P33M

That's a clue that it might be an upstream issue (compounded by whatever's different on Pi 5[1]). Does a Pi 4 exhibit the same when updated to either version? Note...

I'm insufficiently convinced this isn't a regression between Pi 4 and Pi 5 host controller hardware. @sninaber can you email info < at > raspberrypi.com to start a discussion about...

Apologies for the lag in response - I've received your BT adapter and had a first pass at reproducing your issue - which I can't. It's worked correctly for me...

Do you have any other USB devices simultaneously plugged in, and if so what are they? (lsusb -v)

I can reproduce this. There's a dependency between Vbus being connected to the device and the time it takes for initialisation to complete. If the time is too long, the...

In this case I would expect USB_INTE_ERROR_DATA_SEQ_BITS to be set, followed by a panic("Data Seq Error \n") - which is admittedly a bad way to handle it, but not an...

Unfortunately I don't have any STM32 parts. I do have a simulation model of a CDC UART, which I can corrupt in almost the same fashion.

With a modified model that transmits a short DATA1 packet in response to the first IN for the endpoint, it does look like the host controller does something bad and...

I believe the issue stems from the fact that the endpoint type is ignored and the endpoint index is used as a proxy as to whether or not to treat...

By inspection the path the EP1-EP15 transfers take is the same whether they are to bulk or interrupt endpoints. Isochronous should not be affected by this bug. There are workarounds...