[PATCH 3/5] documentation/iommu: update description of ARM System MMU binding

[Will Deacon]

On Fri, Feb 28, 2014 at 05:07:31PM +0000, Timur Tabi wrote:
> On 02/28/2014 10:21 AM, Will Deacon wrote:
> >>> > >+- calxeda,smmu-secure-config-access : Enable proper handling of buggy
> >>> > >+                  implementations that always use secure access to
> >>> > >+                  SMMU configuration registers. In this case non-secure
> >>> > >+                 aliases of secure registers have to be used during
> >>> > >+                 SMMU configuration.
> >> >
> >> >I'm confused.  Why does this property have a "calxeda" prefix?  How is
> >> >it a Calxeda-specific property?
> 
> > Because they wired up their SMMU backwards. It's basically an
> > implementation-specific erratum workaround.
> 
> Hmmmm....
> 
> Other than making the same wiring mistake, is there any reason any other 
> ARM chip would need this property set?

Not that I can think of, no.

> The reason I ask is that it's kinda weird (well, to me at least) that we 
> have an property named for a specific SoC, but the implementation and 
> documentation tries so hard to hide that fact.  I would think that the 
> binding document would provide some explanation as to why the property 
> has a "calxeda" prefix.

I don't agree. We're making sensible use of a vendor prefix to isolate
errata workarounds. The description clearly states that the implementation
is buggy.

Given that there aren't any Calxeda engineers left working on this stuff,
I'm heavily inclined to leave this patch as-is.

Will