Bug 1213365

Is it a built-in port or a thunderbolt dock?

As always, you can try the kernel in OBS Kernel:stable.
Also, it'd be helpful to check whether the problem persists with 6.5-rc2 kernel in OBS Kernel:HEAD, too.

(In reply to Takashi Iwai from comment #1)
> Is it a built-in port or a thunderbolt dock?

"which is an integrated usb hub in a DELL U4924DW monitor"

> 
> As always, you can try the kernel in OBS Kernel:stable.
> Also, it'd be helpful to check whether the problem persists with 6.5-rc2
> kernel in OBS Kernel:HEAD, too.

I'll try next weekend, if time permits.

(In reply to Michael Hirmke from comment #2)
> (In reply to Takashi Iwai from comment #1)
> > Is it a built-in port or a thunderbolt dock?
> 
> "which is an integrated usb hub in a DELL U4924DW monitor"
> 
> > 
> > As always, you can try the kernel in OBS Kernel:stable.
> > Also, it'd be helpful to check whether the problem persists with 6.5-rc2
> > kernel in OBS Kernel:HEAD, too.
> 
> I'll try next weekend, if time permits.

Time didn't permit, yet.
I have to fight massive problems with KDE. Will open another bug.

I was able to test a few things this weekend.
As it seems, I was wrong - usb 9-1 isn't mouse or keyboard, it is my Native Instruments external audio box:

Jul 30 11:08:20 client kernel: usb 9-1: new high-speed USB device number 16 using xhci_hcd
Jul 30 11:08:20 client kernel: usb 9-1: New USB device found, idVendor=17cc, idProduct=1001, bcdDevice= 0.32
Jul 30 11:08:20 client kernel: usb 9-1: New USB device strings: Mfr=12, Product=7, SerialNumber=13
Jul 30 11:08:20 client kernel: usb 9-1: Product: Komplete Audio 6
Jul 30 11:08:20 client kernel: usb 9-1: Manufacturer: Native Instruments
Jul 30 11:08:20 client kernel: usb 9-1: SerialNumber: 6D892EF0
Jul 30 11:08:20 client mtp-probe[2227]: checking bus 9, device 16: "/sys/devices/pci0000:00/0000:00:07.2/0000:39:00.0/0000:3a:02.0/0000:3d:00.0/usb9/9-1"
Jul 30 11:08:20 client mtp-probe[2227]: bus: 9, device: 16 was not an MTP device
Jul 30 11:08:20 client mtp-probe[2242]: checking bus 9, device 16: "/sys/devices/pci0000:00/0000:00:07.2/0000:39:00.0/0000:3a:02.0/0000:3d:00.0/usb9/9-1"
Jul 30 11:08:20 client mtp-probe[2242]: bus: 9, device: 16 was not an MTP device

After a reboot or resume I always have to unplug the box an plug it in again to get it to work.
This happens with Linux 6.4.6-1-default on openSUSE Tumbleweed 20230727.

Having said that, I have noticed many usb problems with kernel 6.4.x.
Many usb devices seem to behave unreliable compared to kernel 6.3.x, probably there is something different/problematic with my usb/usb-c/thunderbolt setup.

My setup looks like this:

Notebook DELL XPS13
-> usb-c/thunderbolt cable
   -> tunderbolt box CalDigit TS3 Plus
     -> usb Native Instruments audio box, usb scanner,
        usb printer, usb card reader, ...
     -> displayport cable + usb-c cable
        -> monitor DELL U4924DW (with usb/usb-c hub)
           -> usb mouse, usb keyboard

And does it happen *always* when NI box is plugged?  The error -71 is -EPROTO, and it often indicates some firmware problem...

(In reply to Takashi Iwai from comment #6)
> And does it happen *always* when NI box is plugged?  The error -71 is
> -EPROTO, and it often indicates some firmware problem...

It always happens, yes.
And maybe it is a firmware problem, but it worked for a long time until I installed 6.4.x. And - the box is too old to get new firmware.
And - it always works after unplugging the box and plugging it in again.

And - as I wrote: There are problems with many of my usb devices since I installed 6.4.x. For example the keyboad isn't available for a few seconds in the plymouth passphrase dialog for the encrypted root partition.

It looks to me as if in 6.4.x the code for usb/usb-c/thunderbolt devices and boxes has been modified heavily.

OK, and it happens only with that connection over the TB hub?

(In reply to Takashi Iwai from comment #8)
> OK, and it happens only with that connection over the TB hub?

This notebook only has two thunderbolt connectors.
I could try to plugin the box with an adapter for a short test, but most parts of my system won't work then - I have to choose between no network, no power and no external monitor then.
I'll try later this day.

I rebooted my notebook with only the usb audio box connected.
The problem didn't occur in this configuration.
I then updated the notbook BIOS and Tumbleweed to the latest versions, but it didn't help to solve the problem, when all my usb boxes and devices are connected.

If only NI box is connected to the hub, and only the hub is connected to the machine, the bug still happens?  Just for narrowing down the condition.

(In reply to Takashi Iwai from comment #11)
> If only NI box is connected to the hub, and only the hub is connected to the
> machine, the bug still happens?  Just for narrowing down the condition.

If only the Thunderbolt box is connected, there are at least the internal devices of the box itself connected, too, i.e. a network adapter, a sound device and a card reader.
I'll try to strip it down as much as possible.

(In reply to Michael Hirmke from comment #12)
> (In reply to Takashi Iwai from comment #11)
> > If only NI box is connected to the hub, and only the hub is connected to the
> > machine, the bug still happens?  Just for narrowing down the condition.
> 
> If only the Thunderbolt box is connected, there are at least the internal
> devices of the box itself connected, too, i.e. a network adapter, a sound
> device and a card reader.
> I'll try to strip it down as much as possible.

Ok, now it really gets mysterious:
I shutdown the notebook, unplugged everything from the Thunderbolt box besides the audio box and started the notebook - problem was still there.
I then shutdown the machine again, plugged everything back in (in exactly the same way as it was before), started the notebook - and the problem was gone. Even after a few reboots the problem didn't reoccur.
I don't understand why, but as it works, I am satisfied :)

Thx!