Re: [Linux-ia64] Re: [PATCH] early machvec init for 2.5

On Thu, Jun 05, 2003 at 08:54:32AM -0700, Jesse Barnes wrote:
> Well, in talking to them, they didn't seem to think that they were
> violating the spec in any way.  They said that since our device wasn't a
> PCI to PCI bridge, DMA coherence vs. reads wasn't strictly necessary

The argument that the native chip set is not subject to PCI ordering
rules because they aren't PCI-PCI Bridges sounds specious to me.
IMHO, asserting the box is PCI 2.2 compliant implies the chip set
also complies. Unfortunately, for most people compliance is
defined by whether <insert favorite MS OS here> runs with the
add-on card they bought at local surplus electronics store.

To be fair, ZX1 chip set (and PARISC chip sets) can break some of
the rules too.  I'll present a few bits about that at OLS 2003.

...
> Yeah, I agree, but for awhile I wasn't sure if we could accomodate the
> Linux driver API semantics at all, so I'm just happy that we can.  Even
> if it is ugly and kludgy, it works.

ok. C'est la Vie.

thanks,
grant