spresense: fix setup processed flag #680
Conversation
|
@hathach I merged it. I hope that's not a problem. |
| @@ -329,6 +329,7 @@ bool dcd_edpt_xfer(uint8_t rhport, uint8_t ep_addr, uint8_t *buffer, uint16_t to | |||
| { | |||
| if (total_bytes == 0) | |||
| { | |||
| usbdcd_driver.setup_processed = true; | |||
There was a problem hiding this comment.
if you want to detect status packet with this logic, it can mis-inteprete other event. For example when receiving 64 (or 128) bytes of a descriptor, to end the data transfer, ZLP is sent in the DATA stage on the IN endpoint. Afterwards there is another ZLP on OUT endpoint for status. This cause the usbdcd_driver.setup_processed set to true before the ZLP is scheduled (and set to true twice).
There was a problem hiding this comment.
There is a better way to catch status, it is an optional callback dcd_edpt0_status_complete() which is added later after your port. The usbd will call this function right after status phase is complete (full transferred). You could use this to pull setup packet without the need of usbdcd_driver.setup_processed)
https://github.com/hathach/tinyusb/blob/master/src/device/dcd.h#L124
Not a problem at all, you understand the dcd much better than me. Though looking through the chance, I think there is probably a better approach with PS: There is another breaking API as well with supporting multiple hid interfaces #678, that could break your circuitpython ci. It is a easy fix, I will comment on the circuitpython PR. |
|
Thanks for your comments! This is very useful feedback. I'll try to use |
Describe the PR
Fix usb setup flag which sometimes caused DTR bit detection problem.