1 |
sean <tech.junk@×××××××××××.net> posted 4A492614.1050603@×××××××××××.net, |
2 |
excerpted below, on Mon, 29 Jun 2009 16:37:40 -0400: |
3 |
|
4 |
> I use xine for both CD and DVD, and a xine-out.wav is generated if |
5 |
> alsaconig has not been run again after a system startup. |
6 |
|
7 |
This sounds to me like one of the sound modules isn't loaded. ALSA has |
8 |
always started muted, until it's unmuted, but that can't be it, as xine |
9 |
should then play, it'd just be muted. But if one of the necessary |
10 |
modules wasn't loaded, xine would presumably detect that by the failure |
11 |
to open the device at all, thus the xine-out.wav. |
12 |
|
13 |
But running the config loads the modules, which apparently are NOT auto- |
14 |
loaded by the alsa service or whatever you have running at start to bring |
15 |
back sound -- that's assuming that you DO have something running to load |
16 |
sound at start. |
17 |
|
18 |
FWIW, I've never had an issue of that nature here. For quite some time I |
19 |
didn't run the alsa service at boot, since I only used sound in KDE, and |
20 |
KDE would start it when I started KDE. However, when I switched to an |
21 |
mpd based music system, I switched to starting sound at boot, and it |
22 |
"just worked". |
23 |
|
24 |
BUT, my kernel policy has for YEARS been to build-in everything possible |
25 |
that I use enough to keep loaded. Stuff like the floppy and loopback |
26 |
drivers don't get built-in, as I very seldom use them and thus might as |
27 |
well not load them unless I DO use them. But all my sound drivers get |
28 |
built-in, because if I'm loading them at boot and never unloading them, I |
29 |
might as well. So presuming the problem is what I think it is, I |
30 |
couldn't have it here, because the sound drivers are built-in, not loaded |
31 |
as modules, so they're always loaded. |
32 |
|
33 |
Of course, that can get a bit complicated for those drivers that need |
34 |
parameters fed to them at load. Loading the modules is then easiest. |
35 |
However, most of them can have the parameters fed to the kernel directly |
36 |
on the kernel command line, as fed to it from grub (or with newer |
37 |
kernels, 2.6.29 maybe??, from the built-in kernel commandline option, |
38 |
which sure simplified /my/ grub kernel line!). The problem then is |
39 |
finding the documentation telling you exactly what form to use on the |
40 |
kernel commandline, since the options take a bit different form there, |
41 |
including the driver name, etc, as it's otherwise not obvious what they |
42 |
refer to. But in most cases it can work, you just have to find out |
43 |
exactly what to feed the kernel commandline. |
44 |
|
45 |
So that's what I'd suggest. If your sound card drivers do NOT require |
46 |
driver commandline options, build them in. If they DO, it's a bit more |
47 |
complicated as you'll have to find documentation covering the exact |
48 |
format of the kernel commandline options to feed, but look around in the |
49 |
kernel documentation dir and do some googling and see what comes up. If |
50 |
you can find what to feed the kernel, do so, build-in the drivers, and |
51 |
that will hopefully eliminate the issue. Of course, this assumes that |
52 |
you are loading the drivers at boot and not unloading them, so building |
53 |
in the drivers doesn't increase the running size of the kernel |
54 |
uselessly. But I expect that's the case for the folks having the problem |
55 |
here, or they'd already be used to loading and unloading their sound |
56 |
drivers as needed and thus wouldn't be having the problem with the |
57 |
failure of the automatic system to load them. |
58 |
|
59 |
-- |
60 |
Duncan - List replies preferred. No HTML msgs. |
61 |
"Every nonfree program has a lord, a master -- |
62 |
and if you use the program, he is your master." Richard Stallman |