[alsa-devel] PATCH - MIDI on ice1724 - real-time kernel problem SOLVED(?)

[Pavel Hofman]

Martin Krüger wrote:
> Pavel Hofman schrieb:
>> Hi Martin,
>>
...............
> 
> Hi Pavel,
> 
> now i'm confused. I just compiled the current daily snapshot with
> "./configure --with-isapnp=no --with-sequencer=yes --with-oss=no 
> --with-debug=verbose --with-cards=ice1724".
> 
> That should give as much debug messages as i need, as i understood. I 
> get only one Message, telling me that ice1724 is using the defined 
> eeprom. After using the midi output port, there is no more message in 
> any log at all.

That is OK, it means the loop-detection code does not kick in, i.e. 
there is no problem with clearing interrupt status bits.

> 
> Due to this, it is really hard to understand the problem, for me it 
> looks like a complete freeze. If i would get "just" a heavy overload, 
> the numlock should still react within 10 minutes, and the cpu cooler 
> should be louder. Both of this should also happen during an irq flood, 
> if i understood that right.

My IRQ flood loaded the PC heavily, but I could still switch windows and 
move mouse around, with long delays. Yours looks like a real freeze.

> 
> Do you have any idea how to find out whats happening there? 

I just add debug printk's to major code points all the way down to 
bottom functions. It takes many compiling and rebooting (after freezes). 
  Unfortunately I do not know of any other way to hunt the bug down.

Good luck,

Pavel.