Lionel Sambuc 6631803105 usb/hcd: asserts should not have side effects
They can be deactivated with -DNDEBUG, which is why they can't contain
code which should always be run.

It seems the local implementation of USB_ASSERT doesn't honor the
CPP flags NDEBUG, this will be corrected in a later patch.

Change-Id: Iac56b09fd563f1b8c4c7be15a468a05b9cc86a18
2014-09-12 13:20:43 +02:00
..
2014-08-29 18:52:49 +02:00
2014-07-31 16:00:30 +02:00
2014-07-31 16:00:30 +02:00
2014-07-31 16:00:30 +02:00

-------------------------------------------------------------------------------
*                           INFORMATION:                                      *
-------------------------------------------------------------------------------
README file for "USBD" USB host controller driver.

created march-june 2014, JPEmbedded (info@jpembedded.eu)

-------------------------------------------------------------------------------
*                           KNOWN LIMITATIONS:                                *
-------------------------------------------------------------------------------
- Only first configuration can be selected for attached device
- Only one device can be handled at a time, no hub functionality
- Driver assumes that there is no preemption for DDEKit threading
- URBs are enqueued in DDEKit but not in USBD itself
- DDEKit way of handling interface numbers is not explicitly defined, bitmask
  formatting for it, is therefore hardcoded into USBD
- Waiting for USB0 clock to leave IDLEST.Disable state, by nanosleep, was
  removed, as this should be implemented for all clocks in clkconf_set
- Control transfers can only be performed with EP0