black screen with “AMD RADEON RX 6700 XT”

RES

New Member
Joined
Jan 12, 2023
Messages
6
Reaction score
4
Credits
52
Hi.

I installed a new Fedora 37 on a new system with an “AMD RADEON RX 6700 XT”.

Since the installation software cannot handle the graphics card, the installation was carried out with the reduced graphics. When the system is installed, “nomodeset” is then automatically set as a kernel parameter.

After the installation, I performed all updates and checked whether all the necessary packages (firmware, mesa, …) were installed. Although they appear to be, as soon as “nomodeset” is disabled, the screen turns off when the system boots.

So there seems to be a problem with the firmware or the driver in the kernel.

Does anyone have an idea how to solve the problem?
 


OP
RES

RES

New Member
Joined
Jan 12, 2023
Messages
6
Reaction score
4
Credits
52
I also made the following observation:

As soon as the system runs “modprobe amdgpu”, the signal to the screen is interrupted, the GPU logo illumination and the GPU fans are disabled. After a few seconds, the lighting of the GPU logo is activated again, but the GPU fans remain off.

Does anyone have any information on how the GPU logo lighting and the GPU fans normally behave when “modprobe amdgpu” is run?
 

f33dm3bits

Gold Member
Gold Supporter
Joined
Dec 11, 2019
Messages
5,306
Reaction score
3,925
Credits
38,697
Before my current graphics card I was using an RX 6700XT as well but on Fedora 36, I didn't install any drivers just the opensource driver that was provided by the kernel. So I would think if it worked on Fedora 36 that graphics card should also still work on Fedora 37. Some dmesg output or syslog out put would be useful. The package xorg-x11-drv-amdgpu shouldn't be needed but you can try installing it and then rebooting to see what happens.
 
Last edited:

osprey

Well-Known Member
Joined
Apr 15, 2022
Messages
414
Reaction score
334
Credits
3,819
That's significant info f33dm3bits. I'm not running fedora at the moment, but also had no trouble with earlier radeon cards, but also had success with AMD's drivers. My query now would be: Is the firmware-amd-graphics installed?
 

f33dm3bits

Gold Member
Gold Supporter
Joined
Dec 11, 2019
Messages
5,306
Reaction score
3,925
Credits
38,697
It would ben an idea to run dmesg to see if that returns any output.
Code:
dmesg | egrep "amdgpu|firmeware"
inxi -G
 
OP
RES

RES

New Member
Joined
Jan 12, 2023
Messages
6
Reaction score
4
Credits
52
The problem already occurs when the "amdgpu" module is loaded. So before loading a graphical user interface that could use other drivers.

Partial content of "/var/log/messages":
Jan 9 20:20:52 resws4 kernel: [drm] amdgpu kernel modesetting enabled. Jan 9 20:20:52 resws4 kernel: amdgpu: Ignoring ACPI CRAT on non-APU system Jan 9 20:20:52 resws4 kernel: amdgpu: Virtual CRAT table created for CPU Jan 9 20:20:52 resws4 kernel: amdgpu: Topology: Add CPU node Jan 9 20:20:52 resws4 kernel: amdgpu 0000:0b:00.0: vgaarb: deactivate vga console Jan 9 20:20:52 resws4 kernel: amdgpu 0000:0b:00.0: enabling device (0006 -> 0007) Jan 9 20:20:52 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: Fetched VBIOS from VFCT Jan 9 20:20:52 resws4 kernel: amdgpu: ATOM BIOS: 113-67XA6SSB1-D01 Jan 9 20:20:52 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: Trusted Memory Zone (TMZ) feature disabled as experimental (default) Jan 9 20:20:52 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: VRAM: 12272M 0x0000008000000000 - 0x00000082FEFFFFFF (12272M used) Jan 9 20:20:52 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: GART: 512M 0x0000000000000000 - 0x000000001FFFFFFF Jan 9 20:20:52 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: AGP: 267894784M 0x0000008400000000 - 0x0000FFFFFFFFFFFF Jan 9 20:20:52 resws4 kernel: [drm] amdgpu: 12272M of VRAM memory ready Jan 9 20:20:52 resws4 kernel: [drm] amdgpu: 16003M of GTT memory ready. Jan 9 20:20:52 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: PSP runtime database doesn't exist Jan 9 20:20:52 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: PSP runtime database doesn't exist Jan 9 20:20:54 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: STB initialized to 2048 entries Jan 9 20:20:54 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: Will use PSP to load VCN firmware Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: RAS: optional ras ta ucode is not available Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: SECUREDISPLAY: securedisplay ta ucode is not available Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: smu driver if version = 0x0000000e, smu fw if version = 0x00000012, smu fw program = 0, version = 0x00413900 (65.57.0) Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: SMU driver if version not matched Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: use vbios provided pptable Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: SMU is initialized successfully! Jan 9 20:20:55 resws4 kernel: kfd kfd: amdgpu: Allocated 3969056 bytes on gart Jan 9 20:20:55 resws4 kernel: amdgpu: sdma_bitmap: ffff Jan 9 20:20:55 resws4 kernel: amdgpu: HMM registered 12272MB device memory Jan 9 20:20:55 resws4 kernel: amdgpu: SRAT table not found Jan 9 20:20:55 resws4 kernel: amdgpu: Virtual CRAT table created for GPU Jan 9 20:20:55 resws4 kernel: amdgpu: Topology: Add dGPU node [0x73df:0x1002] Jan 9 20:20:55 resws4 kernel: kfd kfd: amdgpu: added device 1002:73df Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: SE 2, SH per SE 2, CU per SH 10, active_cu_number 40 Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: ring gfx_0.0.0 uses VM inv eng 0 on hub 0 Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: ring comp_1.0.0 uses VM inv eng 1 on hub 0 Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: ring comp_1.1.0 uses VM inv eng 4 on hub 0 Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: ring comp_1.2.0 uses VM inv eng 5 on hub 0 Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: ring comp_1.3.0 uses VM inv eng 6 on hub 0 Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: ring comp_1.0.1 uses VM inv eng 7 on hub 0 Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: ring comp_1.1.1 uses VM inv eng 8 on hub 0 Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: ring comp_1.2.1 uses VM inv eng 9 on hub 0 Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: ring comp_1.3.1 uses VM inv eng 10 on hub 0 Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: ring kiq_2.1.0 uses VM inv eng 11 on hub 0 Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: ring sdma0 uses VM inv eng 12 on hub 0 Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: ring sdma1 uses VM inv eng 13 on hub 0 Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: ring vcn_dec_0 uses VM inv eng 0 on hub 1 Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: ring vcn_enc_0.0 uses VM inv eng 1 on hub 1 Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: ring vcn_enc_0.1 uses VM inv eng 4 on hub 1 Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: ring jpeg_dec uses VM inv eng 5 on hub 1 Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: amdgpu: Using BACO for runtime pm Jan 9 20:20:55 resws4 kernel: [drm] Initialized amdgpu 3.48.0 20150101 for 0000:0b:00.0 on minor 0 Jan 9 20:20:55 resws4 kernel: amdgpu 0000:0b:00.0: [drm] Cannot find any crtc or sizes
 
OP
RES

RES

New Member
Joined
Jan 12, 2023
Messages
6
Reaction score
4
Credits
52
That's significant info f33dm3bits. I'm not running fedora at the moment, but also had no trouble with earlier radeon cards, but also had success with AMD's drivers. My query now would be: Is the firmware-amd-graphics installed?

Yes:

| amd-gpu-firmware-20221214-145.fc37.noarch
 

f33dm3bits

Gold Member
Gold Supporter
Joined
Dec 11, 2019
Messages
5,306
Reaction score
3,925
Credits
38,697
Did this happen before you updated your system with the most recent kernel? Did this happen while you were live booted into Fedora 37 before installing? If not it might worth trying to boot into an older kernel or too live boot from Fedora 36 or 37 to check what kernel you have there and then try installing that kernel on your current system.
 
OP
RES

RES

New Member
Joined
Jan 12, 2023
Messages
6
Reaction score
4
Credits
52
Did this happen before you updated your system with the most recent kernel? Did this happen while you were live booted into Fedora 37 before installing? If not it might worth trying to boot into an older kernel or too live boot from Fedora 36 or 37 to check what kernel you have there and then try installing that kernel on your current system.

Yes, the problem also occurred with the first installed kernel. Yes, the problem also occurs with the "live-system" which is started via USB, so "nomodeselect" must be set in order for the "live-system" to be used.

I could try to start a "live-system" with Fedora 36 on the machine tomorrow.
 

osprey

Well-Known Member
Joined
Apr 15, 2022
Messages
414
Reaction score
334
Credits
3,819
How is the card connected to the monitor, hdmi, vga, dmi etc.? It may or may not matter, but following the
dmesg output line:
Code:
[drm] Cannot find any crtc or sizes
lead to this rather lengthy bug investigation: https://lore.kernel.org/all/[email protected]//T/.

It's of an earlier kernel, but two items of interest emerged in kernel parameters:
Code:
amdgpu.dc=0
and
Code:
amdgpu.cik_support=1
where one repondent had success with the first option alone, and another respondent had success using both options on the kernel command line. Maybe worth a try.
 

f33dm3bits

Gold Member
Gold Supporter
Joined
Dec 11, 2019
Messages
5,306
Reaction score
3,925
Credits
38,697
It sounds more like driver problem than a firmware problem though, if it was a firmware issue the card wouldn't probably even initialize which would cause your graphic environment to fail to start. Can you also share the output of the following?
Code:
lspci | grep -i vga
inxi -G
 

osprey

Well-Known Member
Joined
Apr 15, 2022
Messages
414
Reaction score
334
Credits
3,819
f33dm3bits wrote:
if it was a firmware issue the card wouldn't probably even initialize which would cause your graphic environment to fail to start

AIUI if the proprietary firmware is not installed, a generic open source firmware is used by the kernel, which is sub-optimal, but doesn't necessarily cause the graphics card to totally fail, rather just limit it, in some cases severely to just text, and differently in other cases. In any case though, unless the user is concerned to only use open source firmware, then the proprietary firmware is best installed, so checking for it makes sense.
 

f33dm3bits

Gold Member
Gold Supporter
Joined
Dec 11, 2019
Messages
5,306
Reaction score
3,925
Credits
38,697
Even though AMD graphics cards have opensource drivers the firmware is still closed source, the firmware needed for that card is the "navy_flounder_sos.bin".
That is available on both Fedora 36 and Fedora 37.
Code:
amd-gpu-firmware-20221214-145.fc36.noarch : Firmware for AMD GPUs
Repo        : updates
Matched from:
Filename    : /usr/lib/firmware/amdgpu/navy_flounder_sos.bin.xz

amd-gpu-firmware-20221214-145.fc37.noarch : Firmware for AMD GPUs
Repo        : updates
Matched from:
Filename    : /usr/lib/firmware/amdgpu/navy_flounder_sos.bin.xz
So if the "amd-gpu-firmware-20221214-145" packages is installed(and that package is installed by default) the needed firmware is available on the system for that graphics card to work. But I agree with you @osprey is worth a try installing the closed source AMD drivers. I'm just not a fan of using the closed source drivers for a graphics card that I know works with the opensource drivers(even for gaming) and I know someone else that has an RX 6900 and it is working fine with the opensource drivers on Fedora 37.
 
Last edited:
OP
RES

RES

New Member
Joined
Jan 12, 2023
Messages
6
Reaction score
4
Credits
52
How is the card connected to the monitor, hdmi, vga, dmi etc.? It may or may not matter, but following the
dmesg output line:
Code:
[drm] Cannot find any crtc or sizes

That pointed the way in the right direction.

The problem got solved:

The display was connected to the graphics card with a DVI cable through a DVI/HDMI adapter. Apparently "amdgpu" had a problem recognizing the screen (correctly) with this configuration, which led to the error message "[drm] Cannot find any crtc or sizes".

I have now connected another display directly to DVI and now everything works as it should.

Thank you very much for your help, your efforts and your time.
 

f33dm3bits

Gold Member
Gold Supporter
Joined
Dec 11, 2019
Messages
5,306
Reaction score
3,925
Credits
38,697
Glad you got it solved!
 
$100 Digital Ocean Credit
Get a free VM to test out Linux!

Linux.org Hosting Donations
Consider making a donation

Staff online


Top