As I said, I was running the sources from said archive, compiled with their config through make bindeb-pkg. No other major modifications have been done. And per their config, that setting is set to y.
So if it happens again with 6.16.10, we’ll have to go back to the patch files. Though of course the best way to tell if the kernel alone was the fix, you’d have to be able to just take the kernel, all its modules und headers from e.g. a Fedora 42 ISO and put that into Debian, to make sure no differences at compile time can be the cause. But I’m not certain how exactly this would best be done.
I just found this thread. I’m having the same issues for months on Fedora 42 with fedora stock + locally configured/compiled kernels. I had reported it in the following thread, not aware that there was another already:
Just wanted to mention that if you believe Fedora somehow fixed it, I disagree with that. Unfortunately Fedora users seem to suffer from the same problem
I have already opened a support ticket, but after a few messages exchanged, they wanted me to reproduce with Ubuntu which I’ve never been using and I did not have the time to get into that next step …
Our current line of thinking is that it is a patch or kernel config option, so since you’re using a custom kernel it still doesn’t rule it out I’m afraid. Do you remember what flags you changed when you built your kernel?
You can simply live boot Ubuntu. For me with (almost, except for adding some grub command lines to handle the various amdgpu glitches that can’t be applied in a different way) unmodified ISOs, F41 and 42 haven’t displayed this issue in at least two weaks of testing each, while with Ubuntu 25.04 it happened within a day.
What exactly did you do for “locally configured/compiled kernels”?
I have also seen it happen with the standard fedora kernel at least since the 6.15 series. As it does not make a difference whether booting with the standard kernel or a custom one, I’m afraid my config (most importantly lots of stuff disabled which I don’t need, compiliation with -march=native, some other small changes) does not really help.
Yes, I know that. But since the issue does not always occur and booting from USB is slow, I would have to spend a significant amount of time on that while I know from this forum that others have seen the issue on ubuntu. So I don’t see the point in reproducing it on ubuntu myself. Especially as I then would not know as well how to run diagnostics on a system I’m not used to.
Well, that this only started out with the 6.15 kernel is a start, I think F42 shipped with 6.14 and that was what I successfully tested. So in theory we would just have to find what changed to 6.15 on the Fedora side that made it break there too.
And actually booting from USB isn’t that slow. Sure, you can’t use just any garbage USB 2.0 stick, but beyond that, depending on what you’re doing, Linux is light weight enough to make it usable. I’m mostly using the headphone jack in my free time after work, so during work hours I’d just use my Debian installation and at home switch to live booting, where I wasn’t doing that much anymore.
Also, there are basically no diagnostics you’d need to run, and the ones you can run are identical across basically any Linux distro. For me, FW support sent me a script that would automate things, that was first making sure all necessary tools were installed depending on your distro and collect the logs on its own.
So, I noticed that the latest stable debian kernel(6.12.48 at time of writing) had some fixes applied to uac3 that aren’t present in the vanilla kernel until 6.17 and that fedora hasn’t backported. I’ve been running it for a few hours now and it’s not failed yet. I make no claims as to this being a solution, but considering my previous experiences with the 6.12 kernel are that it fails consistently in 5-30 minutes, this may be notable
I fear I have to disappoint you. That version has been the latest for a couple of months. And not only does it make testing a lot harder for me, as it is one of the worst kernels when it comes to UI freezes due to a amdgpu bug, but I think I’ve already verified that it didn’t help, but at this point I can’t say for sure anymore. I’ll stay on Fedora’s 6.16.10 until the issue reappears and then go back to their 6.14.11 to see if it still can prevent the issue.
It seems Fedora’s 6.14.11 is still carrying the fix and it’s not in a patch file. I’m at 9 days of testing. At one time it seemed to happen again, but the logs didn’t show anything and just unplugging the module and plugging it back in way enough to resolve it, no reboots needed. I’ll make the 14 days full and then look at the latest 6.12 Debian Kernel.
Unfortunately Debian 6.12.48 as well as vanilla 6.17.3 and 6.17.5 are a no-go, though in my testing the former certainly did better than earlier versions of 6.12. Confirmation wouldn’t be remiss however. What I’ll probably do next is try and build vanilla 6.14.11 with debian configs and see where that gets me.
6.17.6-200.fc42.x86_64 #1 SMP PREEMPT_DYNAMIC Wed Oct 29 18:58:05 UTC 2025 x86_64 GNU/Linux
Summary
[115302.833391] usbhid 1-2.3:1.2: can’t add hid device: -71
[115302.833465] usbhid 1-2.3:1.2: probe with driver usbhid failed with error -71
[115302.861747] usb 1-2.3: uac_clock_source_is_valid(): cannot get clock validity for id 9
[115302.861752] usb 1-2.3: clock source 9 is not valid, cannot use
[115302.862116] usb 1-2.3: 1:1: cannot get freq (v2/v3): err -71
[115302.862492] usb 1-2.3: 1:1: cannot set freq 48000 (v2/v3): err -71
[115302.862868] usb 1-2.3: uac_clock_source_is_valid(): cannot get clock validity for id 9
[115302.862871] usb 1-2.3: clock source 9 is not valid, cannot use
[115302.863245] usb 1-2.3: 1:1: cannot get freq (v2/v3): err -71
[115302.863613] usb 1-2.3: 1:1: cannot set freq 48000 (v2/v3): err -71
6 days and no issues on Debian’s 6.12.48. I’ll have to go back to 6.14.11 for now because the amdgpu issue of the whole UI just freezing up is very bad on 6.12 and for the next few days I need more stability than it offers.
Reinserting the expansion card doesn’t change anything, so this is probably an issue with the entire USB stack on the FW16, not the audio expansion card itself.
Summary
Nov 06 13:46:23 framework kernel: usb 1-1: new high-speed USB device number 14 using xhci_hcd
Nov 06 13:46:23 framework kernel: usb 1-1: New USB device found, idVendor=32ac, idProduct=0010, bcdDevice= 0.02
Nov 06 13:46:23 framework kernel: usb 1-1: New USB device strings: Mfr=1, Product=2, SerialNumber=0
Nov 06 13:46:23 framework kernel: usb 1-1: Product: Audio Expansion Card
Nov 06 13:46:23 framework kernel: usb 1-1: Manufacturer: Framework
Nov 06 13:46:23 framework kernel: input: Framework Audio Expansion Card Consumer Control as /devices/pci0000:00/0000:00:08.1/0000:c5:00.3/usb1/1-1/1-1:1.2/0003:32AC:0010.0011/input/input29
Nov 06 13:46:24 framework kernel: hid-generic 0003:32AC:0010.0011: input,hidraw0: USB HID v1.11 Device [Framework Audio Expansion Card] on usb-0000:c5:00.3-1/input2
Nov 06 13:46:24 framework mtp-probe[7804]: checking bus 1, device 14: “/sys/devices/pci0000:00/0000:00:08.1/0000:c5:00.3/usb1/1-1”
Nov 06 13:46:24 framework mtp-probe[7804]: bus: 1, device: 14 was not an MTP device
Nov 06 13:46:24 framework boltd[1988]: probing: started [1000]
Nov 06 13:46:24 framework touchegg[1818]: Warning: Error opening device /dev/input/event2
Nov 06 13:46:24 framework kded6[2647]: kf.notifications: Failed to notify “Created too many similar notifications in quick succession”
Nov 06 13:46:27 framework boltd[1988]: probing: timeout, done: [2960075] (2000000)
Nov 06 13:46:34 framework boltd[1988]: probing: started [1000]
Nov 06 13:46:36 framework boltd[1988]: probing: timeout, done: [2003313] (2000000)
Nov 06 13:46:40 framework tlp[7868]: Error: TLP’s power saving will not apply on boot because tlp.service is not enabled → Invoke ‘systemctl enable tlp.service’ to ensure the full functionality of TLP.
Nov 06 13:46:50 framework fwupd[4055]: 10:46:50.166 FuEngine failed to add device /sys/devices/pci0000:00/0000:00:08.1/0000:c5:00.3/usb1/1-1: failed to setup: failed to get firmware version info: f
ailed to get GET_REPORT: failed to GetReport: USB error: Operation timed out [-7]
Nov 06 13:46:50 framework kernel: usbhid 1-1:1.2: can’t add hid device: -110
Nov 06 13:46:50 framework kernel: usbhid 1-1:1.2: probe with driver usbhid failed with error -110
Nov 06 13:46:50 framework mtp-probe[7910]: checking bus 1, device 14: “/sys/devices/pci0000:00/0000:00:08.1/0000:c5:00.3/usb1/1-1”
Nov 06 13:46:50 framework mtp-probe[7910]: bus: 1, device: 14 was not an MTP device
Nov 06 13:46:55 framework kernel: usb 1-1: uac_clock_source_is_valid(): cannot get clock validity for id 9
Nov 06 13:46:55 framework kernel: usb 1-1: clock source 9 is not valid, cannot use
Nov 06 13:47:00 framework kernel: usb 1-1: 1:1: cannot get freq (v2/v3): err -110
Nov 06 13:47:05 framework kernel: usb 1-1: 1:1: cannot set freq 48000 (v2/v3): err -110
Nov 06 13:47:10 framework kernel: usb 1-1: uac_clock_source_is_valid(): cannot get clock validity for id 9
Nov 06 13:47:10 framework kernel: usb 1-1: clock source 9 is not valid, cannot use
Nov 06 13:47:15 framework kernel: usb 1-1: 1:1: cannot get freq (v2/v3): err -110
I just tested the audio expansion card by connecting it to an Android phone, and it worked perfectly. But when I reinsert it into the FW16, even in a different expansion slot, it fails.
