thesofproject
[BUG] Dell XPS 13 9315: No sound cards detected
On the Dell XPS 13 9315 with Debian sid/unstable and Linux 5.19 from the suite experimental, very often no soundcards are found.
00:1f.3 Multimedia audio controller [0401]: Intel Corporation Device [8086:51cc] (rev 01)
[ 14.611798] soundwire sdw:2:025d:1316:01: Probe not complete, timed out
[ 14.611813] soundwire sdw:2:025d:1316:01: Update Slave status failed:-110
[ 14.611853] soundwire sdw:3:025d:0714:01: Probe not complete, timed out
[ 14.611855] soundwire sdw:3:025d:0714:01: Update Slave status failed:-110
As reported in issue https://github.com/thesofproject/linux/issues/3777 unloading the module snd_sof_pci_intel_tgl and loading it again works around the issue.
know issue duplicate of https://github.com/thesofproject/linux/issues/3777
Solved by upstream commits, see https://lore.kernel.org/alsa-devel/[email protected]/
Thank you for the prompt reply. I am happy fixes are in Linus’ master branch.
Two questions:
- Is that a regression in Linux 5.18?
- Did Dell send one of these devices to Intel, so they can be used for testing?
- Is that a regression in Linux 5.18?
Unclear. The probe message has been there for multiple cycles, since commit 528be501b7d4a ("soundwire: sdw_slave: add probe_complete structure and new fields")
it's not clear what caused the timeout in practice. Possibly PCI power management?
- Did Dell send one of these devices to Intel, so they can be used for testing?
Yes, and we didn't see the issue with the Canonical OEM image. It's likely a configuration that causes delays in the probe and timeouts.
- Did Dell send one of these devices to Intel, so they can be used for testing?
Yes, and we didn't see the issue with the Canonical OEM image. It's likely a configuration that causes delays in the probe and timeouts.
Yes, me neither, but I’d thought the Intel (FLOSS) QA is going to test mainly with the upstream versions.
I was the author of that broken solution in commit 528be50 ("soundwire: sdw_slave: add probe_complete structure and new fields") and we never saw any issues in all our tests. It's just luck that we had a solution ready: I was doing some cleanups to make the driver more compliant with the device model, and remove all sorts of timing dependencies to rely instead of the deferred probe mechanism. I had no idea that cleanup would actually solve a real problem, it was only necessary in academic corner cases with some drivers blacklisted and manually inserted after a long delay.
Hi, I running kernel version 5.19.17 and I encountered the same problem on Dell XPS 9320, does the patch only work above 5.19.17?
sof-firmware version is 2.2
The recommendation is to update to version 6.0.6 and later.
I see, thanks.
I need to wait the kernel version 6 of lts come out so zfs could bump the Linux compatible.
@plbossart, only in Arch Linux the patch was backported to the Linux kernel package. It’s not in the official stable release (or Linus’ master) to my knowledge.
Hi, I am on 6.0.7 but the issue seems to still persist.
$ uname -a
Linux debian 6.0.0-3-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.0.7-1 (2022-11-05) x86_64 GNU/Linux
The version of firmware-sof-signed is 2.2.2-1
And output of /usr/sbin/alsa-info shows:
[ 10.913863] mei_hdcp 0000:00:16.0-b638ab7e-94e2-4ea2-a552-d1c54b627f04: bound 0000:00:02.0 (ops i915_hdcp_component_ops [i915])
[ 10.914021] sof_sdw sof_sdw: snd_soc_register_card failed -517
[ 10.919220] sof_sdw sof_sdw: snd_soc_register_card failed -517
[ 15.081027] wlp0s20f3: authenticate with b4:a7:c6:52:b9:50
--
[ 15.499648] IPv6: ADDRCONF(NETDEV_CHANGE): wlp0s20f3: link becomes ready
[ 21.018708] sof_sdw sof_sdw: snd_soc_register_card failed -517
[ 21.018753] platform sof_sdw: deferred probe pending
I am quite clueless at the moment as to what should be done. I'd appreciate any inputs.
Hi, I am on 6.0.7 but the issue seems to still persist.
no you're not, see below
$ uname -a Linux debian 6.0.0-3-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.0.7-1 (2022-11-05) x86_64 GNU/Linux
that's 6.0.0?
@plbossart, see the string at the end: 6.0.7-1.
@nileshpatra, the upstream stable series and Debian does not cherry-pick/carry the patch fixing the issue in contrast to Arch Linux, cf issue https://github.com/thesofproject/linux/issues/3777.
@paulmenzel
Thanks for replying. Is there any workaround for it other than compiling the kernel w/ that patch? Unloading/loading that module does not seem to work any more with the newer kernel :(
It's not just me that's confused, see alsa-info
!!Kernel Information
!!------------------
Kernel release: 6.0.0-3-amd64
The patch was added to soundwire/fixes but I don't see it in Linus' tree just yet, so it can't be in linux-stable yet. must be stuck between maintainers. I asked for an update.
Is there new information about weather the fix has been merge into Linux-stable?
I don't see the change as of v6.0.9 unfortunately, I sent a backport to linux-stable.
I see, I would contact gentoo kernel maintainers to see weather they can help to patch the kernel.
The patch is in here right?
https://lore.kernel.org/alsa-devel/[email protected]/
@IchikaZou, cf issue https://github.com/thesofproject/linux/issues/3777, and merge/pull request https://github.com/thesofproject/linux/pull/3911.
- https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?h=master&id=13c30a755847c7e804e1bf755e66e3ff7b7f9367
- https://lore.kernel.org/stable/[email protected]/