Custom Fedora OCI images for Framework laptops

[Matt_Hartley]

@Aman_Chhabra

I wanted to share that I have been running my own custom Kinoite image for the Framework 13 AMD. It’s still a work in progress, but it is at a pretty decent state thus far. This is not meant to be an official distribution per se, but thought I would share what I did with the group in case it’s helpful. It is based off of the awesome work that the universal blue team has done creating the Kinoite-main image. Huge thanks to Jorge for this project.

Awesome! Appreciate the share!

[Jorge_Castro]

@Aman_Chhabra

I can get ~3.5-4 watts at idle, ~6.3 watts during offline video playback. And ~5-8 watts web browsing.

Hah this is so awesome! Thanks for sharing this!

[klaus_k]

got bluefin up and running on my FW13 AMD - really like it so far!
the only thing that is not working out of the box is the fingerprint scanner. is there anything special to do on the immutable system?

downloaded the cab file from: Updating Fingerprint Reader Firmware on Linux for 13th Gen and AMD Ryzen 7040 Series Laptops

tried to install it with fwupdtool, but it fails:

 sudo fwupdtool install --allow-reinstall --allow-older goodix-moc-609c-v01000330.cab 1e8c8470-a59c-571a-82fd-19c9fa32b9c3
Laden …                  [********                               ]11:58:06.743 FuEngine             failed to add device /sys/devices/pci0000:00/0000:00:08.1/0000:c1:00.0: ioctl error: Ungültige Adresse [14]
Laden …                  [***********                            ]11:58:08.722 FuUdevDevice         fu_udev_device_get_sysfs_path: assertion 'FU_IS_UDEV_DEVICE(self)' failed
11:58:08.722 FuDevice             fu_device_add_parent_backend_id: assertion 'backend_id != NULL' failed
Laden …                  [************                           ]11:58:08.723 FuUdevDevice         fu_udev_device_get_sysfs_path: assertion 'FU_IS_UDEV_DEVICE(self)' failed
11:58:08.723 FuDevice             fu_device_add_parent_backend_id: assertion 'backend_id != NULL' failed
11:58:08.723 FuUdevDevice         fu_udev_device_get_sysfs_path: assertion 'FU_IS_UDEV_DEVICE(self)' failed
11:58:08.723 FuDevice             fu_device_add_parent_backend_id: assertion 'backend_id != NULL' failed
11:58:08.723 FuUdevDevice         fu_udev_device_get_sysfs_path: assertion 'FU_IS_UDEV_DEVICE(self)' failed
11:58:08.723 FuDevice             fu_device_add_parent_backend_id: assertion 'backend_id != NULL' failed
11:58:08.723 FuUdevDevice         fu_udev_device_get_sysfs_path: assertion 'FU_IS_UDEV_DEVICE(self)' failed
11:58:08.723 FuDevice             fu_device_add_parent_backend_id: assertion 'backend_id != NULL' failed
11:58:08.723 FuUdevDevice         fu_udev_device_get_sysfs_path: assertion 'FU_IS_UDEV_DEVICE(self)' failed
11:58:08.723 FuDevice             fu_device_add_parent_backend_id: assertion 'backend_id != NULL' failed
11:58:08.723 FuUdevDevice         fu_udev_device_get_sysfs_path: assertion 'FU_IS_UDEV_DEVICE(self)' failed
11:58:08.723 FuDevice             fu_device_add_parent_backend_id: assertion 'backend_id != NULL' failed
Laden …                  [************                           ]11:58:08.723 FuUdevDevice         fu_udev_device_get_sysfs_path: assertion 'FU_IS_UDEV_DEVICE(self)' failed
11:58:08.723 FuDevice             fu_device_add_parent_backend_id: assertion 'backend_id != NULL' failed
Schreiben …              [*******************                    ]
Ausgewähltes Gerät: Fingerprint Sensor
Entpacken …              [ -                                     ]
generic GUID requires a CHID, child, parent or sibling requirement

[Mario_Limonciello]

@klaus_k

got bluefin up and running on my FW13 AMD - really like it so far!
the only thing that is not working out of the box is the fingerprint scanner. is there anything special to do on the immutable system?

The problem is the CAB file doesn’t express the requirements appropriately for which systems it can apply to. This is a requirement with newer fwupd versions. Framework needs to respin it.

See Fedora 39 Kinoite Fingerprint Reader Not Working - #3 by Mario_Limonciello for more details.

[Matt_Hartley]

After the last beta test with LVFS went sideways as I lacked the same input kits found on 13th gen and AMD Ryzen configs, I had to rely on the community…then was drawn into other more immediate issues.

Refocusing on this now. Request for input kits matching what current users are getting, will be testing our our LVFS beta (after some internal changes are done). It’s on the list, I promise!

[RiverShen]

@Aman_Chhabra

I wanted to share that I have been running my own custom Kinoite image for the Framework 13 AMD. It’s still a work in progress, but it is at a pretty decent state thus far. This is not meant to be an official distribution per se, but thought I would share what I did with the group in case it’s helpful. It is based off of the awesome work that the universal blue team has done creating the Kinoite-main image. Huge thanks to Jorge for this project.

Very cool! Definitely get significant power savings using this, but the epp_performance command errors out, and with epp_balance_performance or lower (maybe epp_performance too, but can’t verify) I get random mouse dropout while gaming. Not sure what could cause this, but had to switch back to ublue main. Still a great image to rebase to whenever I need a full day of productivity without a recharge. Thanks for sharing, and I’d bet if this image gets more attention from the knowledgeable community here to polish stuff like that, it could become the defacto AMD framework image.

[Aman_Chhabra]

@RiverShen

Very cool! Definitely get significant power savings using this, but the epp_performance command errors out, and with epp_balance_performance or lower (maybe epp_performance too, but can’t verify) I get random mouse dropout while gaming. Not sure what could cause this, but had to switch back to ublue main. Still a great image to rebase to whenever I need a full day of productivity without a recharge. Thanks for sharing, and I’d bet if this image gets more attention from the knowledgeable community here to polish stuff like that, it could become the defacto AMD framework image.

Thanks for sharing your results. I will have to look into why the epp_performance setting isn’t working for you. Regarding the mouse maybe it’s due to the usb autosuspend configurations? I’m not sure I haven’t run into any issues with that myself yet.

But also wanted to give you a heads up that you shouldn’t need to manually set the epp settings anymore. I created a COPR with a patched power-profiles-daemon which is now integrated into the image. So when you select power saving / balanced / performance it should automatically set the epp setting for you (no need to use the just script anymore). I am planning on doing a more detailed post / tutorial for the community.

[Jorge_Castro]

@RiverShen

the knowledgeable community here to polish stuff like that, it could become the defacto AMD framework image.

Yes! I’m glad people are finding the power of the fork! I’m on an Intel Framework so I can’t take lead on this, but if we figure it out and get the fixes in then we can just slipstream them in. If you wanna take lead we could set up a testing branch in the repo so we could stage things like this and then just land them when they’re ready.

I am hesistant to carry a patched g-p-p but if someone’s also talking to Fedora/GNOME and there’s communication effort to bring that upstream then it could be a great idea. I’m sure Matt and Fedora talk regularly so perhaps a testing branch might do some good here?

I think if we had a handful of people in the repo then the launch of the 16 could be smooth(er) if we set up branches for the models. My hope is that more Framework owners just dive in, we do shared ownership of the repos so all we need to figure out is a structure in the git repo to sort this and start to think about how to support future models.

EDIT: I’m addressing @Aman_Chhabra too!