Re: CVS commit: src/sys/arch

[Jean-Yves Migeon <jeanyves.migeon%free.fr@localhost>]

On Tue, 17 Aug 2010 18:18:22 +0100, David Laight <david%l8s.co.uk@localhost> 
wrote:
>> Unifying sizes of certain types up to 64-bits would be a good choice
>> from engineering standpoint.  Actually measuring the overhead would
>> answer whether it is worth.
> 
> Why not use a 64bit type that is the union of a 64bit number and two
> 32bit numbers.
> A kernel that uses 32bit paddr_t would use tha approriate (for the
> endianness) 32bit value, having initialised the unused half to zero.
> 
> That would remove almost everything except the structure size overhead

Which is precisely the "problem" here :)

> while maintaining binary compatibility.

I don't think so; modules that manipulates 32 bits bus_addr_t will likely
fail with PAE when you cross the 4GB boundary. I don't see how the union
could solve that.

-- 
Jean-Yves Migeon
jeanyves.migeon%free.fr@localhost