Any update on a UEFI shell updater tool for us Linux folk?
A couple of friendly reminders:
The “I lost my boot options” is easily mitigated by having a valid .efi stub in the fallback location.
In my case it’s $ESP/EFI/BOOT/BOOTX64.EFI, or literally /boot/efi/EFI/BOOT/BOOTX64.EFI . I just copied my normal rEFInd install .efi to that location. Then, on the first boot after the BIOS update obliterates the boot options, I use the “fallback rEFInd” (that doesn’t have my custom options associated with it) to chain load the real rEFInd location, and into my main partition I boot. Reinstall rEFInd over itself in my OS (just to restore the BIOS boot options memory), then I’m all back to normal. No separate boot device needed.
Also, the Mediatek powered RZ608 (mt7921k) and RZ616 (mt7922) are not affected by Intel’s garbage LAR bugs. 6ghz is available (marked no IR in iw list but usable) for me on my RZ608 using in kernel tree modules + linux-firmware package on arch. See GitHub - morrownr/USB-WiFi: USB WiFi Adapter Information for Linux for more details. Agreed that Framework really should fix this regression to an advertised feature that was available at 11th Gen launch. But it’s a ~$20 fix to go to an RZ616 if it’s really eating you up to not have 6ghz on Linux.
I have a bug to report - unless someone tells me it is something else - I just installed 3.19 in my 11 gen i5. Install was effortless on windows - but now when I click any of the shift, ctl, fn, win, or alt keys there would be a beep (not from the speakers but the crappy thing on the mobo that makes post noises. Repeat clicks would do two notes of the same pitch followed by 1 note of a different pitch. And after about 25 total noises the sound has now gone away. Anyone else have this?
I have rebooted and it persists, I can upload a video if that helps. This is a weird one!
Wild guess based on quite dated experiences with Windows: did sticky keys get enabled somehow? I recall that in Win7, pressing shift five times in a row would enable it, you’d get some BIOS beeps of various tones as you pressed modifier keys, and if you pressed shift five more times in a row, it would disable.
Please do that.
Does the same happen in the BIOS?
I updated my Linux-only system with the Ventoy+Win10 installer path, which was successful, but it looks like the Intel CSME did not update (fwupdmgr security
warning is still showing the old version).
I was thinking about running it again, but I’m a bit hesitant without some explanation about why it didn’t update first time.
SOLVED Thanks all - the sticky keys was the culprit - I’ve been using windows since Windows 3 and have never turned that on before! Learn something every day…
I installed the update through my hirens usb tool. It updated the EC and the bios version but did not update the intel CSME version like @Jeremy_Fitzhardinge above. I use Arch Linux so the only other option would be to wait for the UEFI shell installer to come out since it won’t be available via lvfs. After update it booted just fine and all ports seem to work just fine as well.
Is there a process to install the BIOS update without having a battery connected? I have the case that Framework sells (cooler master), but as it doesn’t have a battery, how is Framework supporting their other device setups?
Need to do a check with the community - anyone that updated 11th Gen Intel Core BIOS 3.19 Beta Release to this seeing issues at all?
If so, please list them as follows:
- Updating from which previous BIOS release? 3.17
- How did the update go, any issues experienced? Can’t update as I’m using the cooler master case from Framework
- Any bugs experienced (not feature requests, bugs)? Battery “requirement” halts from applying bios & CVE fixes, once that’s addressed and I can provide more
As I stated above, I updated from 3.17 to 3.19 using the hirens rescue disk and not a full windows install. It updated the bios but didn’t update the CSME to the new version. everything works just as it did before the update. It should also be noted that 3.19 is no longer in beta status but is in full release state.
Looking through this thread, I see several reports of installations by Linux users where CSME didn’t update, and no reports where it did, but it may be that in cases where it did, people aren’t posting, so:
Has anyone found a method to install this update, without having a full Windows installation, where CSME does update correctly?
And with a full Windows installation, does it update correctly?
Has anyone found a method to install this update, without having a full Windows installation, where CSME does update correctly?
Is there any method without a Windows installation?
So here’s what I did to update mine:
I extracted the isflash.bin
file (3.19 EFI) from the windows zip file with
7z x Framework_Laptop_11th_Gen_Intel_Core_BIOS_3.19.exe
I have a directory on my EFI partition that has a bootx64.efi so I can get to an EFI shell and other files:
# ls -l /boot/efi/EFI/Insyde/
total 74052
-rwx------ 1 root root 951744 Jul 31 2022 bootx64.efi
-rwx------ 1 root root 1892992 Oct 6 22:53 FwUpdLcl.efi
-rwx------ 1 root root 1574208 Jul 29 2022 H2OFFT-Sx64.efi
-rwx------ 1 root root 35699128 Oct 6 22:22 hx20.3.19.bin
-rwx------ 1 root root 35699096 Jul 29 2022 hx20_capsule_3.10.bin
I had to get FwUpdLcl.efi from elsewhere but you can maybe find it online, once you have told grub to boot from your partition, eg:
insmod part_gpt
insmod ext2
insmod ntfs
insmod chain
set root=(hd0,1)
chainloader (hd0,1)/efi/Insyde/bootx64.efi
(or the correct partition) for your system - then you can
fs0:
cd efi\insyde\
H2OFFT-Sx64.efi isflash.bin
where the isflash.bin is the file you extracted from the exe
I didn’t verify this updates the ME yet, but wanted to share this, this didn’t require me to do anything weird with grub etc if you just copy the files to your /boot/efi space
I did toss the set together in a zip here:
https://puck.nether.net/~jared/Framework_Intel_BIOS.3.19.zip
If you grab the FWUpdate.bin from the exe as well I believe that’s the ME firmware and you could run that with FwdUpdLcl.efi to update your ME by rebooting into the EFI CLI shell again
ETA: The FwdUpdLcl.efi that I have is for Alder Lake in that zip, so won’t work, but if you don’t have one in the directory it won’t complete the update
Also: you need to boot the updater (H2OFFT-Sx64.efi
) directly before you boot the bootx64.efi, it will error but will then let you run the bootx64.efi - don’t leave grub
Welcome to the community @Jared_Mauch1 and thanks for the write up.
It’s been over a month since this BIOS update has been released, and already out of beta. But we have people from the community reverse engineering the windows update so they can update their laptops.
I appreciate that framework has stood behind their 11th gen motherboards and kept releasing these updates. But why would it take so long to re-package this to EFI? If feels like people having to hack on their laptops to release a vendor provided patch is not optimal when the same vendor has managed to release multiple linux compatible updates already.
End of rant. Keeping things positive, I think this is a superb product from a great company with awesome support. Griping about bios update packaging instead of them existing at all is clearly a first world problem. This is framework delivering on their promise of lowering e-waste by keeping the 11th gen motherboards alive.
I think the problem is that the Firmware team at Framework is still small, and while they were able to get the windows update out, they’ve now had to pivot all their time to the firmware issues on the AMD systems that are about to launch. They have made it clear that the EFI update for 3.19 is still important, and they will get to it soon, but the new product that is about to release has become a higher priority than the 1st gen product that works fine on BIOS 3.17 for most people, and that CAN be updated to 3.19 with a workaround if needed.
I agree that it is bothersome to not have the easy method of doing this available, but the Framework team is doing exceptionally at supporting all of their hardware despite the generation, which I think is highly commendable.
Completely agree with you. I think they should even shutdown / pause those forum threads…before they get boiling.
What they did with 12th gen bios update.
They just have no answer, discussing “when will it arrive” just make people angry.
Let it be
I would have to disagree on that one with you. I think it’s important to keep these threads open for people to report any BIOS issues so that Framework is aware of issues.
Not necessarily, the reasons why I believe people got mad is because there isn’t any official explanation from Framework and there isn’t regular communication. A lot of the tools that Framework’s partners use are Windows based such as the HDMI and DisplayPort updater. Framework even had to ask for permission to release them to the community. So from that, porting things over the Linux can take time.
I agree with Azure’s statement that Framework is a small team and given the resources they have, it’s a careful balancing act between supporting older products and ensuring the new products have their best foot forward. I think it was ultimately a tough but correct decision for the company to prioritize the AMD launch. But this comes back to the issue of lack of communication. Framework hasn’t explicitly mentioned that is the case so all this is just speculation.
Honestly, having some kind of monthly or regular post on software updates and changes in focus would alleviate a lot of tension in the community. Looking at the 12th BIOS update thread, while some may say the 1-2 month frequency of official responses is regular, I would slightly disagree as it’s out of the blue. No one knows when they’re going to happen.
For a small team, a lot of things will move slowly and that’s just the nature of it. There’s just not enough resources to go around. But I think that’s what makes regular and predictable communication so much more important to the community and vital to maintaining community sentiment.