Dell XPS 17 latest apt update results in no sound (via internal speakers and mic)

[Briain]

Hi Folks

I'm running a Dell XPS17 9700 with Debian 12 and apart from the fingerprint reader, everything has been working perfectly for months. Amongst other things, I do use it for audio work (so sometimes with a USB DAC) but for normal day to day stuff, I do use the laptop's internal mic and speakers.

The latest update has just 'broken' the internal speaker and microphone functionality; the speaker icon on the icons only task manager shows a diagonal line through it and when I click on it, the pop-up states that 'no output or input devices found' (and indeed, PulseAudio Volume Control shows no devices as being present) so I now have no 'internal' sound functionality at all.

From journalctl -b the below shows some sound related entries with the last three lines being in red:

Code: Select all

Apr 17 10:43:25 dell kernel: sof-audio-pci-intel-cnl 0000:00:1f.3: DSP detected with PCI class/subclass/prog-if info 0x040100
Apr 17 10:43:25 dell kernel: sof-audio-pci-intel-cnl 0000:00:1f.3: enabling device (0000 -> 0002)
Apr 17 10:43:25 dell kernel: sof-audio-pci-intel-cnl 0000:00:1f.3: DSP detected with PCI class/subclass/prog-if 0x040100
Apr 17 10:43:25 dell kernel: sof-audio-pci-intel-cnl 0000:00:1f.3: bound 0000:00:02.0 (ops i915_audio_component_bind_ops [i915])
Apr 17 10:43:25 dell kernel: iwlwifi 0000:00:14.3: Detected RF HR B3, rfid=0x10a100
Apr 17 10:43:25 dell kernel: sof-audio-pci-intel-cnl 0000:00:1f.3: use msi interrupt mode
Apr 17 10:43:25 dell kernel: sof-audio-pci-intel-cnl 0000:00:1f.3: hda codecs found, mask 4
Apr 17 10:43:25 dell kernel: sof-audio-pci-intel-cnl 0000:00:1f.3: firmware: direct-loading firmware intel/sof/sof-cml.ri
Apr 17 10:43:25 dell kernel: sof-audio-pci-intel-cnl 0000:00:1f.3: error: firmware intel/sof/sof-cml.ri contains unsupported or invalid extended manifest: -22
Apr 17 10:43:25 dell kernel: sof-audio-pci-intel-cnl 0000:00:1f.3: error: failed to load DSP firmware -22
Apr 17 10:43:25 dell kernel: sof-audio-pci-intel-cnl 0000:00:1f.3: error: sof_probe_work failed err: -22

The Dell also received a bios/efi update (via Discover Software Centre) and where it was on 1.28.0 it's now on 1.29.0. I did wonder about whether that was the cause. I did ponder installing the previous bios/efi, but rather than try that, I instead just booted the laptop from the Debian 12 Live edition (from a USB stick) and the sound worked fine (I also tried booting into Windows 10 and the sound's working fine with that) so to me it looks suspiciously like one of the recent apt update packages has caused the problem.

If anyone has any thoughts, they'd be gratefully received as at the moment I have no sound (unless using an external DAC and monitor, which isn't available where I daily drive my laptop) so somewhat horrifically (to me), I'm going to have to use Windows 10 until I find a fix.

Briain

PS When I first installed Debian 12, the microphone didn't work (the speaker was always fine) but I found fix for that; see here: viewtopic.php?p=797264#p797264 (I'm just positing that in case anyone with a similar issue finds this post) and it's been working ever since. With that fault, the hardware all appeared (and no errors in journalctl -b) but the mic didn't pick up any sound, so a very different to the recent update issue described in the main part of my above post.

[fabien]

Hello,

PS When I first installed Debian 12, the microphone didn't work (the speaker was always fine) but I found fix for that; see here: viewtopic.php?p=797264#p797264

It is worth noting that the script that you used to fix your initial problem basically removes files from the official firmware-sof-signed package and replaces them with new versions. These unsupported versions probably explain the error you encounter.

[Briain]

Hi fabien

Thank you for responding so quickly and yes, that makes total sense (though it has survived many updates since I initially ran it, not long after Bookworm was initially released). I just tried re-running the script (hey, nothing's working, so why not) but as expected, it didn't fix anything, however, I have progressed things and it looks like it was the update to 6.1.20 that's 'broken' my system as I've just tried re-booting and selecting version 6.1.0-18-amd64; the sound (both internal microphone and speaker) is now working perfectly.

Of course, that leaves me potentially now being forever stuck on 6.1.0-18-amd64, but hey, I now have my sound functionality back and won't have to start using Win 10 as a daily driver; you have no idea just how relieved I am about that! I do have a VirtualBox instance of Win 10 installed and I do sometimes have to use that (for a couple of SDR packages; SDR Console being the main one I need to use) but that means I can simply swipe it (on the second workspace) out of my way and do everything else on Debian.

Briain

PS I must contact the KDE devs and ask if they can implement a slap gesture to change workspaces; it'd be kinda nice to be able to slap Windows out of the way!

[Briain]

Hi

Just for everybody's information, I've just checked back and found that I'd last used the script (successfully) after reinstalling Debian at version 6.1.0-16-amd64 and it's carried on working through the 6.1.0-17-amd64 and 6.1.0-18-amd64 updates (without needing to re-run the scrip), breaking only with after very latest (6.1.20-amd64) update. As I say, for now I'm just going to drop back to using 6.1.0-18-amd64 as I need to get things working in a hurry, but I'll look deeper into it in a few weeks.

Briain

[Linuxgaming1824]

That's typically how things work with linux. You've identified you have issues, on your system, when you use the 6.1.20 kernel. The solution? Use a different kernel. You don't have to worry so much that you lack the upgrades granted by a 0.0.2 point upgrade to the kernel. Just use a different kernel. Problem solved!

[Briain]

Yes absolutely; I'm more than happy to reside on 6.1.0-18 as everything is working beautifully! The only reason that I'm planning to eventually try restoring the files from the official firmware-sof-signed package and then re-trying the latest kernel is that I'd be really rather curious to see if the latest kernel changes have sorted the original problem (the internal mic not working) which was what induced me to finding and running that script in the first place, but as I say, it's just me being curious, so though I will certainly do that, likely it will be a good while before I get round to trying it (it's a rainy day job).