[LAU] new rtirq release

[Fernando Lopez-Lezcano]

On 05/04/2012 02:57 PM, Rui Nuno Capela wrote:
> On 05/04/2012 08:08 PM, Fernando Lopez-Lezcano wrote:
>> On 05/04/2012 05:39 AM, Rui Nuno Capela wrote:
>>> On 05/04/2012 12:39 AM, Fernando Lopez-Lezcano wrote:
>>>> Rui: this addition/hack in rtirq takes care of snd_hda_intel in my
>>>> 3.2.16-rt27 system and will presumably do the right thing for any other
>>>> pci soundcards:
>>>
>> [MUNCH]
>>>
>>> given that, which got slightly re-hacked somehow, a new rtirq release
>>> has bumped in:
>>> http://www.rncbc.org/jack#rtirq
>>
>> Hmm, sorry Rui but that does not seem to work, you can't rely on the
>> name as provided in /proc/asound/cards, it will not match in the generic
>> case, so this:
>>
>> ----
>> PIDS=`ps -eo pid,comm | egrep -i "irq.${IRQ}.snd.${NAME2:0:4}" | awk
>> '{print $1}'`
>> ----
>>
>> should be something like:
>>
>> ----
>> PIDS=`ps -eo pid,comm | egrep -i "irq.${IRQ}.snd_.*" | awk '{print $1}'`
>> ----
>>
>> that should be enough, it would be very unlikely (impossible?) that
>> there is another soundcard on the same interrupt so a match with snd_
>> should be enough, and even if that were the case it would be processed
>> twice I guess, no harm done.
>
> what exactly happens in your particular case?

Nothing happens (ie: all processes with that particular irq number get 
the high priority). I did not _really_ investigate, it just did not work 
and I reverted to what I had tried before.

> maybe "snd.${NAME2:0:4}" comes as a 1st best try (where NAME2 stands for
> the name provided from /proc/asound/cards), then "snd.*" for the
> fallback (2nd try) ?

Sure, that would work...

> check(this)out
> https://www.rncbc.org/svn/rtirq/trunk
> posing for an immediate 20120505-29 stamp :)

Hmm, I don't see a difference in that url, maybe you did not copy things 
over?
-- Fernando