1 |
On Saturday 10 January 2009, Alan McKinnon wrote: |
2 |
> On Saturday 10 January 2009 12:26:09 Mick wrote: |
3 |
> > On Saturday 10 January 2009, Dale wrote: |
4 |
> > > Neil Bothwick wrote: |
5 |
> > > > On Fri, 9 Jan 2009 22:37:44 +0000, Mick wrote: |
6 |
> > > >> Filed a bug and it was suggested to me that I try |
7 |
> > > >> building the alsa drivers as modules. I tried it on for size and |
8 |
> > > >> guess what, it worked! |
9 |
> > > > |
10 |
> > > > AFAIR the Gentoo ALSA docs have always recommended building as |
11 |
> > > > modules. |
12 |
> > > |
13 |
> > > And yet I have never used modules for anything except Nvidia. That has |
14 |
> > > worked for me over 5 years now. What a puzzle. |
15 |
> > |
16 |
> > Same here. With three different machines I had alsa built in the kernel, |
17 |
> > less that six months or so after alsa first became part of the kernel |
18 |
> > (can't remember if I was running stable back then). |
19 |
> |
20 |
> My money says that if you track this down to the exact lines of code that |
21 |
> cause the problem, you'll find something along the lines of incorrect |
22 |
> loading order. As in, module A must load before module B, but you compiled |
23 |
> B into the kernel. Plus no dev picked it up and nobody coded a check for |
24 |
> it. |
25 |
|
26 |
I noticed this when I built a kernel with sound as modules: |
27 |
|
28 |
CC [M] sound/core/pcm.o |
29 |
CC [M] sound/core/pcm_native.o |
30 |
sound/core/pcm_native.c: In function 'snd_pcm_fasync': |
31 |
sound/core/pcm_native.c:3255: warning: label 'out' defined but not used |
32 |
CC [M] sound/core/pcm_lib.o |
33 |
CC [M] sound/core/pcm_timer.o |
34 |
|
35 |
also noticed: |
36 |
|
37 |
LD [M] sound/pci/ac97/snd-ac97-codec.o |
38 |
LD sound/synth/built-in.o |
39 |
LD sound/usb/built-in.o |
40 |
LD vmlinux.o |
41 |
MODPOST vmlinux.o |
42 |
WARNING: modpost: Found 1 section mismatch(es). |
43 |
To see full details build your kernel with: |
44 |
'make CONFIG_DEBUG_SECTION_MISMATCH=y' |
45 |
|
46 |
> I believe the correct technical term for this is "a software bug" :-) |
47 |
> |
48 |
> It's usually worth the effort to report a bug, the least that can happen is |
49 |
> someone else's life is easier in the future |
50 |
|
51 |
I have filed bug 253110 for my problem (sound) but not for the OP. |
52 |
-- |
53 |
Regards, |
54 |
Mick |