[Cbe-oss-dev] IOC translation & firmware question

[Arnd Bergmann]

On Tuesday 07 November 2006 08:02, Benjamin Herrenschmidt wrote:
> Is the firmware doing any setting of the IOC cache or IOC translation at
> all ? Is it left into the default power up state ? That is disabled ?

AFAIK, old firmware versions set it up the same way Linux used to do,
anything later than 5.50 should use a linear mapping. All versions
however set up the spider mappings to match the programmed DMA window.

> I've fully implemented the IOMMU support using Jeremy's work and
> adapting it to my new stuff, and it seems to work fine. However, I'd
> like to measure the performance impact of the IOMMU, so I have
> implemented support for IOMMU off, in which case, I leave the IOC in the
> state left by the firmware and pass direct memory pages addresses (with
> just the top bit set for spider) to devices.

I believe that breaks with old DD2 boards that have only 512MB of RAM
and are booted with yaboot instead of netboot, because those may
not have set up the IOC correctly. Hartmut, can you confirm that?

I don't think we care much about those old boards any more, but
if they break, we should document it somewhere.

	Arnd <><