Re: [Nasm-devel] [PATCH 6/7] AVX-512: Change the data type for instruction flags

[Cyrill G. <gor...@gm...>]

On Wed, Aug 28, 2013 at 10:33:19AM +0000, Song, Jin Kyu wrote:
> > > > Why you've changed flags here and a couple of other places?
> > >
> > > The reason is actually written in the commit message. "Since they are
> > not bit masks, only one
> > > instruction set is allowed for each instruction." So both SSE and MMX
> > could not be set for
> > > one instruction. As nasm64developer mentioned in his email, this may not
> > be a proper way
> > > to expand and define bits for instruction sets. And it needs a major
> > restructuring of
> > > instruction flags not a simple fix of increasing data type size.
> > 
> > Yes, feature flag is broken, i know. Can't such change affect existing
> > code?
> 
> I thought this would not affect existing code because these feature flags were
> set to zero until now. The only place using this flag I could find was
> 'preferred disasm' part. So I slightly modified the mask bits to make it
> compatible with that part of code.
> 
> Please correct me if I am wrong since I do not think I fully understood
> the whole code history of NASM.

Yup, sorry for warning, it were all zeros previously, but still I think what
we need here is

 - fix this 64 bitfield to generic bitmap (I've it in my todo list, actually
   for too long alredy :-)

 - keep all IF_ flags you don't need at moment to zeros but change IF_ value
   for those instructions which you need to distinguish in avx512 sake, nothing
   else. Once generic bitmaps will be ready we simply update all IF_ flags
   to have different bits set and this resolve all problems, no?