Re: Arm drivers: are device tree descriptions required for every device?

[Brook Milligan <brook%nmsu.edu@localhost>]

> On Apr 14, 2023, at 7:23 PM, Brook Milligan <brook%nmsu.edu@localhost> wrote:
> 
> I want to circle back to the motivation for starting this thread.
> 
> I have a BeagleBone Black that boots NetBSD fine.  However, it is not detecting anything in the USB port.  For example, connecting an SD card either prior to booting or after booting during runtime leads to no storage medium being detected.  In contrast, a NetBSD/amd64 machine, laptops, cameras, etc. detect the very same card.
> 
> This led me to think that something was off with the device tree files, the kernel config, or the device probing. From this thread, it seems that there should be no need to include a USB storage device in the device tree files, and the GENERIC kernel config does include them.  Thus, from that I would have expected the USB storage to be detected by the kernel.  I am now at a loss regarding where to start.
> 
> I have attached the kernel message buffer (from a boot with the USB storage already installed) in hopes that someone can provide advice on where to start tracking this down.  I have all the firmware properties from ofctl as well, but that is quite lengthy so I’m leaving it until someone thinks it might be useful.

To provide a bit more data, I have tried the suggestions here: 

	https://www.netbsd.org/docs/kernel/#usb-debugging

Looking at vmstat -u usbhist output, I see pairs of lines like the following.

1681529412.773499 usb_event_thread#2@0: sc 0x90e61080 woke up
1681529412.773501 usb_discover#32@0: called!

These occur when USB devices are plugged in or removed, but they do not occur every time.  I think all the removal events are caught (but am not quite sure), but some insertion events are definitely not.  There is no other indication of activity.

Does this mean that usb_discover() is not discovering anything?  That is, in a working USB system, should I expect to see more messages after the usb_discover call message?

How can this be debugged?

Thanks for your help.

Cheers,
Brook