1 |
On Sun, Jan 22, 2012 at 8:06 AM, Michael Mol <mikemol@×××××.com> wrote: |
2 |
> On Sun, Jan 22, 2012 at 10:54 AM, Mark Knecht <markknecht@×××××.com> wrote: |
3 |
>> I've seen reports for years about folks having problems with some KVMs |
4 |
>> under Linux. I've never personally had one myself. However I've been |
5 |
>> helping a Windows friend break his Redmond addiction over the last few |
6 |
>> months using Gentoo. He has a nice 3 monitor KDE-based system that's |
7 |
>> been working fine but there was one monitor that refused to set up |
8 |
>> with the right resolution. We left it alone for a long time as it was |
9 |
>> usable but finally yesterday got together to figure out what was |
10 |
>> happening. From the title it should be clear that the problem was a |
11 |
>> KVM hooked to that one monitor. Removing the KVM completely solved the |
12 |
>> problem. |
13 |
>> |
14 |
>> Now, what I'm wondering is why this same video card/KVM/monitor |
15 |
>> combination which apparently worked in Windows should have any |
16 |
>> problems in Linux? Anyone know why? |
17 |
>> |
18 |
>> In the spirit of full discloser I don't really know that this |
19 |
>> _specific_ video card was tested in Windows, but he owns multiple |
20 |
>> NVidia 8400GS cards and it's my understanding that other 8400GS cards |
21 |
>> did work with this KVM & monitor, so unless it's this specific card |
22 |
>> having a defect, or even being just a bit weak in some way, it would |
23 |
>> seem to be the insertion of the KVM itself that upset things. |
24 |
>> |
25 |
>> Looking at the monitor's specs/requirements for running the higher |
26 |
>> resolutions it uses, as should not be a surprise, higher frequencies |
27 |
>> to do higher resolutions. If the KVM was filtering those a bit then |
28 |
>> it's possible things wouldn't work, but that doesn't explain why it |
29 |
>> did work in Windows. |
30 |
>> |
31 |
>> Basically, I looked around in Google for anyone that had real info |
32 |
>> about why this problem occurs, couldn't find any that made sense, and |
33 |
>> am wondering how to choose a KVM that's going to work out of the box |
34 |
>> short of asking for model numbers, etc. |
35 |
> |
36 |
> I assume these are VGA displays? |
37 |
> |
38 |
> I've noticed that the CRTs attached to my Win7 box at work don't get |
39 |
> configured for the highest refresh rate unless I force it. Also, I've |
40 |
> noticed it decide that '1280x1024' is the 'recommended' resolution for |
41 |
> my displays, though they'll do 1600x1200@60Hz. |
42 |
> |
43 |
> It could just be a matter of Windows using 75Hz instead of 85Hz, or |
44 |
> 60Hz instead of 75Hz. |
45 |
> |
46 |
> |
47 |
> -- |
48 |
> :wq |
49 |
> |
50 |
|
51 |
That could certainly somehow be part of it, although in the manual for |
52 |
the monitor (Acer 2216W) said the resolution of interest (1680x1050) |
53 |
only runs at one horizontal/vertical set of dot clock so it isn't like |
54 |
there was a choice there of down shifting and X just chose the lower |
55 |
rate. According to the manual, if Windows set up 1680x1050 then it |
56 |
must have been using the only rates, etc. (I think!) |
57 |
|
58 |
And yes, the interface on that monitor is the old-style VGA. |
59 |
|
60 |
It's fairly clear that X kept saying there wasn't a resolution |
61 |
available from the monitor to support what I had requested in the |
62 |
xorg.conf file. I was asking for 1680x1050, being told the monitor |
63 |
didn't support it, and then given 1280x1024 instead. When he removed |
64 |
the KVM all those messages went away, X & KDE said the monitor was |
65 |
running 1680x1050, and the OSD on the monitor itself said it was doing |
66 |
the requested setup. |
67 |
|
68 |
I played a bit with get-edid | parse-edid. Logically that stuff even |
69 |
working says the VGA monitor cable is bidirectional. I started |
70 |
wondering if the KVM messes up the data coming back, or what else |
71 |
might be going on. |
72 |
|
73 |
Thanks for the ideas, |
74 |
Mark |