RE: [PATCH] cxl/region: use region (not root decoder) granularity for calculations

[Dan Williams <dan.j.williams@intel.com>]

Jim Harris wrote:
> Root decoder granularity must match value from CFWMS, which may not
> be the region's granularity for non-interleaved root decoders.
> 
> So when calculating granularities for host bridge decoders, use the
> region's granularity instead of the root decoder's granularity to ensure
> the correct granularities are set for the host bridge decoders and any
> downstream switch decoders.
[..]
> diff --git a/drivers/cxl/core/region.c b/drivers/cxl/core/region.c
> index 6d63b8798c29..70f7c66ee2ce 100644
> --- a/drivers/cxl/core/region.c
> +++ b/drivers/cxl/core/region.c
> @@ -1133,7 +1133,7 @@ static int cxl_port_setup_targets(struct cxl_port *port,
>  	}
>  
>  	if (is_cxl_root(parent_port)) {
> -		parent_ig = cxlrd->cxlsd.cxld.interleave_granularity;
> +		parent_ig = p->interleave_granularity;
>  		parent_iw = cxlrd->cxlsd.cxld.interleave_ways;

So the reason I went looking for the interleave_granularity_store()
version of the patch was only because this one when viewed in isolation
has the problem that the root-decoder granularity can only be overridden
to the region granularity when root-decoder interleave ways is 1. 

Is this assumed safe because interleave_granularity_store() is enforcing
that regions must match root-decoder granularity? If "yes", then this
likely wants a "/* See interleave_granularity_store() ... */" like
comment to describe why this override is safe, if "no" then I am missing
something.

What I liked about the idea of putting this in
interleave_granularity_store() is keeping all the simplification
shortcuts and commentary in one place, in case the region interleave !=
root-decoder interleave ban is lifted.

I don't mind doing this fixup in cxl_port_setup_targets(), but would
want the comment per above.