1 |
Marco Matthies posted <4352EDBE.7030503@×××.net>, excerpted below, on |
2 |
Mon, 17 Oct 2005 02:18:06 +0200: |
3 |
|
4 |
> Here's someone who is also getting these huge reported memory sizes: |
5 |
> http://lkml.org/lkml/2005/9/12/35 |
6 |
> Found it by googling for 983552MB, there's more there. |
7 |
> |
8 |
> There's is some info on mtrr's with vesafb here: |
9 |
> /usr/src/linux-2.6.13-gentoo-r3/Documentation/fb/vesafb.txt, line 176 |
10 |
> Seems to suggest trying video=vesafb:nomtrr if you're using vesafb. |
11 |
> |
12 |
> Also some other reports seem to hint at Xorg as the culprit: |
13 |
> https://bugs.freedesktop.org/show_bug.cgi?id=4310 |
14 |
> and others found by googling "mtrr: type mismatch for". |
15 |
> |
16 |
> You might want to also grep Xorg.log for the addresses provided by |
17 |
> 'cat /proc/mtrr' or just the prefix such as |
18 |
> 'grep -in 0xd /var/log/Xorg.0.log' in my case, that should some results |
19 |
> (Xorg.log doesn't seem to mention mtrr literally, as far as i can see). |
20 |
|
21 |
Same here. The address range corresponds, but mtrr isn't mentioned by |
22 |
name in xorg.conf. |
23 |
|
24 |
You can check video memory (and a couple other items) by running lspci -v. |
25 |
Look for the line that has a size in megs, under the correct video device, |
26 |
of course... <g> |
27 |
|
28 |
This is all great fun and I sure learned a lot yesterday researching, then |
29 |
putting it into words therefore anchoring it in my memory as well as |
30 |
forcing myself to contemplate angles for the purpose of writing that I |
31 |
wouldn't necessarily see, otherwise. However, Saturday nite is my Monday, |
32 |
and I spent all of my Monday nite researching and writing this and a |
33 |
couple other technical replies... and got only an hour or so of sleep! |
34 |
<g> So... tonite I'm going to try to avoid getting too deeply involved in |
35 |
this and ignore the additional research I now want to do given the above |
36 |
links! <g> Maybe that way I can get some sleep today (day=nite for me, |
37 |
nite=day)! |
38 |
|
39 |
... One of the things that is making this interesting for me, is that |
40 |
I've been trying to run dual Radeons for some time, one AGP one PCI, each |
41 |
with two outputs so I can run FOUR 21" monitors instead of only TWO, like |
42 |
I used to run an NVidia (Twinview) with the closed drivers and an old |
43 |
cheap S3 Virge (giving me three monitors, then a 19" and two 17"), a |
44 |
couple years ago. However, I keep getting conflicting resources when I |
45 |
try. I've always wanted to take the time to see if I could grok things |
46 |
well enough to try and reconfigure around the issue, and this is giving me |
47 |
the motivation (and additional brain power) to grasp at least one angle of |
48 |
it. I'm not sure that's actually what's conflicting, as I haven't tried |
49 |
running it that way for awhile, but I'm at least developing enough |
50 |
understanding now to make a bit of sense of things and figure out from the |
51 |
errors where the problem actually is, now, when I get a chance to get back |
52 |
to it. That's /miles/ farther than I was, previously! |
53 |
|
54 |
In addition, few enough folks actually grok this stuff that if I can get |
55 |
it, I'm sure the knowledge will be of help to quite a few others, as well. |
56 |
|
57 |
Oh... Previously I was trying to run a separate "screen" config for each |
58 |
monitor, thus, two per card since each card has two outputs. I'm now |
59 |
running xorg's merged framebuffer for Radeons, so it's actually possible |
60 |
the problem will have disappeared when I actually try it again... |
61 |
|
62 |
So anyway, that's why I'm spending so much time on this myself... The |
63 |
knowledge gained at this point can be of help to at least two, and later |
64 |
others, so it's not just me gaining immediate benefit, nor is it just you. |
65 |
However, the research will work better when I'm rested, so tomorrow or |
66 |
later this week... |
67 |
|
68 |
-- |
69 |
Duncan - List replies preferred. No HTML msgs. |
70 |
"Every nonfree program has a lord, a master -- |
71 |
and if you use the program, he is your master." Richard Stallman in |
72 |
http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html |
73 |
|
74 |
|
75 |
-- |
76 |
gentoo-amd64@g.o mailing list |