Framework Laptop 13 Ryzen 7040 BIOS 3.05 Release and Driver Bundle

framewwork laptop 13 ryzen bios 3.05 successfully updated yesterday via linux and fwupdmgr. the actual boot/POST with the update flashing process took quite a while but finished without errors. back into linux with 3.05 uefi firmware showing and present. thanks.

1 Like

I updated my Ryzen 7040 FW 13 to 3.05 (running Ubuntu 22.04) this weekend and it went well, but my system no longer stays asleep. Today I tried restarting the laptop, and without starting any programs I put it to sleep. It woke up by itself within 15 minutes. It is connected to a Dell monitor that provides power and USB connections to my mouse and keyboard, but this that’s how it was before 3.05 too. This issue seems to have been introduced by the upgrade to the 3.05 BIOS.

Is anyone else experiencing this?

Hey,
thanks for the update and support. I have updated the BIOS using fwupdmgr on Fedora 39 and there was no issue during the process. However, on Windows 11 I noticed that the Device Encryption option has vanished. I have so far checked the following things:

System Information
Device Encryption Support: Reasons for failed automatic device encryption: PCR7 binding is not supported, Un-allowed DMA-capable bus/device(s) detected, WinRE is not configured

PCR7 Configuration: Binding Not Possible

Secure Boot State: On

BIOS Mode: UEFI

powercfg /a
As far as I understand, PCR7 requires S0 to be supported, which is the case, confirmed by powercfg /a.

tpm.msc
Reports the TPM to be available.

BIOS settings
Secure Boot is enforced
TPM 1.1 is enabled and set to available

I have read that e.g. Lenove firmware updates disable Secure Boot during updating and sometimes do not re-enable it, but both the BIOS and Windows report Secure Boot to be enabled.

Has anyone else ran into this issue?

This sounds like the IRQ1 issue very likely. You can use a udev rule to turn off keyboard wakeup for now.

Thanks to you both for posting these. Worked well for me.

Btw: I’m on Ubuntu 22.04.4 and didn’t need to replace apt’s fwupd with a snap.

Apt shows its version as:
fwupd/jammy-updates,now 1.7.9-1~22.04.3 amd64 [installed]

Snap does show a newer version available:
$ snap find fwupd

Name Version Publisher Notes Summary
fwupd 1.9.16 richard-hughes - Firmware updates for Linux

Thought this may help other’s not sold on snap packages.

1 Like

@druneau thanks for posting this! What method did you use to collect your data and graph it?

I also perceived some sluggish fan ramp up after upgrading to bios 3.05.

This kind of thing is precisely why I like to benchmark my systems.

Running the same test again with the new bios appears to indicate lower CPU Temp at a cost of < 3% reduced throughput for both tests.

For anyone else experiencing this recently it doesn’t appear to be a result of the 3.05 BIOS. I reverted to the 3.03 BIOS and continue to have this problem.

I believe it must be a linux-kernel (6.8) or linux-firmware issue and I have filed a bug report with my distro for the problem Bug #2063143 “Frequent boot to black display” : Bugs : linux-meta package : Ubuntu

Might be worth booting with a prior kernel to see if that an acceptable workaround?

I have tried but unfortunately booting a prior kernel version just hangs even after downgrading linux-firmware to an (what I think is) acceptable version.

Not really sure what my next steps would be besides just wait and see what Ubuntu responds with and hope that some users here begin to see the issues I am having once Ubuntu Noble releases very very soon.

I used an ectool binary from the fw-fanctrl (never used those scripts) in order to run the command

ectool pwmgetfanrpm

in a 1 sec loop. Also gathered average cpu clock speeds with this command (thanks chat gpt lol)

awk '{ sum += $0 } END { print int((sum / NR) / 1000 + 0.5) }' /sys/devices/system/cpu/cpu*/cpufreq/scaling_cur_freq

Then just used google sheet to create a chart. A chart with bad tickers on x axis… not very usefull.

It looks like you were right! The udev rule described here seemed to do the trick for me: Framework Laptop 13 Ryzen 7040 BIOS 3.05 Release and Driver Bundle - #14 by dimitris.

Thanks for the suggestion, @Mario_Limonciello!

2 Likes

That’s rough. Hope you find a tenable solution pronto.

Not sure if this is your daily driver. I’ve booted from an usb drive to have a running system and access my files while dealing with issues.

I’ve found this line is a good balance of performance/cost/space/size.

Thanks, I was checking out ectool & fw-fanctrl, however disabling secure boot breaks company policy at my job, so I haven’t run them.

I would be ecstatic if FW provided fan speed data via the next BIOS update.

It’s cool how both Asus and Gigabyte’s systems reveal this via the it87 module.

We investigated this a while back, and fan speed reporting was not part of the ACPI specification. However it looks like this has been added recently with _FST!
https://uefi.org/htmlspecs/ACPI_Spec_6_4_html/11_Thermal_Management/fan-device.html?highlight=fan

We will investigate if we can add this at some point in the future.

For now if you do want fan speed, we recommend grabbing this data using ectool, etc.

4 Likes

If you’re allowed custom modules (which requires a custom signing key with secure boot), I recently added fan control into Dustin’s framework-laptop-kmod.

Hey Kieran, if you guys do get a chance to modify the ACPI thermal stuff, would you be able to lower the thermal trip points to something below 175°C (448K)? Linux kernel 6.8 and above doesn’t like it otherwise.

Kernel bug report

Edit: Forum Thread: [RESPONDED] Firmware bug: ACPI table error causes missing sensors on Linux 6.7+ - #8 by Quentin

If the computer belongs to your company, better not to mess with it too much. If the computer is yours, it’s a red flag that your company is effectively doing long-arm control over your own device

Sure! We will review this issue to see if we need to fix it in our next BIOS release.

3 Likes

In addition to fan speed data in the BIOS, I would love to have the ability to set, at the very least, a fan speed minimum in the BIOS, which can go up from there, of course. My preference is to use a fan at low RPM continuously to keep things a bit cooler whenever possible. I used to do this on my mac when it was supported via a third party app, but that’s not possible any longer to my knowledge.

The third party app for framework is fw-fanctrl