poll: problems with MTRRs?

[D. Hugh Redelmeier]

| From: Lennart Sorensen <lsorense-1wCw9BSqJbv44Nm34jS7GywD8/FfD2ys at public.gmane.org>

| On Tue, Mar 17, 2009 at 11:23:39AM -0400, D. Hugh Redelmeier wrote:

| > I don't remember any Athlon 64 without an IOMMU.  Not that I would
| > necessarily know.  Do you have any more solid information?
| 
| Well I have an Athlon 64 3500+ under my desk which is one of the early
| ones that does not support AMD-V and hence has no IOMMU.  It is an old
| socket 939 chip.  It only has 1GB of ram though so it never mattered.
| The other disadvantage that CPU has is that you can not run a 64bit
| guest in a virtual machine on it in say vmware due to the lack of AMD-V.
| 
| The Cx, Dx, and Ex spec Athlon 64's don't have AMD-V.  The Fx, Gx and
| newer do.
| 
| As far as I can tell, no socket 754 or 939 Athlon64 had the IOMMU.  I am
| not entirely convinced that is right though (and it is from wikipedia
| so who knows).  All AM2 chips should have it.
| 
| On the opteron's, it seems the 1xx doesn't have it, while the 2xx, 8xx,
| 1xxx, 2xxx and 8xxx all have it.
| 
| I am pretty sure some of the socket 939 chips had some virtualization
| support that predates the AMD-V stuff, which allowed at least the 64bit
| guests to work in vmware, but didn't really do much performance wise.
| AMD-V apparently is much better.

OK, I was confused.  Now I am less confused, but more aware of my
confusion.

Linux calls the AMD GART an IOMMU (because it is, albeit a limited one).  
All the AMD 64-bit processors have the GART.  Intel CPUs do not.  The GART 
appears to be quite effective in eliminating the need for bounce buffers 
but not in segregating virtual machines' access to hardware.

AMD and Intel both have much fancier IOMMUs as part of the newish support 
for virtualization.  The Intel one (VT-D) requires chipset support.  I 
now think the AMD one probably does as well, but I'm not sure.  Anyway, 
I think that means that few of us have either unless we bought recent 
server hardware (and that might not be sufficient).

The best explanation that I've come across is in section 1 and 2 of:
  http://developer.amd.com/Assets/IOMMU-ben-yehuda.pdf
This paper was presented at OLS in 2006 but the OLS site does not
provide the papers as separate files.
  http://www.linuxsymposium.org/2006/linuxsymposium_procv1.pdf

Here's the spec for AMD's IOMMU.  Unfortunately (but properly) it does
not say where to find implementations:
  http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/34434.pdf
I don't even know how to tell if you have one.
--
The Toronto Linux Users Group.      Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists