PATCH: support for NEC SX architecture

[Dave Korn]

Jaka Močnik wrote:
> hi!
> 
> On Wed, 2009-06-03 at 15:39 +0100, Dave Korn wrote:
>> All of a sudden I'm uncertain whether it's suitable to include support
>> for a somewhat private undocumented target in FSF sources.  I don't know
>> if there's a policy about it but it stands out from other targets in this
>> regard.  Do you have any idea how many of these supercomputers there are
>> in use out there?
> well, the sources are GPLed, implemented independently of any existing, 
> proprietary code, and well documented. the submission has also been 
> approved by erich (cc-ed above).

  Sorry, I've only barely started looking at them so far.  I took a browse
through the sx-gcc website wiki and only found skeletal documentation.  I now
see you have quite a comprehensive instruction listing in include/opcode/sx.h,
but it's not something that anyone could just compile binutils and start
writing assembler code for.

> so as far as the legal issues (and 
> probably philosophical ones as well) are concerned, I'd say the patch is 
> suitable.
> 
> also, I suppose that having a free toolchain for these machines, which is
> at the base of a free system software stack that we're aiming at, can't be
> a bad thing.
> 
> however, erich, who is directly affiliated with NEC will be a better 
> address for this and related questions.

  Erich, is there any chance at all NEC could be persuaded to change policy
and release the CPU and ISA documentation?  That would be the ideal situation.

    cheers,
      DaveK