Is xilinx source code license ok with rtems?

[Keith Robertson]

Joel Sherrill wrote:
> What is the code in question used for?

The code is an RTOS agnostic low level driver for the xilinx trimode 
(10/100/1000) emac.  If designed correctly, a thin layer on top of the 
xilinx code could adapt it to the bsd ip stack.

> And where do you propose it be located?

If we were to take this route, all xilinx code would be entirely 
contained in the virtex bsp.

> As a practical matter, if it become part of the xilinx BSP and was NOT 
> in a position
> to be shared with anything else, it shouldn't be a problem.

Yes, part of the virtex bsp.

 >  No worse
> than using some
> other board specific vendor disclaimed file in a BSP.  But if must be 
> clearly hidden
> in a BSP specific directory or (maybe) powerpc/shared/xilinx with a BIG 
> WARNING
> file.

Ok.  I wasn't even thinking it would be in the shared directory.  Just 
the powerpc/virtex/network directory.  This code is only useful to 
someone with a virtex4 FX.

I'm still not entirely sure what direction I'll take:  Write the driver 
from scratch, or use the xilinx low level code.  Both have benefits and 
detractions.  The main possible detractions of the xilinx code are the 
license (which seems to be ok) and whether I can sensibly extract, with 
minimal modifications, the appropriate source files from their 
distribution structure and import them into a bsp directory in rtems. 
If I have to heavily modify the code, I think we'll be better in the 
long term if I just write a driver from scratch.

I, like Ralf, am totally against having code of this type outside the 
rtems tree.

Cheers.

Keith