| 1 |
On Wed, 2002-01-23 at 11:57, Nero24 (Planet 3DNow!) wrote: |
| 2 |
> Hello Chris, |
| 3 |
> |
| 4 |
> regarding the Large Pages Bug with Windows 2000 and Linux there are |
| 5 |
> different version out, whether the bug is fixed in Palomino or not. So we |
| 6 |
> decided to stop rumoring and asked AMD public relations manager Jan Gütter. |
| 7 |
> What he said you can read at Planet 3DNow! |
| 8 |
> www.planet3dnow.de/cgi-bin/newspub/viewnews.cgi?category=1&id=1011807957 |
| 9 |
|
| 10 |
The INVPLG CPU bug is not the particular issue being discussed here |
| 11 |
(yes, this has all been very confusing). The issue we are talking about |
| 12 |
is a subtle interaction between speculative writes, extended paging and |
| 13 |
the GART (not a CPU bug, this was a misunderstanding). This quirk |
| 14 |
results in corruption of AGP and system memory, and it's really the |
| 15 |
responsibility of the Linux kernel and drivers to make sure that this |
| 16 |
issue doesn't happen. See my post at |
| 17 |
http://www.geocrawler.com/lists/3/Linux/35/75/7626960/ which contains an |
| 18 |
update including an official explantion from AMD of a problem that can |
| 19 |
happen and may be happening to Linux people. Follow the thread and |
| 20 |
you'll see that it's a real issue. It is a very interesting problem. |
| 21 |
|
| 22 |
The Linux kernel developers are currently investigating this speculative |
| 23 |
write/GART interaction to see if it could affect Linux users. So far, |
| 24 |
the consensus seems to be that it *does* affect Linux AGP users and |
| 25 |
developers are currently trying to find a solution that doesn't not hurt |
| 26 |
performance as "mem=nopentium" (turning off extended translation) does. |
| 27 |
|
| 28 |
Best Regards, |
| 29 |
|
| 30 |
-- |
| 31 |
Daniel Robbins <drobbins@g.o> |
| 32 |
Chief Architect/President http://www.gentoo.org |
| 33 |
Gentoo Technologies, Inc. |
Archives