[Gta04-owner] Headset jack adapter

[NeilBrown]

On Fri, 05 Oct 2012 20:20:52 +0100 Neil Jerram <neil at ossau.homelinux.net>
wrote:

> Neil Jerram <neil at ossau.homelinux.net> writes:
> 
> >>> When I was playing with a headset at the weekend, I noticed stack traces
> >>> in the kernel log, in the area of ALSA (snd_pcm_...), I think at the
> >>> same times as I was plugging and unplugging the headset.
> >>
> >> Can you post them?
> >
> > I'll do that.  I tried reproducing them just after my last email, but
> > unfortunately they weren't obliging.  I'll have some more attempts over
> > the weekend.
> 
> I just looked at my dmesg for something else and noticed 3 copies of
> this:
> 
> [ 5795.269104] omap_dma: You need to stop the DMA channels before unlinking
> [ 5795.276306] [<c0012250>] (unwind_backtrace+0x0/0xe0) from [<c0025e80>] (omap_dma_unlink_lch+0x6c/0x8c)
> [ 5795.287811] [<c0025e80>] (omap_dma_unlink_lch+0x6c/0x8c) from [<c030b708>] (omap_pcm_hw_free+0x28/0x54)
> [ 5795.298522] [<c030b708>] (omap_pcm_hw_free+0x28/0x54) from [<c03070c4>] (soc_pcm_hw_free+0x88/0xd8)
> [ 5795.308837] [<c03070c4>] (soc_pcm_hw_free+0x88/0xd8) from [<c02e0ddc>] (snd_pcm_release_substream+0x44/0xa4)
> [ 5795.319946] [<c02e0ddc>] (snd_pcm_release_substream+0x44/0xa4) from [<c02e0e64>] (snd_pcm_release+0x28/0x6c)
> [ 5795.331146] [<c02e0e64>] (snd_pcm_release+0x28/0x6c) from [<c00a7c94>] (fput+0x104/0x23c)
> [ 5795.340484] [<c00a7c94>] (fput+0x104/0x23c) from [<c00a4adc>] (filp_close+0x70/0x7c)
> [ 5795.349395] [<c00a4adc>] (filp_close+0x70/0x7c) from [<c00a4bb0>] (sys_close+0xc8/0x120)
> [ 5795.359222] [<c00a4bb0>] (sys_close+0xc8/0x120) from [<c000d800>] (ret_fast_syscall+0x0/0x30)
> 
> I don't know if that's the same as I saw before when playing with the
> headset, but probably it's worth reporting anyway.
> 
>       Neil

Thanks....  somewhere deep in OMAP-magic land.
DMA still has a feel of Dragons May Abide to me.  The theory is easy enough,
but I'm a long way from being sufficiently immersed in the details.

Maybe adding
   omap_stop_dma(prtd->dma_ch);
just before
   omap_dma_unlink_lch(...)
in omap_pcm_hw_free() in sound/soc/omap/omap-pcm.c
would make  the message go away ... or might make it crash more colourfully.
Either might be helpful...

Hmmm... google finds:
  http://mailman.alsa-project.org/pipermail/alsa-devel/2008-December/013563.html

I wonder if it is relevant.

NeilBrown
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 828 bytes
Desc: not available
URL: <http://lists.goldelico.com/pipermail/gta04-owner/attachments/20121008/953a51a6/attachment.bin>