Re: [PATCH v2] cxl/region: Allow 6 & 12 way regions on 3-way HB interleaves

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

On Mon, 17 Mar 2025 20:08:24 -0700
Alison Schofield <alison.schofield@intel.com> wrote:

> On Fri, Mar 14, 2025 at 12:00:36PM +0000, Jonathan Cameron wrote:
> > On Thu,  6 Mar 2025 15:22:37 -0800
> > alison.schofield@intel.com wrote:
> >   
> > > From: Alison Schofield <alison.schofield@intel.com>
> > > 
> > > The CXL driver requires the granularity of a region and its root
> > > decoder to be the same. This is particularly restrictive for 3-way
> > > host bridge interleaves where the only spec defined interleave
> > > configurations for creating 6-way and 12-way regions on a 3-way HB
> > > interleave require mixed granularities.
> > > 
> > > CXL 3.2 Specification 9.13.1.1:
> > > Legal Interleaving Configurations: 12-way, 6-way, and 3-way  
> > 
> > Ah this is finally a valid reason to do coarser interleave first
> > (going away from host). That was subject of long discussions way back
> > when original interleaving code was discussed (mostly because
> > my mental model did it that way around and I couldn't follow what
> > the kernel code was doing).
> > 
> > We 'could' revisit allowing this more generally - at least for
> > already configured set ups.  Only real reason for that is
> > either that someone ships a config that does it for a different
> > case, or that we want to avoid special casing 3*x cases.  
> 
> Hi Jonathan,
> 
> I'll bring these configs up as an OPEN at the upcoming community collab,
> and hope you can elaborate more there.
> 

Sure. I'll try and remember how all this worked!

I'm not really suggesting we do handle those other cases unless we know
they are in the wild or it simplifies the code by removing special cases
(I'm not sure it does!)

Jonathan

> --Alison
> 
> 
> > 
> > Jonathan  
>