Re: [PATCH] MIPS32 DSP intrinsics

[Thiemo Seufer <ths at networkno dot de>]

Richard Sandiford wrote:
> Eric Christopher <echristo@redhat.com> writes:
> > Nigel: Provided Richard has no objections and I get a promise from you
> > to submit simulator support as soon as the description is released
> > (first patch should be submitted no more than a week after...) then I'll
> > withdraw my objection to putting the patches in on this basis. (You
> > still have to get them up to snuff for both Richard and I on other
> > fronts).
> 
> FWIW, I was kind-of sympathetic to your position ;)
> 
> We seem to have a situation in which MIPS still consider the details of
> the ASE to be "too secret" for either the simulator code or user-level
> documentation to be released.  That means:
> 
>   (a) no-one outside MIPS can really use the code at this stage.
>   (b) no-one outside MIPS can test the code properly at this stage.
>   (c) no-one outside MIPS can sensibly change the code at this stage.

That's not necessarily true, since MIPS might have early adopters which
signed an NDA. I can't see how forcing MIPS into providing them with a
gcc fork is helpful. About the testing issue, I think MIPS is interested
in keeping it work, so it isn't an immediate problem.

Anyways, the worst case would be a reversion of the relevant patches,
and that's AFAICS less work than this whole discussion.


Thiemo