Re: [RFC PATCH 1/4] PCI/TSM: Create MMIO descriptors via TDISP Report

[Ankit Agrawal]

Thanks for the review. Comments inline.

>>
>>  While not wrong per-se having a strict check would be far clearer, no?
>
> I agree, if (i  == mmio->nr) is clearer.

Ok sure, I'll change it.

> Could you help specify in the changelog how to ensure the obfuscated
> address BAR aligned. I believe this is not mandated in PCI TDISP spec,
> so we need to let all TSM vendors know about the rule.
>
> I remember it is done by TSM setting the MMIO_REPORTING_OFFSET as the
> largest BAR size on LOCK_INTERFACE_REQUEST, is it?

Will add. DEN0137-2.0-bet2 A9.6.2 has "This algorithm assumes that the
MRO (MMIO reporting Offset) value is a multiple of the BAR size.". It asks to
derive in-BAR offset as region_base % bar_size.

ARM also recommends that RMM pick MRO as x + 2^n, 2^n >= largest BAR.
SO it is a pseudo random multiple of it. I'll mention about this assumption on
TSM implementation alongwith the spec section.