You are not logged in.
I was having trouble with Wayland and saw that there seemed to be a lot of issues with nVidia. So, I did a search about changing to AMD.
Many of the results said I should just change the card and all would just magically work. So I just changed the cards and it turns out not to be the case.
I get the grub screen and choose the option I normally use. I get as far as ARCH-ROOT: clean, <number of files and blocks. No further.
I can't get to a terminal. I'm not sure what I should do at the advanced options selection.
I also noticed that while doing an update using pacman, a message stating that the mesa driver was not being updated. Ignored it as I was using nvidia.
I can't find anything about this on the web, everything is too old and wouldn't address Wayland.
Last edited by mt_arch_user (2025-08-09 01:24:43)
Offline
Nothing here makes sense. There aren't 'a lot of issue' with nvidia and wayland. I've never heard of 'atm' in relation to graphic cards. Ignoring pacman warnings when you don't know what they even said is bad.
I'm guessing since you're using grub, you have 'quiet' on the kernel command line? Get rid of that. If you're trying to boot to graphical mode, you can boot to multi-user instead using the kernel command line. https://wiki.archlinux.org/title/System … _boot_into
Online
Ops! I meant AMD. I remember when they were ATI. Sorry.
Last edited by mt_arch_user (2025-06-29 20:30:17)
Offline
The last sentence seems to require I use systemctl, which I can't use as I can't get a prompt.
Offline
As far as ignoring the pacman message, I know what it said. It said it was not updating the mesa driver and showed the version numbers. I didn't bother with it as I was using the nVidia driver, not the mesa driver, so it didn't matter.
Offline
The last sentence seems to require I use systemctl, which I can't use as I can't get a prompt.
you can boot to multi-user instead using the kernel command line. https://wiki.archlinux.org/title/System … _boot_into
Read the entire section
Online
I uninstalled the AMD card and reinstalled the nVidia card and the system comes up as it should.
I did some research on the mesa driver and it seems it was blacklisted as it conflicts with mesa-amber.
Everything I can find is tyhat mesa-amber is the correct module for the AMD card.
I installed mesa-amber, which uninstalled mesa.
The Arch wiki says that any needed modules will be loaded at boot, so I swapped the cards again.
Still, no joy.
Can I force thye mesa-amber module to be loaded at boot using /etc/modules-load.d/?
Do I still have the wrong driver?
As far as the kernel boot parameters, should I assume that it would be at the edit portion of grub?
I don't see any occurence of quiet in the grub lines
Last edited by mt_arch_user (2025-06-29 22:53:12)
Offline
Everything I can find is tyhat mesa-amber is the correct module for the AMD card.
How old is the card? In decades?
Still, no joy.
you can boot to multi-user instead using the kernel command line. https://wiki.archlinux.org/title/System … _boot_into
In any event we'll need some hard data from the system - if you didn't reboot w/ the power button, you can access older journals "sudo journalctl -b -δ", larger δ means older journal
Edit:
As far as the kernel boot parameters, should I assume that it would be at the edit portion of grub?
Yes.
Last edited by seth (2025-06-29 23:17:42)
Offline
Everything I can find is tyhat mesa-amber is the correct module for the AMD card.
How old is the card? In decades?
Not sure. I bought it at the same time I bought the nVidia card. It's pretty old.
Still, no joy.
you can boot to multi-user instead using the kernel command line. https://wiki.archlinux.org/title/System … _boot_into
In any event we'll need some hard data from the system - if you didn't reboot w/ the power button, you can access older journals "sudo journalctl -b -δ", larger δ means older journal
I had to power down the whole system to change the card. What's that character at the end of journalctl?
Edit:
As far as the kernel boot parameters, should I assume that it would be at the edit portion of grub?
Yes.
Let me see if I can get to a prompt. I should then be able to run <the command to recreate the startup file>
Offline
https://en.wikipedia.org/wiki/Delta_(letter) is meant as placeholder for a digit (1,2,3,4,5,…)
Not sure. I bought it at the same time I bought the nVidia card. It's pretty old.
What does the box say about the model?
Offline
It's a Sapphire card with an AMD Radeon R9 series 290X
It's a little less than 10 years old (!!!!)
When I power on, the bios display is good, the grub menu comes up as it should, but upon boot I get only the line that shows it checked the files and then just stops.
I tried adding the message systemd.unit=multi-user.target and the system crashed! Blue screen with a box to scan.
So, I have the nvidia card back in. Can I load the mesa-amber driver while it's running nvidia? Because looking at lsmod | grep mesa shows nothing, so that driver isn't being loaded. Or, would it not load as there's no AMD card installed?
Offline
mesa-amber isn't correct unless your card is older than r300, which is over 20 years old. I don't know where you found that information, but forget it.
Online
I reinstalled the nvidia card and logged in
I did
sudo pacman -S mesa-amber. pacman tole me it was reinstalling and when finished I got ::Running post transaction hooks... followed by Arming ConditionNeedsUpdate... A search on that message showed only conflicting information, so I rebooted.
I did
modprobe -n mesa-amberand was told Module mesa-amber not found in directory /lib/modules/6.12.35-1-lts
I opened tty3 (Ctrl-Alt-F3), logged in as root and did
find / -name mesa=amberThe only result was the license in usr/lib/licenses
I did
pacman -Q -i mesa-amberand received a page of information including the install date of today.
I did
pacman -Q -ikl mesa-amberand received a page of information including the install date of today. I got this list of files:
mesa-amber /usr/
mesa-amber /usr/include/
mesa-amber /usr/include/EGL/
mesa-amber /usr/include/EGL/eglextchromium.h
mesa-amber /usr/include/EGL/eglmesaext.h
mesa-amber /usr/include/GL/
mesa-amber /usr/include/GL/internal/
mesa-amber /usr/include/GL/internal/dri_interface.h
mesa-amber /usr/include/GL/osmesa.h
mesa-amber /usr/include/gbm.h
mesa-amber /usr/lib/
mesa-amber /usr/lib/dri/
mesa-amber /usr/lib/dri/i830_dri.so
mesa-amber /usr/lib/dri/i915_dri.so
mesa-amber /usr/lib/dri/i965_dri.so
mesa-amber /usr/lib/dri/kms_swrast_dri.so
mesa-amber /usr/lib/dri/nouveau_vieux_dri.so
mesa-amber /usr/lib/dri/r200_dri.so
mesa-amber /usr/lib/dri/radeon_dri.so
mesa-amber /usr/lib/dri/swrast_dri.so
mesa-amber /usr/lib/libEGL_amber.so
mesa-amber /usr/lib/libEGL_amber.so.0
mesa-amber /usr/lib/libEGL_amber.so.0.0.0
mesa-amber /usr/lib/libGLX_amber.so
mesa-amber /usr/lib/libGLX_amber.so.0
mesa-amber /usr/lib/libGLX_amber.so.0.0.0
mesa-amber /usr/lib/libGLX_indirect.so.0
mesa-amber /usr/lib/libOSMesa.so
mesa-amber /usr/lib/libOSMesa.so.8
mesa-amber /usr/lib/libOSMesa.so.8.0.0
mesa-amber /usr/lib/libgbm.so
mesa-amber /usr/lib/libgbm.so.1
mesa-amber /usr/lib/libgbm.so.1.0.0
mesa-amber /usr/lib/libglapi.so
mesa-amber /usr/lib/libglapi.so.0
mesa-amber /usr/lib/libglapi.so.0.0.0
mesa-amber /usr/lib/pkgconfig/
mesa-amber /usr/lib/pkgconfig/dri.pc
mesa-amber /usr/lib/pkgconfig/gbm.pc
mesa-amber /usr/lib/pkgconfig/osmesa.pc
mesa-amber /usr/share/
mesa-amber /usr/share/drirc.d/
mesa-amber /usr/share/drirc.d/00-mesa-defaults.conf
mesa-amber /usr/share/glvnd/
mesa-amber /usr/share/glvnd/egl_vendor.d/
mesa-amber /usr/share/glvnd/egl_vendor.d/50_amber.json
mesa-amber /usr/share/licenses/
mesa-amber /usr/share/licenses/mesa-amber/
mesa-amber /usr/share/licenses/mesa-amber/LICENSE
mesa-amber: 49 total files, 0 missing filesI did
insmod mesa-amberand was told ERROR: could not load module mesa-amber: No such file or directory.
I have come to the conclusion that mesa-amber is on the machine but the processes used to activate it are not aware of its location.
And that's way the system won't start with the AMD card installed.
So, how can I correct this?
Offline
1. forget about mesa-amber and reinstall mesa
2. We aren't talking about kernel modules here, insmod isn't going to do anything.
Online
mesa-amber isn't correct unless your card is older than r300, which is over 20 years old. I don't know where you found that information, but forget it.
OK. It's a R290. Pacman tells me there is no amdcpu and no radeon. So what is the correct module for this card?
Offline
1. forget about mesa-amber and reinstall mesa
2. We aren't talking about kernel modules here, insmod isn't going to do anything.
I'll have to remove mesa-amber, yes?
And do I need to do anything to get mesa to load/work/be installed?
Offline
No, it's R9 290, that is not R290.
It's a GCN 2.0 card, so it'll use radeon by default and you can switch it to AMDGPU. Those modules are part of the kernel package, they're not separate packages.
Reinstalling mesa will automatically remove mesa-amber.
Last edited by Scimmia (2025-06-30 02:34:54)
Online
No, it's R9 290, that is not R290.
It's a GCN 2.0 card, so it'll use radeon by default and you can switch it to AMDGPU. Those modules are part of the kernel package, they're not separate packages.
Reinstalling mesa will automatically remove mesa-amber.
OK Here goes!
Offline
Nope! Didn't work.
Is there something I need to do to make the system look for a new card before booting?
Last edited by mt_arch_user (2025-06-30 02:54:48)
Offline
have you restored your mkinitcpio.conf to how it was before you edited according to the wiki?
usually for nvid gpus you have to remove the kms hook and add the nvid modules - to restore that back the default remove the nvid modules and readd the kms hook - look at the wiki to figure out where to place it
regenerate your initrd afterwards
also: since the recent split of the linux-firmware package make sure to have linux-firmware-amdgpu installed
maybe you have to add
radeon.si_support=0 radeon.cik_support=0 amdgpu.si_support=1 amdgpu.cik_support=1to the kernel to enforce amdgpu over radeon - I'm not sure if this is required when linux-firmware-radeon is not installed
Offline
Nothing here makes sense. There aren't 'a lot of issue' with nvidia and wayland. I've never heard of 'atm' in relation to graphic cards. Ignoring pacman warnings when you don't know what they even said is bad.
I'm guessing since you're using grub, you have 'quiet' on the kernel command line? Get rid of that. If you're trying to boot to graphical mode, you can boot to multi-user instead using the kernel command line. https://wiki.archlinux.org/title/System … _boot_into
Nvidia drivers notoriously suck. That's probably what those people meant, especially for Linux. Same thing with Intel.
Last edited by PastelWeb_00s (2025-06-30 05:33:21)
Offline
First give us output to
uname -r
cat /etc/default/grub
Then try
sudo pacman -S mesa
sudo pacman -Syyu
sudo nano /etc/default/grub > add the following to kernel parameters
radeon.cik_support=0 amdgpu.cik_support=1
sudo update-grub
Please also make sure /etc/default/grub does not have nomodeset or vga= as a parameter
sudo nano /etc/mkinitcpio.conf
In modules do
MODULES=(amdgpu radeon) make sure to put it in this order, so >
MODULES=(amdgpu radeon, potentially other stuff here)
sudo nano /etc/mkinitcpio.conf
MODULES=(amdgpu radeon)
sudo mkinitcpio -p
Give us output of
zcat /proc/config.gz | grep -i CONFIG_DRM_AMDGPU_CIK=y
sudo pacman -S linux-firmware-amdgpu
Offline
@PastelWeb_00s Please use [code][/code] tags. Edit your post in this regard.
Also
pacman -Syyu
check the pacman manpage for what the second y does and then stop doing this unless you *really* mean to. You put undue load on the mirrors to jeopardize your pacman database integrity.
@mt_arch_user we will need hard data from the system
If you don't get the AMD GPU to work yourself, please add "systemd.unit=multi-user.target" to the https://wiki.archlinux.org/title/Kernel_parameters and if this still doesn't work also "nomodeset"
Then please post your complete system journal for the boot:
sudo journalctl -b | curl -F 'file=@-' 0x0.stThis entire mess is most likely for the display server to be hardcoded to use the nvidia GPU.
If you're running X11 and you've a file /etc/X11/xorg.conf - remove that resp. move it somewhere else (not! into /etc/X11/xorg.conf.d)
Incidentally:
grep -ri nvidia /{etc,usr/share}/X11/xorg.conf*you can also run this while using the nvidia card, if there's a static server config, it will show up.
Offline
@PastelWeb_00s Please use [code][/code] tags. Edit your post in this regard.
Alsopacman -Syyu
check the pacman manpage for what the second y does and then stop doing this unless you *really* mean to. You put undue load on the mirrors to jeopardize your pacman database integrity.
@mt_arch_user we will need hard data from the system
If you don't get the AMD GPU to work yourself, please add "systemd.unit=multi-user.target" to the https://wiki.archlinux.org/title/Kernel_parameters and if this still doesn't work also "nomodeset"
Then please post your complete system journal for the boot:sudo journalctl -b | curl -F 'file=@-' 0x0.stThis entire mess is most likely for the display server to be hardcoded to use the nvidia GPU.
If you're running X11 and you've a file /etc/X11/xorg.conf - remove that resp. move it somewhere else (not! into /etc/X11/xorg.conf.d)
Incidentally:grep -ri nvidia /{etc,usr/share}/X11/xorg.conf*you can also run this while using the nvidia card, if there's a static server config, it will show up.
Just making sure he has all the latest packages, I don't use Syyu all the time.
Offline
@seth
If you don't get the AMD GPU to work yourself, please add "systemd.unit=multi-user.target" to the https://wiki.archlinux.org/title/Kernel_parameters and if this still doesn't work also "nomodeset"
Then please post your complete system journal for the boot:
This did not work. Just to make sure I'm doing this right,
When I boot, I see the bios for the system, then grub comes up. I press e to edit.
In that popup window is what I changing. It says:
set params 'Arch Linux'
savedefault
load video
set gfxpayload=keep
insmod gzio
insmod part gpt
insmod ex2
set root='hd4,gpt3'
if [ x$feature platform search hint =xy ]; then
search --no-floppy --fs-uuid --set=root --hint-bios=hd4:gpt3 --hint-efi=hd4,gpt3 --hint-baremetal ahci4,gpt3 6a96a49d-32a3-4b03-9b8a-56d502647016
else
search --no-floppy --fs-uuid --set=root 6a96a49d-32a3-4b03-9b8a-56d502647016
fi
echo 'Loading Linux linux-lts ...
linux /boot/vmlinuz-linux-lts root=UUID=6a96a49d-32a3-4b03-9b8a-56d502647016 rw loglevel=3 quiet
echo 'loading initial ramdisk ...'
initrd /boot/intel-ucode.img /boot/initramfs-linux-lts.imgWhere exactly do I add the
systemd.unit=multi-user.target?
I reinstalled the nvidia card and ran the command you suggested.
The command:
grep -ri nvidia /{etc,usr/share}/X11/xorg.conf*returned nothing.
When booting with the nvidia card, I am using Wayland, not X11. That's what prompted this whole mess.
Wayland is just not working correctly at present. The mouse pointer grows to a huge size and then shrinks back down.
The items in the panel will not respond to the mouse until I first right click and then left click. The mouse pointer hangs when going from one screen to the other.
The word I get is X11 is going away and I'm stuck with Wayland, like it or not. So...
Offline