video and usb driver (rk3588)

[Norman Feske]

Hi Michael,

>> BTW, the official way to obtain a contiguous RAM buffer (for the use as
>> framebuffer) is platform driver. For a 'Dma_buffer' object, you can
>> request the 'dma_addr' [2] to be poked into the graphics device register
>> for the framebuffer base address.
>>
> Will  try this after every other option :-) I doubt it will solve
> alignment fault.

here is my guess:

- The boot loader sets up the framebuffer in a known RAM area.
- This framebuffer is mapped as uncached memory, similar one would map
   an memory-mapped I/O resource.
- Unaligned memory accesses by the user land are allowed in principle,
   but
- Unaligned memory accesses to uncached memory are always forbidden.
- The blitting code assumes that unaligned accesses at 32-bit boundaries
   are fine.
- Your boot-fb driver reuses the original framebuffer mapping,
   hence it inherits the way of how the framebuffer is originally mapped.
   Consequently, the blitting code issues unaligned bus accesses to
   uncached memory.

If this is the case, the use of a fresh allocated DMA buffer using the 
platform driver would indeed solve this issue because the blitting code 
would not touch uncached memory.

Should this work and you see caching artifacts on screen, you may need 
to explicitly call 'cache_clean_invalidate_data' to write back the pixel 
data to RAM after blitting.

Keeping my fingers crossed...

Cheers
Norman

-- 
Dr.-Ing. Norman Feske
Genode Labs

https://www.genode-labs.com · https://genode.org

Genode Labs GmbH · Amtsgericht Dresden · HRB 28424 · Sitz Dresden
Geschäftsführer: Dr.-Ing. Norman Feske, Christian Helmuth