^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) ===========
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) DWC3 driver
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) ===========
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) TODO
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) ~~~~
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) Please pick something while reading :)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) - Convert interrupt handler to per-ep-thread-irq
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) As it turns out some DWC3-commands ~1ms to complete. Currently we spin
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) until the command completes which is bad.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) Implementation idea:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) - dwc core implements a demultiplexing irq chip for interrupts per
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) endpoint. The interrupt numbers are allocated during probe and belong
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) to the device. If MSI provides per-endpoint interrupt this dummy
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) interrupt chip can be replaced with "real" interrupts.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) - interrupts are requested / allocated on usb_ep_enable() and removed on
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) usb_ep_disable(). Worst case are 32 interrupts, the lower limit is two
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) for ep0/1.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) - dwc3_send_gadget_ep_cmd() will sleep in wait_for_completion_timeout()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) until the command completes.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) - the interrupt handler is split into the following pieces:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) - primary handler of the device
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) goes through every event and calls generic_handle_irq() for event
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) it. On return from generic_handle_irq() in acknowledges the event
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) counter so interrupt goes away (eventually).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) - threaded handler of the device
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) none
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) - primary handler of the EP-interrupt
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) reads the event and tries to process it. Everything that requires
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) sleeping is handed over to the Thread. The event is saved in an
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) per-endpoint data-structure.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) We probably have to pay attention not to process events once we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) handed something to thread so we don't process event X prio Y
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) where X > Y.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) - threaded handler of the EP-interrupt
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) handles the remaining EP work which might sleep such as waiting
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) for command completion.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) Latency:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) There should be no increase in latency since the interrupt-thread has a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) high priority and will be run before an average task in user land
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) (except the user changed priorities).