Closed brtchip-gdm closed 9 months ago
This change fixes an explicit case of transfer timeouts for interrupt/isochronous endpoints. It is not effective for bulk endpoints. The modifications here are carried into issue/branch #47. The printfs in this branch are useful but not necessary to carry forward. Not going to merge this at the moment pending work on #47.
When a timeout occurs on full speed device interrupt endpoints with a short bInterval it would predictably fail to recover on alternate transfers. This was normally masked by either short timeouts or infinite timeout periods in the USBH_transfer.
Include extra detail for debugging.