[alsa-devel] [RFC] SGI O2 MACE audio ALSA module

[TJ]

Hi,

Thank you for trying it.

On 11/07/07, sknauert at wesleyan.edu <sknauert at wesleyan.edu> wrote:
> Thanks for the quick work. Gave it a try.
>
> Shiva:/usr/src/linux-2.6.21.6# make CROSS_COMPILE=mips-linux-gnu- all
> ...
>   Building modules, stage 2.
>   MODPOST 3 modules
> WARNING: "snd_pcm_period_elapsed" [sound/mips/snd-mace-audio.ko] undefined!
> WARNING: "snd_pcm_set_ops" [sound/mips/snd-mace-audio.ko] undefined!
> WARNING: "snd_pcm_lib_ioctl" [sound/mips/snd-mace-audio.ko] undefined!
> WARNING: "snd_pcm_new" [sound/mips/snd-mace-audio.ko] undefined!
> make[1]: *** [__modpost] Error 1
> make: *** [modules] Error 2
>

I have never seen this before, it looks like the ALSA core/support
modules have not been build on which my module depends on. Here's my
SND/ALSA config:

#
# Advanced Linux Sound Architecture
#
CONFIG_SND=m
CONFIG_SND_TIMER=m
CONFIG_SND_PCM=m
CONFIG_SND_HWDEP=m
CONFIG_SND_RAWMIDI=m
CONFIG_SND_SEQUENCER=m
CONFIG_SND_SEQ_DUMMY=m
CONFIG_SND_OSSEMUL=y
CONFIG_SND_MIXER_OSS=m
CONFIG_SND_PCM_OSS=m
CONFIG_SND_PCM_OSS_PLUGINS=y
CONFIG_SND_SEQUENCER_OSS=y
CONFIG_SND_RTCTIMER=m
CONFIG_SND_SEQ_RTCTIMER_DEFAULT=y
CONFIG_SND_DYNAMIC_MINORS=y
CONFIG_SND_SUPPORT_OLD_API=y
CONFIG_SND_VERBOSE_PROCFS=y
CONFIG_SND_VERBOSE_PRINTK=y
CONFIG_SND_DEBUG=y
CONFIG_SND_DEBUG_DETECT=y
CONFIG_SND_PCM_XRUN_DEBUG=y

CONFIG_SND_MACE_AUDIO=m
CONFIG_SND_MACE_AUDIO_SPY=m

I guess I should be adding more depends to my Kconfig, I've never
tried to see what the bare minimum alsa modules needed are.

> Surprisingly, make modules before trying to compile the kernel didn't give
> an error, but these warnings:

'make all' includes 'make modules'. I only ever use make all, which is
fine since make does what it's supposed to do and only rebuilds stuff
whose source files have changed.


>   LD [M]  sound/core/snd.o
>   CC [M]  sound/mips/mace_audio.o
> sound/mips/mace_audio.c: In function 'sma_pcm_open':
> sound/mips/mace_audio.c:398: warning: unused variable 'chi'
> sound/mips/mace_audio.c: In function 'sma_gain_put':
> sound/mips/mace_audio.c:893: warning: unused variable 'flags'
> sound/mips/mace_audio.c: In function 'sma_proc_mace_read':
> sound/mips/mace_audio.c:1023: warning: format '%lX' expects type 'long
> unsigned int', but argument 3 has type 'unsigned char *'
> sound/mips/mace_audio.c:1023: warning: format '%lu' expects type 'long
> unsigned int', but argument 4 has type 'unsigned char *'
> sound/mips/mace_audio.c: In function 'sma_create':
> sound/mips/mace_audio.c:1169: warning: passing argument 2 of 'request_irq'
> from incompatible pointer type
> sound/mips/mace_audio.c:1136: warning: unused variable 'mringbase'
> sound/mips/mace_audio.c: At top level:
> sound/mips/mace_audio.c:1060: warning: 'sma_proc_debug_append' defined but
> not used
> sound/mips/mace_audio.c:372: warning: 'print_pointers' defined but not used

This my test/dev code expect complaints about printf (%X for unsigned
char), I could put in explicit type casts, but I do plan to remove all
the spurious printk calls at some point and this helps to remind me.

Out of all of those I have to admit the 'unused variable flags' was an
oversight by me, I meant to remove that..

> This seems like a minor oversight like a patch from the wrong directory.
>
> However, if this is a real bug, I'm using gcc 4.1.1-21 with the MIPS
> cross-compiler and build system from people.debian.org. I haven't tried
> native compilation on the O2 yet. Let me know what you're using to get
> something working and I can be more helpful.

I compile both on the O2 (occasionally) and on a Debian cross dev
environment, built using:
http://people.debian.org/~ths/toolchain/..../buildcross

> I applied the patch -p1 < (let me know if it needs something extra) and
> tried both the Linux-MIPS 2.6.21.6 (which isn't linked on the homepage for
> some reason) and mainline from kernel.org with my working .config from
> 2.6.21.3 with make oldconfig.

I would be surprised if it patched a kernel.org kernel, but it should
apply fine the linux-mip.org kernel. I will have a look to check. Did
you use the patch from the email, that may have been mangled? (It was
a bit of a test to see have Gmail behaves) or the one I linked to on
lm.o? (below)

> Again thanks for all the hard work on this driver. I can't wait to get it
> working on my O2.

Same here, and thank you for looking at the patch/module.

Thorben

http://www.linux-mips.org/pub/linux/mips/people/trevelyan/mace_audio/mace_audio-20070709.patch