On Sat, 2016-11-19 at 15:14 +1100, Paul Mackerras wrote: > > > These should be a device-tree property. We can fallback to hard wired > > values if it doesn't exist but we should at least look for one. > > Tell me what the property is called and I'll add code to use it. :) > That's the whole reason why I moved this to C code. > > > > > Note: P8 firmwares all have a bug creating a bogus "tlb-sets" property > > in the CPU node, so let's create a new one instead, with 2 entries > > (hash vs. radix) or 2 new ones, one for hash and one for radix (when > > available).
Well, as I said above, there's a defined one but it has bogus values on almost all P8 firwmares. So I think we need the core code to export values for use by both the core mm and KVM which can then be picked up from the DT with "quirks" to fixup the DT values. (A bit like I did for the never-applied cache geometry patches) That or we make up new names. The question remains whether we need a separate property for radix vs. hash though, we probably should as the "radix is half of hash" might not be true on future chips. Cheers, Ben.
