RP2040(RPi Pico) with PIO USB Host Fails at Get 8 byte of Device Descriptor

[Helloyunho]

I'm using my Raspberry Pi Pico to convert MIDI input(PIO host)(MIDI device specifically Launchpad MK2) to a custom USB signal(micro-USB port on Pico).

USBD works fine, but when I plug my Launchpad to Pico, it fails at Get 8 byte of Device Descriptor.

I cannot connect it to debugger right now.

Below is the log(of USBH only):

[1:] USBH Device Attach
Full Speed
[1:0] Open EP0 with Size = 8
Get 8 byte of Device Descriptor
[1:0] Get Descriptor: 80 06 00 01 00 00 08 00 
[:0] on EP 00 with 0 bytes: FAILED
[1:0] Control FAILED 1/3, retrying
[:0] on EP 00 with 0 bytes: FAILED
[1:0] Control FAILED 2/3, retrying
[:0] on EP 00 with 0 bytes: FAILED
[1:0] Control FAILED 3/3, retrying
[:0] on EP 00 with 0 bytes: FAILED
[1:0] Control FAILED, xferred_bytes = 0
80 06 00 01 00 00 08 00 

Enumeration attempt 2/3
[1:0] Get Descriptor: 80 06 00 01 00 00 08 00 
[:0] on EP 00 with 0 bytes: FAILED
[1:0] Control FAILED 1/3, retrying
[:0] on EP 00 with 0 bytes: FAILED
[1:0] Control FAILED 2/3, retrying
[:0] on EP 00 with 0 bytes: FAILED
[1:0] Control FAILED 3/3, retrying
[:0] on EP 00 with 0 bytes: FAILED
[1:0] Control FAILED, xferred_bytes = 0
80 06 00 01 00 00 08 00 

Enumeration attempt 3/3
[1:0] Get Descriptor: 80 06 00 01 00 00 08 00 
[:0] on EP 00 with 0 bytes: FAILED
[1:0] Control FAILED 1/3, retrying
[:0] on EP 00 with 0 bytes: FAILED
[1:0] Control FAILED 2/3, retrying
[:0] on EP 00 with 0 bytes: FAILED
[1:0] Control FAILED 3/3, retrying
[:0] on EP 00 with 0 bytes: FAILED
[1:0] Control FAILED, xferred_bytes = 0
80 06 00 01 00 00 08 00 

[1:0:0] USBH DEVICE REMOVED

[Helloyunho]

So an update: my random flash drive actually passes that part(it's not a midi device tho so it doesn't work obviously)
But it still fails at getting full descriptor... maybe related to #3149?

[Helloyunho]

Another update: my USB mouse receiver works and doesnt fail on descriptor
Plugging Launchkey MK3(another MIDI device from same company) makes pi stuck at Get 8 bytes of Device Descriptor(different issue than above, it gets the control data then it stuck at there until I unplug it)

[hathach]

use master branch of both tinyus and pio-usb to see if it help, you can also try sekigon-gonnoc/Pico-PIO-USB#186 as well. If that does not work, unfortunately it is device specific issue i.e your device has some timing expectation that pio-usb does not meet. It is big banging afterall.

[Helloyunho]

It still fails... Is there no way to fix this? Like manually adjust timing or etc?

[hathach]

yeah, you can try, https://github.com/sekigon-gonnoc/Pico-PIO-USB/blob/main/src/pio_usb_host.c there is no docs and no easy way though. You also need a decent analyzer to make check what is wrong with the timing. Unfortunately as said above, this is device specific issue, and need actual (or similar) device to test/troubleshoot.

[Helloyunho]

Sadly I don't have equipments like logic analyzer so I probably can't fix the issue without someone else's help...

[hathach]

you can try to run rp2040 at 240mhz to see if that helps, it often make the timing faster. Other than, you have to wait for more future update on either tinyus or pio-usb to see if that work with your device.

[Helloyunho]

Again, doesn't work.. I think I should wait until someone else help this problem

[SerVB]

Hey, I have pretty much the same problem with the same log. My MIDI keyboard yamaha piaggero np-15 isn't enumerated (so no mount callback is called; but umount is called). My board is Feather RP2040 USB Host.

I will try later and share if it helped:

• 240 MHz speed.
• RP2040(RPi Pico) with PIO USB Host Fails at `Get 8 byte of Device Descriptor` #3196 (comment) – try using newer versions – I guess to overwrite current Arduino libs I have (installed the latest releases) with the files from the main branches, right?

Please tell me if I can do anything else...

Just for context to avoid false expectations: the experience is quite frustrating so far: I'm pretty new to this Arduino Libraries world (or how can I call it), and getting logs from tinyusb was quite an achievement for me 😅 Btw, is there an official guide about getting logs, maybe I struggled without a reason?

[SerVB]

With 240 MHz, the behavior is different: it doesn't fail right away but still doesn't finish the enumeration because of some assert:

  0000:  26 03                                            |&.|
[:1] on EP 00 with 0 bytes: OK

[1:1] Get Descriptor: 80 06 01 03 09 04 26 00 
[:1] on EP 00 with 8 bytes: OK
[:1] on EP 80 with 38 bytes: OK
[1:1] Control data:
  0000:  26 03 59 00 61 00 6D 00 61 00 68 00 61 00 20 00  |&.Y.a.m.a.h.a. .|
  0010:  43 00 6F 00 72 00 70 00 6F 00 72 00 61 00 74 00  |C.o.r.p.o.r.a.t.|
  0020:  69 00 6F 00 6E 00                                |i.o.n.|
[:1] on EP 00 with 0 bytes: OK

[1:1] Get Descriptor: 80 06 02 03 09 04 02 00 
[:1] on EP 00 with 8 bytes: OK
[:1] on EP 80 with 2 bytes: OK
[1:1] Control data:
  0000:  12 03                                            |..|
[:1] on EP 00 with 0 bytes: OK

[1:1] Get Descriptor: 80 06 02 03 09 04 12 00 
[:1] on EP 00 with 8 bytes: OK
[:1] on EP 80 with 18 bytes: OK
[1:1] Control data:
  0000:  12 03 50 00 69 00 61 00 67 00 67 00 65 00 72 00  |..P.i.a.g.g.e.r.|
  0010:  6F 00                                            |o.|
[:1] on EP 00 with 0 bytes: OK

Get Configuration[0] Descriptor (9 bytes)
[1:1] Get Descriptor: 80 06 00 02 00 00 09 00 
[:1] on EP 00 with 8 bytes: OK
[:1] on EP 80 with 9 bytes: OK
[1:1] Control data:
  0000:  09 02 2F 01 04 01 00 C0 01                       |../......|
[:1] on EP 00 with 0 bytes: OK

process_enumeration 1738: ASSERT FAILED

So the code is:

tinyusb/src/host/usbh.c

Lines 1737 to 1738 in 79445c2

	// TODO not enough buffer to hold configuration descriptor
	TU_ASSERT(total_len <= CFG_TUH_ENUMERATION_BUFSIZE,);

. I've increased CFG_TUH_ENUMERATION_BUFSIZE from 256 to 512 in tusb_config_rp2040.h.

And now it's Device configured // Parsing Configuration descriptor (wTotalLength = 303)!

So I now have the example working:

TinyUSB Dual Device Info Example
Device attached, address = 1
Device 1: ID 0499:1748
Device Descriptor:
  bLength             18
  bDescriptorType     1
  bcdUSB              0200
  bDeviceClass        0
  bDeviceSubClass     0
  bDeviceProtocol     0
  bMaxPacketSize0     64
  idVendor            0x0499
  idProduct           0x1748
  bcdDevice           0100
  iManufacturer       1     Yamaha Corporation
  iProduct            2     Piaggero
  iSerialNumber       0     
  bNumConfigurations  1
Device 1: ID 0499:1748 Yamaha Corporation Piaggero

I even haven't tried a master branch. Thanks for the tip about 240 MHz! I will try to continue in this state for now.

I hope my message helps somebody with the same problem. And it will be great if developers paid attention to it as well, e.g.:

• Consider documenting how to debug such a problem (mainly, how to activate logs to Serial, not to Serial1).
• Consider documenting some kind of Faq: use 240 MHz if possible, increase CFG_TUH_ENUMERATION_BUFSIZE if possible.

[Helloyunho]

yeah... it still doesn't work on my side...
it actually doesn't even give different result when i use set_sys_clock_khz(240000, true);

[SerVB]

Just wanted to share one more finding when I tried the following example from the EZ_USB_MIDI_HOST lib: https://github.com/rppicomidi/EZ_USB_MIDI_HOST/blob/main/examples/arduino/EZ_USB_MIDI_HOST_PIO_example/EZ_USB_MIDI_HOST_PIO_example.ino.

It failed for me to connect the same midi keyboard to the same port of the same board with the good old Control FAILED, xferred_bytes = 0 even with 240 MHz and CFG_TUH_ENUMERATION_BUFSIZE=512.

What helped to make this example working was to switch from optimization -O0 (disabled) to the standard -Os (small). I haven't tried others because it made it work already.

Does it mean that the board is close to the minimal performance of supporting my usb midi keyboard?

Then I'm wondering whether switching from rp2040 to rp2350 helps to make it more reliable...