[LAU] Google TV device w/ AV out

Patrick Shirkey pshirkey at boosthardware.com
Sat Jun 29 21:57:31 UTC 2013


On Sat, June 29, 2013 10:23 pm, Jeremy Jongepier wrote:
> On 06/29/2013 02:58 AM, Patrick Shirkey wrote:
>>
>> On Fri, June 28, 2013 11:52 pm, Jeremy Jongepier wrote:
>>> On 06/19/2013 03:39 PM, Jeremy Jongepier wrote:
>>>> On 06/19/2013 12:08 AM, Dan MacDonald wrote:
>>>>> That's what I expected, seeing how little success I've had on my
>>>>> Pandaboard
>>>>> with J2 git. Have you tried J1 too?
>>>>
>>>> Ok, got Jack1 running. Recompiled the kernel without debugging and
>>>> disabled WiFi. I now see I can also tweak the IRQ threads, nice.
>>>>
>>>> Jeremy
>>>
>>> Just an update. Getting things stable is quite some work actually but
>>> I'm slowly getting there. I've stripped the kernel config a bit further
>>> and have a good starting point now. Some findings:
>>> * RT patch doesn't apply cleanly. I've hacked my way through it but
>>> can't get it to compile.
>>
>> What are the error messages?
>>
>
> An undefined reference to printascii in some RockChip related driver:
>
> drivers/built-in.o: In function `dbg':
> mtk_wcn_cmb_stub_rockchip.c:(.text+0x3a598): undefined reference to
> `printascii'
> make: *** [.tmp_vmlinux1] Error 1
>


According to this thread it's an ARM specific debug method.

http://www.spinics.net/lists/arm-kernel/msg19006.html

Someone here might be able to help if you pastebin the revised patch


>>
>>> * I've tried using threadirqs but then the stick wouldn't boot
>>> (couldn't
>>> find any rootfs to load from MMC). Applied this patch:
>>> http://www.spinics.net/lists/linux-rt-users/msg08062.html
>>> Then I get threaded IRQs but it's not much of an improvement, the USB
>>> IRQ that has the WiFi connected to it consumes a lot of CPU and I can't
>>> get JACK to run stable. I did try prioritizing IRQs with rtirq but it
>>> didn't make any difference, the board seems to run better without
>>> threaded IRQs.
>>> * Tried dwc_otg.speed=1 as a kernel parameter, no improvement.
>>> * This thing doesn't have serial console access, I've tried all
>>> possible
>>> pads, vias and pins but no joy :( I did manage to solder a button on
>>> the
>>> board that allows me to directly boot into flash mode which is already
>>> quite convenient.
>>> * No ethernet access, only WiFi. But enabling WiFi means JACK won't run
>>> properly. Also tried an external USB dongle, same problems basically.
>>
>> What is the wifi chipset?
>>
>
> Of the rk3066 based device? Realtek 8188eu. The dongle is using a Ralink
> rt53xx chipset.
>

When does the CPU load from the wifi devices start to spike?

When you enable the wifi driver
After you ifup the NIC
While you have a remote connection

Did you test the bandwidth to confirm that the config  dwc_otg.speed=1 was
reducing the throughput to 12Mb/s

Did you also try:   dwc_otg.fiq_fix_enable=0/1


>>
>>> * So I have to work using the HDMI connection and a keyboard. Had to
>>> disable blanking otherwise JACK would crash.
>>>
>>> Well, that last example sums it all up I guess, this board isn't really
>>> made for doing real-time audio. Unless people prove me wrong, I'd love
>>> to hear what other people's experiences are. Any tips, ideas,
>>> suggestions? I'm now glazing at amsynth 1.4.0 running headless while
>>> JACK occasionally spits out 'delay of .... usecs exceeds estimated
>>> spare
>>> time of .....; restart ...' messages and that with jackd -P84 -p32
>>> -t2000 -dalsa -n3 -p128 -s
>>
>> What happens at higher period values and if you disable duplex mode?
>>
>
> I could test that but haven't done so yet because my goal is to have
> JACK running with -p64 and full duplex. JACK runs stable at -p128 but
> then the audio output is distorted and can't fix it with resetting the
> JACK buffersize (this 'trick' works with the RPi though). So far I'm
> getting the idea that at least the RK3066 based board I have is unusable
> for my purpose. And I have a feeling it won't be much better with a
> RK3188 based board.
>

I think it is early days yet.  It took a couple of years to get the maudio
quattro USB to play nice. It turned out it was using LE on the output and
BE on the input.

Just for reference sake, if you try it with just -P enabled does it
perform any better?



> Regards,
>
> Jeremy
>
>>
>>
>>> I HAVE to use the ALSA -s (softmode) option otherwise the whole USB
>>> stack locks up. Pffff, maybe I should buy a board that does work, like
>>> the pcDuino or BeagleBone Black.
>>>
>>> Regards,
>>>
>>> Jeremy
>>> _______________________________________________
>>> Linux-audio-user mailing list
>>> Linux-audio-user at lists.linuxaudio.org
>>> http://lists.linuxaudio.org/listinfo/linux-audio-user
>>>
>>
>>
>> --
>> Patrick Shirkey
>> Boost Hardware Ltd
>> _______________________________________________
>> Linux-audio-user mailing list
>> Linux-audio-user at lists.linuxaudio.org
>> http://lists.linuxaudio.org/listinfo/linux-audio-user
>>
>
> _______________________________________________
> Linux-audio-user mailing list
> Linux-audio-user at lists.linuxaudio.org
> http://lists.linuxaudio.org/listinfo/linux-audio-user
>


--
Patrick Shirkey
Boost Hardware Ltd


More information about the Linux-audio-user mailing list