USB Passthrough on MacBook Pro (Retina, 2015) running macOS Sierra

[sam.li]

I'm attempting to set up USB passthrough for an Ubuntu LTS image that is set up for embedded development. I managed to get USB passthrough (using this image) to work on my desktop running Ubuntu 16.04LTS (after adding my user to the vboxusers group), but I can't for the life of me seem to get it to work on my MacBook.

Here's the output of

VBoxManage list usbhost

with the STLink plugged in.

Code: Select all

UUID:               f055d220-67a2-4c0e-b15b-aca612b7a983
VendorId:           0x05ac (05AC)
ProductId:          0x8290 (8290)
Revision:           1.64 (0164)
Port:               3
USB version/speed:  0/Full
Manufacturer:       Broadcom Corp.
Product:            Bluetooth USB Host Controller
Address:            p=0x8290;v=0x05ac;s=0x00000000c8ea82e5;l=0x14300000
Current State:      Available

UUID:               4fbb626e-52f5-456e-a86e-2431384f9494
VendorId:           0x05ac (05AC)
ProductId:          0x0274 (0274)
Revision:           6.36 (0636)
Port:               4
USB version/speed:  0/Full
Manufacturer:       Apple Inc.
Product:            Apple Internal Keyboard / Trackpad
SerialNumber:       D3H639626N1GHMFAV6RF
Address:            p=0x0274;v=0x05ac;s=0x000000008e081e1d;l=0x14400000
Current State:      Unavailable

UUID:               b37bb569-8b41-4293-97f2-22fb909fd24b
VendorId:           0x0483 (0483)
ProductId:          0x3748 (3748)
Revision:           1.0 (0100)
Port:               2
USB version/speed:  0/Full
Manufacturer:       STMicroelectronics
Product:            STM32 STLink
SerialNumber:       UÿopƒRH9g
Address:            p=0x3748;v=0x0483;s=0x0000004f575f4ff6;l=0x14200000
Current State:      Busy

And running

lsusb

on my Virtual Machine

Code: Select all

Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 002 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub

I have a USB Filter enabled with the following fields

Code: Select all

Name: STLink
Vendor ID: 0483
Product ID: 3748
Remote: Any

And Enable USB controller is checked, with USB 2.0 (EHCI Controller) enabled. I'm running VirtualBox Version 5.1.14 r112924 (Qt5.6.2) and have Oracle VM VirtualBox Extension Pack 5.1.14r112924 installed.

Any help is greatly appreciated. My hunch is that all my USB ports are USB 3.0, and something about that is causing problems, seeing as I was able to get USB passthrough to work on my desktop using the same image and settings.

[sam.li]

I think I've managed to come up with a workaround.

So it seems like this is an issue with the latest version of VirtualBox. I was looking in the bug tracker this afternoon, and I came across this ticket, which prompted me to attempt a downgrade to VirtualBox 5.0.8. After downgrading VirtualBox and the extension pack, the STLink device is detected. So I believe there's some sort of regression between 5.0.8 and VirtualBox 5.1.14

[mpack]

There is another workaround that someone just reminded me of in another thread.

viewtopic.php?f=2&t=56750#p362667

[michaln]

It would be really useful if someone pinpointed between which revisions exactly the device stopped working. Or sent us one.

[mpack]

I'm not sure anything stopped working as such. I'm not sure if you're familiar with the ST-Link, but it's a JTAG like debug interface for embedded devices containing MCU's made by ST. I wouldn't really expect USB JTAG-alikes to work reliably from a VM, they're likely to be too fussy about timing. A paravirtual type solution ought to work better, and that's sort of what I believe the link I provided is about. Of course I'm using a very broad definition of "paravirtual": I simply mean simulating a physical interface using a hidden communition protocol.

[michaln]

Anything fussy about timing shouldn't be using USB in the first place... although not all designers of USB devices seem to be aware of that. But no, I've never seen a ST-Link device before, and I have no idea how they work on the USB level.

[mpack]

For debug and flash programming links USB is the only practical option these days, and I can assure you that they work far better than the old LPT dongles ever did.

[michaln]

At least to me, flash programming and debugging does not imply "fussy about timing". So yeah, of course it works better than parallel ports that no one has anymore

As for the original problem, I think it should be fixed, and it affected specific USB devices with "creative" names, ones including ASCII control characters to be specific. Not sure what they were thinking.