Re: [Jack-Devel] Debugging Jack Clients

On Tue, Sep 6, 2011 at 7:37 PM, Gabriel Beddingfield <[hidden]>wrote:

> On 09/06/2011 01:29 PM, ccrome wrote:
>
>> Ah, perhaps I've found the issue.  I guess when Jack try to re-start, it
>> reconfigures the codec.  I'm using a custom audio card that takes some
>>
>
> No... it shouldn't try to reconfigure the codec.


Huh, it's definitely touching the I2C bus.  Dunno why though.


>
>
>  time to re-configure, and it seems to enter a continuous loop where it
>> tries to configure the codec, then it gets an xrun, so it tries again.
>>
>
> If this loop is happening... it's a side-effect of the Real Problem. This
> is OMAP3, right?


Why yes.  Yes it is.  What do you suspect the Real Problem is ?



>
>
>         creating alsa driver ...
>>        hw:0|hw:0|1024|2|48000|0|0|**nomon|swmeter|-|16bit
>>        control device hw:0
>>        configuring for 48000Hz, period = 1024 frames (21.3 ms), buffer
>>        = 2 periods
>>
>
> Try setting larger buffers.  E.g. -p 4096 -n 4


Will do when I get back to work.
We're trying to keep latency very low though.



>
>
>         # gdb client
>>        gdb> break process
>>        gdb> run
>>        Starting program:
>>        /home/caleb/mmfx/platform/**platform_beagle/puppy_driver/**
>> jack_scripts/jack_sine
>>
>>        [Thread debugging using libthread_db enabled]
>>        [New Thread 0x40a73470 (LWP 22052)]
>>        engine sample rate: 48000
>>        [New Thread 0x412f6470 (LWP 22055)]
>>           ^C
>>        Program received signal SIGINT, Interrupt.
>>        0x401e9f44 in nanosleep () from /lib/arm-linux-gnueabi/libc.**so.6
>>        (gdb) bad status (254) for client sine handling event (type = 8)
>>        bad status (254) for client sine handling event (type = 8)
>>        bad status (254) for client sine handling event (type = 8)
>>        gdb>
>>        gdb> continue
>>
>>
>>    The callback 'process' is never called again, and the jackd_log file
>>    is filled with alsa xruns.
>>
>>    Any thoughts?
>>
>
> Even though you set the timeout at 50 seconds... it appears that this error
> came because you didn't respond to an EVENT NOTIFICATION within 2 seconds.
>  (The 2secs is not configurable.)  Event type 8 is an xrun. Did you perhaps
> set an xrun callback?  If you did, I recommend that you disable it while
> debugging.
>

No, I didn't intentionally set an xrun callback. I do recall that I'm using
a deprecated jack_init api (will take a look when I get back to work).  I
don't recall the exact function call that was deprecated.  I'll take a look.

Thanks,
  -Caleb



> -gabriel
>

1315363858.4822_0.ltw:2,a <CAG5mAdyKHCFy-grERYZQgXZK5MOYnCTXX94A7zQyGwQY-eLWuA at mail dot gmail dot com>