For i.MX8X (QXP and others), can A35 cores be programmed in big-endian ?
ARMv8 documention says:
- instruction fetches are always treated as little-endian,
- data-access is IMPLEMENTATION DEFINED whether both little-endian and big-little modes are supported (and if only one which one is supported is also IMPLEMENTATION DEFINED - here obviously little-endian).
ARM architecture itself indeed supports both big and small endianness. However, only little-endian implementations have been used and implemented on the i.MX8 processors to date. All software releases supported will use little endianness, which is the endianness currently supported.
The i.MX8X in is still in pre-production so some documentation may not be available. But on the available preliminary documentation you can see which interfaces provide endianness settings in case have your application requires it. For further technical information on this processor in pre-production status please refer to your FAE or Distributor.
My apologies for the inconvenience!