Already on it 
Iāll post back in a few days with any findings.
But it seems promising so far, my easy reproduction of changing gnomeās scaling factor did not trigger the flickering.
Also very promising, is this being the only error I see in dmesg:
cros_ec_lpcs cros_ec_lpcs.0: EC ID not detected
The touchpad is still laggy though, Iām starting to think thatās just the touchpadās refresh rate not being proportional to 60hz.
I donāt see a laggy touchpad, maybe this is indicative of another issue? Check whether something is pegging a CPU, like a constant interrupt.
Sorry should have been clear I donāt associate the touchpad with anything on AMDās side.
I actually attributed it to the i2c errors in dmesg on <=3.03.
This is more speaking for the topic of the 3.05b update and no longer seeing those i2c errors.
Both the flickering, and the i2c errors Iāve suspected were bios / IOMMU issues so I was very eager a week ago when they announced the beta.
Installed the BIOS and drivers on Windows 11 without issue.
Installed the new BIOS and the driver bundle under Windows 11 on my 7840U FW13, and now the internal display turns off at the login screen. The splash boot screen displays fine, then it goes blank, and then a few seconds later the backlight turns off.
Iām able to connect external displays and use the machine, but the laptop display remains dark (no backlight on) and even though Windows detects the internal display (as āIntegrated Monitorā) in Device Manager, it does not appear as an available display for extension or mirroring in Display settings.
troubleshooting steps taken:
This is clearly a windows/driver issue, as the display works fine in while booting, when in safe mode, and when booted from a live linux USB disk.
Please help!
EDIT: rolling back the display driver fixes the issue. I have opened a support ticket and will post the outcome of that here.
EDIT 2: Running DDU and reinstalling the new driver from scratch has fixed the problem. Thank you Zeke and the other support staff!
I installed the update with the EFI updater. Iām running Arch Linux with kernel 6.8.2.
No problems so far, dmesg looks good.
But, the first boot-up after the efi update, only 512MB ram was reserved for the iGPU, even though I had configured UMA_GAME mode in the bios. Hmm did it reset? Checking the bios config, no, it was correctly configured there, as were the all the rest of my changes. I did notice that the name and help text for this option were greatly improved! Itās now called āiGPU memoryā and explains how much is assigned in āautoā vs āgameā mode, depending on total memory installed. Nice. So ā¦ I toggled it to auto, and back to game, saved and booted, and now itās allocated 4GB as expected out of my 32GB total. So was it a first boot problem, and a plain reboot would have fixed it, or did toggling and re-saving the setting fix it? Iām not sure ā¦
Update worked for me after trying it with two old pendrives that failed.
I used the microSD expansion card and a fresh microSD instead.
It took about 15 minutes and required mains connected.
Iām in for the Beta especially because of the new password regulations.
And it works like a charm! Thanks!
also successfully updated bios with efi
Game mode is set for iGPU and now the available ram is 29 GB. I am curious since it isnāt quite the 4GB the bios claims to allocate out of my 32ā¦
I got 27.2GiB(read from btop)RAM and 4005M VRAM(read from radeontop), where did you read 29GB
I use Pop!_OS, thereās a built-in āsystem monitorā
edit:
top gives me MiB mem: 27868.7 total.
Think of Z States a bit like Intel C States. They are particular modes a processor can drop down into that disable features not necessary at a certain time for power management.
Please take anything below this with a grain of salt as my understanding is very far from good.
AMD 6000 series and up (as of writing this 06/04/2024) have two Z states:
This is almost certainly an oversimplification, but in one sentence, Z states are a way of radically reducing how much power your processor is drawing. If Iāve got anything wrong, please let me know.
I got an error trying to upgrade through the Gnome Firmware application. The update is listed after I ran fwupdmgr enable-remote lvfs-testing, but when pressing the upgrade button, I get:
Failed to create ā/boot/EFI/UpdateCapsuleā: Read-only file system
Might be a Gnome issue, though.
Solution:
Follow the steps outlined here: fwupd - ArchWiki
In my case sudo mkdir /boot/EFI, and then restarting fwupd.service.
I have just upgraded and I am still getting:
ucsi_acpi USBC000:00: ucsi_handle_connector_change: ACK failed (-110)
after resuming. Nothing is even connected to the laptop (except for the modules: 2ĆUSB-C, 1ĆHDMI, 1xMicroSD).
How do I disable sg_display=0? Will āsudo grubby --update-kernel=ALL --args=āamdgpu.sg_display=1āā do the job? Iām really new to this
I havenāt experienced the white screen bug since Iāve re-enabled scatter/gather and disabled game mode
Can you please tell me how can I re-enable it? Thank you
Wrong thread, but that should do the job. Better yet is to not set amdgpu.sg_display at all, and remove it from your configuration. However, I do not use GRUB, so Iām not sure how to do that.
If memory serves, you could just edit the file /etc/default/grub, remove all references to amdgpu.sg_display there, and finally regenerate using sudo grub-mkconfig -o /boot/grub/grub.cfg.
I think this is the command youāre looking for:
sudo grubby --update-kernel=ALL --remove-args=āamdgpu.sg_display=0
Iāve had a couple BSODs when resuming from standby. I assume this is likely related to the IRQ1 known issue, since I didnāt have any such issues on the previous 3.0.3 EC fix beta.
BSOD is āCritical process diedā.
After about a dayās worth of use, the only alarming issue I see is that dmesg is spammed with
xhci_hcd 0000:c1:00.3: Refused to change power state from D0 to D3hot
Is this the IRQ1 failure mentioned in the known issues?