[PlanetCCRMA] Hammerfall-DSP Multiface problems

[Fernando Lopez-Lezcano]

On 05/29/2012 02:56 PM, plutek wrote:
> From: Fernando Lopez-Lezcano<nando at ccrma.Stanford.EDU>
>>> my system details:
>>> T520 ThinkPad
>>> 6 Gb RAM
>>> i7 quad-core CPU @ 2.8GHz
>>> RME Hammerfall DSP Multiface on ExpressCard adapter

(BTW, my current laptop is a T510 and runs a Multiface with the same 
adapter fine with 3.2.x rt, but on fc14 i386 for now)

>>> PlanetCCRMA on FC16 (x86-64)
>>>
>>> my observations so far:
>>>
>>> kernel 3.1.0-7.fc16.x86-64 ----
>>> jack runs reliably down to 64x2
>>> hdspmixer runs reliably
>>>
>>> kernels 3.2.17-1.rt28.1.fc16.ccrma.x86-64.rt AND 3.3.7-1.fc16.x86-64
>>> jack starts, but locks up very quickly
>>> hdspmixer locks with error: "HwDep ioctl failed. Metering stopped"
>>>
>>>
>>> any help sorting this out would be most welcome... let me know if there are error messages or other console output which would be useful!
>>
>> Hmmm, weird. Which version of alsa-tools do you have installed? (rpm -q
>> alsa-tools to find out).
>
> 1.0.25-1.fc16.x86_64
>
> but i'm not really sure if that's the culprit, or if it's a kernel thing that's throwing the hdsp, and the hdspmixer fail is just a symptom of that... on some other systems (with older kernels, mind you!) 1.0.23 was better for hdspmixer, so *maybe*. however, i do know there's still something else going on, because jack (jack2, btw) fails even if i haven't touched hdspmixer since a reboot.
>
>> I'm running 3.2.17+ rt here at ccrma with RME hardware, but this is on
>> old Fedora 14 i386 installs. I'm also running fc16 x86_64 + rt kernels
>> on a test laptop (old Lenovo T61) I have and had no problems with jack
>> using the built-in card.
>
> you haven't tested *RME hardware* on *that* setup (fc16 x86_64 + rt kernels on the old Lenovo T61)?

Hmmmm, I think not. I can try to remember to bring an rme pcmcia card 
tomorrow from home and test on my test laptop (with a Multiface II).

For the hdspmixer problem I used to have a patched alsa-tools package in 
Planet CCRMA, 1.0.24 was not good enough for the latest versions of the 
hdsp driver. I would have assumed that 1.0.25 fixed those problems - the 
patch came from alsa - but it seems it did not, which is weird.

> hmmmmm.... what would you recommend as my next step?

It is strange that jack itself is not happy with newer kernels. That 
would seem to suggest a kernel bug with your particular hardware.

What do you mean by "jack locks up very quickly"? Do you mean the whole 
machine locks up? Or just the jackd process? Does jackd print anything 
before it locks up?

-- Fernando