No
such file or directory /var/tmp/portage/x11-libs/libdrm-2.4.120/temp/environment: line 93:
warning: setlocale: LC_TIME: cannot change locale (en_US.UTF8): No such
file or directory
*
* The following package has failed to build, install, or execute postinst:
*
* (x11-libs/libdrm-2.4.120:0/0::gentoo, ebuild scheduled for merge),
Log file:
* '/var/log/portage/x11-libs:libdrm-2.4.120:20240624-140627.log'
The key part is that part about 8 lines or so up. If the locale is not
set, it will fail. Once it is set, it compiles just fine.
I don't know if the docs should be changed or not. It could be that in
most cases, doing it early could cause other problems.
Michael wrote:
For an ESP on a disk partition, it would be FAT32 (while FAT12 or FAT16
can be used for removable media). VFAT is an extension to FAT allowing long filenames. In any case it's just a symlink:
~ $ ls -la /usr/bin/mkfs.vfat
lrwxrwxrwx 1 root root 8 Mar 24 08:52 /usr/bin/mkfs.vfat -> mkfs.fat
I followed the guide. I copied the mkfs command and the option then
replaced with my partition info. The command in the docs was a sda
drive not nvme. This part of the docs worked fine last time. I just repeated the same.
I also ran into that locale thing again. Using your export command,
fixed it, again. I find it odd that the commands to reset the
environment does not reset that somehow. Anyway, it works. I have that LC_ALL set on my main rig. It's been that way for years. Should I
change it or leave it since it works??
My CPU seems to be running about 20F cooler now.
Have you seen this before?
(chroot) livecd /usr/src/linux # dracut --kver=$(cat include/config/kernel.release)
dracut[I]: Executing: /usr/bin/dracut --kver=6.9.4-gentoo
dracut[F]: Can't write to
/efi/dba652170b7a716f303c3c5966799436/6.9.4-gentoo: Directory /efi/dba652170b7a716f303c3c5966799436/6.9.4-gentoo does not exist or is
not accessible.
(chroot) livecd /usr/src/linux #
The directory inside /efi does not exist.
This is from the boot media,
not the chroot environment.
livecd ~ # mount | grep efi
/dev/nvme0n1p1 on /mnt/gentoo/efi type vfat (rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=ascii,shortname=mi xed,errors=remount-ro) livecd ~ #
I removed the efi parts for the live media that is booted. The efi
partition is mounted rw and is vfat. Right?
I used touch to create a
test file and it created the file in /efi from within the chroot
environment just fine. It seems dracut has a issue tho. I went back
through the docs and don't see anything I missed but I don't see what it
is that creates what dracut is looking for either. This is the correct
file structure??
(chroot) livecd / # ls -al /
total 447012
drwxr-xr-x 17 root root 4096 Jun 23 12:06 .
drwxr-xr-x 17 root root 4096 Jun 23 12:06 ..
lrwxrwxrwx 1 root root 7 Jun 23 12:06 bin -> usr/bin
drwxr-xr-x 4 root root 4096 Jun 24 14:29 boot
drwxr-xr-x 19 root root 4760 Jun 24 11:58 dev
drwxr-xr-x 3 root root 4096 Dec 31 1969 efi
drwxr-xr-x 49 root root 4096 Jun 24 14:32 etc
drwxr-xr-x 2 root root 4096 Jun 23 12:06 home
lrwxrwxrwx 1 root root 7 Jun 23 12:06 lib -> usr/lib
lrwxrwxrwx 1 root root 9 Jun 23 12:06 lib64 -> usr/lib64
drwxr-xr-x 2 root root 4096 Jun 23 12:06 media
drwxr-xr-x 2 root root 4096 Jun 23 12:06 mnt
drwxr-xr-x 5 root root 4096 Jun 24 05:19 opt
dr-xr-xr-x 461 root root 0 Jun 24 04:39 proc
drwx------ 2 root root 4096 Jun 23 12:06 root
drwxr-xr-x 3 root root 60 Jun 24 04:40 run
lrwxrwxrwx 1 root root 7 Jun 23 12:06 sbin -> usr/bin
-rw-r--r-- 1 root root 457688576 Jun 23 12:44 stage3-amd64-desktop-openrc-20240623T164908Z.tar.xz
dr-xr-xr-x 13 root root 0 Jun 24 04:44 sys
drwxrwxrwt 3 root root 60 Jun 24 14:32 tmp
drwxr-xr-x 11 root root 4096 Jun 23 12:09 usr
drwxr-xr-x 10 root root 4096 Jun 24 10:19 var
Thoughts? I miss something??
Dale
:-) :-)
Michael wrote:
On Monday, 24 June 2024 20:47:15 BST Dale wrote:
Have you seen this before?
No, because I've never used dracut.
I just had a thought. I have /usr on the root partition now. Do I even
need a init thingy?
(chroot) livecd /usr/src/linux # dracut --kver=$(cat
include/config/kernel.release)
dracut[I]: Executing: /usr/bin/dracut --kver=6.9.4-gentoo
dracut[F]: Can't write to
/efi/dba652170b7a716f303c3c5966799436/6.9.4-gentoo: Directory
/efi/dba652170b7a716f303c3c5966799436/6.9.4-gentoo does not exist or is
not accessible.
(chroot) livecd /usr/src/linux #
The directory inside /efi does not exist.
The long string is either a PARTUUID, or a fs UUID.
Run blikid to find out what it is.
/dev/nvme0n1p1: UUID="6488-1019" BLOCK_SIZE="512" TYPE="vfat" PARTLABEL="efi-part" PARTUUID="50add3c0-4ab3-4453-85e5-93af643a586e"
I don't see anything, UUID or anything else that starts with dba6 for
any partition. I don't know why dracut is looking for that.
Do you have a directory called EFI in your /efi partition?
Have you mounted your /mnt/gentoo/boot partition when you called dracut?
Yep. There is. Should that be there? This is what is there.
(chroot) livecd / # ls -al /efi/
total 12
drwxr-xr-x 3 root root 4096 Dec 31 1969 .
drwxr-xr-x 17 root root 4096 Jun 23 12:06 ..
drwxr-xr-x 3 root root 4096 Jun 24 14:29 EFI
-rwxr-xr-x 1 root root 0 Jun 24 14:01 test
(chroot) livecd / # ls -al /efi/EFI/gentoo/grubx64.efi
-rwxr-xr-x 1 root root 143360 Jun 24 14:33 /efi/EFI/gentoo/grubx64.efi
I mounted /boot when I mounted the others. I already have a kernel and
such in /boot. When I did a search, I found where someone posted they
used force with dracut to get it to install a init thingy. I used it
and it did build and put one in /boot. Thing is, I've never had to use
force before and figure something is wrong somewhere.
This is /boot.
(chroot) livecd / # ls -al /boot/
total 22512
drwxr-xr-x 4 root root 4096 Jun 24 14:29 .
drwxr-xr-x 17 root root 4096 Jun 23 12:06 ..
-rw-r--r-- 1 root root 77312 Jun 24 11:40 amd-uc.img
-rw-r--r-- 1 root root 158391 Jun 24 12:02 config-6.9.4-6
drwxr-xr-x 6 root root 4096 Jun 24 14:29 grub
-rw------- 1 root root 7740012 Jun 24 14:06 initramfs-6.9.4-6.img -rw-r--r-- 1 root root 15000576 Jun 24 12:02 kernel-6.9.4-6
drwx------ 2 root root 16384 Jun 24 04:03 lost+found
That give you any clues? Or am I starting over again. ROFL
Dale
:-) :-)
Michael wrote:
The first option in the man page explains what you did:
https://linux.die.net/man/8/dracut
-f, --force
overwrite existing initramfs file.
Did you have an initramfs already in there?
I had the kernel and the config file on /boot.
I did not have anything
else there except grub and the CPU microcode file. I didn't save the
init thingy when I was copying files over from previous install. I just saved the config and kernel.
[snip ...]No don't start over! Have you read through this:
https://wiki.gentoo.org/wiki/Dracut
Yep. I saw that too. Thing is, it confuses me. On the main install
page, it shows /efi mounted on the / partition. In other words, the
same place /boot, /usr and /var are mounted too. In the page you link
to it seems to show the efi partition mounted inside /boot. Like this: /boot/efi. The main page I think says this is no longer recommended.
Which am I to follow? If it being inside /boot is discouraged, someone
needs to update the page.
I'm continuing on with the install but still puzzled about the dracut
error. Is this what /efi should look like?
(chroot) livecd / # tree /efi/
/efi/
└── EFI
└── gentoo
└── grubx64.efi
3 directories, 1 file
(chroot) livecd / #
I never looked in the directory on the last install. Nothing reported a error so I just went with it. ;-)
Picking here to get a fresh sub thread going. Attempt two.
I got all the packages installed. Made sure the Nvidia video drivers
were loaded. I remembered to make sure elogind was running. It wasn't,
so I started it. Why doesn't display-manager pull that in???
Anyway, I
started display-manager and got the sddm screen. I selected Plasma x11 instead of Plasma wayland. I then logged in. It took a little bit but finally the screen came up. All I get is the 'Welcome to KDE plasma"
window. Everything else is black. I like the screen that says "simple
by default". It's simple alright. A black screen is pretty simple.
The mouse pointer moves so it sees the mouse. Right click should bring[snip ..]
up a menu. Nothing. The little panel thing that is usually at the
bottom, nothing. Once I close the welcome window, it's black.
This is the messages file from the point of me starting display-manager.
Does this yield any clues?
Michael wrote:
You need to have USE="elogind -systemd" in your make.conf, then add the elogind service to the *boot* runlevel as shown here:
https://wiki.gentoo.org/wiki/Elogind
I read down through that. I did find that acl had made it into the USE
flag line. I removed it.
It's not on my main rig so no idea where that
came from.
[snip ...]Can you please save and attach as plain text files your:
1. dmesg
2. Xorg.0.log
3. ~/.local/share/sddm/xorg-session.log
4. /var/log/sddm.log
after you end up in a black screen, in case they reveal something.
Should be attached. I blanked the files and then rebooted and started display-manager, (DM). You should have only the most recent info. I'm
also putting a chunk of messages below. It might help. It isn't much.
Same as before it seems. I still say this is something simple but hard
to find. :/
Dale
:-) :-)
Messages:
Jun 25 13:31:18 Gentoo-1 kernel: nvidia-modeset: WARNING: GPU:0: Unable
to read EDID for display device DP-3
Update. I played around a bit. I figured I didn't have a lot to lose
here. It either works, or it doesn't. After playing around a bit, I
got it to work. I have not restarted it to see if it will work again,
yet. I wanted to grab some log info first, while it is working. So,
this part is about when KDE comes up completely but could include some attempts that failed. Comes up completely means, in the correct
resolution, background image and the panel thing on the bottom, which
means plasma is running as it should. I'm doing these inside the email instead of as attachments. Sorry for the length. I just want to share
this while I have it available.
[ 1236.810] (--) NVIDIA(GPU-0):[Snip ...]
[ 1237.031] (--) NVIDIA(GPU-0): LG Electronics W2253 (DFP-5): connected
[ 1237.031] (--) NVIDIA(GPU-0): LG Electronics W2253 (DFP-5): Internal TMDS [ 1237.031] (--) NVIDIA(GPU-0): LG Electronics W2253 (DFP-5): 300.0 MHz maximum pixel clock
This is xorg.conf.
Gentoo-1 ~ # cat /etc/X11/xorg.conf[Snip ...]
# nvidia-xconfig: X configuration file generated by nvidia-xconfig
# nvidia-xconfig: version 550.90.07
Section "Monitor"
Identifier "Monitor0"
VendorName "Unknown"
ModelName "Unknown"
Option "DPMS"
EndSection
#Section "Device"
# Identifier "Device0"
# Driver "nvidia"
# VendorName "NVIDIA Corporation"
#EndSection
Section "Device"
Identifier "Device0"
# Identifier "nvidia"
Driver "nvidia"
BusID "PCI:9:0:0"
Option "UseEDID" "false" ## Comment out this entry for now <==
EndSection
Section "Screen"
Identifier "Screen0"
Device "Device0"
Monitor "Monitor0"
DefaultDepth 24
SubSection "Display"
Depth 24
EndSubSection
EndSection
Gentoo-1 ~ #
I took the info you sent and sort of combined it with what
nvidia-xconfig created. This is when it worked, this time anyway.
Other things I tried did nothing. Now that I have documented all this,
I'm going to reboot the rig and see if it works again. I think we
already have failed logs but will share what happens.
Given I get different results even with the same settings, I'm wondering about that monitor. I've changed video cards so that should eliminate
that. The only common thing is the monitor. Thing is, that monitor
worked for a long time on my main rig, it also has worked fine on the
NAS box and the old Dell system as well. That was very recent I might
add. This new system is the only one that has issues with that monitor.
The new monitor should give us clues. If it just works when it gets
here, then it is the monitor acting weird. If it does the same thing,
there is a config error somewhere. I can't think of anything else.
Open to ideas still. I'd like to get this working. If for no other
reason, the new monitor could have the same issue and require some
special settings somewhere.
Thanks for all the help. Sorry to have so much info in one email. :/
At least we have details of when it is working now tho. :-D
Mark Knecht wrote:
On Thu, Jun 27, 2024 at 12:03 AM Dale <rdalek1967@gmail.com
<mailto:rdalek1967@gmail.com>> wrote:
Update. I played around a bit. I figured I didn't have a lot to lose here. It either works, or it doesn't. After playing around a bit, I
got it to work. I have not restarted it to see if it will work again, yet. I wanted to grab some log info first, while it is working. So, this part is about when KDE comes up completely but could include some attempts that failed. Comes up completely means, in the correct resolution, background image and the panel thing on the bottom, which means plasma is running as it should. I'm doing these inside the email instead of as attachments. Sorry for the length. I just want to share this while I have it available.
This first one, I had a few failures before it succeeded. I couldn't figure out when the working bit started so it is the complete log.
Hi Dale,
I have returned home and can once again bottom post.
There's no way I'm going to read and understand this
whole thread but I did have one question and one comment:
1) Did you ever actually try the Kubunu option that didn't
require an install? I saw you mentions Knoppix and maybe
one other option. Just curious as to what the results were.
2) I see you discussing xorg.conf file which I don't use
here but did you generate this file - if you are really using
it - using the nvidia-settings app, or by hand?
If your system will stay up I believe nvidia-settings is
recommended by NVidia, or was anyway. It does a good
job of showing the layout and handling options that NVidia
says make their cards work better.
Good luck,
Mark
I booted the Gentoo live image and Knoppix. Knoppix is old. I don't
think it is being maintained anymore but even on this new hardware, even
it worked. I've booted other things to and all work fine except the
Gentoo install, the one thing I need to work. I checked, I did download Kubuntu but I don't remember trying it. With my memory tho, I may have
and just don't remember it. :/ I'll add it to my Ventoy stick and try
it shortly. It's hot, humid and my energy level isn't much. The family visit to the hospital drained me good. And she is still sick. I took
her some tomatoes this morning and she likes the peaches I got for her
too. Not much she can eat.
I've tried with no xorg.conf at first. Then I tried with one that tells
it to use the nvidia driver, even tho lsmod shows it loaded and lspci -k shows it being used. Then I used nvidia-xconfig to create a conf file.
Then I tried some options that Michael suggested. On occasion, it
works. Most of the time, it doesn't. To be honest, I'm not sure if
anything we do is affecting it. I think sometimes things just drop into place and it works. Most of the time, things don't drop into place and
it fails or only partially works.
Given I've used different boot media, a different video card and
different config options, I'm thinking it is the monitor and it just
clashes with this one install but works with others, which may figure
out to ignore what the monitor fails to do. I find it odd but it is
logical.
I've tried to use nvidia-settings but the man page is like Greek. The
one thing I did figure out, -q all. On my main rig, it spits out TONS
of info. On the new rig, it just says something like no display found
or something. It's like two lines, maybe one.
Had several interruptions so managed to try Kubuntu "Try" option. It
comes up just fine, correct resolution, plasma is working and all. I'm attaching the Xorg log from Kubuntu.
Maybe the Kubuntu log will shed some light.
Dale
:-) :-)
Michael wrote:
[snip ...]
[ 30.345] (II) modeset(0): [DRI2] DRI driver: nouveau <== Not nvidia ==
[snip ...]
[ 30.295] (II) modeset(0): Output DP-1 disconnected
[ 30.295] (II) modeset(0): Output DP-2 disconnected
[ 30.295] (II) modeset(0): Output DP-3 disconnected
[ 30.295] (II) modeset(0): Output DP-4 connected
What does it take to connect the cable on the FIRST port of the video
card?
This reply may be a little odd. I wrote some then went back and tried
some stuff and added info to it. Trying to make it make sense.
At first, when it really wouldn't work, I had it on the bottom port. Atsome point we thought that was the last port, #4 or DP-3 in the logs,
and not the first port. I moved it to the top port which is #1, we
thought. It's still on that port but that port did work once and
resulted in "solved" being added to the subject line. I just double
checked. It is plugged in the top port. Unless the bottom port is #1
like I originally thought, then it should be on port #1. I did a search
and found a image that shows it puts the port number on the metal
bracket. I removed the card so I could see the numbers. The bottom
port is number 1 like I originally thought. So, I had it in the right
port to begin with, which wasn't working either. I'll put it back on
what the metal bracket says is port #1, or bottom port. I booted up,
started DM, correct resolution but no plasma and background is black.
Still not a functional desktop. Partially works tho. Time to reboot.
On reboot, sddm and KDE are low resolution. It does have a background
image and plasma is working. Keep in mind, all I did is reboot. I
didn't change any config file or run any commands. Reboot again. This
time, correct resolution but no plasma. Again, all I did was reboot.
No changes to anything at all. As you can see, each time I reboot, it
is like rolling dice. I suspect if I keep rebooting it will eventually
do the black screen and power the monitor off.
I did originally try to use the nouveau drivers. It kinda worked, once
at least, but the screen was very slow to respond and the mouse was very jerky. It just wasn't good enough for whatever reason. I recompiled
the kernel without those drivers and emerged nvidia.
I booted the rig up and decided to try something. Once it was booted, I logged in from my main rig via ssh. I then typed in the command to
start DM. It started and looked OK. Then I just up arrow and changed
it to restart the DM. I restarted DM back to back several times, more
than a dozen. Sometimes it wouldn't work, sometimes it would be low resolution and sometimes it would come up and look like it should, hi
res and all. I also tried logging in when it was working and I could login. The biggest thing I noticed, it never came up fully. Most of
the time it came up in hi res but no plasma. A few times it was a low
res screen and no plasma. Looked like maybe 720P or less.
The thing is, it didn't fully come up even once. It was always lacking plasma at least. Some of the time, it was low res. Several times, the monitor would go black and cut off completely. It would go to sleep.
If the new monitor works, I'm thinking Micheal is right. The monitor
works with slower systems and the nouveau drivers on boot media. With Nvidia on the install, hit or miss, mostly miss. I think it has only
worked fully twice.
The last update showed the monitor a couple states away. FedEx is
pretty speedy. I think if things move well, it could be here Monday.
It still shows Tuesday tho. Given I been dealing with this for a week
or so now, another few days isn't a big deal. Plus, I'll have a really
nice large monitor for these old eyes. o_o
I took meds last night. I didn't wake up in time to pick my basil
again. I got three planters of it that need picking. It's hot and
humid outside. Try again in the morning. That basil sure is good.
Even opening the jar smells awesome. Nothing like the store bought stuff.
I finally uploaded some pics. Some are while the rig is running. You
can see the LEDs on the memory stick lit up. Some are components.
Also, there is a couple of the little m.2 stick cooler. I think that
little thing is so cute. Kinda looks like the CPU cooler, just
smaller. m.2 sticks runs at just above 100F even when pretty busy.
Very effective. :-D This is a link to the gallery or whatever.
https://postimg.cc/gallery/w6HQp83
Just imagine that in a Fractal Design Define XL case now. Dang that
case is big. It's a hair larger than my Cooler Master HAF-932 which is
a awesome case. Oh, I also finally fixed the power on light on the
front. I hooked the wires up backwards. LEDs never work well when connected up wrong. LOL
Dale
:-) :-)
I booted the rig up and decided to try something. Once it was booted, I logged in from my main rig via ssh. I then typed in the command to
start DM. It started and looked OK. Then I just up arrow and changed
it to restart the DM. I restarted DM back to back several times, more
than a dozen. Sometimes it wouldn't work, sometimes it would be low resolution and sometimes it would come up and look like it should, hi
res and all. I also tried logging in when it was working and I could
login. The biggest thing I noticed, it never came up fully. Most of
the time it came up in hi res but no plasma. A few times it was a low
res screen and no plasma. Looked like maybe 720P or less.
The thing is, it didn't fully come up even once. It was always lacking plasma at least. Some of the time, it was low res. Several times, the monitor would go black and cut off completely. It would go to sleep.
If the new monitor works, I'm thinking Micheal is right. The monitor
works with slower systems and the nouveau drivers on boot media. With
Nvidia on the install, hit or miss, mostly miss. I think it has only
worked fully twice.
Here's a little update. First, Kubuntu wiped my Gentoo install from
being able to boot it. I had to boot the Gentoo USB image, mount
everything and reinstall grub/EFI stuff. I got it back and now I can
boot either one by selecting it in the BIOS.
Second. I decided to annoy the heck out of that thing. I logged in
over ssh and I kept restarting DM. Took me several dozen tries but eventually I got KDE to come up, plasma and all. I opened the Nvidia
GUI and got it to save the xorg.conf file. It's different but it has a
lot of info now.
Michael wrote:
This is good! :D
Edit the section below by adding the PreferredMode line:
Section "Monitor"
Identifier "Monitor0"
VendorName "Unknown"
ModelName "LG Electronics W2253"
HorizSync 30.0 - 83.0
VertRefresh 56.0 - 61.0
Option "PreferredMode" "1920x1080_60.00"
Option "DPMS"
EndSection
Save it, then check if you get a consistent result when you restart the
DM, or reboot.
If it is still playing up you need to show the Xorg.0.log while the
monitor is working with the nvidia driver, to see which Modeline it uses and add this in the "Monitor" section too, above the "PreferredMode"
entry.
If none of this works reliably, then you have to capture the EDID table while the monitor is working, save it in a file and set nvidia-settings
to use it hereafter. However, if you will NOT be using this monitor for much longer this would be an exercise only to make you feel good for beating it into submission! LOL!
OK. First time, it worked. Second time, it came up but in low res.
Plasma was working to, both times. So, I opened the Nvidia GUI and told
it to safe a new config with it in hi res mostly just to see what it
would add if anything. It did. The only change I could see was it
added metamodes options like this:
Section "Screen"
# Removed Option "metamodes" "nvidia-auto-select +0+0"
Identifier "Screen0"
Device "Device0"
Monitor "Monitor0"
DefaultDepth 24
Option "Stereo" "0"
Option "nvidiaXineramaInfoOrder" "DP-3"
Option "metamodes" "1920x1080 +0+0"
# Option "metamodes" "1920x1080 +0+0; 1024x768_60 +0+0"
Option "SLI" "Off"
Option "MultiGPU" "Off"
Option "BaseMosaic" "off"
SubSection "Display"
Depth 24
EndSubSection
EndSection
As you can see, I removed the low res part, don't want that anyway, and
it seems to work more often but still fails a lot. So, I removed the metamodes Nvidia added. Then it didn't come up at all, sddm or
anything, when I restarted DM. I thought it adding your setting to
another section might help. Guess not. With the following removed, it
seems to do a little better but still can't depend on it to work.
Section "Screen"
# Option "nvidiaXineramaInfoOrder" "DP-3"
# Option "metamodes" "1920x1080 +0+0"
# Option "metamodes" "1920x1080 +0+0; 1024x768_60 +0+0"
Basically, it is doing like it always has. Most of the time, it doesn't work. Sometimes it does. To be honest tho, I still wonder if it even
reads the file. When I change something, I'd expect something
different. Generally, it doesn't seem to affect anything unless I
reboot completely. Sometimes I'm not sure it does then either. So, I
think this old monitor just has a issue. Maybe I need to rebuild some
more of the power supply or something in case it is providing bad power
to the parts that talks to the puter. Either way, if the new one works,
I just don't think this old monitor is going to work correctly no matter
what we throw at it.
I will saw this, we threw the kitchen sink at it. I think this is one
of the longest threads I've seen in a long while. o_O
Dale
:-)
P. S. Monitor updated. Now it says tomorrow. It's in the right place
to do so. It's at the Memphis hub. Usually, it comes from there to the local distribution point and on the truck. They sometimes deliver by
noon. Could be a little late given it is Monday.
| Sysop: | Keyop |
|---|---|
| Location: | Huddersfield, West Yorkshire, UK |
| Users: | 546 |
| Nodes: | 16 (2 / 14) |
| Uptime: | 00:29:47 |
| Calls: | 10,385 |
| Calls today: | 2 |
| Files: | 14,057 |
| Messages: | 6,416,567 |