Grant Grundler wrote:
BTW, this discussion really belongs on parisc-linux.org mailing list.
It's not specific to debian-hppa and is very kernel centric.
On Fri, Aug 08, 2003 at 09:10:44AM +0000, Joel Soete wrote:
IIRC, "non-equivalently mapped aliases"
Is there much more detailed public doc some where about this stuff?
sorry - I don't understand the problems well enough.
The PA 2.0 arch book might describe the cache behaviors sufficiently
to determine this will be a problem.
BTW< this topic has been thoroughly discussed on parisc-linux
mailing list. See www.parisc-linux.org mailing list archives.
google search for
equivalent mapped [parisc-linux site:lists.parisc-linux.org]
yielded the original thread from Jerry Huck:
http://lists.parisc-linux.org/pipermail/parisc-linux/1999-December/008101.html
I see I was not yet in the list; at this very moment I was still battle
to build a kernel from hpux and the first cd safe me :)
I only partially understand what Jerry talking about and have no idea
how linux VM implementation collides with the aliasing rules.
Bjorn Helgaas posted a another thread two years later that might
be more helpful since it's specific to superdome (AFAIK same problems
as with N-class and L3000):
http://lists.parisc-linux.org/pipermail/parisc-linux/2001-April/012388.html
I didn't see any direct followup to Bjorn's mail in the archive though
some discussion did occur in other forums (much later).
Well better starting point I will gona check,
Thanks a lot,
Joel
