Ok.
Here is a report after simple debugging with "traps" like pa_log("test");
The call stack:
pa_alsa_sink_new -> alsa-sink.c
pa_sink_put(u->sink) -> sink.c
pa_assert_se(sink_set_state(s, PA_SINK_IDLE) == 0) -> sink.c -----> it
hangs here.
I'll continue debugging tomorrow.
2012/11/21 Tanu Kaskinen <tanuk at iki.fi>:
> On Wed, 2012-11-21 at 10:53 +0000, Colin Guthrie wrote:
>> 'Twas brillig, and Stefan Stefanov at 21/11/12 07:35 did gyre and gimble:
>> > Colin ask me to post output from "pulseaudio -vvvvv". Here is the
>> > result as attachment.
>>
>> So, the attachment looks mostly OK, things start up sink is stated etc.
>> all is looking well...
>>
>> Can you do some other tests?
>>
>> 1. Can you make sure no pulseaudio processes are running (keep killing
>> them until they die!), then run: "pulseaudio --start" This should start
>> PA in the background and give you control back to your shell. From what
>> we've seen so far, it might be that it "sticks" and does not give you
>> your shell back.
>
> This has been already debugged a bit further in IRC. PulseAudio is
> getting stuck in the module loading phase, so the daemon startup never
> completes, and thus "pulseaudio --start" never finishes. The last
> messages are about the alsa sink thread starting up, or if
> module-udev-detect is commented out, then about the null sink thread
> starting up. The problem seems to be either pa_thread_new() never
> returning or something else getting stuck that is done both by the alsa
> sink and the null sink soon after starting the thread. I suggested
> putting a log message right after the pa_thread_new() call (to see if
> pa_thread_new() returns or not), results from testing that suggestion
> have not been reported so far.
>
> --
> Tanu
>
> _______________________________________________
> pulseaudio-discuss mailing list
> pulseaudio-discuss at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/pulseaudio-discuss
