[Letux-kernel] new v5.17 bug on GTA04 with musb

[H. Nikolaus Schaller]

Hi,
anyone else also observing this issue?

I have booted letux-5.17-rc2 and get console log:

[   23.131195] using random self ethernet address
[   23.135894] using random host ethernet address
[   23.142639] using host ethernet address: 32:70:05:18:ff:78
[   23.142639] using self ethernet address: 46:10:3a:b3:af:d9
[   23.161499] usb0: HOST MAC 32:70:05:18:ff:78
[   23.186492] usb0: MAC 46:10:3a:b3:af:d9
[   23.196533] using random self ethernet address
[   23.201293] using random host ethernet address
[   23.206085] g_ether gadget: Ethernet Gadget, version: Memorial Day 2008
[   23.216064] g_ether gadget: g_ether ready
[   23.286529] musb-hdrc musb-hdrc.1.auto: Could not enable: -22
[   23.485809] musb-hdrc musb-hdrc.1.auto: Could not enable: -22
[   23.597503] musb-hdrc musb-hdrc.1.auto: Could not enable: -22
[   23.603607] musb-hdrc musb-hdrc.1.auto: Could not enable: -22
[   23.926086] musb-hdrc musb-hdrc.1.auto: Could not enable: -22
[   23.965881] musb-hdrc musb-hdrc.1.auto: Could not enable: -22
[   23.976470] musb-hdrc musb-hdrc.1.auto: Could not enable: -22
[   24.680023] musb-hdrc musb-hdrc.1.auto: pm runtime get failed in musb_gadget_queue
[   24.688232] musb-hdrc musb-hdrc.1.auto: Could not enable: -22
[   24.694335] musb-hdrc musb-hdrc.1.auto: Could not enable: -22
[   24.706481] musb-hdrc musb-hdrc.1.auto: Could not enable: -22
[   25.742034] musb-hdrc musb-hdrc.1.auto: pm runtime get failed in musb_gadget_queue
[   25.750061] musb-hdrc musb-hdrc.1.auto: pm runtime get failed in musb_gadget_queue
[   25.757995] musb-hdrc musb-hdrc.1.auto: pm runtime get failed in musb_gadget_queue
[   25.766113] musb-hdrc musb-hdrc.1.auto: Could not enable: -22
[   25.772399] IPv6: ADDRCONF(NETDEV_CHANGE): usb0: link becomes ready
[   25.779479] musb-hdrc musb-hdrc.1.auto: pm runtime get failed in musb_gadget_queue
...
[   49.359893] musb-hdrc musb-hdrc.1.auto: pm runtime get failed in musb_gadget_queue
usw.

This is as soon/as long as the USB cable is connected.

I have also tried letux-5.16.0,  letux-5.16.4 which do not have the issue
while letux-5.17-rc1 also has it.

So it is once again some bug introduced from upstream :(

Direct musb related commits seem to be:

iMac:master hns$ git log --oneline v5.16..v5.17-rc1 --grep musb
ce1d37cb7697a usb: musb: dsps: Use platform_get_irq_byname() to get the interrupt
cf081d009c447 usb: musb: Set the DT node on the child device
9879c81b68079 usb: musb: Drop unneeded resource copying
iMac:master hns$ 

Trying to reverting them one after one on top of letux-5.17-rc2 makes
me believe that reverting cf081d009c447 is partially sufficient...

There are still spurious stack trace exceptions from irq_exit.

Can anyone please confirm before we report an official upstream bug?

BR,
Nikolaus