You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
@jfedor2 At least with commit d17f8ef, a single hub works fine most of the time. With some drivers, usb_midi_host for example, unplug detection can fail almost all of the time. The polling of hub status immediately after a failed IN or OUT transfer on the unplugged device also fails, and the TinyUSB hub driver never tries again. See issue hathach/tinyusb#2971. As this issue does not happen using the native RP2040 hardware, I would say the issue is most likely in this driver.
If I roll back the library version to 0.5.3, hathach/tinyusb#2971 does not happen anymore. I agree with @jfedor2. Something went wrong after release 0.5.3. I tested main (442af43), release 0.6.1, and 0.6.0 and see the same issue. The last commit that correctly handled a disconnect from a hub while there was a pending transfer was 0f747aa. I know this happens all of the time with a USB MIDI host, and I suspect it happens with USB CDC ACM host, too. @sekigon-gonnoc, any advice on where to look? The change from release 0.5.3 to 0.6.0 was pretty major.
I noticed that USB hubs stopped working some time recently (when plugged into a Pico acting as a host).
I have bisected the issue down to these two commits: d735828, 36d5be4.
The text was updated successfully, but these errors were encountered: