SAMA5 EHCI:  Did not work with DEBUG off.  Appears to be because of some D-Cache flushing that was performed only with DEBUG ON.  Now is unconditional

USB MSC host class driver: Don't bother retrying to initialize the FLASH if the interface is returning fatal transfer errors

SAMA5:  Add support EHCI/OHCI to sama5d3x-ek/demo (does not work yet); Fix some EHCI/OHCI compilation issues when DEBUG is disabled

Fix warnings in USB MSC when compiled for a high speed device; also fix USB MSC default VID/PID configuration

Extend USB trace logic.  When we are configured to decode and stringify trace data, de-stringification support is not included for class driver trace output

Back-port some bugfixes from the ENCX24J600 to the ENC28J60 driver (unverified).  From Max Holtzberg

Networking:  Important bug fix.  The backlog buffer pointer was not being incrmented so that only the first backlog buffer was getting added to the freelist.  From Max Holtzberg

Networking:  Fix race condition that can cause missing loss-of-connection events.  From Max Holtzberg

CDC/ACM and PL2303 device drivers:  Don't use the max packet size assigned to an endpoint in order to determine the request buffer size.  The endpoint has not yet been configured that max packet size may be wrong.

SAMA5 UDPHS:  More USB fixes mostly related to byte counts, endpoint configuration, and dma configuration

Initial cut of a driver for the TI CC3000 network module with support on the Freescale KL25Z board from Alan Carvalho de Assis

STM32 Timer Register Bit Definitions: Some CCER bit settings changed per SourceForge bug #18 submitted by CCCTSAO