brintos

brintos / linux-shallow public Read only

0
0
Text · 1.8 KiB · f94a7ba Raw
54 lines · plain
1===========2DWC3 driver3===========4 5 6TODO7~~~~8 9Please pick something while reading :)10 11- Convert interrupt handler to per-ep-thread-irq12 13  As it turns out some DWC3-commands ~1ms to complete. Currently we spin14  until the command completes which is bad.15 16  Implementation idea:17 18  - dwc core implements a demultiplexing irq chip for interrupts per19    endpoint. The interrupt numbers are allocated during probe and belong20    to the device. If MSI provides per-endpoint interrupt this dummy21    interrupt chip can be replaced with "real" interrupts.22  - interrupts are requested / allocated on usb_ep_enable() and removed on23    usb_ep_disable(). Worst case are 32 interrupts, the lower limit is two24    for ep0/1.25  - dwc3_send_gadget_ep_cmd() will sleep in wait_for_completion_timeout()26    until the command completes.27  - the interrupt handler is split into the following pieces:28 29    - primary handler of the device30      goes through every event and calls generic_handle_irq() for event31      it. On return from generic_handle_irq() in acknowledges the event32      counter so interrupt goes away (eventually).33 34    - threaded handler of the device35      none36 37    - primary handler of the EP-interrupt38      reads the event and tries to process it. Everything that requires39      sleeping is handed over to the Thread. The event is saved in an40      per-endpoint data-structure.41      We probably have to pay attention not to process events once we42      handed something to thread so we don't process event X prio Y43      where X > Y.44 45    - threaded handler of the EP-interrupt46      handles the remaining EP work which might sleep such as waiting47      for command completion.48 49  Latency:50 51   There should be no increase in latency since the interrupt-thread has a52   high priority and will be run before an average task in user land53   (except the user changed priorities).54