1 |
It turned out the issue with the IOMMU was a BIOS setting -- and a totally |
2 |
unsuspected one at that! Of course, the AMD IOMMU device is part of the |
3 |
AGP-GART, and I had never been able to get DRM/DRI working, either. The |
4 |
problem with both was one and the same -- I had Fast-writes (FW) enabled |
5 |
in the AGP Hypertransport configuration. Turning it off allowed both the |
6 |
hardware IOMMU and DRM to work. As I had earlier figured out that I |
7 |
couldn't boot with the hardware IOMMU enabled and had to use the software |
8 |
one, at least past kernel 2.6.15 as they killed the SCSI bouncebuffers for |
9 |
.16, I tested first with that. Once I got it working, I had a good idea |
10 |
the DRM/DRI would work as well, and so it did! No more being stuck with |
11 |
Mesa software 3D! Hallelujah and much rejoicing! I had been fighting |
12 |
with that on and off for quite awhile (over a year, tho it got much more |
13 |
urgent when I upgraded from a gig to 8 gig of RAM and needed the IOMMU), |
14 |
and had almost come to believe that part of my mobo chipset was borken. |
15 |
Glad it wasn't! =8^) |
16 |
|
17 |
-- |
18 |
Duncan - List replies preferred. No HTML msgs. |
19 |
"Every nonfree program has a lord, a master -- |
20 |
and if you use the program, he is your master." Richard Stallman in |
21 |
http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html |
22 |
|
23 |
|
24 |
-- |
25 |
gentoo-amd64@g.o mailing list |