Re: [PATCH v5] cxl/region: check interleave capability

[Jonathan Cameron <Jonathan.Cameron@Huawei.com>]

> >   
> > > +	 * interleave bits is 0, there is no interleaving, the following
> > > +	 * check is ignored.  
> > 
> > If the interleave is 3 there is no interleave?  That seems an odd statement
> > perhaps make that comment more detailed.
> > My understanding is that it's just more complex and all bits are relevant.  
> yes, but it is difficult to describe the interleave bits. 
> In the SPCE " 8.2.4.20.13 Decoder Protection ",  only when IW < 8, it called 
> [IG + IW +7: IG+8] as IW bits.
> but when IW >= 8, what is the IW bits definition? 
> 
> In my opinion:
> if IW = 8, all the bits of HPAOFFSET[51:0] is used,  so the IW bits is none.
> if IW >8, the bits [IG+IW-1: IG+8] of HPAOFFSET is not used, so the IW bits should be  [IG+IW-1: IG+8],

I'd just add a note that things are more complex for multiples of 3 and
drop the 'no interleave' comment.
It's an early validation function to give more useful feedback than
the current no feedback. I'm not sure it needs to be perfect in catching
issues as long as we don't think it will give false rejections of valid
parameters.