1 |
OK, so let's have a go at the ATI drivers again. |
2 |
|
3 |
First, get rid of the open source ati drivers |
4 |
emerge --unmerge -av radeon-ucode xf86-video-ati |
5 |
|
6 |
Editing the /etc/make.conf file to change the last line to |
7 |
VIDEO_CARDS="fglrx" |
8 |
then running emerge -NDuav system changes the mesa driver so that I have |
9 |
no valid selections for "VIDEO_CARDS=", so I'll unmerge mesa |
10 |
|
11 |
emerge --unmerge -av mesa |
12 |
|
13 |
running emerge -NDuav world identified the xorg-drivers and pm-utils |
14 |
packages that have changed, and attempts to re-emerge mesa, so I'll |
15 |
unmerge those two |
16 |
|
17 |
emerge --unmerge -av xorg-drivers pm-utils |
18 |
|
19 |
moving on to the kernel, I verify that I have the correct kernel |
20 |
parameters: |
21 |
|
22 |
Device drivers -> |
23 |
Graphics Support -> |
24 |
<*> dev/agpgart (AGP support) |
25 |
|
26 |
None of the other modules below the agpgart appear to be appropriate, so |
27 |
I'll not enable any of these. |
28 |
|
29 |
Disable the direct rendering manager and framebuffer support, so this |
30 |
should take care of the radeon driver module. I'll do a locate search |
31 |
for this before I reboot. |
32 |
|
33 |
> Device Drivers -> |
34 |
> Generic Driver Options -> |
35 |
> < > Userspace firmware loading support |
36 |
> [ ] Include in-kernel firmware blobs in kernel binary |
37 |
|
38 |
Recompile the kernel and update using grub. |
39 |
|
40 |
emerge ati-drivers |
41 |
|
42 |
It seems this was still present from the previous attempt, so I'll |
43 |
re-emerge it. |
44 |
|
45 |
Running emerge -NDuav system still pulls in mesa, so back in it goes. |
46 |
Running emerge -NDuav world pulls back in xorg-drivers and pm-utils, so |
47 |
these are also reinstalled. |
48 |
|
49 |
emerge --depclean pulls out a few packages, |
50 |
radeontool |
51 |
gst-plugins-v4l |
52 |
gst-plugins-v4l2 |
53 |
gentoo-sources-2.6.34-r6 |
54 |
|
55 |
since gentoo-sources-2.6.34-r6 is the kernel I'm presently operating |
56 |
with, I'll re-emerge that and verify that it is still current via |
57 |
eselect kernel list. |
58 |
|
59 |
revdep-rebuild doesn't add anything else. |
60 |
|
61 |
Prevent boot-up into X-windows, then reboot. |
62 |
|
63 |
rc-update del xdm default |
64 |
reboot |
65 |
|
66 |
run aticonfig --initial to create an xorg.conf file. |
67 |
|
68 |
startx |
69 |
|
70 |
total system crash - hard reset required to reboot. |
71 |
|
72 |
Hmm - the xorg.log file has this error. |
73 |
atiddxDriScreenInit failed, GPS not been initialized. |
74 |
|
75 |
Can anyone shed light on this? |
76 |
|
77 |
Thanks |
78 |
|
79 |
Jeff |
80 |
|
81 |
|
82 |
On Fri, 2010-10-15 at 20:11 -0400, Jeff Cranmer wrote: |
83 |
> On Thu, 2010-10-14 at 10:32 +0300, Nikos Chantziaras wrote: |
84 |
> |
85 |
> > |
86 |
> > You mentioned that you used "fglrxinfo" in your post, which assumes |
87 |
> > you're using the closed source driver :-/ Another hint was that you're |
88 |
> > using an HD5000 series card, which is not supported correctly by the |
89 |
> > open source drivers anyway at this point. |
90 |
> > |
91 |
> Yeah - I realised afterwards that I had left a few things in place from |
92 |
> a failed attempt at installing the closed source drivers. |
93 |
> |
94 |
> > Anyway, if you're on open source drivers, I think that's your problem. |
95 |
> > Only the closed source driver works correctly with 3D for HD5000 cards |
96 |
> > at this moment. |
97 |
> > |
98 |
> > |
99 |
> Thanks for the advice. I'll take another crack at the closed source |
100 |
> drivers, and if that doesn't work, I'll rip out the $%^& thing and |
101 |
> replace it with an Nvidia card<g> |
102 |
> |
103 |
> Jeff |
104 |
> |
105 |
> |
106 |
> |