Re: MIPS32 / MIPS64 release 2 support in simulator

[Andrew Cagney <cagney@gnu.org>]

David Ung wrote:
> On Mon, 2004-11-08 at 16:43, Andrew Cagney wrote:
> 
>>cgd@broadcom.com wrote:
>>
>>>At Thu, 4 Nov 2004 13:53:38 +0000 (UTC), "David Ung" wrote:
>>>
>>>
>>>>I shall update mips.igen with the new r2 instructions and then submit it
>>>>as a patch.
>>>
>>>
>>>cool, thanks.
>>
>>Don't.  These new changes go in a new file, and that file gets a very 
>>clear FSF copyright.
>>
>>Andrew

David, this isn't a technical decision (if it were, Chris would be 
right).  The new changes, or more exactly the new instructions, go in a 
new file clearly marked (C) FSF (and for that matter clearly stating 
that it was contributed by MIPS Inc).  As for the mark-up of the 
existing instructions, that's mindnumming tedium.

Sorry about this pedentary.  I'm off to pester another lawyer.

Andrew

> arhh.  Chris's argument of that the rest of the tools consider rev 2
> instructions as an ISA is quite a valid one.  And for consistency with
> the existing structure, it looks to be the right thing to do.  
> But when I got to merging in the r2 instructions into mips.igen, I do
> see that for every mips32 / 64 instruction, I now have to also tag them
> mips32r2 / 64r2.  It would probably be nicer to just tell it to include
> all the mips32 instructions (which would be like treating it as an
> ASE).  
> But then again it brings us back to the original argument, and I
> probably would lean slightly towards the ISA model and be part of
> mips.igen depite polluting the whole file with *mips32r2: and *mips64r2:
> tags.
> 
> David.
> 
>