1 |
Jeff Cranmer writes: |
2 |
|
3 |
> Regarding eselect, I used eselect opengl xorg-xll to select the open |
4 |
> source driver instead of ati. I think this is correct? |
5 |
|
6 |
Sounds right. |
7 |
|
8 |
> The kernel options are configured as follows |
9 |
> |
10 |
> Device Drivers -> |
11 |
> Generic Driver Options -> |
12 |
> <*> Userspace firmware loading support |
13 |
> [*] Include in-kernel firmware blobs in kernel binary |
14 |
> Graphics support -> |
15 |
> <*> /dev/agpgart (AGP Support) |
16 |
> <*> Direct Rendering Manager (XFree86 4.1.0 and higher DRI support) |
17 |
> -> <*> ATI Radeon |
18 |
> [*] Enable modesetting on radeon by default |
19 |
|
20 |
The help for this last setting tells: |
21 |
|
22 |
This is a completely new driver. It's only part of the existing drm |
23 |
for compatibility reasons. It requires an entirely different graphics |
24 |
stack above it and works very differently from the old drm stack. |
25 |
i.e. don't enable this unless you know what you are doing it may |
26 |
cause issues or bugs compared to the previous userspace driver stack. |
27 |
[...] |
28 |
This code support the following hardware : R1XX,R2XX,R3XX,R4XX,R5XX |
29 |
radeon up to X1950). Works is underway to provide support for R6XX, |
30 |
R7XX and newer hardware (radeon from HD2XXX to HD4XXX). |
31 |
|
32 |
You seem to have a HD 5700 card, even newer than the HD4XXX which is not |
33 |
yet supported. So I'd turn this option off and try again. |
34 |
|
35 |
I have the driver compiled as a module, and get these messages in dmesg |
36 |
when loading: |
37 |
[drm] Initialized drm 1.1.0 20060810 |
38 |
[drm] radeon defaulting to userspace modesetting. |
39 |
fglrx_pci 0000:01:05.0: setting latency timer to 64 |
40 |
[drm] Initialized radeon 1.33.0 20080528 for 0000:01:05.0 on minor 0 |
41 |
[drm] Setting GART location based on new memory map |
42 |
[drm] Loading RS780 CP Microcode |
43 |
[drm] Resetting GPU |
44 |
[drm] writeback test succeeded in 1 usecs |
45 |
[drm] Resetting GPU |
46 |
|
47 |
I suggest compiling as a module, so you can easier check for kernel |
48 |
messages when loading the module, and it allows to try the closed-source |
49 |
ati-drivers. I had a memory problem with them (but did not try the last |
50 |
version), and had to use an older X server, so I gave the open-source |
51 |
drivers a try. They work, with less memory and CPU usage, but I get |
52 |
distortions. I also have to use the older xorg-server 1.7, because with |
53 |
1.8 X crashed every few minutes. |
54 |
|
55 |
Whoops, I just see that I also have the fglrx module running, this should |
56 |
not happen. Could this be the cause for my graphics distortions? Let's |
57 |
see. |
58 |
|
59 |
Wonko |