While debugging, I found that on RP2040 the first looped background
write on a StateMachine would be garbled: It the first few outputs
would be as expected, but then the data would go back to the start
of the buffer.
I realized that this could be due to there already being a pending
interrupt on this DMA channel that had occurred previously but never
been acknowledged. In fact, by printing the value of `dma_hw->intr`
before this, I could see that it was the case.
After this change, both my special case reproducer from the related
issue works, and the odd gaps in LED data transmission seen in the
WIP code for TM1814 LEDs was fixed.
I never saw this problem on Feather RP2350, only on Metro RP2040,
but I don't know why. It would depend on what had previously
happened to the DMA channel that ends up allocated to this PIO
state machine. Less likely, it could depend on details of the DMA
peripheral that changed between the chips.
Closes: #9678
This is the result of a `pre-commit run --all-files` to fix the
translation file changes which were causing a CI check fail. I'm
not convinced this change is a good idea. I would prefer to set
USBError.errno, if possible, but I don't know how to do that.
This commit makes sure that functions in usb.core.Device return as
soon as they raise a MicroPython exception rather than continuing
on with whatever code would normally run. This will hopefully help
to avoid things like dereferencing null pointers.
shared-module/usb/core/Device.c had several spots where it would
raise a MicroPython exception with a NULL argument. That made it
very difficult for CircuitPython code to attempt to recover from
fault conditions caused by quirky USB device behavior. This adds
short error strings to distinguish the different types of faults.
